{ "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": 4.265281677246094, "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": 4.475153923034668, "learning_rate": 2.11954217888936e-08, "loss": 1.0326, "mean_token_accuracy": 0.7414199113845825, "num_tokens": 79961.0, "step": 2 }, { "epoch": 0.0003816308357715303, "ewc_loss": 1.3642420526593924e-11, "ewc_loss_diag": 1.4016565685892601e-15, "ewc_loss_parallel": 1.354472090042691e-14, "grad_norm": 4.328714847564697, "learning_rate": 4.23908435777872e-08, "loss": 0.9023, "mean_token_accuracy": 0.7781416177749634, "num_tokens": 118063.0, "step": 3 }, { "epoch": 0.0005088411143620405, "ewc_loss": 4.220055416226387e-10, "ewc_loss_diag": 5.440092820663267e-14, "ewc_loss_parallel": 4.1744385725905886e-13, "grad_norm": 4.460257053375244, "learning_rate": 6.358626536668081e-08, "loss": 0.9764, "mean_token_accuracy": 0.7542060017585754, "num_tokens": 149306.0, "step": 4 }, { "epoch": 0.0006360513929525506, "ewc_loss": 2.9685907065868378e-09, "ewc_loss_diag": 4.0678571622265736e-13, "ewc_loss_parallel": 2.9274360713316128e-12, "grad_norm": 4.255980968475342, "learning_rate": 8.47816871555744e-08, "loss": 0.9984, "mean_token_accuracy": 0.7507978081703186, "num_tokens": 192477.0, "step": 5 }, { "epoch": 0.0007632616715430606, "ewc_loss": 8.381903171539307e-09, "ewc_loss_diag": 8.739675649849232e-13, "ewc_loss_parallel": 8.29913915367797e-12, "grad_norm": 4.611693859100342, "learning_rate": 1.05977108944468e-07, "loss": 0.9604, "mean_token_accuracy": 0.7627652883529663, "num_tokens": 226246.0, "step": 6 }, { "epoch": 0.0008904719501335708, "ewc_loss": 4.307366907596588e-08, "ewc_loss_diag": 4.519051799434237e-12, "ewc_loss_parallel": 4.2518877307884395e-11, "grad_norm": 4.389065265655518, "learning_rate": 1.2717253073336161e-07, "loss": 1.0128, "mean_token_accuracy": 0.7552992105484009, "num_tokens": 263593.0, "step": 7 }, { "epoch": 0.001017682228724081, "ewc_loss": 7.031485438346863e-08, "ewc_loss_diag": 7.048583938740194e-12, "ewc_loss_parallel": 6.957634468562901e-11, "grad_norm": 5.192540168762207, "learning_rate": 1.4836795252225522e-07, "loss": 0.9098, "mean_token_accuracy": 0.7726883888244629, "num_tokens": 298202.0, "step": 8 }, { "epoch": 0.001144892507314591, "ewc_loss": 2.775341272354126e-07, "ewc_loss_diag": 3.115019353572279e-11, "ewc_loss_parallel": 2.7466739993542433e-10, "grad_norm": 4.455545902252197, "learning_rate": 1.695633743111488e-07, "loss": 1.0355, "mean_token_accuracy": 0.7470693588256836, "num_tokens": 340163.0, "step": 9 }, { "epoch": 0.0012721027859051012, "ewc_loss": 4.2654573917388916e-07, "ewc_loss_diag": 4.5702108764089644e-11, "ewc_loss_parallel": 4.220055416226387e-10, "grad_norm": 4.644032955169678, "learning_rate": 1.907587961000424e-07, "loss": 1.0413, "mean_token_accuracy": 0.7432145476341248, "num_tokens": 372703.0, "step": 10 }, { "epoch": 0.0013993130644956112, "ewc_loss": 5.587935447692871e-07, "ewc_loss_diag": 6.093614501878619e-11, "ewc_loss_parallel": 5.529727786779404e-10, "grad_norm": 4.180925369262695, "learning_rate": 2.11954217888936e-07, "loss": 0.9762, "mean_token_accuracy": 0.7518008947372437, "num_tokens": 414187.0, "step": 11 }, { "epoch": 0.0015265233430861213, "ewc_loss": 1.9818544387817383e-06, "ewc_loss_diag": 2.4010660126805305e-10, "ewc_loss_parallel": 1.949956640601158e-09, "grad_norm": 4.57502555847168, "learning_rate": 2.3314963967782957e-07, "loss": 1.0059, "mean_token_accuracy": 0.7502951622009277, "num_tokens": 446133.0, "step": 12 }, { "epoch": 0.0016537336216766315, "ewc_loss": 2.7120113372802734e-06, "ewc_loss_diag": 2.9831426218152046e-10, "ewc_loss_parallel": 2.6775524020195007e-09, "grad_norm": 4.683113098144531, "learning_rate": 2.5434506146672323e-07, "loss": 1.0488, "mean_token_accuracy": 0.7320963144302368, "num_tokens": 477130.0, "step": 13 }, { "epoch": 0.0017809439002671415, "ewc_loss": 3.5315752029418945e-06, "ewc_loss_diag": 3.892637323588133e-10, "ewc_loss_parallel": 3.4924596548080444e-09, "grad_norm": 4.240339279174805, "learning_rate": 2.755404832556168e-07, "loss": 1.0059, "mean_token_accuracy": 0.7581409811973572, "num_tokens": 521754.0, "step": 14 }, { "epoch": 0.0019081541788576518, "ewc_loss": 4.470348358154297e-06, "ewc_loss_diag": 6.402842700481415e-10, "ewc_loss_parallel": 4.423782229423523e-09, "grad_norm": 4.401723384857178, "learning_rate": 2.9673590504451043e-07, "loss": 0.9518, "mean_token_accuracy": 0.766878604888916, "num_tokens": 560311.0, "step": 15 }, { "epoch": 0.002035364457448162, "ewc_loss": 1.239776611328125e-05, "ewc_loss_diag": 1.709850039333105e-09, "ewc_loss_parallel": 1.2223608791828156e-08, "grad_norm": 4.27218770980835, "learning_rate": 3.17931326833404e-07, "loss": 1.0246, "mean_token_accuracy": 0.7412195801734924, "num_tokens": 595214.0, "step": 16 }, { "epoch": 0.002162574736038672, "ewc_loss": 1.7523765563964844e-05, "ewc_loss_diag": 2.0372681319713593e-09, "ewc_loss_parallel": 1.7229467630386353e-08, "grad_norm": 4.669015884399414, "learning_rate": 3.391267486222976e-07, "loss": 1.0326, "mean_token_accuracy": 0.7427088618278503, "num_tokens": 631427.0, "step": 17 }, { "epoch": 0.002289785014629182, "ewc_loss": 2.086162567138672e-05, "ewc_loss_diag": 2.4010660126805305e-09, "ewc_loss_parallel": 2.0605511963367462e-08, "grad_norm": 4.699876308441162, "learning_rate": 3.603221704111912e-07, "loss": 1.0829, "mean_token_accuracy": 0.7350613474845886, "num_tokens": 671155.0, "step": 18 }, { "epoch": 0.0024169952932196924, "ewc_loss": 2.5033950805664062e-05, "ewc_loss_diag": 2.9831426218152046e-09, "ewc_loss_parallel": 2.4796463549137115e-08, "grad_norm": 4.34279203414917, "learning_rate": 3.815175922000848e-07, "loss": 0.9856, "mean_token_accuracy": 0.7541123628616333, "num_tokens": 704515.0, "step": 19 }, { "epoch": 0.0025442055718102024, "ewc_loss": 2.8848648071289062e-05, "ewc_loss_diag": 3.4051481634378433e-09, "ewc_loss_parallel": 2.852175384759903e-08, "grad_norm": 4.260576248168945, "learning_rate": 4.027130139889784e-07, "loss": 0.9604, "mean_token_accuracy": 0.7602264285087585, "num_tokens": 745414.0, "step": 20 }, { "epoch": 0.0026714158504007124, "ewc_loss": 4.8160552978515625e-05, "ewc_loss_diag": 8.32369551062584e-09, "ewc_loss_parallel": 4.7497451305389404e-08, "grad_norm": 4.191277027130127, "learning_rate": 4.23908435777872e-07, "loss": 1.0015, "mean_token_accuracy": 0.7442262172698975, "num_tokens": 779384.0, "step": 21 }, { "epoch": 0.0027986261289912225, "ewc_loss": 8.821487426757812e-05, "ewc_loss_diag": 1.280568540096283e-08, "ewc_loss_parallel": 8.66129994392395e-08, "grad_norm": 4.072340488433838, "learning_rate": 4.451038575667656e-07, "loss": 0.9812, "mean_token_accuracy": 0.7519391179084778, "num_tokens": 812994.0, "step": 22 }, { "epoch": 0.0029258364075817325, "ewc_loss": 0.00011348724365234375, "ewc_loss_diag": 1.501757651567459e-08, "ewc_loss_parallel": 1.1222437024116516e-07, "grad_norm": 3.8792619705200195, "learning_rate": 4.6629927935565915e-07, "loss": 0.9174, "mean_token_accuracy": 0.7692388892173767, "num_tokens": 849351.0, "step": 23 }, { "epoch": 0.0030530466861722425, "ewc_loss": 0.000133514404296875, "ewc_loss_diag": 1.6298145055770874e-08, "ewc_loss_parallel": 1.3131648302078247e-07, "grad_norm": 4.583385944366455, "learning_rate": 4.874947011445528e-07, "loss": 0.957, "mean_token_accuracy": 0.7511836290359497, "num_tokens": 880319.0, "step": 24 }, { "epoch": 0.003180256964762753, "ewc_loss": 0.000141143798828125, "ewc_loss_diag": 1.9208528101444244e-08, "ewc_loss_parallel": 1.387670636177063e-07, "grad_norm": 3.9712276458740234, "learning_rate": 5.086901229334465e-07, "loss": 1.0517, "mean_token_accuracy": 0.7296730279922485, "num_tokens": 921902.0, "step": 25 }, { "epoch": 0.003307467243353263, "ewc_loss": 0.00015544891357421875, "ewc_loss_diag": 2.200249582529068e-08, "ewc_loss_parallel": 1.5366822481155396e-07, "grad_norm": 4.039771556854248, "learning_rate": 5.2988554472234e-07, "loss": 0.9822, "mean_token_accuracy": 0.7460971474647522, "num_tokens": 959531.0, "step": 26 }, { "epoch": 0.003434677521943773, "ewc_loss": 0.0001659393310546875, "ewc_loss_diag": 2.3981556296348572e-08, "ewc_loss_parallel": 1.6298145055770874e-07, "grad_norm": 3.5706350803375244, "learning_rate": 5.510809665112336e-07, "loss": 0.9008, "mean_token_accuracy": 0.7683305740356445, "num_tokens": 997280.0, "step": 27 }, { "epoch": 0.003561887800534283, "ewc_loss": 0.0001697540283203125, "ewc_loss_diag": 2.5960616767406464e-08, "ewc_loss_parallel": 1.6670674085617065e-07, "grad_norm": 3.991682291030884, "learning_rate": 5.722763883001272e-07, "loss": 1.0311, "mean_token_accuracy": 0.7352985143661499, "num_tokens": 1034608.0, "step": 28 }, { "epoch": 0.003689098079124793, "ewc_loss": 0.000270843505859375, "ewc_loss_diag": 5.3551048040390015e-08, "ewc_loss_parallel": 2.644956111907959e-07, "grad_norm": 3.2497782707214355, "learning_rate": 5.934718100890209e-07, "loss": 0.8947, "mean_token_accuracy": 0.7677261233329773, "num_tokens": 1075907.0, "step": 29 }, { "epoch": 0.0038163083577153036, "ewc_loss": 0.0004730224609375, "ewc_loss_diag": 7.310882210731506e-08, "ewc_loss_parallel": 4.6566128730773926e-07, "grad_norm": 3.1127984523773193, "learning_rate": 6.146672318779143e-07, "loss": 0.9272, "mean_token_accuracy": 0.7624205350875854, "num_tokens": 1114492.0, "step": 30 }, { "epoch": 0.003943518636305814, "ewc_loss": 0.00057220458984375, "ewc_loss_diag": 8.940696716308594e-08, "ewc_loss_parallel": 5.662441253662109e-07, "grad_norm": 2.853756904602051, "learning_rate": 6.35862653666808e-07, "loss": 0.8338, "mean_token_accuracy": 0.7817703485488892, "num_tokens": 1151596.0, "step": 31 }, { "epoch": 0.004070728914896324, "ewc_loss": 0.00063323974609375, "ewc_loss_diag": 9.96515154838562e-08, "ewc_loss_parallel": 6.221234798431396e-07, "grad_norm": 2.9924685955047607, "learning_rate": 6.570580754557016e-07, "loss": 0.9077, "mean_token_accuracy": 0.7619364261627197, "num_tokens": 1191475.0, "step": 32 }, { "epoch": 0.004197939193486834, "ewc_loss": 0.000644683837890625, "ewc_loss_diag": 1.0943040251731873e-07, "ewc_loss_parallel": 6.332993507385254e-07, "grad_norm": 2.828888177871704, "learning_rate": 6.782534972445952e-07, "loss": 0.8937, "mean_token_accuracy": 0.7645761370658875, "num_tokens": 1230726.0, "step": 33 }, { "epoch": 0.004325149472077344, "ewc_loss": 0.00063323974609375, "ewc_loss_diag": 1.1734664440155029e-07, "ewc_loss_parallel": 6.221234798431396e-07, "grad_norm": 2.802513837814331, "learning_rate": 6.994489190334889e-07, "loss": 0.8565, "mean_token_accuracy": 0.7733057141304016, "num_tokens": 1273523.0, "step": 34 }, { "epoch": 0.004452359750667854, "ewc_loss": 0.000598907470703125, "ewc_loss_diag": 1.2945383787155151e-07, "ewc_loss_parallel": 5.885958671569824e-07, "grad_norm": 2.69305157661438, "learning_rate": 7.206443408223824e-07, "loss": 0.8522, "mean_token_accuracy": 0.77350252866745, "num_tokens": 1312013.0, "step": 35 }, { "epoch": 0.004579570029258364, "ewc_loss": 0.000583648681640625, "ewc_loss_diag": 1.424923539161682e-07, "ewc_loss_parallel": 5.699694156646729e-07, "grad_norm": 2.7648863792419434, "learning_rate": 7.41839762611276e-07, "loss": 0.8724, "mean_token_accuracy": 0.765031635761261, "num_tokens": 1350139.0, "step": 36 }, { "epoch": 0.004706780307848874, "ewc_loss": 0.00057220458984375, "ewc_loss_diag": 1.5273690223693848e-07, "ewc_loss_parallel": 5.550682544708252e-07, "grad_norm": 3.3498759269714355, "learning_rate": 7.630351844001696e-07, "loss": 0.8383, "mean_token_accuracy": 0.773446261882782, "num_tokens": 1387670.0, "step": 37 }, { "epoch": 0.004833990586439385, "ewc_loss": 0.00055694580078125, "ewc_loss_diag": 1.6391277313232422e-07, "ewc_loss_parallel": 5.401670932769775e-07, "grad_norm": 2.462393283843994, "learning_rate": 7.842306061890633e-07, "loss": 0.85, "mean_token_accuracy": 0.7711904048919678, "num_tokens": 1431658.0, "step": 38 }, { "epoch": 0.004961200865029895, "ewc_loss": 0.000621795654296875, "ewc_loss_diag": 1.862645149230957e-07, "ewc_loss_parallel": 6.034970283508301e-07, "grad_norm": 2.302238941192627, "learning_rate": 8.054260279779568e-07, "loss": 0.7682, "mean_token_accuracy": 0.7934924364089966, "num_tokens": 1467751.0, "step": 39 }, { "epoch": 0.005088411143620405, "ewc_loss": 0.00145721435546875, "ewc_loss_diag": 2.4400651454925537e-07, "ewc_loss_parallel": 1.430511474609375e-06, "grad_norm": 2.2597286701202393, "learning_rate": 8.266214497668503e-07, "loss": 0.8461, "mean_token_accuracy": 0.7668367624282837, "num_tokens": 1507078.0, "step": 40 }, { "epoch": 0.005215621422210915, "ewc_loss": 0.00238037109375, "ewc_loss_diag": 2.998858690261841e-07, "ewc_loss_parallel": 2.3543834686279297e-06, "grad_norm": 1.9281625747680664, "learning_rate": 8.47816871555744e-07, "loss": 0.805, "mean_token_accuracy": 0.7772718071937561, "num_tokens": 1544856.0, "step": 41 }, { "epoch": 0.005342831700801425, "ewc_loss": 0.0025787353515625, "ewc_loss_diag": 3.650784492492676e-07, "ewc_loss_parallel": 2.5480985641479492e-06, "grad_norm": 1.921697735786438, "learning_rate": 8.690122933446376e-07, "loss": 0.7684, "mean_token_accuracy": 0.785409152507782, "num_tokens": 1576776.0, "step": 42 }, { "epoch": 0.005470041979391935, "ewc_loss": 0.0025482177734375, "ewc_loss_diag": 4.2282044887542725e-07, "ewc_loss_parallel": 2.5033950805664062e-06, "grad_norm": 1.672420859336853, "learning_rate": 8.902077151335312e-07, "loss": 0.7622, "mean_token_accuracy": 0.7841333746910095, "num_tokens": 1613819.0, "step": 43 }, { "epoch": 0.005597252257982445, "ewc_loss": 0.0024261474609375, "ewc_loss_diag": 4.805624485015869e-07, "ewc_loss_parallel": 2.384185791015625e-06, "grad_norm": 1.7224812507629395, "learning_rate": 9.114031369224248e-07, "loss": 0.802, "mean_token_accuracy": 0.7765865325927734, "num_tokens": 1654834.0, "step": 44 }, { "epoch": 0.005724462536572955, "ewc_loss": 0.0023040771484375, "ewc_loss_diag": 5.066394805908203e-07, "ewc_loss_parallel": 2.2649765014648438e-06, "grad_norm": 1.841170072555542, "learning_rate": 9.325985587113183e-07, "loss": 0.7885, "mean_token_accuracy": 0.778503954410553, "num_tokens": 1691238.0, "step": 45 }, { "epoch": 0.005851672815163465, "ewc_loss": 0.0022735595703125, "ewc_loss_diag": 5.252659320831299e-07, "ewc_loss_parallel": 2.2202730178833008e-06, "grad_norm": 1.4265581369400024, "learning_rate": 9.53793980500212e-07, "loss": 0.6978, "mean_token_accuracy": 0.7980125546455383, "num_tokens": 1734446.0, "step": 46 }, { "epoch": 0.005978883093753975, "ewc_loss": 0.0020599365234375, "ewc_loss_diag": 5.699694156646729e-07, "ewc_loss_parallel": 1.996755599975586e-06, "grad_norm": 1.4914591312408447, "learning_rate": 9.749894022891056e-07, "loss": 0.7389, "mean_token_accuracy": 0.7866091132164001, "num_tokens": 1770266.0, "step": 47 }, { "epoch": 0.006106093372344485, "ewc_loss": 0.0019073486328125, "ewc_loss_diag": 5.960464477539062e-07, "ewc_loss_parallel": 1.8477439880371094e-06, "grad_norm": 1.5782800912857056, "learning_rate": 9.961848240779993e-07, "loss": 0.7999, "mean_token_accuracy": 0.771947979927063, "num_tokens": 1807979.0, "step": 48 }, { "epoch": 0.006233303650934996, "ewc_loss": 0.00193023681640625, "ewc_loss_diag": 6.295740604400635e-07, "ewc_loss_parallel": 1.8700957298278809e-06, "grad_norm": 1.3627146482467651, "learning_rate": 1.017380245866893e-06, "loss": 0.6951, "mean_token_accuracy": 0.7982243895530701, "num_tokens": 1851061.0, "step": 49 }, { "epoch": 0.006360513929525506, "ewc_loss": 0.0018310546875, "ewc_loss_diag": 6.631016731262207e-07, "ewc_loss_parallel": 1.7657876014709473e-06, "grad_norm": 1.325443983078003, "learning_rate": 1.0385756676557863e-06, "loss": 0.6584, "mean_token_accuracy": 0.8136093616485596, "num_tokens": 1887647.0, "step": 50 }, { "epoch": 0.006487724208116016, "ewc_loss": 0.0017547607421875, "ewc_loss_diag": 7.003545761108398e-07, "ewc_loss_parallel": 1.6838312149047852e-06, "grad_norm": 1.4162333011627197, "learning_rate": 1.05977108944468e-06, "loss": 0.7846, "mean_token_accuracy": 0.7783876657485962, "num_tokens": 1924843.0, "step": 51 }, { "epoch": 0.006614934486706526, "ewc_loss": 0.00177001953125, "ewc_loss_diag": 7.413327693939209e-07, "ewc_loss_parallel": 1.691281795501709e-06, "grad_norm": 1.3577414751052856, "learning_rate": 1.0809665112335736e-06, "loss": 0.7082, "mean_token_accuracy": 0.7952818870544434, "num_tokens": 1967116.0, "step": 52 }, { "epoch": 0.006742144765297036, "ewc_loss": 0.00182342529296875, "ewc_loss_diag": 7.972121238708496e-07, "ewc_loss_parallel": 1.7434358596801758e-06, "grad_norm": 1.301903486251831, "learning_rate": 1.1021619330224672e-06, "loss": 0.7067, "mean_token_accuracy": 0.7926398515701294, "num_tokens": 1999575.0, "step": 53 }, { "epoch": 0.006869355043887546, "ewc_loss": 0.0019683837890625, "ewc_loss_diag": 8.344650268554688e-07, "ewc_loss_parallel": 1.8924474716186523e-06, "grad_norm": 1.5223225355148315, "learning_rate": 1.1233573548113608e-06, "loss": 0.7655, "mean_token_accuracy": 0.7779439687728882, "num_tokens": 2039396.0, "step": 54 }, { "epoch": 0.006996565322478056, "ewc_loss": 0.0023345947265625, "ewc_loss_diag": 9.08970832824707e-07, "ewc_loss_parallel": 2.250075340270996e-06, "grad_norm": 1.2870354652404785, "learning_rate": 1.1445527766002544e-06, "loss": 0.7988, "mean_token_accuracy": 0.7724531888961792, "num_tokens": 2080258.0, "step": 55 }, { "epoch": 0.007123775601068566, "ewc_loss": 0.002685546875, "ewc_loss_diag": 9.834766387939453e-07, "ewc_loss_parallel": 2.592802047729492e-06, "grad_norm": 1.251711368560791, "learning_rate": 1.165748198389148e-06, "loss": 0.7703, "mean_token_accuracy": 0.7788785099983215, "num_tokens": 2116552.0, "step": 56 }, { "epoch": 0.007250985879659076, "ewc_loss": 0.0029449462890625, "ewc_loss_diag": 1.0356307029724121e-06, "ewc_loss_parallel": 2.8312206268310547e-06, "grad_norm": 1.0966187715530396, "learning_rate": 1.1869436201780417e-06, "loss": 0.6758, "mean_token_accuracy": 0.8010223507881165, "num_tokens": 2161097.0, "step": 57 }, { "epoch": 0.007378196158249586, "ewc_loss": 0.0030975341796875, "ewc_loss_diag": 1.0654330253601074e-06, "ewc_loss_parallel": 2.995133399963379e-06, "grad_norm": 1.118251919746399, "learning_rate": 1.2081390419669353e-06, "loss": 0.6384, "mean_token_accuracy": 0.8121455311775208, "num_tokens": 2201190.0, "step": 58 }, { "epoch": 0.007505406436840096, "ewc_loss": 0.00335693359375, "ewc_loss_diag": 1.1399388313293457e-06, "ewc_loss_parallel": 3.248453140258789e-06, "grad_norm": 1.2238147258758545, "learning_rate": 1.2293344637558287e-06, "loss": 0.733, "mean_token_accuracy": 0.7848732471466064, "num_tokens": 2233725.0, "step": 59 }, { "epoch": 0.007632616715430607, "ewc_loss": 0.00372314453125, "ewc_loss_diag": 1.214444637298584e-06, "ewc_loss_parallel": 3.606081008911133e-06, "grad_norm": 1.1454943418502808, "learning_rate": 1.2505298855447225e-06, "loss": 0.696, "mean_token_accuracy": 0.7892272472381592, "num_tokens": 2274459.0, "step": 60 }, { "epoch": 0.007759826994021117, "ewc_loss": 0.0038909912109375, "ewc_loss_diag": 1.2814998626708984e-06, "ewc_loss_parallel": 3.769993782043457e-06, "grad_norm": 1.0745954513549805, "learning_rate": 1.271725307333616e-06, "loss": 0.6493, "mean_token_accuracy": 0.806334912776947, "num_tokens": 2317530.0, "step": 61 }, { "epoch": 0.007887037272611627, "ewc_loss": 0.003997802734375, "ewc_loss_diag": 1.3932585716247559e-06, "ewc_loss_parallel": 3.844499588012695e-06, "grad_norm": 1.1813477277755737, "learning_rate": 1.2929207291225096e-06, "loss": 0.7387, "mean_token_accuracy": 0.7862181067466736, "num_tokens": 2355260.0, "step": 62 }, { "epoch": 0.008014247551202136, "ewc_loss": 0.00433349609375, "ewc_loss_diag": 1.4975666999816895e-06, "ewc_loss_parallel": 4.172325134277344e-06, "grad_norm": 1.1288197040557861, "learning_rate": 1.3141161509114032e-06, "loss": 0.6553, "mean_token_accuracy": 0.8082386255264282, "num_tokens": 2397135.0, "step": 63 }, { "epoch": 0.008141457829792647, "ewc_loss": 0.00433349609375, "ewc_loss_diag": 1.55717134475708e-06, "ewc_loss_parallel": 4.172325134277344e-06, "grad_norm": 1.1383388042449951, "learning_rate": 1.3353115727002968e-06, "loss": 0.6909, "mean_token_accuracy": 0.796122670173645, "num_tokens": 2440206.0, "step": 64 }, { "epoch": 0.008268668108383158, "ewc_loss": 0.004364013671875, "ewc_loss_diag": 1.6167759895324707e-06, "ewc_loss_parallel": 4.202127456665039e-06, "grad_norm": 1.229943037033081, "learning_rate": 1.3565069944891903e-06, "loss": 0.7045, "mean_token_accuracy": 0.785762369632721, "num_tokens": 2469565.0, "step": 65 }, { "epoch": 0.008395878386973667, "ewc_loss": 0.004669189453125, "ewc_loss_diag": 1.6838312149047852e-06, "ewc_loss_parallel": 4.500150680541992e-06, "grad_norm": 1.0785280466079712, "learning_rate": 1.3777024162780841e-06, "loss": 0.6707, "mean_token_accuracy": 0.7989567518234253, "num_tokens": 2514314.0, "step": 66 }, { "epoch": 0.008523088665564178, "ewc_loss": 0.004486083984375, "ewc_loss_diag": 1.7508864402770996e-06, "ewc_loss_parallel": 4.291534423828125e-06, "grad_norm": 1.0949269533157349, "learning_rate": 1.3988978380669777e-06, "loss": 0.6988, "mean_token_accuracy": 0.7894227504730225, "num_tokens": 2558466.0, "step": 67 }, { "epoch": 0.008650298944154687, "ewc_loss": 0.004486083984375, "ewc_loss_diag": 1.8328428268432617e-06, "ewc_loss_parallel": 4.291534423828125e-06, "grad_norm": 1.0989441871643066, "learning_rate": 1.420093259855871e-06, "loss": 0.6634, "mean_token_accuracy": 0.8043680191040039, "num_tokens": 2596843.0, "step": 68 }, { "epoch": 0.008777509222745198, "ewc_loss": 0.004547119140625, "ewc_loss_diag": 1.8998980522155762e-06, "ewc_loss_parallel": 4.351139068603516e-06, "grad_norm": 1.0850286483764648, "learning_rate": 1.4412886816447649e-06, "loss": 0.6353, "mean_token_accuracy": 0.8072134256362915, "num_tokens": 2632933.0, "step": 69 }, { "epoch": 0.008904719501335707, "ewc_loss": 0.0045166015625, "ewc_loss_diag": 1.9669532775878906e-06, "ewc_loss_parallel": 4.32133674621582e-06, "grad_norm": 1.0446014404296875, "learning_rate": 1.4624841034336584e-06, "loss": 0.6185, "mean_token_accuracy": 0.8171088099479675, "num_tokens": 2673526.0, "step": 70 }, { "epoch": 0.009031929779926218, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 2.0265579223632812e-06, "ewc_loss_parallel": 4.231929779052734e-06, "grad_norm": 1.0983628034591675, "learning_rate": 1.483679525222552e-06, "loss": 0.6412, "mean_token_accuracy": 0.807606041431427, "num_tokens": 2707770.0, "step": 71 }, { "epoch": 0.009159140058516728, "ewc_loss": 0.004547119140625, "ewc_loss_diag": 2.0563602447509766e-06, "ewc_loss_parallel": 4.32133674621582e-06, "grad_norm": 1.051400065422058, "learning_rate": 1.5048749470114456e-06, "loss": 0.6092, "mean_token_accuracy": 0.8191919922828674, "num_tokens": 2745236.0, "step": 72 }, { "epoch": 0.009286350337107238, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 2.1010637283325195e-06, "ewc_loss_parallel": 4.231929779052734e-06, "grad_norm": 1.0788224935531616, "learning_rate": 1.5260703688003392e-06, "loss": 0.6811, "mean_token_accuracy": 0.7966352701187134, "num_tokens": 2780291.0, "step": 73 }, { "epoch": 0.009413560615697748, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 2.1457672119140625e-06, "ewc_loss_parallel": 4.231929779052734e-06, "grad_norm": 1.0916452407836914, "learning_rate": 1.5472657905892327e-06, "loss": 0.7214, "mean_token_accuracy": 0.7899674773216248, "num_tokens": 2814771.0, "step": 74 }, { "epoch": 0.009540770894288259, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 2.16066837310791e-06, "ewc_loss_parallel": 4.231929779052734e-06, "grad_norm": 0.9853731989860535, "learning_rate": 1.5684612123781265e-06, "loss": 0.6206, "mean_token_accuracy": 0.8129335045814514, "num_tokens": 2861375.0, "step": 75 }, { "epoch": 0.00966798117287877, "ewc_loss": 0.0042724609375, "ewc_loss_diag": 2.2202730178833008e-06, "ewc_loss_parallel": 4.0531158447265625e-06, "grad_norm": 1.0086241960525513, "learning_rate": 1.58965663416702e-06, "loss": 0.6135, "mean_token_accuracy": 0.8105727434158325, "num_tokens": 2896139.0, "step": 76 }, { "epoch": 0.009795191451469279, "ewc_loss": 0.00433349609375, "ewc_loss_diag": 2.2649765014648438e-06, "ewc_loss_parallel": 4.112720489501953e-06, "grad_norm": 0.980162501335144, "learning_rate": 1.6108520559559137e-06, "loss": 0.5826, "mean_token_accuracy": 0.8216656446456909, "num_tokens": 2940269.0, "step": 77 }, { "epoch": 0.00992240173005979, "ewc_loss": 0.00433349609375, "ewc_loss_diag": 2.339482307434082e-06, "ewc_loss_parallel": 4.082918167114258e-06, "grad_norm": 1.0358502864837646, "learning_rate": 1.6320474777448073e-06, "loss": 0.6133, "mean_token_accuracy": 0.8112741112709045, "num_tokens": 2971115.0, "step": 78 }, { "epoch": 0.010049612008650299, "ewc_loss": 0.004364013671875, "ewc_loss_diag": 2.3692846298217773e-06, "ewc_loss_parallel": 4.112720489501953e-06, "grad_norm": 0.996283769607544, "learning_rate": 1.6532428995337006e-06, "loss": 0.5341, "mean_token_accuracy": 0.8353405594825745, "num_tokens": 3003727.0, "step": 79 }, { "epoch": 0.01017682228724081, "ewc_loss": 0.00433349609375, "ewc_loss_diag": 2.3990869522094727e-06, "ewc_loss_parallel": 4.082918167114258e-06, "grad_norm": 0.9578836560249329, "learning_rate": 1.6744383213225942e-06, "loss": 0.5641, "mean_token_accuracy": 0.8271893262863159, "num_tokens": 3045557.0, "step": 80 }, { "epoch": 0.010304032565831319, "ewc_loss": 0.004302978515625, "ewc_loss_diag": 2.473592758178711e-06, "ewc_loss_parallel": 4.0531158447265625e-06, "grad_norm": 1.0119707584381104, "learning_rate": 1.695633743111488e-06, "loss": 0.6309, "mean_token_accuracy": 0.8066471815109253, "num_tokens": 3080478.0, "step": 81 }, { "epoch": 0.01043124284442183, "ewc_loss": 0.004364013671875, "ewc_loss_diag": 2.518296241760254e-06, "ewc_loss_parallel": 4.112720489501953e-06, "grad_norm": 0.9945659041404724, "learning_rate": 1.7168291649003816e-06, "loss": 0.6489, "mean_token_accuracy": 0.8045002222061157, "num_tokens": 3119205.0, "step": 82 }, { "epoch": 0.010558453123012339, "ewc_loss": 0.004364013671875, "ewc_loss_diag": 2.60770320892334e-06, "ewc_loss_parallel": 4.082918167114258e-06, "grad_norm": 1.0306048393249512, "learning_rate": 1.7380245866892751e-06, "loss": 0.6499, "mean_token_accuracy": 0.7978031635284424, "num_tokens": 3151816.0, "step": 83 }, { "epoch": 0.01068566340160285, "ewc_loss": 0.00439453125, "ewc_loss_diag": 2.637505531311035e-06, "ewc_loss_parallel": 4.112720489501953e-06, "grad_norm": 1.007057547569275, "learning_rate": 1.759220008478169e-06, "loss": 0.6347, "mean_token_accuracy": 0.8064342141151428, "num_tokens": 3189765.0, "step": 84 }, { "epoch": 0.010812873680193359, "ewc_loss": 0.00439453125, "ewc_loss_diag": 2.6971101760864258e-06, "ewc_loss_parallel": 4.112720489501953e-06, "grad_norm": 0.9796878695487976, "learning_rate": 1.7804154302670625e-06, "loss": 0.6631, "mean_token_accuracy": 0.7996473908424377, "num_tokens": 3228988.0, "step": 85 }, { "epoch": 0.01094008395878387, "ewc_loss": 0.004364013671875, "ewc_loss_diag": 2.7567148208618164e-06, "ewc_loss_parallel": 4.082918167114258e-06, "grad_norm": 0.9702089428901672, "learning_rate": 1.801610852055956e-06, "loss": 0.6661, "mean_token_accuracy": 0.7987935543060303, "num_tokens": 3270648.0, "step": 86 }, { "epoch": 0.01106729423737438, "ewc_loss": 0.004364013671875, "ewc_loss_diag": 2.8014183044433594e-06, "ewc_loss_parallel": 4.082918167114258e-06, "grad_norm": 0.9579720497131348, "learning_rate": 1.8228062738448496e-06, "loss": 0.6335, "mean_token_accuracy": 0.8110852241516113, "num_tokens": 3312134.0, "step": 87 }, { "epoch": 0.01119450451596489, "ewc_loss": 0.004364013671875, "ewc_loss_diag": 2.8759241104125977e-06, "ewc_loss_parallel": 4.082918167114258e-06, "grad_norm": 0.9538876414299011, "learning_rate": 1.844001695633743e-06, "loss": 0.6347, "mean_token_accuracy": 0.8081629276275635, "num_tokens": 3351904.0, "step": 88 }, { "epoch": 0.0113217147945554, "ewc_loss": 0.004425048828125, "ewc_loss_diag": 2.950429916381836e-06, "ewc_loss_parallel": 4.112720489501953e-06, "grad_norm": 1.198214054107666, "learning_rate": 1.8651971174226366e-06, "loss": 0.6031, "mean_token_accuracy": 0.8166182637214661, "num_tokens": 3389610.0, "step": 89 }, { "epoch": 0.01144892507314591, "ewc_loss": 0.0048828125, "ewc_loss_diag": 2.9802322387695312e-06, "ewc_loss_parallel": 4.589557647705078e-06, "grad_norm": 1.0262924432754517, "learning_rate": 1.8863925392115304e-06, "loss": 0.6758, "mean_token_accuracy": 0.7961544990539551, "num_tokens": 3425779.0, "step": 90 }, { "epoch": 0.01157613535173642, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 3.039836883544922e-06, "ewc_loss_parallel": 4.1425228118896484e-06, "grad_norm": 0.9913787245750427, "learning_rate": 1.907587961000424e-06, "loss": 0.6493, "mean_token_accuracy": 0.8050867319107056, "num_tokens": 3463938.0, "step": 91 }, { "epoch": 0.01170334563032693, "ewc_loss": 0.004547119140625, "ewc_loss_diag": 3.11434268951416e-06, "ewc_loss_parallel": 4.231929779052734e-06, "grad_norm": 2.0918736457824707, "learning_rate": 1.9287833827893175e-06, "loss": 0.6162, "mean_token_accuracy": 0.8125557899475098, "num_tokens": 3500494.0, "step": 92 }, { "epoch": 0.01183055590891744, "ewc_loss": 0.006256103515625, "ewc_loss_diag": 3.1888484954833984e-06, "ewc_loss_parallel": 5.9604644775390625e-06, "grad_norm": 1.0868303775787354, "learning_rate": 1.949978804578211e-06, "loss": 0.5999, "mean_token_accuracy": 0.8189985752105713, "num_tokens": 3540901.0, "step": 93 }, { "epoch": 0.01195776618750795, "ewc_loss": 0.00469970703125, "ewc_loss_diag": 3.248453140258789e-06, "ewc_loss_parallel": 4.351139068603516e-06, "grad_norm": 0.9586123824119568, "learning_rate": 1.9711742263671047e-06, "loss": 0.6239, "mean_token_accuracy": 0.8082180023193359, "num_tokens": 3579275.0, "step": 94 }, { "epoch": 0.012084976466098461, "ewc_loss": 0.00457763671875, "ewc_loss_diag": 3.293156623840332e-06, "ewc_loss_parallel": 4.231929779052734e-06, "grad_norm": 0.9820171594619751, "learning_rate": 1.9923696481559987e-06, "loss": 0.592, "mean_token_accuracy": 0.8164796829223633, "num_tokens": 3614755.0, "step": 95 }, { "epoch": 0.01221218674468897, "ewc_loss": 0.004791259765625, "ewc_loss_diag": 3.3527612686157227e-06, "ewc_loss_parallel": 4.470348358154297e-06, "grad_norm": 1.0374003648757935, "learning_rate": 2.0135650699448923e-06, "loss": 0.6484, "mean_token_accuracy": 0.8014882206916809, "num_tokens": 3655744.0, "step": 96 }, { "epoch": 0.012339397023279481, "ewc_loss": 0.004791259765625, "ewc_loss_diag": 3.427267074584961e-06, "ewc_loss_parallel": 4.4405460357666016e-06, "grad_norm": 1.0229041576385498, "learning_rate": 2.034760491733786e-06, "loss": 0.6388, "mean_token_accuracy": 0.8092427253723145, "num_tokens": 3697257.0, "step": 97 }, { "epoch": 0.012466607301869992, "ewc_loss": 0.00482177734375, "ewc_loss_diag": 3.5017728805541992e-06, "ewc_loss_parallel": 4.470348358154297e-06, "grad_norm": 0.9797555208206177, "learning_rate": 2.055955913522679e-06, "loss": 0.5857, "mean_token_accuracy": 0.8205819129943848, "num_tokens": 3735737.0, "step": 98 }, { "epoch": 0.012593817580460501, "ewc_loss": 0.00482177734375, "ewc_loss_diag": 3.5315752029418945e-06, "ewc_loss_parallel": 4.4405460357666016e-06, "grad_norm": 1.01290762424469, "learning_rate": 2.0771513353115726e-06, "loss": 0.6089, "mean_token_accuracy": 0.8138123750686646, "num_tokens": 3774212.0, "step": 99 }, { "epoch": 0.012721027859051012, "ewc_loss": 0.004852294921875, "ewc_loss_diag": 3.56137752532959e-06, "ewc_loss_parallel": 4.500150680541992e-06, "grad_norm": 1.0108587741851807, "learning_rate": 2.098346757100466e-06, "loss": 0.5881, "mean_token_accuracy": 0.8205299377441406, "num_tokens": 3807640.0, "step": 100 }, { "epoch": 0.012848238137641521, "ewc_loss": 0.00482177734375, "ewc_loss_diag": 3.591179847717285e-06, "ewc_loss_parallel": 4.4405460357666016e-06, "grad_norm": 0.9739696383476257, "learning_rate": 2.11954217888936e-06, "loss": 0.6164, "mean_token_accuracy": 0.814426600933075, "num_tokens": 3848908.0, "step": 101 }, { "epoch": 0.012975448416232032, "ewc_loss": 0.0047607421875, "ewc_loss_diag": 3.680586814880371e-06, "ewc_loss_parallel": 4.380941390991211e-06, "grad_norm": 0.9841924905776978, "learning_rate": 2.1407376006782537e-06, "loss": 0.615, "mean_token_accuracy": 0.8085083961486816, "num_tokens": 3886341.0, "step": 102 }, { "epoch": 0.013102658694822541, "ewc_loss": 0.00482177734375, "ewc_loss_diag": 3.7401914596557617e-06, "ewc_loss_parallel": 4.4405460357666016e-06, "grad_norm": 1.00825834274292, "learning_rate": 2.1619330224671473e-06, "loss": 0.5229, "mean_token_accuracy": 0.8345839977264404, "num_tokens": 3917111.0, "step": 103 }, { "epoch": 0.013229868973413052, "ewc_loss": 0.004852294921875, "ewc_loss_diag": 3.874301910400391e-06, "ewc_loss_parallel": 4.470348358154297e-06, "grad_norm": 0.9514006972312927, "learning_rate": 2.183128444256041e-06, "loss": 0.5948, "mean_token_accuracy": 0.8176305294036865, "num_tokens": 3959642.0, "step": 104 }, { "epoch": 0.013357079252003561, "ewc_loss": 0.004852294921875, "ewc_loss_diag": 3.904104232788086e-06, "ewc_loss_parallel": 4.4405460357666016e-06, "grad_norm": 0.9465989470481873, "learning_rate": 2.2043238660449344e-06, "loss": 0.6122, "mean_token_accuracy": 0.8125813007354736, "num_tokens": 4001028.0, "step": 105 }, { "epoch": 0.013484289530594072, "ewc_loss": 0.004852294921875, "ewc_loss_diag": 3.9637088775634766e-06, "ewc_loss_parallel": 4.4405460357666016e-06, "grad_norm": 0.960536777973175, "learning_rate": 2.225519287833828e-06, "loss": 0.6305, "mean_token_accuracy": 0.8059889078140259, "num_tokens": 4045249.0, "step": 106 }, { "epoch": 0.013611499809184581, "ewc_loss": 0.004913330078125, "ewc_loss_diag": 3.993511199951172e-06, "ewc_loss_parallel": 4.5299530029296875e-06, "grad_norm": 1.0000903606414795, "learning_rate": 2.2467147096227216e-06, "loss": 0.6347, "mean_token_accuracy": 0.8087677955627441, "num_tokens": 4082369.0, "step": 107 }, { "epoch": 0.013738710087775092, "ewc_loss": 0.005035400390625, "ewc_loss_diag": 4.023313522338867e-06, "ewc_loss_parallel": 4.649162292480469e-06, "grad_norm": 1.0383390188217163, "learning_rate": 2.267910131411615e-06, "loss": 0.6199, "mean_token_accuracy": 0.8083995580673218, "num_tokens": 4117448.0, "step": 108 }, { "epoch": 0.013865920366365603, "ewc_loss": 0.00506591796875, "ewc_loss_diag": 4.082918167114258e-06, "ewc_loss_parallel": 4.678964614868164e-06, "grad_norm": 0.9769485592842102, "learning_rate": 2.2891055532005087e-06, "loss": 0.5668, "mean_token_accuracy": 0.829140305519104, "num_tokens": 4159132.0, "step": 109 }, { "epoch": 0.013993130644956112, "ewc_loss": 0.004974365234375, "ewc_loss_diag": 4.172325134277344e-06, "ewc_loss_parallel": 4.559755325317383e-06, "grad_norm": 0.9667256474494934, "learning_rate": 2.3103009749894023e-06, "loss": 0.5977, "mean_token_accuracy": 0.813626766204834, "num_tokens": 4199800.0, "step": 110 }, { "epoch": 0.014120340923546623, "ewc_loss": 0.00506591796875, "ewc_loss_diag": 4.202127456665039e-06, "ewc_loss_parallel": 4.649162292480469e-06, "grad_norm": 0.9663078188896179, "learning_rate": 2.331496396778296e-06, "loss": 0.6026, "mean_token_accuracy": 0.8172672986984253, "num_tokens": 4244513.0, "step": 111 }, { "epoch": 0.014247551202137132, "ewc_loss": 0.005035400390625, "ewc_loss_diag": 4.26173210144043e-06, "ewc_loss_parallel": 4.6193599700927734e-06, "grad_norm": 0.9712136387825012, "learning_rate": 2.3526918185671895e-06, "loss": 0.5505, "mean_token_accuracy": 0.8308150768280029, "num_tokens": 4282858.0, "step": 112 }, { "epoch": 0.014374761480727643, "ewc_loss": 0.005096435546875, "ewc_loss_diag": 4.26173210144043e-06, "ewc_loss_parallel": 4.678964614868164e-06, "grad_norm": 0.9749730825424194, "learning_rate": 2.3738872403560835e-06, "loss": 0.5943, "mean_token_accuracy": 0.8162032961845398, "num_tokens": 4325092.0, "step": 113 }, { "epoch": 0.014501971759318152, "ewc_loss": 0.005096435546875, "ewc_loss_diag": 4.32133674621582e-06, "ewc_loss_parallel": 4.678964614868164e-06, "grad_norm": 0.9509042501449585, "learning_rate": 2.395082662144977e-06, "loss": 0.5317, "mean_token_accuracy": 0.8333113789558411, "num_tokens": 4363459.0, "step": 114 }, { "epoch": 0.014629182037908663, "ewc_loss": 0.00506591796875, "ewc_loss_diag": 4.351139068603516e-06, "ewc_loss_parallel": 4.649162292480469e-06, "grad_norm": 0.9885274767875671, "learning_rate": 2.4162780839338706e-06, "loss": 0.6107, "mean_token_accuracy": 0.8123911619186401, "num_tokens": 4400065.0, "step": 115 }, { "epoch": 0.014756392316499172, "ewc_loss": 0.00531005859375, "ewc_loss_diag": 4.410743713378906e-06, "ewc_loss_parallel": 4.887580871582031e-06, "grad_norm": 0.989808976650238, "learning_rate": 2.4374735057227638e-06, "loss": 0.5903, "mean_token_accuracy": 0.8169189691543579, "num_tokens": 4437828.0, "step": 116 }, { "epoch": 0.014883602595089683, "ewc_loss": 0.00537109375, "ewc_loss_diag": 4.500150680541992e-06, "ewc_loss_parallel": 4.9173831939697266e-06, "grad_norm": 1.006516456604004, "learning_rate": 2.4586689275116574e-06, "loss": 0.561, "mean_token_accuracy": 0.8237448930740356, "num_tokens": 4475556.0, "step": 117 }, { "epoch": 0.015010812873680193, "ewc_loss": 0.005462646484375, "ewc_loss_diag": 4.559755325317383e-06, "ewc_loss_parallel": 5.0067901611328125e-06, "grad_norm": 0.998740553855896, "learning_rate": 2.479864349300551e-06, "loss": 0.5644, "mean_token_accuracy": 0.8266042470932007, "num_tokens": 4512087.0, "step": 118 }, { "epoch": 0.015138023152270703, "ewc_loss": 0.005462646484375, "ewc_loss_diag": 4.589557647705078e-06, "ewc_loss_parallel": 5.0067901611328125e-06, "grad_norm": 0.9803270697593689, "learning_rate": 2.501059771089445e-06, "loss": 0.5435, "mean_token_accuracy": 0.8300206065177917, "num_tokens": 4552116.0, "step": 119 }, { "epoch": 0.015265233430861214, "ewc_loss": 0.005401611328125, "ewc_loss_diag": 4.649162292480469e-06, "ewc_loss_parallel": 4.947185516357422e-06, "grad_norm": 0.983306884765625, "learning_rate": 2.5222551928783385e-06, "loss": 0.5856, "mean_token_accuracy": 0.8212295770645142, "num_tokens": 4591723.0, "step": 120 }, { "epoch": 0.015392443709451724, "ewc_loss": 0.005523681640625, "ewc_loss_diag": 4.649162292480469e-06, "ewc_loss_parallel": 5.066394805908203e-06, "grad_norm": 0.9409831166267395, "learning_rate": 2.543450614667232e-06, "loss": 0.5258, "mean_token_accuracy": 0.8393486142158508, "num_tokens": 4635800.0, "step": 121 }, { "epoch": 0.015519653988042234, "ewc_loss": 0.00543212890625, "ewc_loss_diag": 4.738569259643555e-06, "ewc_loss_parallel": 4.976987838745117e-06, "grad_norm": 0.9623906016349792, "learning_rate": 2.5646460364561257e-06, "loss": 0.5828, "mean_token_accuracy": 0.8202117681503296, "num_tokens": 4679908.0, "step": 122 }, { "epoch": 0.015646864266632744, "ewc_loss": 0.005645751953125, "ewc_loss_diag": 4.857778549194336e-06, "ewc_loss_parallel": 5.155801773071289e-06, "grad_norm": 0.9894478917121887, "learning_rate": 2.5858414582450192e-06, "loss": 0.5673, "mean_token_accuracy": 0.8238950967788696, "num_tokens": 4720120.0, "step": 123 }, { "epoch": 0.015774074545223254, "ewc_loss": 0.005706787109375, "ewc_loss_diag": 4.9173831939697266e-06, "ewc_loss_parallel": 5.21540641784668e-06, "grad_norm": 1.0044302940368652, "learning_rate": 2.607036880033913e-06, "loss": 0.5901, "mean_token_accuracy": 0.8128325939178467, "num_tokens": 4758777.0, "step": 124 }, { "epoch": 0.015901284823813765, "ewc_loss": 0.005767822265625, "ewc_loss_diag": 4.947185516357422e-06, "ewc_loss_parallel": 5.27501106262207e-06, "grad_norm": 1.0201690196990967, "learning_rate": 2.6282323018228064e-06, "loss": 0.5551, "mean_token_accuracy": 0.8300791382789612, "num_tokens": 4794781.0, "step": 125 }, { "epoch": 0.016028495102404273, "ewc_loss": 0.00579833984375, "ewc_loss_diag": 5.0067901611328125e-06, "ewc_loss_parallel": 5.304813385009766e-06, "grad_norm": 0.9992542862892151, "learning_rate": 2.6494277236117e-06, "loss": 0.6028, "mean_token_accuracy": 0.8122930526733398, "num_tokens": 4836391.0, "step": 126 }, { "epoch": 0.016155705380994784, "ewc_loss": 0.0057373046875, "ewc_loss_diag": 5.0067901611328125e-06, "ewc_loss_parallel": 5.245208740234375e-06, "grad_norm": 1.0198649168014526, "learning_rate": 2.6706231454005935e-06, "loss": 0.5742, "mean_token_accuracy": 0.8235432505607605, "num_tokens": 4874193.0, "step": 127 }, { "epoch": 0.016282915659585295, "ewc_loss": 0.005889892578125, "ewc_loss_diag": 5.066394805908203e-06, "ewc_loss_parallel": 5.364418029785156e-06, "grad_norm": 1.032028317451477, "learning_rate": 2.691818567189487e-06, "loss": 0.5791, "mean_token_accuracy": 0.8180660009384155, "num_tokens": 4909311.0, "step": 128 }, { "epoch": 0.016410125938175806, "ewc_loss": 0.005889892578125, "ewc_loss_diag": 5.125999450683594e-06, "ewc_loss_parallel": 5.364418029785156e-06, "grad_norm": 1.018909215927124, "learning_rate": 2.7130139889783807e-06, "loss": 0.5775, "mean_token_accuracy": 0.8227211833000183, "num_tokens": 4947409.0, "step": 129 }, { "epoch": 0.016537336216766316, "ewc_loss": 0.00592041015625, "ewc_loss_diag": 5.21540641784668e-06, "ewc_loss_parallel": 5.3942203521728516e-06, "grad_norm": 1.0331991910934448, "learning_rate": 2.7342094107672743e-06, "loss": 0.5827, "mean_token_accuracy": 0.8182331323623657, "num_tokens": 4982792.0, "step": 130 }, { "epoch": 0.016664546495356824, "ewc_loss": 0.006072998046875, "ewc_loss_diag": 5.27501106262207e-06, "ewc_loss_parallel": 5.543231964111328e-06, "grad_norm": 1.0069475173950195, "learning_rate": 2.7554048325561683e-06, "loss": 0.5595, "mean_token_accuracy": 0.8242459297180176, "num_tokens": 5020795.0, "step": 131 }, { "epoch": 0.016791756773947335, "ewc_loss": 0.00592041015625, "ewc_loss_diag": 5.334615707397461e-06, "ewc_loss_parallel": 5.364418029785156e-06, "grad_norm": 1.108953833580017, "learning_rate": 2.776600254345062e-06, "loss": 0.5959, "mean_token_accuracy": 0.8162384629249573, "num_tokens": 5061519.0, "step": 132 }, { "epoch": 0.016918967052537846, "ewc_loss": 0.006439208984375, "ewc_loss_diag": 5.3942203521728516e-06, "ewc_loss_parallel": 5.900859832763672e-06, "grad_norm": 1.0686503648757935, "learning_rate": 2.7977956761339554e-06, "loss": 0.5934, "mean_token_accuracy": 0.8161289691925049, "num_tokens": 5094608.0, "step": 133 }, { "epoch": 0.017046177331128357, "ewc_loss": 0.006134033203125, "ewc_loss_diag": 5.453824996948242e-06, "ewc_loss_parallel": 5.5730342864990234e-06, "grad_norm": 1.0421741008758545, "learning_rate": 2.8189910979228486e-06, "loss": 0.5962, "mean_token_accuracy": 0.8159689903259277, "num_tokens": 5131130.0, "step": 134 }, { "epoch": 0.017173387609718864, "ewc_loss": 0.0062255859375, "ewc_loss_diag": 5.513429641723633e-06, "ewc_loss_parallel": 5.662441253662109e-06, "grad_norm": 1.0121541023254395, "learning_rate": 2.840186519711742e-06, "loss": 0.5695, "mean_token_accuracy": 0.8238610029220581, "num_tokens": 5168921.0, "step": 135 }, { "epoch": 0.017300597888309375, "ewc_loss": 0.00616455078125, "ewc_loss_diag": 5.5730342864990234e-06, "ewc_loss_parallel": 5.602836608886719e-06, "grad_norm": 1.0448473691940308, "learning_rate": 2.8613819415006357e-06, "loss": 0.5788, "mean_token_accuracy": 0.8211151361465454, "num_tokens": 5206417.0, "step": 136 }, { "epoch": 0.017427808166899886, "ewc_loss": 0.006378173828125, "ewc_loss_diag": 5.602836608886719e-06, "ewc_loss_parallel": 5.841255187988281e-06, "grad_norm": 1.0763996839523315, "learning_rate": 2.8825773632895297e-06, "loss": 0.6146, "mean_token_accuracy": 0.8125318288803101, "num_tokens": 5239984.0, "step": 137 }, { "epoch": 0.017555018445490397, "ewc_loss": 0.00653076171875, "ewc_loss_diag": 5.662441253662109e-06, "ewc_loss_parallel": 5.990266799926758e-06, "grad_norm": 1.0458848476409912, "learning_rate": 2.9037727850784233e-06, "loss": 0.595, "mean_token_accuracy": 0.8167170286178589, "num_tokens": 5276444.0, "step": 138 }, { "epoch": 0.017682228724080904, "ewc_loss": 0.00640869140625, "ewc_loss_diag": 5.7220458984375e-06, "ewc_loss_parallel": 5.841255187988281e-06, "grad_norm": 1.025828242301941, "learning_rate": 2.924968206867317e-06, "loss": 0.5252, "mean_token_accuracy": 0.8354911804199219, "num_tokens": 5312064.0, "step": 139 }, { "epoch": 0.017809439002671415, "ewc_loss": 0.0064697265625, "ewc_loss_diag": 5.811452865600586e-06, "ewc_loss_parallel": 5.900859832763672e-06, "grad_norm": 1.0998436212539673, "learning_rate": 2.9461636286562104e-06, "loss": 0.6184, "mean_token_accuracy": 0.8173549771308899, "num_tokens": 5350405.0, "step": 140 }, { "epoch": 0.017936649281261926, "ewc_loss": 0.006866455078125, "ewc_loss_diag": 5.841255187988281e-06, "ewc_loss_parallel": 6.288290023803711e-06, "grad_norm": 1.0638245344161987, "learning_rate": 2.967359050445104e-06, "loss": 0.5656, "mean_token_accuracy": 0.8272473812103271, "num_tokens": 5388338.0, "step": 141 }, { "epoch": 0.018063859559852437, "ewc_loss": 0.006591796875, "ewc_loss_diag": 5.9604644775390625e-06, "ewc_loss_parallel": 6.020069122314453e-06, "grad_norm": 1.0325511693954468, "learning_rate": 2.9885544722339976e-06, "loss": 0.5756, "mean_token_accuracy": 0.8242830634117126, "num_tokens": 5432228.0, "step": 142 }, { "epoch": 0.018191069838442948, "ewc_loss": 0.0067138671875, "ewc_loss_diag": 5.990266799926758e-06, "ewc_loss_parallel": 6.109476089477539e-06, "grad_norm": 1.0357011556625366, "learning_rate": 3.009749894022891e-06, "loss": 0.5543, "mean_token_accuracy": 0.829386293888092, "num_tokens": 5469969.0, "step": 143 }, { "epoch": 0.018318280117033455, "ewc_loss": 0.006805419921875, "ewc_loss_diag": 6.020069122314453e-06, "ewc_loss_parallel": 6.198883056640625e-06, "grad_norm": 1.0383045673370361, "learning_rate": 3.0309453158117847e-06, "loss": 0.5444, "mean_token_accuracy": 0.8288853168487549, "num_tokens": 5506262.0, "step": 144 }, { "epoch": 0.018445490395623966, "ewc_loss": 0.006866455078125, "ewc_loss_diag": 6.0498714447021484e-06, "ewc_loss_parallel": 6.258487701416016e-06, "grad_norm": 1.0693985223770142, "learning_rate": 3.0521407376006783e-06, "loss": 0.5628, "mean_token_accuracy": 0.8242295980453491, "num_tokens": 5542870.0, "step": 145 }, { "epoch": 0.018572700674214477, "ewc_loss": 0.007080078125, "ewc_loss_diag": 6.079673767089844e-06, "ewc_loss_parallel": 6.467103958129883e-06, "grad_norm": 1.061974287033081, "learning_rate": 3.073336159389572e-06, "loss": 0.5762, "mean_token_accuracy": 0.8199236989021301, "num_tokens": 5579958.0, "step": 146 }, { "epoch": 0.018699910952804988, "ewc_loss": 0.006988525390625, "ewc_loss_diag": 6.079673767089844e-06, "ewc_loss_parallel": 6.377696990966797e-06, "grad_norm": 1.0576270818710327, "learning_rate": 3.0945315811784655e-06, "loss": 0.5929, "mean_token_accuracy": 0.8145914077758789, "num_tokens": 5623306.0, "step": 147 }, { "epoch": 0.018827121231395495, "ewc_loss": 0.007049560546875, "ewc_loss_diag": 6.16908073425293e-06, "ewc_loss_parallel": 6.4373016357421875e-06, "grad_norm": 1.0436288118362427, "learning_rate": 3.115727002967359e-06, "loss": 0.5239, "mean_token_accuracy": 0.8388384580612183, "num_tokens": 5660801.0, "step": 148 }, { "epoch": 0.018954331509986006, "ewc_loss": 0.00701904296875, "ewc_loss_diag": 6.198883056640625e-06, "ewc_loss_parallel": 6.407499313354492e-06, "grad_norm": 1.0684795379638672, "learning_rate": 3.136922424756253e-06, "loss": 0.5515, "mean_token_accuracy": 0.8279135227203369, "num_tokens": 5696625.0, "step": 149 }, { "epoch": 0.019081541788576517, "ewc_loss": 0.007110595703125, "ewc_loss_diag": 6.22868537902832e-06, "ewc_loss_parallel": 6.496906280517578e-06, "grad_norm": 1.0647361278533936, "learning_rate": 3.1581178465451466e-06, "loss": 0.572, "mean_token_accuracy": 0.8213365077972412, "num_tokens": 5733491.0, "step": 150 }, { "epoch": 0.019208752067167028, "ewc_loss": 0.00714111328125, "ewc_loss_diag": 6.288290023803711e-06, "ewc_loss_parallel": 6.496906280517578e-06, "grad_norm": 1.0628689527511597, "learning_rate": 3.17931326833404e-06, "loss": 0.5945, "mean_token_accuracy": 0.8146520853042603, "num_tokens": 5773550.0, "step": 151 }, { "epoch": 0.01933596234575754, "ewc_loss": 0.0072021484375, "ewc_loss_diag": 6.318092346191406e-06, "ewc_loss_parallel": 6.556510925292969e-06, "grad_norm": 1.069075345993042, "learning_rate": 3.2005086901229338e-06, "loss": 0.5839, "mean_token_accuracy": 0.8178383111953735, "num_tokens": 5813114.0, "step": 152 }, { "epoch": 0.019463172624348046, "ewc_loss": 0.007232666015625, "ewc_loss_diag": 6.3478946685791016e-06, "ewc_loss_parallel": 6.586313247680664e-06, "grad_norm": 1.0801304578781128, "learning_rate": 3.2217041119118274e-06, "loss": 0.6122, "mean_token_accuracy": 0.8121961951255798, "num_tokens": 5852702.0, "step": 153 }, { "epoch": 0.019590382902938557, "ewc_loss": 0.007354736328125, "ewc_loss_diag": 6.407499313354492e-06, "ewc_loss_parallel": 6.705522537231445e-06, "grad_norm": 1.0801398754119873, "learning_rate": 3.242899533700721e-06, "loss": 0.6736, "mean_token_accuracy": 0.7923115491867065, "num_tokens": 5894481.0, "step": 154 }, { "epoch": 0.019717593181529068, "ewc_loss": 0.00738525390625, "ewc_loss_diag": 6.496906280517578e-06, "ewc_loss_parallel": 6.735324859619141e-06, "grad_norm": 1.0786099433898926, "learning_rate": 3.2640949554896145e-06, "loss": 0.5471, "mean_token_accuracy": 0.8248840570449829, "num_tokens": 5929501.0, "step": 155 }, { "epoch": 0.01984480346011958, "ewc_loss": 0.00750732421875, "ewc_loss_diag": 6.586313247680664e-06, "ewc_loss_parallel": 6.8247318267822266e-06, "grad_norm": 1.0798110961914062, "learning_rate": 3.2852903772785085e-06, "loss": 0.5488, "mean_token_accuracy": 0.8277626037597656, "num_tokens": 5966407.0, "step": 156 }, { "epoch": 0.019972013738710086, "ewc_loss": 0.007537841796875, "ewc_loss_diag": 6.616115570068359e-06, "ewc_loss_parallel": 6.854534149169922e-06, "grad_norm": 1.073267936706543, "learning_rate": 3.3064857990674012e-06, "loss": 0.5003, "mean_token_accuracy": 0.8424673676490784, "num_tokens": 6004059.0, "step": 157 }, { "epoch": 0.020099224017300597, "ewc_loss": 0.007537841796875, "ewc_loss_diag": 6.67572021484375e-06, "ewc_loss_parallel": 6.854534149169922e-06, "grad_norm": 1.0671788454055786, "learning_rate": 3.327681220856295e-06, "loss": 0.5452, "mean_token_accuracy": 0.8266563415527344, "num_tokens": 6040516.0, "step": 158 }, { "epoch": 0.020226434295891108, "ewc_loss": 0.00762939453125, "ewc_loss_diag": 6.765127182006836e-06, "ewc_loss_parallel": 6.943941116333008e-06, "grad_norm": 1.0986684560775757, "learning_rate": 3.3488766426451884e-06, "loss": 0.6185, "mean_token_accuracy": 0.8069310188293457, "num_tokens": 6077801.0, "step": 159 }, { "epoch": 0.02035364457448162, "ewc_loss": 0.007781982421875, "ewc_loss_diag": 6.8247318267822266e-06, "ewc_loss_parallel": 7.12275505065918e-06, "grad_norm": 1.0735411643981934, "learning_rate": 3.3700720644340824e-06, "loss": 0.5431, "mean_token_accuracy": 0.8301776647567749, "num_tokens": 6114102.0, "step": 160 }, { "epoch": 0.020480854853072127, "ewc_loss": 0.00775146484375, "ewc_loss_diag": 6.854534149169922e-06, "ewc_loss_parallel": 7.092952728271484e-06, "grad_norm": 1.0749865770339966, "learning_rate": 3.391267486222976e-06, "loss": 0.5669, "mean_token_accuracy": 0.8238477110862732, "num_tokens": 6157342.0, "step": 161 }, { "epoch": 0.020608065131662637, "ewc_loss": 0.0078125, "ewc_loss_diag": 6.9141387939453125e-06, "ewc_loss_parallel": 7.152557373046875e-06, "grad_norm": 1.1159112453460693, "learning_rate": 3.4124629080118695e-06, "loss": 0.5639, "mean_token_accuracy": 0.823174238204956, "num_tokens": 6193058.0, "step": 162 }, { "epoch": 0.02073527541025315, "ewc_loss": 0.00799560546875, "ewc_loss_diag": 6.943941116333008e-06, "ewc_loss_parallel": 7.331371307373047e-06, "grad_norm": 1.1677496433258057, "learning_rate": 3.433658329800763e-06, "loss": 0.5645, "mean_token_accuracy": 0.827482283115387, "num_tokens": 6223616.0, "step": 163 }, { "epoch": 0.02086248568884366, "ewc_loss": 0.0081787109375, "ewc_loss_diag": 6.943941116333008e-06, "ewc_loss_parallel": 7.510185241699219e-06, "grad_norm": 1.0856256484985352, "learning_rate": 3.4548537515896567e-06, "loss": 0.5627, "mean_token_accuracy": 0.82852703332901, "num_tokens": 6264920.0, "step": 164 }, { "epoch": 0.02098969596743417, "ewc_loss": 0.0078125, "ewc_loss_diag": 6.973743438720703e-06, "ewc_loss_parallel": 7.12275505065918e-06, "grad_norm": 1.0871143341064453, "learning_rate": 3.4760491733785503e-06, "loss": 0.6013, "mean_token_accuracy": 0.8129422664642334, "num_tokens": 6305270.0, "step": 165 }, { "epoch": 0.021116906246024678, "ewc_loss": 0.0081787109375, "ewc_loss_diag": 7.0035457611083984e-06, "ewc_loss_parallel": 7.450580596923828e-06, "grad_norm": 1.0837191343307495, "learning_rate": 3.497244595167444e-06, "loss": 0.5075, "mean_token_accuracy": 0.8386249542236328, "num_tokens": 6341945.0, "step": 166 }, { "epoch": 0.02124411652461519, "ewc_loss": 0.00811767578125, "ewc_loss_diag": 7.063150405883789e-06, "ewc_loss_parallel": 7.420778274536133e-06, "grad_norm": 1.0757571458816528, "learning_rate": 3.518440016956338e-06, "loss": 0.5159, "mean_token_accuracy": 0.8412646651268005, "num_tokens": 6381565.0, "step": 167 }, { "epoch": 0.0213713268032057, "ewc_loss": 0.0081787109375, "ewc_loss_diag": 7.152557373046875e-06, "ewc_loss_parallel": 7.4803829193115234e-06, "grad_norm": 1.0980170965194702, "learning_rate": 3.5396354387452314e-06, "loss": 0.5473, "mean_token_accuracy": 0.8277859687805176, "num_tokens": 6419771.0, "step": 168 }, { "epoch": 0.02149853708179621, "ewc_loss": 0.00830078125, "ewc_loss_diag": 7.18235969543457e-06, "ewc_loss_parallel": 7.569789886474609e-06, "grad_norm": 1.1442586183547974, "learning_rate": 3.560830860534125e-06, "loss": 0.5148, "mean_token_accuracy": 0.8367180228233337, "num_tokens": 6451413.0, "step": 169 }, { "epoch": 0.021625747360386718, "ewc_loss": 0.0084228515625, "ewc_loss_diag": 7.18235969543457e-06, "ewc_loss_parallel": 7.68899917602539e-06, "grad_norm": 1.1079176664352417, "learning_rate": 3.5820262823230186e-06, "loss": 0.5286, "mean_token_accuracy": 0.8350605964660645, "num_tokens": 6488883.0, "step": 170 }, { "epoch": 0.02175295763897723, "ewc_loss": 0.00830078125, "ewc_loss_diag": 7.212162017822266e-06, "ewc_loss_parallel": 7.569789886474609e-06, "grad_norm": 1.1352920532226562, "learning_rate": 3.603221704111912e-06, "loss": 0.5708, "mean_token_accuracy": 0.8227958679199219, "num_tokens": 6523669.0, "step": 171 }, { "epoch": 0.02188016791756774, "ewc_loss": 0.0084228515625, "ewc_loss_diag": 7.241964340209961e-06, "ewc_loss_parallel": 7.68899917602539e-06, "grad_norm": 1.103186845779419, "learning_rate": 3.6244171259008057e-06, "loss": 0.5359, "mean_token_accuracy": 0.8287124633789062, "num_tokens": 6561795.0, "step": 172 }, { "epoch": 0.02200737819615825, "ewc_loss": 0.00836181640625, "ewc_loss_diag": 7.331371307373047e-06, "ewc_loss_parallel": 7.62939453125e-06, "grad_norm": 1.1073002815246582, "learning_rate": 3.6456125476896993e-06, "loss": 0.5227, "mean_token_accuracy": 0.8357725143432617, "num_tokens": 6604102.0, "step": 173 }, { "epoch": 0.02213458847474876, "ewc_loss": 0.0084228515625, "ewc_loss_diag": 7.420778274536133e-06, "ewc_loss_parallel": 7.68899917602539e-06, "grad_norm": 1.1343944072723389, "learning_rate": 3.6668079694785933e-06, "loss": 0.606, "mean_token_accuracy": 0.813065767288208, "num_tokens": 6645412.0, "step": 174 }, { "epoch": 0.02226179875333927, "ewc_loss": 0.008544921875, "ewc_loss_diag": 7.450580596923828e-06, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.1363154649734497, "learning_rate": 3.688003391267486e-06, "loss": 0.5113, "mean_token_accuracy": 0.8412123918533325, "num_tokens": 6678407.0, "step": 175 }, { "epoch": 0.02238900903192978, "ewc_loss": 0.008544921875, "ewc_loss_diag": 7.450580596923828e-06, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.1025477647781372, "learning_rate": 3.7091988130563796e-06, "loss": 0.5453, "mean_token_accuracy": 0.8285747170448303, "num_tokens": 6720968.0, "step": 176 }, { "epoch": 0.02251621931052029, "ewc_loss": 0.00848388671875, "ewc_loss_diag": 7.510185241699219e-06, "ewc_loss_parallel": 7.748603820800781e-06, "grad_norm": 1.1416358947753906, "learning_rate": 3.730394234845273e-06, "loss": 0.549, "mean_token_accuracy": 0.8299175500869751, "num_tokens": 6757286.0, "step": 177 }, { "epoch": 0.0226434295891108, "ewc_loss": 0.0086669921875, "ewc_loss_diag": 7.510185241699219e-06, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.1767897605895996, "learning_rate": 3.751589656634167e-06, "loss": 0.5861, "mean_token_accuracy": 0.8174928426742554, "num_tokens": 6790260.0, "step": 178 }, { "epoch": 0.02277063986770131, "ewc_loss": 0.0086669921875, "ewc_loss_diag": 7.539987564086914e-06, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.118586778640747, "learning_rate": 3.7727850784230608e-06, "loss": 0.5275, "mean_token_accuracy": 0.8336976766586304, "num_tokens": 6828882.0, "step": 179 }, { "epoch": 0.02289785014629182, "ewc_loss": 0.0086669921875, "ewc_loss_diag": 7.68899917602539e-06, "ewc_loss_parallel": 7.867813110351562e-06, "grad_norm": 1.113728642463684, "learning_rate": 3.7939805002119543e-06, "loss": 0.5368, "mean_token_accuracy": 0.8335002064704895, "num_tokens": 6867332.0, "step": 180 }, { "epoch": 0.02302506042488233, "ewc_loss": 0.00872802734375, "ewc_loss_diag": 7.748603820800781e-06, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.1283541917800903, "learning_rate": 3.815175922000848e-06, "loss": 0.6493, "mean_token_accuracy": 0.7981195449829102, "num_tokens": 6914101.0, "step": 181 }, { "epoch": 0.02315227070347284, "ewc_loss": 0.00872802734375, "ewc_loss_diag": 7.808208465576172e-06, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.135806918144226, "learning_rate": 3.8363713437897415e-06, "loss": 0.5467, "mean_token_accuracy": 0.8270754814147949, "num_tokens": 6952981.0, "step": 182 }, { "epoch": 0.02327948098206335, "ewc_loss": 0.00872802734375, "ewc_loss_diag": 7.867813110351562e-06, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.1567318439483643, "learning_rate": 3.857566765578635e-06, "loss": 0.5305, "mean_token_accuracy": 0.8347234725952148, "num_tokens": 6987534.0, "step": 183 }, { "epoch": 0.02340669126065386, "ewc_loss": 0.0087890625, "ewc_loss_diag": 7.867813110351562e-06, "ewc_loss_parallel": 7.987022399902344e-06, "grad_norm": 1.1421045064926147, "learning_rate": 3.878762187367529e-06, "loss": 0.595, "mean_token_accuracy": 0.8150124549865723, "num_tokens": 7029343.0, "step": 184 }, { "epoch": 0.02353390153924437, "ewc_loss": 0.00872802734375, "ewc_loss_diag": 7.927417755126953e-06, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.1270803213119507, "learning_rate": 3.899957609156422e-06, "loss": 0.5374, "mean_token_accuracy": 0.8321290612220764, "num_tokens": 7067558.0, "step": 185 }, { "epoch": 0.02366111181783488, "ewc_loss": 0.00872802734375, "ewc_loss_diag": 8.046627044677734e-06, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.15101957321167, "learning_rate": 3.921153030945316e-06, "loss": 0.5179, "mean_token_accuracy": 0.8321793079376221, "num_tokens": 7106164.0, "step": 186 }, { "epoch": 0.023788322096425393, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 8.106231689453125e-06, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.1421270370483398, "learning_rate": 3.942348452734209e-06, "loss": 0.5722, "mean_token_accuracy": 0.8239493370056152, "num_tokens": 7145549.0, "step": 187 }, { "epoch": 0.0239155323750159, "ewc_loss": 0.0089111328125, "ewc_loss_diag": 8.225440979003906e-06, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.1844390630722046, "learning_rate": 3.963543874523103e-06, "loss": 0.5393, "mean_token_accuracy": 0.8314747214317322, "num_tokens": 7179697.0, "step": 188 }, { "epoch": 0.02404274265360641, "ewc_loss": 0.009033203125, "ewc_loss_diag": 8.285045623779297e-06, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.1682357788085938, "learning_rate": 3.984739296311997e-06, "loss": 0.6012, "mean_token_accuracy": 0.8111205101013184, "num_tokens": 7220103.0, "step": 189 }, { "epoch": 0.024169952932196922, "ewc_loss": 0.00897216796875, "ewc_loss_diag": 8.285045623779297e-06, "ewc_loss_parallel": 8.106231689453125e-06, "grad_norm": 1.1331967115402222, "learning_rate": 4.005934718100891e-06, "loss": 0.4792, "mean_token_accuracy": 0.8523812890052795, "num_tokens": 7258059.0, "step": 190 }, { "epoch": 0.024297163210787433, "ewc_loss": 0.00897216796875, "ewc_loss_diag": 8.285045623779297e-06, "ewc_loss_parallel": 8.106231689453125e-06, "grad_norm": 1.1796824932098389, "learning_rate": 4.0271301398897845e-06, "loss": 0.5812, "mean_token_accuracy": 0.8206424713134766, "num_tokens": 7295111.0, "step": 191 }, { "epoch": 0.02442437348937794, "ewc_loss": 0.00909423828125, "ewc_loss_diag": 8.404254913330078e-06, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.207746148109436, "learning_rate": 4.048325561678678e-06, "loss": 0.5887, "mean_token_accuracy": 0.815436840057373, "num_tokens": 7332703.0, "step": 192 }, { "epoch": 0.02455158376796845, "ewc_loss": 0.00909423828125, "ewc_loss_diag": 8.404254913330078e-06, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.1572965383529663, "learning_rate": 4.069520983467572e-06, "loss": 0.5884, "mean_token_accuracy": 0.8155784606933594, "num_tokens": 7376446.0, "step": 193 }, { "epoch": 0.024678794046558962, "ewc_loss": 0.009033203125, "ewc_loss_diag": 8.463859558105469e-06, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.1596455574035645, "learning_rate": 4.090716405256464e-06, "loss": 0.5366, "mean_token_accuracy": 0.8298442959785461, "num_tokens": 7416378.0, "step": 194 }, { "epoch": 0.024806004325149473, "ewc_loss": 0.00909423828125, "ewc_loss_diag": 8.58306884765625e-06, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.172319769859314, "learning_rate": 4.111911827045358e-06, "loss": 0.5471, "mean_token_accuracy": 0.8315893411636353, "num_tokens": 7458178.0, "step": 195 }, { "epoch": 0.024933214603739984, "ewc_loss": 0.00909423828125, "ewc_loss_diag": 8.761882781982422e-06, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.1293538808822632, "learning_rate": 4.1331072488342515e-06, "loss": 0.4874, "mean_token_accuracy": 0.8472422957420349, "num_tokens": 7502436.0, "step": 196 }, { "epoch": 0.02506042488233049, "ewc_loss": 0.00909423828125, "ewc_loss_diag": 8.821487426757812e-06, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.17777419090271, "learning_rate": 4.154302670623145e-06, "loss": 0.5472, "mean_token_accuracy": 0.826442539691925, "num_tokens": 7539625.0, "step": 197 }, { "epoch": 0.025187635160921002, "ewc_loss": 0.00921630859375, "ewc_loss_diag": 8.940696716308594e-06, "ewc_loss_parallel": 8.285045623779297e-06, "grad_norm": 1.203145980834961, "learning_rate": 4.175498092412039e-06, "loss": 0.5248, "mean_token_accuracy": 0.8343236446380615, "num_tokens": 7570904.0, "step": 198 }, { "epoch": 0.025314845439511513, "ewc_loss": 0.00921630859375, "ewc_loss_diag": 8.940696716308594e-06, "ewc_loss_parallel": 8.285045623779297e-06, "grad_norm": 1.1678516864776611, "learning_rate": 4.196693514200932e-06, "loss": 0.5202, "mean_token_accuracy": 0.8369565010070801, "num_tokens": 7608652.0, "step": 199 }, { "epoch": 0.025442055718102024, "ewc_loss": 0.00921630859375, "ewc_loss_diag": 9.000301361083984e-06, "ewc_loss_parallel": 8.285045623779297e-06, "grad_norm": 1.2175370454788208, "learning_rate": 4.217888935989827e-06, "loss": 0.5996, "mean_token_accuracy": 0.8158578872680664, "num_tokens": 7647759.0, "step": 200 }, { "epoch": 0.02556926599669253, "ewc_loss": 0.00927734375, "ewc_loss_diag": 9.119510650634766e-06, "ewc_loss_parallel": 8.344650268554688e-06, "grad_norm": 1.216282844543457, "learning_rate": 4.23908435777872e-06, "loss": 0.5135, "mean_token_accuracy": 0.8347392082214355, "num_tokens": 7678918.0, "step": 201 }, { "epoch": 0.025696476275283042, "ewc_loss": 0.00921630859375, "ewc_loss_diag": 9.179115295410156e-06, "ewc_loss_parallel": 8.285045623779297e-06, "grad_norm": 1.176664113998413, "learning_rate": 4.260279779567614e-06, "loss": 0.4906, "mean_token_accuracy": 0.8437362909317017, "num_tokens": 7715797.0, "step": 202 }, { "epoch": 0.025823686553873553, "ewc_loss": 0.00921630859375, "ewc_loss_diag": 9.179115295410156e-06, "ewc_loss_parallel": 8.285045623779297e-06, "grad_norm": 1.1845871210098267, "learning_rate": 4.281475201356507e-06, "loss": 0.5975, "mean_token_accuracy": 0.8145182132720947, "num_tokens": 7763053.0, "step": 203 }, { "epoch": 0.025950896832464064, "ewc_loss": 0.00927734375, "ewc_loss_diag": 9.298324584960938e-06, "ewc_loss_parallel": 8.344650268554688e-06, "grad_norm": 1.2696220874786377, "learning_rate": 4.302670623145401e-06, "loss": 0.5621, "mean_token_accuracy": 0.8278034925460815, "num_tokens": 7793956.0, "step": 204 }, { "epoch": 0.026078107111054575, "ewc_loss": 0.00946044921875, "ewc_loss_diag": 9.357929229736328e-06, "ewc_loss_parallel": 8.52346420288086e-06, "grad_norm": 1.1901949644088745, "learning_rate": 4.3238660449342946e-06, "loss": 0.542, "mean_token_accuracy": 0.8326339721679688, "num_tokens": 7833997.0, "step": 205 }, { "epoch": 0.026205317389645082, "ewc_loss": 0.00927734375, "ewc_loss_diag": 9.47713851928711e-06, "ewc_loss_parallel": 8.344650268554688e-06, "grad_norm": 1.1901168823242188, "learning_rate": 4.345061466723188e-06, "loss": 0.5885, "mean_token_accuracy": 0.8201048374176025, "num_tokens": 7877711.0, "step": 206 }, { "epoch": 0.026332527668235593, "ewc_loss": 0.0093994140625, "ewc_loss_diag": 9.47713851928711e-06, "ewc_loss_parallel": 8.463859558105469e-06, "grad_norm": 1.234164834022522, "learning_rate": 4.366256888512082e-06, "loss": 0.5053, "mean_token_accuracy": 0.8451974987983704, "num_tokens": 7918136.0, "step": 207 }, { "epoch": 0.026459737946826104, "ewc_loss": 0.009765625, "ewc_loss_diag": 9.5367431640625e-06, "ewc_loss_parallel": 8.761882781982422e-06, "grad_norm": 1.1938376426696777, "learning_rate": 4.387452310300975e-06, "loss": 0.5488, "mean_token_accuracy": 0.8306267261505127, "num_tokens": 7960847.0, "step": 208 }, { "epoch": 0.026586948225416615, "ewc_loss": 0.0096435546875, "ewc_loss_diag": 9.655952453613281e-06, "ewc_loss_parallel": 8.64267349243164e-06, "grad_norm": 1.2193188667297363, "learning_rate": 4.408647732089869e-06, "loss": 0.5162, "mean_token_accuracy": 0.8352984189987183, "num_tokens": 7996567.0, "step": 209 }, { "epoch": 0.026714158504007122, "ewc_loss": 0.0098876953125, "ewc_loss_diag": 9.655952453613281e-06, "ewc_loss_parallel": 8.881092071533203e-06, "grad_norm": 1.2230148315429688, "learning_rate": 4.4298431538787625e-06, "loss": 0.5397, "mean_token_accuracy": 0.8296481966972351, "num_tokens": 8032441.0, "step": 210 }, { "epoch": 0.026841368782597633, "ewc_loss": 0.0098876953125, "ewc_loss_diag": 9.715557098388672e-06, "ewc_loss_parallel": 8.881092071533203e-06, "grad_norm": 1.213107943534851, "learning_rate": 4.451038575667656e-06, "loss": 0.5655, "mean_token_accuracy": 0.8233979344367981, "num_tokens": 8072020.0, "step": 211 }, { "epoch": 0.026968579061188144, "ewc_loss": 0.0098876953125, "ewc_loss_diag": 9.775161743164062e-06, "ewc_loss_parallel": 8.940696716308594e-06, "grad_norm": 1.2534103393554688, "learning_rate": 4.47223399745655e-06, "loss": 0.6153, "mean_token_accuracy": 0.8122574090957642, "num_tokens": 8105736.0, "step": 212 }, { "epoch": 0.027095789339778655, "ewc_loss": 0.00994873046875, "ewc_loss_diag": 9.834766387939453e-06, "ewc_loss_parallel": 9.000301361083984e-06, "grad_norm": 1.2138304710388184, "learning_rate": 4.493429419245443e-06, "loss": 0.4724, "mean_token_accuracy": 0.8479066491127014, "num_tokens": 8139021.0, "step": 213 }, { "epoch": 0.027222999618369163, "ewc_loss": 0.0098876953125, "ewc_loss_diag": 9.953975677490234e-06, "ewc_loss_parallel": 8.940696716308594e-06, "grad_norm": 1.2011648416519165, "learning_rate": 4.514624841034337e-06, "loss": 0.5389, "mean_token_accuracy": 0.8282807469367981, "num_tokens": 8182361.0, "step": 214 }, { "epoch": 0.027350209896959674, "ewc_loss": 0.010009765625, "ewc_loss_diag": 1.0013580322265625e-05, "ewc_loss_parallel": 9.059906005859375e-06, "grad_norm": 1.2392780780792236, "learning_rate": 4.53582026282323e-06, "loss": 0.5553, "mean_token_accuracy": 0.8259034156799316, "num_tokens": 8218131.0, "step": 215 }, { "epoch": 0.027477420175550184, "ewc_loss": 0.01007080078125, "ewc_loss_diag": 1.0013580322265625e-05, "ewc_loss_parallel": 9.119510650634766e-06, "grad_norm": 1.2329761981964111, "learning_rate": 4.557015684612124e-06, "loss": 0.5336, "mean_token_accuracy": 0.8375018835067749, "num_tokens": 8255886.0, "step": 216 }, { "epoch": 0.027604630454140695, "ewc_loss": 0.0101318359375, "ewc_loss_diag": 1.0132789611816406e-05, "ewc_loss_parallel": 9.119510650634766e-06, "grad_norm": 1.2029303312301636, "learning_rate": 4.5782111064010175e-06, "loss": 0.4962, "mean_token_accuracy": 0.8426978588104248, "num_tokens": 8297196.0, "step": 217 }, { "epoch": 0.027731840732731206, "ewc_loss": 0.0101318359375, "ewc_loss_diag": 1.0192394256591797e-05, "ewc_loss_parallel": 9.119510650634766e-06, "grad_norm": 1.244044542312622, "learning_rate": 4.599406528189911e-06, "loss": 0.5302, "mean_token_accuracy": 0.8329624533653259, "num_tokens": 8331857.0, "step": 218 }, { "epoch": 0.027859051011321714, "ewc_loss": 0.0103759765625, "ewc_loss_diag": 1.0251998901367188e-05, "ewc_loss_parallel": 9.357929229736328e-06, "grad_norm": 1.2200068235397339, "learning_rate": 4.620601949978805e-06, "loss": 0.5251, "mean_token_accuracy": 0.835071861743927, "num_tokens": 8371120.0, "step": 219 }, { "epoch": 0.027986261289912225, "ewc_loss": 0.01019287109375, "ewc_loss_diag": 1.0251998901367188e-05, "ewc_loss_parallel": 9.179115295410156e-06, "grad_norm": 1.2748485803604126, "learning_rate": 4.641797371767698e-06, "loss": 0.6365, "mean_token_accuracy": 0.8024067878723145, "num_tokens": 8407502.0, "step": 220 }, { "epoch": 0.028113471568502735, "ewc_loss": 0.010498046875, "ewc_loss_diag": 1.0371208190917969e-05, "ewc_loss_parallel": 9.47713851928711e-06, "grad_norm": 1.2309682369232178, "learning_rate": 4.662992793556592e-06, "loss": 0.5001, "mean_token_accuracy": 0.8449999690055847, "num_tokens": 8444360.0, "step": 221 }, { "epoch": 0.028240681847093246, "ewc_loss": 0.01025390625, "ewc_loss_diag": 1.0371208190917969e-05, "ewc_loss_parallel": 9.238719940185547e-06, "grad_norm": 1.2599971294403076, "learning_rate": 4.684188215345485e-06, "loss": 0.5122, "mean_token_accuracy": 0.837722897529602, "num_tokens": 8476749.0, "step": 222 }, { "epoch": 0.028367892125683754, "ewc_loss": 0.010498046875, "ewc_loss_diag": 1.049041748046875e-05, "ewc_loss_parallel": 9.47713851928711e-06, "grad_norm": 1.2410659790039062, "learning_rate": 4.705383637134379e-06, "loss": 0.5455, "mean_token_accuracy": 0.8294072151184082, "num_tokens": 8517157.0, "step": 223 }, { "epoch": 0.028495102404274265, "ewc_loss": 0.01031494140625, "ewc_loss_diag": 1.049041748046875e-05, "ewc_loss_parallel": 9.298324584960938e-06, "grad_norm": 1.2388724088668823, "learning_rate": 4.7265790589232725e-06, "loss": 0.5137, "mean_token_accuracy": 0.8383884429931641, "num_tokens": 8556677.0, "step": 224 }, { "epoch": 0.028622312682864776, "ewc_loss": 0.0106201171875, "ewc_loss_diag": 1.049041748046875e-05, "ewc_loss_parallel": 9.59634780883789e-06, "grad_norm": 1.2494444847106934, "learning_rate": 4.747774480712167e-06, "loss": 0.5507, "mean_token_accuracy": 0.8296607732772827, "num_tokens": 8594335.0, "step": 225 }, { "epoch": 0.028749522961455286, "ewc_loss": 0.0106201171875, "ewc_loss_diag": 1.055002212524414e-05, "ewc_loss_parallel": 9.59634780883789e-06, "grad_norm": 1.2607604265213013, "learning_rate": 4.7689699025010605e-06, "loss": 0.4793, "mean_token_accuracy": 0.8480616211891174, "num_tokens": 8626259.0, "step": 226 }, { "epoch": 0.028876733240045797, "ewc_loss": 0.01080322265625, "ewc_loss_diag": 1.055002212524414e-05, "ewc_loss_parallel": 9.775161743164062e-06, "grad_norm": 1.2820478677749634, "learning_rate": 4.790165324289954e-06, "loss": 0.626, "mean_token_accuracy": 0.8085088133811951, "num_tokens": 8665481.0, "step": 227 }, { "epoch": 0.029003943518636305, "ewc_loss": 0.0108642578125, "ewc_loss_diag": 1.0669231414794922e-05, "ewc_loss_parallel": 9.834766387939453e-06, "grad_norm": 1.2447781562805176, "learning_rate": 4.811360746078848e-06, "loss": 0.5005, "mean_token_accuracy": 0.8446066379547119, "num_tokens": 8703136.0, "step": 228 }, { "epoch": 0.029131153797226816, "ewc_loss": 0.0107421875, "ewc_loss_diag": 1.0669231414794922e-05, "ewc_loss_parallel": 9.715557098388672e-06, "grad_norm": 1.322231650352478, "learning_rate": 4.832556167867741e-06, "loss": 0.5111, "mean_token_accuracy": 0.8383244872093201, "num_tokens": 8734989.0, "step": 229 }, { "epoch": 0.029258364075817327, "ewc_loss": 0.01104736328125, "ewc_loss_diag": 1.0728836059570312e-05, "ewc_loss_parallel": 9.953975677490234e-06, "grad_norm": 1.2631207704544067, "learning_rate": 4.853751589656634e-06, "loss": 0.4875, "mean_token_accuracy": 0.8462427258491516, "num_tokens": 8768145.0, "step": 230 }, { "epoch": 0.029385574354407838, "ewc_loss": 0.0108642578125, "ewc_loss_diag": 1.0788440704345703e-05, "ewc_loss_parallel": 9.775161743164062e-06, "grad_norm": 1.3224139213562012, "learning_rate": 4.8749470114455275e-06, "loss": 0.5351, "mean_token_accuracy": 0.8331634998321533, "num_tokens": 8804668.0, "step": 231 }, { "epoch": 0.029512784632998345, "ewc_loss": 0.0111083984375, "ewc_loss_diag": 1.0848045349121094e-05, "ewc_loss_parallel": 1.0013580322265625e-05, "grad_norm": 1.251139760017395, "learning_rate": 4.896142433234421e-06, "loss": 0.5589, "mean_token_accuracy": 0.8215550184249878, "num_tokens": 8848343.0, "step": 232 }, { "epoch": 0.029639994911588856, "ewc_loss": 0.01104736328125, "ewc_loss_diag": 1.0848045349121094e-05, "ewc_loss_parallel": 9.953975677490234e-06, "grad_norm": 1.2827281951904297, "learning_rate": 4.917337855023315e-06, "loss": 0.5228, "mean_token_accuracy": 0.8342278003692627, "num_tokens": 8883794.0, "step": 233 }, { "epoch": 0.029767205190179367, "ewc_loss": 0.01116943359375, "ewc_loss_diag": 1.0907649993896484e-05, "ewc_loss_parallel": 1.0073184967041016e-05, "grad_norm": 1.2747321128845215, "learning_rate": 4.938533276812208e-06, "loss": 0.5414, "mean_token_accuracy": 0.8291553854942322, "num_tokens": 8921597.0, "step": 234 }, { "epoch": 0.029894415468769878, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.0907649993896484e-05, "ewc_loss_parallel": 1.0132789611816406e-05, "grad_norm": 1.3510808944702148, "learning_rate": 4.959728698601102e-06, "loss": 0.554, "mean_token_accuracy": 0.826665461063385, "num_tokens": 8956260.0, "step": 235 }, { "epoch": 0.030021625747360385, "ewc_loss": 0.01141357421875, "ewc_loss_diag": 1.1026859283447266e-05, "ewc_loss_parallel": 1.0311603546142578e-05, "grad_norm": 1.3332104682922363, "learning_rate": 4.980924120389996e-06, "loss": 0.631, "mean_token_accuracy": 0.8110279440879822, "num_tokens": 8992492.0, "step": 236 }, { "epoch": 0.030148836025950896, "ewc_loss": 0.01129150390625, "ewc_loss_diag": 1.1146068572998047e-05, "ewc_loss_parallel": 1.0192394256591797e-05, "grad_norm": 1.2973769903182983, "learning_rate": 5.00211954217889e-06, "loss": 0.5388, "mean_token_accuracy": 0.8318448066711426, "num_tokens": 9028300.0, "step": 237 }, { "epoch": 0.030276046304541407, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.1205673217773438e-05, "ewc_loss_parallel": 1.0132789611816406e-05, "grad_norm": 1.2693312168121338, "learning_rate": 5.0233149639677834e-06, "loss": 0.528, "mean_token_accuracy": 0.8354594707489014, "num_tokens": 9066620.0, "step": 238 }, { "epoch": 0.030403256583131918, "ewc_loss": 0.01141357421875, "ewc_loss_diag": 1.1205673217773438e-05, "ewc_loss_parallel": 1.0311603546142578e-05, "grad_norm": 1.2754271030426025, "learning_rate": 5.044510385756677e-06, "loss": 0.4813, "mean_token_accuracy": 0.8481278419494629, "num_tokens": 9105213.0, "step": 239 }, { "epoch": 0.03053046686172243, "ewc_loss": 0.01153564453125, "ewc_loss_diag": 1.1205673217773438e-05, "ewc_loss_parallel": 1.043081283569336e-05, "grad_norm": 1.3148870468139648, "learning_rate": 5.065705807545571e-06, "loss": 0.5444, "mean_token_accuracy": 0.8297669887542725, "num_tokens": 9139673.0, "step": 240 }, { "epoch": 0.030657677140312936, "ewc_loss": 0.01171875, "ewc_loss_diag": 1.1265277862548828e-05, "ewc_loss_parallel": 1.055002212524414e-05, "grad_norm": 1.3025459051132202, "learning_rate": 5.086901229334464e-06, "loss": 0.5429, "mean_token_accuracy": 0.8284420371055603, "num_tokens": 9172435.0, "step": 241 }, { "epoch": 0.030784887418903447, "ewc_loss": 0.0118408203125, "ewc_loss_diag": 1.1265277862548828e-05, "ewc_loss_parallel": 1.0669231414794922e-05, "grad_norm": 1.3173580169677734, "learning_rate": 5.108096651123358e-06, "loss": 0.5609, "mean_token_accuracy": 0.8219592571258545, "num_tokens": 9206015.0, "step": 242 }, { "epoch": 0.030912097697493958, "ewc_loss": 0.0118408203125, "ewc_loss_diag": 1.1324882507324219e-05, "ewc_loss_parallel": 1.0728836059570312e-05, "grad_norm": 1.2741446495056152, "learning_rate": 5.129292072912251e-06, "loss": 0.5044, "mean_token_accuracy": 0.8421614170074463, "num_tokens": 9247224.0, "step": 243 }, { "epoch": 0.03103930797608447, "ewc_loss": 0.0118408203125, "ewc_loss_diag": 1.138448715209961e-05, "ewc_loss_parallel": 1.0669231414794922e-05, "grad_norm": 1.3037207126617432, "learning_rate": 5.150487494701145e-06, "loss": 0.4833, "mean_token_accuracy": 0.8473750948905945, "num_tokens": 9282171.0, "step": 244 }, { "epoch": 0.031166518254674976, "ewc_loss": 0.01214599609375, "ewc_loss_diag": 1.138448715209961e-05, "ewc_loss_parallel": 1.0967254638671875e-05, "grad_norm": 1.281870722770691, "learning_rate": 5.1716829164900385e-06, "loss": 0.5372, "mean_token_accuracy": 0.832693338394165, "num_tokens": 9324078.0, "step": 245 }, { "epoch": 0.03129372853326549, "ewc_loss": 0.0120849609375, "ewc_loss_diag": 1.138448715209961e-05, "ewc_loss_parallel": 1.0907649993896484e-05, "grad_norm": 1.308188557624817, "learning_rate": 5.192878338278932e-06, "loss": 0.5359, "mean_token_accuracy": 0.8338114619255066, "num_tokens": 9364516.0, "step": 246 }, { "epoch": 0.031420938811855995, "ewc_loss": 0.01220703125, "ewc_loss_diag": 1.150369644165039e-05, "ewc_loss_parallel": 1.1026859283447266e-05, "grad_norm": 1.3106383085250854, "learning_rate": 5.214073760067826e-06, "loss": 0.5138, "mean_token_accuracy": 0.8385791778564453, "num_tokens": 9403804.0, "step": 247 }, { "epoch": 0.03154814909044651, "ewc_loss": 0.01226806640625, "ewc_loss_diag": 1.1622905731201172e-05, "ewc_loss_parallel": 1.1086463928222656e-05, "grad_norm": 1.3247050046920776, "learning_rate": 5.235269181856719e-06, "loss": 0.557, "mean_token_accuracy": 0.8239167332649231, "num_tokens": 9441589.0, "step": 248 }, { "epoch": 0.031675359369037016, "ewc_loss": 0.0123291015625, "ewc_loss_diag": 1.1682510375976562e-05, "ewc_loss_parallel": 1.1146068572998047e-05, "grad_norm": 1.3249754905700684, "learning_rate": 5.256464603645613e-06, "loss": 0.5567, "mean_token_accuracy": 0.8261878490447998, "num_tokens": 9480240.0, "step": 249 }, { "epoch": 0.03180256964762753, "ewc_loss": 0.0123291015625, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 1.1146068572998047e-05, "grad_norm": 1.3499834537506104, "learning_rate": 5.277660025434506e-06, "loss": 0.5154, "mean_token_accuracy": 0.8343110084533691, "num_tokens": 9512704.0, "step": 250 }, { "epoch": 0.03192977992621804, "ewc_loss": 0.0125732421875, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 1.138448715209961e-05, "grad_norm": 1.3795902729034424, "learning_rate": 5.2988554472234e-06, "loss": 0.5744, "mean_token_accuracy": 0.8185173273086548, "num_tokens": 9545057.0, "step": 251 }, { "epoch": 0.032056990204808546, "ewc_loss": 0.01251220703125, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 1.1324882507324219e-05, "grad_norm": 1.3162647485733032, "learning_rate": 5.3200508690122935e-06, "loss": 0.5238, "mean_token_accuracy": 0.8360809087753296, "num_tokens": 9585797.0, "step": 252 }, { "epoch": 0.03218420048339906, "ewc_loss": 0.01239013671875, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 1.1205673217773438e-05, "grad_norm": 1.3413009643554688, "learning_rate": 5.341246290801187e-06, "loss": 0.4801, "mean_token_accuracy": 0.8487274050712585, "num_tokens": 9617890.0, "step": 253 }, { "epoch": 0.03231141076198957, "ewc_loss": 0.01251220703125, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 1.1324882507324219e-05, "grad_norm": 1.3206498622894287, "learning_rate": 5.362441712590081e-06, "loss": 0.4825, "mean_token_accuracy": 0.8456436395645142, "num_tokens": 9652895.0, "step": 254 }, { "epoch": 0.03243862104058008, "ewc_loss": 0.0126953125, "ewc_loss_diag": 1.1980533599853516e-05, "ewc_loss_parallel": 1.1444091796875e-05, "grad_norm": 1.345096230506897, "learning_rate": 5.383637134378974e-06, "loss": 0.5404, "mean_token_accuracy": 0.8289921283721924, "num_tokens": 9691804.0, "step": 255 }, { "epoch": 0.03256583131917059, "ewc_loss": 0.01263427734375, "ewc_loss_diag": 1.1980533599853516e-05, "ewc_loss_parallel": 1.138448715209961e-05, "grad_norm": 1.3459738492965698, "learning_rate": 5.404832556167868e-06, "loss": 0.5617, "mean_token_accuracy": 0.8267815113067627, "num_tokens": 9736178.0, "step": 256 }, { "epoch": 0.0326930415977611, "ewc_loss": 0.0126953125, "ewc_loss_diag": 1.2040138244628906e-05, "ewc_loss_parallel": 1.150369644165039e-05, "grad_norm": 1.3395640850067139, "learning_rate": 5.426027977956761e-06, "loss": 0.458, "mean_token_accuracy": 0.8552752733230591, "num_tokens": 9773867.0, "step": 257 }, { "epoch": 0.03282025187635161, "ewc_loss": 0.0126953125, "ewc_loss_diag": 1.2159347534179688e-05, "ewc_loss_parallel": 1.150369644165039e-05, "grad_norm": 1.335180401802063, "learning_rate": 5.447223399745655e-06, "loss": 0.5604, "mean_token_accuracy": 0.8261077404022217, "num_tokens": 9813764.0, "step": 258 }, { "epoch": 0.03294746215494212, "ewc_loss": 0.0126953125, "ewc_loss_diag": 1.2218952178955078e-05, "ewc_loss_parallel": 1.1444091796875e-05, "grad_norm": 1.3260794878005981, "learning_rate": 5.4684188215345485e-06, "loss": 0.5197, "mean_token_accuracy": 0.8369706273078918, "num_tokens": 9855447.0, "step": 259 }, { "epoch": 0.03307467243353263, "ewc_loss": 0.0128173828125, "ewc_loss_diag": 1.2278556823730469e-05, "ewc_loss_parallel": 1.1622905731201172e-05, "grad_norm": 1.3576922416687012, "learning_rate": 5.489614243323442e-06, "loss": 0.4824, "mean_token_accuracy": 0.8499454855918884, "num_tokens": 9889806.0, "step": 260 }, { "epoch": 0.03320188271212314, "ewc_loss": 0.012939453125, "ewc_loss_diag": 1.233816146850586e-05, "ewc_loss_parallel": 1.1742115020751953e-05, "grad_norm": 1.3575104475021362, "learning_rate": 5.5108096651123365e-06, "loss": 0.6148, "mean_token_accuracy": 0.8094427585601807, "num_tokens": 9929257.0, "step": 261 }, { "epoch": 0.03332909299071365, "ewc_loss": 0.012939453125, "ewc_loss_diag": 1.233816146850586e-05, "ewc_loss_parallel": 1.1742115020751953e-05, "grad_norm": 1.3454554080963135, "learning_rate": 5.53200508690123e-06, "loss": 0.5403, "mean_token_accuracy": 0.828926682472229, "num_tokens": 9967822.0, "step": 262 }, { "epoch": 0.03345630326930416, "ewc_loss": 0.01300048828125, "ewc_loss_diag": 1.233816146850586e-05, "ewc_loss_parallel": 1.1801719665527344e-05, "grad_norm": 1.3380929231643677, "learning_rate": 5.553200508690124e-06, "loss": 0.4814, "mean_token_accuracy": 0.8489738702774048, "num_tokens": 10004873.0, "step": 263 }, { "epoch": 0.03358351354789467, "ewc_loss": 0.01300048828125, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 1.1801719665527344e-05, "grad_norm": 1.4132871627807617, "learning_rate": 5.574395930479017e-06, "loss": 0.5606, "mean_token_accuracy": 0.8232994079589844, "num_tokens": 10033761.0, "step": 264 }, { "epoch": 0.03371072382648518, "ewc_loss": 0.01336669921875, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 1.2159347534179688e-05, "grad_norm": 1.3520903587341309, "learning_rate": 5.595591352267911e-06, "loss": 0.5481, "mean_token_accuracy": 0.8269220590591431, "num_tokens": 10072586.0, "step": 265 }, { "epoch": 0.03383793410507569, "ewc_loss": 0.0133056640625, "ewc_loss_diag": 1.2576580047607422e-05, "ewc_loss_parallel": 1.2040138244628906e-05, "grad_norm": 1.3618693351745605, "learning_rate": 5.616786774056804e-06, "loss": 0.5402, "mean_token_accuracy": 0.8313953876495361, "num_tokens": 10112210.0, "step": 266 }, { "epoch": 0.0339651443836662, "ewc_loss": 0.01336669921875, "ewc_loss_diag": 1.2636184692382812e-05, "ewc_loss_parallel": 1.2099742889404297e-05, "grad_norm": 1.345933437347412, "learning_rate": 5.637982195845697e-06, "loss": 0.5253, "mean_token_accuracy": 0.8367031812667847, "num_tokens": 10155908.0, "step": 267 }, { "epoch": 0.03409235466225671, "ewc_loss": 0.013671875, "ewc_loss_diag": 1.2695789337158203e-05, "ewc_loss_parallel": 1.239776611328125e-05, "grad_norm": 1.3800030946731567, "learning_rate": 5.659177617634591e-06, "loss": 0.5806, "mean_token_accuracy": 0.8171882629394531, "num_tokens": 10195944.0, "step": 268 }, { "epoch": 0.03421956494084722, "ewc_loss": 0.01385498046875, "ewc_loss_diag": 1.2874603271484375e-05, "ewc_loss_parallel": 1.2576580047607422e-05, "grad_norm": 1.369752049446106, "learning_rate": 5.680373039423484e-06, "loss": 0.5681, "mean_token_accuracy": 0.8220208287239075, "num_tokens": 10233602.0, "step": 269 }, { "epoch": 0.03434677521943773, "ewc_loss": 0.01385498046875, "ewc_loss_diag": 1.2874603271484375e-05, "ewc_loss_parallel": 1.2576580047607422e-05, "grad_norm": 1.3932791948318481, "learning_rate": 5.701568461212378e-06, "loss": 0.518, "mean_token_accuracy": 0.8364863991737366, "num_tokens": 10267001.0, "step": 270 }, { "epoch": 0.03447398549802824, "ewc_loss": 0.01397705078125, "ewc_loss_diag": 1.2993812561035156e-05, "ewc_loss_parallel": 1.2695789337158203e-05, "grad_norm": 1.3673646450042725, "learning_rate": 5.7227638830012714e-06, "loss": 0.5263, "mean_token_accuracy": 0.8316179513931274, "num_tokens": 10308264.0, "step": 271 }, { "epoch": 0.03460119577661875, "ewc_loss": 0.01397705078125, "ewc_loss_diag": 1.2993812561035156e-05, "ewc_loss_parallel": 1.2695789337158203e-05, "grad_norm": 1.421981692314148, "learning_rate": 5.743959304790166e-06, "loss": 0.6237, "mean_token_accuracy": 0.805856466293335, "num_tokens": 10343440.0, "step": 272 }, { "epoch": 0.034728406055209264, "ewc_loss": 0.01397705078125, "ewc_loss_diag": 1.2993812561035156e-05, "ewc_loss_parallel": 1.2695789337158203e-05, "grad_norm": 1.3896785974502563, "learning_rate": 5.7651547265790594e-06, "loss": 0.5584, "mean_token_accuracy": 0.8278908729553223, "num_tokens": 10383402.0, "step": 273 }, { "epoch": 0.03485561633379977, "ewc_loss": 0.0140380859375, "ewc_loss_diag": 1.3113021850585938e-05, "ewc_loss_parallel": 1.2755393981933594e-05, "grad_norm": 1.3939214944839478, "learning_rate": 5.786350148367953e-06, "loss": 0.5319, "mean_token_accuracy": 0.8315196633338928, "num_tokens": 10421700.0, "step": 274 }, { "epoch": 0.03498282661239028, "ewc_loss": 0.01409912109375, "ewc_loss_diag": 1.3172626495361328e-05, "ewc_loss_parallel": 1.2755393981933594e-05, "grad_norm": 1.395064115524292, "learning_rate": 5.807545570156847e-06, "loss": 0.5709, "mean_token_accuracy": 0.82318115234375, "num_tokens": 10460295.0, "step": 275 }, { "epoch": 0.03511003689098079, "ewc_loss": 0.01434326171875, "ewc_loss_diag": 1.3232231140136719e-05, "ewc_loss_parallel": 1.2993812561035156e-05, "grad_norm": 1.3880956172943115, "learning_rate": 5.82874099194574e-06, "loss": 0.5221, "mean_token_accuracy": 0.835822343826294, "num_tokens": 10501619.0, "step": 276 }, { "epoch": 0.0352372471695713, "ewc_loss": 0.01434326171875, "ewc_loss_diag": 1.33514404296875e-05, "ewc_loss_parallel": 1.2993812561035156e-05, "grad_norm": 1.403422474861145, "learning_rate": 5.849936413734634e-06, "loss": 0.5254, "mean_token_accuracy": 0.835223913192749, "num_tokens": 10540465.0, "step": 277 }, { "epoch": 0.03536445744816181, "ewc_loss": 0.014404296875, "ewc_loss_diag": 1.341104507446289e-05, "ewc_loss_parallel": 1.3053417205810547e-05, "grad_norm": 1.4013409614562988, "learning_rate": 5.871131835523527e-06, "loss": 0.5289, "mean_token_accuracy": 0.8336309790611267, "num_tokens": 10580980.0, "step": 278 }, { "epoch": 0.03549166772675232, "ewc_loss": 0.014404296875, "ewc_loss_diag": 1.341104507446289e-05, "ewc_loss_parallel": 1.3053417205810547e-05, "grad_norm": 1.5901734828948975, "learning_rate": 5.892327257312421e-06, "loss": 0.6065, "mean_token_accuracy": 0.8099920153617859, "num_tokens": 10621332.0, "step": 279 }, { "epoch": 0.03561887800534283, "ewc_loss": 0.01495361328125, "ewc_loss_diag": 1.3470649719238281e-05, "ewc_loss_parallel": 1.3589859008789062e-05, "grad_norm": 1.4105514287948608, "learning_rate": 5.9135226791013145e-06, "loss": 0.5557, "mean_token_accuracy": 0.8303256630897522, "num_tokens": 10654468.0, "step": 280 }, { "epoch": 0.035746088283933344, "ewc_loss": 0.0142822265625, "ewc_loss_diag": 1.3589859008789062e-05, "ewc_loss_parallel": 1.2934207916259766e-05, "grad_norm": 1.4088040590286255, "learning_rate": 5.934718100890208e-06, "loss": 0.4787, "mean_token_accuracy": 0.8502947688102722, "num_tokens": 10692191.0, "step": 281 }, { "epoch": 0.03587329856252385, "ewc_loss": 0.01446533203125, "ewc_loss_diag": 1.3589859008789062e-05, "ewc_loss_parallel": 1.3113021850585938e-05, "grad_norm": 1.3902554512023926, "learning_rate": 5.955913522679102e-06, "loss": 0.5231, "mean_token_accuracy": 0.8401461243629456, "num_tokens": 10731568.0, "step": 282 }, { "epoch": 0.03600050884111436, "ewc_loss": 0.0147705078125, "ewc_loss_diag": 1.3589859008789062e-05, "ewc_loss_parallel": 1.341104507446289e-05, "grad_norm": 1.4329144954681396, "learning_rate": 5.977108944467995e-06, "loss": 0.5854, "mean_token_accuracy": 0.8206642866134644, "num_tokens": 10772166.0, "step": 283 }, { "epoch": 0.036127719119704874, "ewc_loss": 0.0147705078125, "ewc_loss_diag": 1.3709068298339844e-05, "ewc_loss_parallel": 1.341104507446289e-05, "grad_norm": 1.4081830978393555, "learning_rate": 5.998304366256889e-06, "loss": 0.5319, "mean_token_accuracy": 0.8355863094329834, "num_tokens": 10807527.0, "step": 284 }, { "epoch": 0.03625492939829538, "ewc_loss": 0.014892578125, "ewc_loss_diag": 1.3709068298339844e-05, "ewc_loss_parallel": 1.3530254364013672e-05, "grad_norm": 1.4398921728134155, "learning_rate": 6.019499788045782e-06, "loss": 0.5034, "mean_token_accuracy": 0.8400826454162598, "num_tokens": 10846064.0, "step": 285 }, { "epoch": 0.036382139676885895, "ewc_loss": 0.01495361328125, "ewc_loss_diag": 1.3828277587890625e-05, "ewc_loss_parallel": 1.3530254364013672e-05, "grad_norm": 1.3911033868789673, "learning_rate": 6.040695209834676e-06, "loss": 0.4843, "mean_token_accuracy": 0.8524522185325623, "num_tokens": 10884496.0, "step": 286 }, { "epoch": 0.0365093499554764, "ewc_loss": 0.01495361328125, "ewc_loss_diag": 1.3887882232666016e-05, "ewc_loss_parallel": 1.3530254364013672e-05, "grad_norm": 1.453841209411621, "learning_rate": 6.0618906316235695e-06, "loss": 0.5246, "mean_token_accuracy": 0.8351706266403198, "num_tokens": 10917542.0, "step": 287 }, { "epoch": 0.03663656023406691, "ewc_loss": 0.0150146484375, "ewc_loss_diag": 1.4007091522216797e-05, "ewc_loss_parallel": 1.3589859008789062e-05, "grad_norm": 1.4540331363677979, "learning_rate": 6.083086053412463e-06, "loss": 0.5846, "mean_token_accuracy": 0.8236963748931885, "num_tokens": 10949596.0, "step": 288 }, { "epoch": 0.036763770512657425, "ewc_loss": 0.01507568359375, "ewc_loss_diag": 1.4007091522216797e-05, "ewc_loss_parallel": 1.3649463653564453e-05, "grad_norm": 1.4278481006622314, "learning_rate": 6.104281475201357e-06, "loss": 0.4989, "mean_token_accuracy": 0.8410958647727966, "num_tokens": 10986143.0, "step": 289 }, { "epoch": 0.03689098079124793, "ewc_loss": 0.01519775390625, "ewc_loss_diag": 1.4066696166992188e-05, "ewc_loss_parallel": 1.3768672943115234e-05, "grad_norm": 1.4291729927062988, "learning_rate": 6.12547689699025e-06, "loss": 0.536, "mean_token_accuracy": 0.8305087685585022, "num_tokens": 11024678.0, "step": 290 }, { "epoch": 0.03701819106983844, "ewc_loss": 0.0152587890625, "ewc_loss_diag": 1.4066696166992188e-05, "ewc_loss_parallel": 1.3828277587890625e-05, "grad_norm": 1.4209529161453247, "learning_rate": 6.146672318779144e-06, "loss": 0.5551, "mean_token_accuracy": 0.8272398710250854, "num_tokens": 11068950.0, "step": 291 }, { "epoch": 0.037145401348428954, "ewc_loss": 0.015380859375, "ewc_loss_diag": 1.424551010131836e-05, "ewc_loss_parallel": 1.3947486877441406e-05, "grad_norm": 1.4330754280090332, "learning_rate": 6.167867740568037e-06, "loss": 0.502, "mean_token_accuracy": 0.8386868238449097, "num_tokens": 11108697.0, "step": 292 }, { "epoch": 0.03727261162701946, "ewc_loss": 0.01544189453125, "ewc_loss_diag": 1.424551010131836e-05, "ewc_loss_parallel": 1.4066696166992188e-05, "grad_norm": 1.42387855052948, "learning_rate": 6.189063162356931e-06, "loss": 0.5195, "mean_token_accuracy": 0.8409460783004761, "num_tokens": 11148482.0, "step": 293 }, { "epoch": 0.037399821905609976, "ewc_loss": 0.015380859375, "ewc_loss_diag": 1.436471939086914e-05, "ewc_loss_parallel": 1.4007091522216797e-05, "grad_norm": 1.4302659034729004, "learning_rate": 6.2102585841458245e-06, "loss": 0.5297, "mean_token_accuracy": 0.8301655054092407, "num_tokens": 11189139.0, "step": 294 }, { "epoch": 0.03752703218420048, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.4424324035644531e-05, "ewc_loss_parallel": 1.436471939086914e-05, "grad_norm": 1.4566603899002075, "learning_rate": 6.231454005934718e-06, "loss": 0.52, "mean_token_accuracy": 0.8350043892860413, "num_tokens": 11226484.0, "step": 295 }, { "epoch": 0.03765424246279099, "ewc_loss": 0.015625, "ewc_loss_diag": 1.4424324035644531e-05, "ewc_loss_parallel": 1.424551010131836e-05, "grad_norm": 1.4316445589065552, "learning_rate": 6.252649427723612e-06, "loss": 0.4817, "mean_token_accuracy": 0.8478517532348633, "num_tokens": 11263806.0, "step": 296 }, { "epoch": 0.037781452741381505, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.4483928680419922e-05, "ewc_loss_parallel": 1.436471939086914e-05, "grad_norm": 1.4824230670928955, "learning_rate": 6.273844849512506e-06, "loss": 0.5937, "mean_token_accuracy": 0.8163173794746399, "num_tokens": 11299822.0, "step": 297 }, { "epoch": 0.03790866301997201, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.4543533325195312e-05, "ewc_loss_parallel": 1.430511474609375e-05, "grad_norm": 1.4935193061828613, "learning_rate": 6.295040271301399e-06, "loss": 0.5626, "mean_token_accuracy": 0.8256459832191467, "num_tokens": 11330796.0, "step": 298 }, { "epoch": 0.03803587329856253, "ewc_loss": 0.015869140625, "ewc_loss_diag": 1.4662742614746094e-05, "ewc_loss_parallel": 1.436471939086914e-05, "grad_norm": 1.4595140218734741, "learning_rate": 6.316235693090293e-06, "loss": 0.5496, "mean_token_accuracy": 0.8274595737457275, "num_tokens": 11369529.0, "step": 299 }, { "epoch": 0.038163083577153034, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.4841556549072266e-05, "ewc_loss_parallel": 1.424551010131836e-05, "grad_norm": 1.467988133430481, "learning_rate": 6.337431114879186e-06, "loss": 0.5279, "mean_token_accuracy": 0.8411991000175476, "num_tokens": 11405044.0, "step": 300 }, { "epoch": 0.03829029385574354, "ewc_loss": 0.0159912109375, "ewc_loss_diag": 1.4841556549072266e-05, "ewc_loss_parallel": 1.4543533325195312e-05, "grad_norm": 1.4481011629104614, "learning_rate": 6.35862653666808e-06, "loss": 0.522, "mean_token_accuracy": 0.8385902047157288, "num_tokens": 11445409.0, "step": 301 }, { "epoch": 0.038417504134334056, "ewc_loss": 0.0159912109375, "ewc_loss_diag": 1.4841556549072266e-05, "ewc_loss_parallel": 1.4483928680419922e-05, "grad_norm": 1.4412448406219482, "learning_rate": 6.379821958456973e-06, "loss": 0.5193, "mean_token_accuracy": 0.8376444578170776, "num_tokens": 11488285.0, "step": 302 }, { "epoch": 0.03854471441292456, "ewc_loss": 0.0159912109375, "ewc_loss_diag": 1.5020370483398438e-05, "ewc_loss_parallel": 1.4543533325195312e-05, "grad_norm": 1.4866797924041748, "learning_rate": 6.4010173802458676e-06, "loss": 0.5226, "mean_token_accuracy": 0.8357629179954529, "num_tokens": 11520970.0, "step": 303 }, { "epoch": 0.03867192469151508, "ewc_loss": 0.0159912109375, "ewc_loss_diag": 1.5020370483398438e-05, "ewc_loss_parallel": 1.4543533325195312e-05, "grad_norm": 1.4557980298995972, "learning_rate": 6.42221280203476e-06, "loss": 0.5731, "mean_token_accuracy": 0.8212007284164429, "num_tokens": 11563590.0, "step": 304 }, { "epoch": 0.038799134970105585, "ewc_loss": 0.0159912109375, "ewc_loss_diag": 1.5079975128173828e-05, "ewc_loss_parallel": 1.4543533325195312e-05, "grad_norm": 1.5174230337142944, "learning_rate": 6.443408223823655e-06, "loss": 0.6179, "mean_token_accuracy": 0.8063617944717407, "num_tokens": 11598629.0, "step": 305 }, { "epoch": 0.03892634524869609, "ewc_loss": 0.0162353515625, "ewc_loss_diag": 1.5079975128173828e-05, "ewc_loss_parallel": 1.4722347259521484e-05, "grad_norm": 1.4640471935272217, "learning_rate": 6.4646036456125474e-06, "loss": 0.522, "mean_token_accuracy": 0.8387561440467834, "num_tokens": 11635526.0, "step": 306 }, { "epoch": 0.03905355552728661, "ewc_loss": 0.01611328125, "ewc_loss_diag": 1.5139579772949219e-05, "ewc_loss_parallel": 1.4662742614746094e-05, "grad_norm": 1.4972007274627686, "learning_rate": 6.485799067401442e-06, "loss": 0.5365, "mean_token_accuracy": 0.8331388235092163, "num_tokens": 11672974.0, "step": 307 }, { "epoch": 0.039180765805877114, "ewc_loss": 0.01611328125, "ewc_loss_diag": 1.519918441772461e-05, "ewc_loss_parallel": 1.4662742614746094e-05, "grad_norm": 1.4861899614334106, "learning_rate": 6.5069944891903354e-06, "loss": 0.5594, "mean_token_accuracy": 0.8286173939704895, "num_tokens": 11708886.0, "step": 308 }, { "epoch": 0.03930797608446762, "ewc_loss": 0.016357421875, "ewc_loss_diag": 1.52587890625e-05, "ewc_loss_parallel": 1.4841556549072266e-05, "grad_norm": 1.487052321434021, "learning_rate": 6.528189910979229e-06, "loss": 0.5654, "mean_token_accuracy": 0.8292717337608337, "num_tokens": 11749217.0, "step": 309 }, { "epoch": 0.039435186363058136, "ewc_loss": 0.0162353515625, "ewc_loss_diag": 1.52587890625e-05, "ewc_loss_parallel": 1.4722347259521484e-05, "grad_norm": 1.472164511680603, "learning_rate": 6.549385332768123e-06, "loss": 0.5318, "mean_token_accuracy": 0.8361539840698242, "num_tokens": 11788380.0, "step": 310 }, { "epoch": 0.039562396641648644, "ewc_loss": 0.016357421875, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.4841556549072266e-05, "grad_norm": 1.5088038444519043, "learning_rate": 6.570580754557017e-06, "loss": 0.5001, "mean_token_accuracy": 0.8435973525047302, "num_tokens": 11820587.0, "step": 311 }, { "epoch": 0.03968960692023916, "ewc_loss": 0.0164794921875, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.4892499446868896, "learning_rate": 6.59177617634591e-06, "loss": 0.5089, "mean_token_accuracy": 0.8407118320465088, "num_tokens": 11854473.0, "step": 312 }, { "epoch": 0.039816817198829665, "ewc_loss": 0.0164794921875, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.499858021736145, "learning_rate": 6.6129715981348025e-06, "loss": 0.5497, "mean_token_accuracy": 0.832945704460144, "num_tokens": 11894744.0, "step": 313 }, { "epoch": 0.03994402747742017, "ewc_loss": 0.0166015625, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.5020370483398438e-05, "grad_norm": 1.4688067436218262, "learning_rate": 6.634167019923697e-06, "loss": 0.5057, "mean_token_accuracy": 0.8407015204429626, "num_tokens": 11932346.0, "step": 314 }, { "epoch": 0.04007123775601069, "ewc_loss": 0.0167236328125, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.5139579772949219e-05, "grad_norm": 1.5355583429336548, "learning_rate": 6.65536244171259e-06, "loss": 0.5653, "mean_token_accuracy": 0.8216589689254761, "num_tokens": 11967683.0, "step": 315 }, { "epoch": 0.040198448034601195, "ewc_loss": 0.0167236328125, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.519918441772461e-05, "grad_norm": 1.462282419204712, "learning_rate": 6.676557863501484e-06, "loss": 0.5517, "mean_token_accuracy": 0.8259767293930054, "num_tokens": 12010521.0, "step": 316 }, { "epoch": 0.04032565831319171, "ewc_loss": 0.016845703125, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.52587890625e-05, "grad_norm": 1.5126181840896606, "learning_rate": 6.697753285290377e-06, "loss": 0.5282, "mean_token_accuracy": 0.8357778191566467, "num_tokens": 12053317.0, "step": 317 }, { "epoch": 0.040452868591782216, "ewc_loss": 0.016845703125, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.52587890625e-05, "grad_norm": 1.5464491844177246, "learning_rate": 6.718948707079271e-06, "loss": 0.5796, "mean_token_accuracy": 0.8183053731918335, "num_tokens": 12081589.0, "step": 318 }, { "epoch": 0.040580078870372724, "ewc_loss": 0.0169677734375, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.537799835205078e-05, "grad_norm": 1.5270969867706299, "learning_rate": 6.740144128868165e-06, "loss": 0.5199, "mean_token_accuracy": 0.8360527157783508, "num_tokens": 12114826.0, "step": 319 }, { "epoch": 0.04070728914896324, "ewc_loss": 0.016845703125, "ewc_loss_diag": 1.5616416931152344e-05, "ewc_loss_parallel": 1.52587890625e-05, "grad_norm": 1.5091358423233032, "learning_rate": 6.761339550657058e-06, "loss": 0.5065, "mean_token_accuracy": 0.8414556980133057, "num_tokens": 12150769.0, "step": 320 }, { "epoch": 0.040834499427553746, "ewc_loss": 0.016845703125, "ewc_loss_diag": 1.5735626220703125e-05, "ewc_loss_parallel": 1.52587890625e-05, "grad_norm": 1.5026952028274536, "learning_rate": 6.782534972445952e-06, "loss": 0.5265, "mean_token_accuracy": 0.8368262052536011, "num_tokens": 12189035.0, "step": 321 }, { "epoch": 0.04096170970614425, "ewc_loss": 0.0169677734375, "ewc_loss_diag": 1.5735626220703125e-05, "ewc_loss_parallel": 1.537799835205078e-05, "grad_norm": 1.5199980735778809, "learning_rate": 6.803730394234846e-06, "loss": 0.5126, "mean_token_accuracy": 0.837785005569458, "num_tokens": 12225974.0, "step": 322 }, { "epoch": 0.04108891998473477, "ewc_loss": 0.0172119140625, "ewc_loss_diag": 1.5735626220703125e-05, "ewc_loss_parallel": 1.5616416931152344e-05, "grad_norm": 1.5169214010238647, "learning_rate": 6.824925816023739e-06, "loss": 0.5845, "mean_token_accuracy": 0.8203893899917603, "num_tokens": 12266971.0, "step": 323 }, { "epoch": 0.041216130263325275, "ewc_loss": 0.0172119140625, "ewc_loss_diag": 1.5735626220703125e-05, "ewc_loss_parallel": 1.5616416931152344e-05, "grad_norm": 1.524245023727417, "learning_rate": 6.8461212378126335e-06, "loss": 0.501, "mean_token_accuracy": 0.8485267758369446, "num_tokens": 12306365.0, "step": 324 }, { "epoch": 0.04134334054191579, "ewc_loss": 0.01708984375, "ewc_loss_diag": 1.5735626220703125e-05, "ewc_loss_parallel": 1.5497207641601562e-05, "grad_norm": 1.5264586210250854, "learning_rate": 6.867316659601526e-06, "loss": 0.55, "mean_token_accuracy": 0.8372502326965332, "num_tokens": 12345343.0, "step": 325 }, { "epoch": 0.0414705508205063, "ewc_loss": 0.0172119140625, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.5616416931152344e-05, "grad_norm": 1.5485173463821411, "learning_rate": 6.888512081390421e-06, "loss": 0.6146, "mean_token_accuracy": 0.8111693859100342, "num_tokens": 12384532.0, "step": 326 }, { "epoch": 0.041597761099096804, "ewc_loss": 0.017333984375, "ewc_loss_diag": 1.5974044799804688e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.5067827701568604, "learning_rate": 6.909707503179313e-06, "loss": 0.4761, "mean_token_accuracy": 0.8510562181472778, "num_tokens": 12424955.0, "step": 327 }, { "epoch": 0.04172497137768732, "ewc_loss": 0.017333984375, "ewc_loss_diag": 1.5974044799804688e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.5089768171310425, "learning_rate": 6.930902924968208e-06, "loss": 0.496, "mean_token_accuracy": 0.8489694595336914, "num_tokens": 12469583.0, "step": 328 }, { "epoch": 0.041852181656277826, "ewc_loss": 0.017333984375, "ewc_loss_diag": 1.5974044799804688e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.5503695011138916, "learning_rate": 6.9520983467571005e-06, "loss": 0.5457, "mean_token_accuracy": 0.8265141248703003, "num_tokens": 12504997.0, "step": 329 }, { "epoch": 0.04197939193486834, "ewc_loss": 0.0174560546875, "ewc_loss_diag": 1.609325408935547e-05, "ewc_loss_parallel": 1.5854835510253906e-05, "grad_norm": 1.5256750583648682, "learning_rate": 6.973293768545995e-06, "loss": 0.5469, "mean_token_accuracy": 0.8327999711036682, "num_tokens": 12546248.0, "step": 330 }, { "epoch": 0.04210660221345885, "ewc_loss": 0.0174560546875, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.5854835510253906e-05, "grad_norm": 1.5690386295318604, "learning_rate": 6.994489190334888e-06, "loss": 0.5637, "mean_token_accuracy": 0.8260197639465332, "num_tokens": 12580741.0, "step": 331 }, { "epoch": 0.042233812492049355, "ewc_loss": 0.0174560546875, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.5854835510253906e-05, "grad_norm": 1.5240201950073242, "learning_rate": 7.015684612123781e-06, "loss": 0.5742, "mean_token_accuracy": 0.8242843747138977, "num_tokens": 12622539.0, "step": 332 }, { "epoch": 0.04236102277063987, "ewc_loss": 0.0174560546875, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.5854835510253906e-05, "grad_norm": 1.5587711334228516, "learning_rate": 7.036880033912676e-06, "loss": 0.5274, "mean_token_accuracy": 0.8349636793136597, "num_tokens": 12660137.0, "step": 333 }, { "epoch": 0.04248823304923038, "ewc_loss": 0.0174560546875, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.5854835510253906e-05, "grad_norm": 1.522378921508789, "learning_rate": 7.058075455701568e-06, "loss": 0.5542, "mean_token_accuracy": 0.8293956518173218, "num_tokens": 12702341.0, "step": 334 }, { "epoch": 0.04261544332782089, "ewc_loss": 0.017578125, "ewc_loss_diag": 1.633167266845703e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.5732967853546143, "learning_rate": 7.079270877490463e-06, "loss": 0.4942, "mean_token_accuracy": 0.845281720161438, "num_tokens": 12739560.0, "step": 335 }, { "epoch": 0.0427426536064114, "ewc_loss": 0.017578125, "ewc_loss_diag": 1.633167266845703e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.5188084840774536, "learning_rate": 7.1004662992793556e-06, "loss": 0.5307, "mean_token_accuracy": 0.8404848575592041, "num_tokens": 12781534.0, "step": 336 }, { "epoch": 0.042869863885001906, "ewc_loss": 0.017578125, "ewc_loss_diag": 1.633167266845703e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.6089941263198853, "learning_rate": 7.12166172106825e-06, "loss": 0.5341, "mean_token_accuracy": 0.8334126472473145, "num_tokens": 12815840.0, "step": 337 }, { "epoch": 0.04299707416359242, "ewc_loss": 0.017578125, "ewc_loss_diag": 1.633167266845703e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.5459096431732178, "learning_rate": 7.142857142857143e-06, "loss": 0.5451, "mean_token_accuracy": 0.8312187194824219, "num_tokens": 12849201.0, "step": 338 }, { "epoch": 0.04312428444218293, "ewc_loss": 0.017822265625, "ewc_loss_diag": 1.6450881958007812e-05, "ewc_loss_parallel": 1.621246337890625e-05, "grad_norm": 1.5974464416503906, "learning_rate": 7.164052564646037e-06, "loss": 0.5527, "mean_token_accuracy": 0.8319306373596191, "num_tokens": 12888490.0, "step": 339 }, { "epoch": 0.043251494720773435, "ewc_loss": 0.017578125, "ewc_loss_diag": 1.6450881958007812e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.529467225074768, "learning_rate": 7.18524798643493e-06, "loss": 0.5264, "mean_token_accuracy": 0.8361747860908508, "num_tokens": 12927554.0, "step": 340 }, { "epoch": 0.04337870499936395, "ewc_loss": 0.01806640625, "ewc_loss_diag": 1.6570091247558594e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.6214464902877808, "learning_rate": 7.206443408223824e-06, "loss": 0.5667, "mean_token_accuracy": 0.8242692947387695, "num_tokens": 12960801.0, "step": 341 }, { "epoch": 0.04350591527795446, "ewc_loss": 0.017822265625, "ewc_loss_diag": 1.6570091247558594e-05, "ewc_loss_parallel": 1.609325408935547e-05, "grad_norm": 1.5611960887908936, "learning_rate": 7.227638830012717e-06, "loss": 0.4893, "mean_token_accuracy": 0.8460599780082703, "num_tokens": 12993119.0, "step": 342 }, { "epoch": 0.04363312555654497, "ewc_loss": 0.0181884765625, "ewc_loss_diag": 1.6570091247558594e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.5963290929794312, "learning_rate": 7.2488342518016114e-06, "loss": 0.5484, "mean_token_accuracy": 0.830621063709259, "num_tokens": 13033966.0, "step": 343 }, { "epoch": 0.04376033583513548, "ewc_loss": 0.017822265625, "ewc_loss_diag": 1.6570091247558594e-05, "ewc_loss_parallel": 1.609325408935547e-05, "grad_norm": 1.5914511680603027, "learning_rate": 7.270029673590505e-06, "loss": 0.5269, "mean_token_accuracy": 0.835917055606842, "num_tokens": 13063649.0, "step": 344 }, { "epoch": 0.043887546113725986, "ewc_loss": 0.01806640625, "ewc_loss_diag": 1.6570091247558594e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.5798346996307373, "learning_rate": 7.291225095379399e-06, "loss": 0.5787, "mean_token_accuracy": 0.820258617401123, "num_tokens": 13106926.0, "step": 345 }, { "epoch": 0.0440147563923165, "ewc_loss": 0.0181884765625, "ewc_loss_diag": 1.6570091247558594e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.5782890319824219, "learning_rate": 7.312420517168292e-06, "loss": 0.525, "mean_token_accuracy": 0.8376072645187378, "num_tokens": 13142961.0, "step": 346 }, { "epoch": 0.04414196667090701, "ewc_loss": 0.01806640625, "ewc_loss_diag": 1.6927719116210938e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.578997254371643, "learning_rate": 7.333615938957187e-06, "loss": 0.512, "mean_token_accuracy": 0.8404483795166016, "num_tokens": 13183544.0, "step": 347 }, { "epoch": 0.04426917694949752, "ewc_loss": 0.018310546875, "ewc_loss_diag": 1.6927719116210938e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.5790656805038452, "learning_rate": 7.354811360746079e-06, "loss": 0.5234, "mean_token_accuracy": 0.8363844752311707, "num_tokens": 13222609.0, "step": 348 }, { "epoch": 0.04439638722808803, "ewc_loss": 0.018310546875, "ewc_loss_diag": 1.6927719116210938e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.5966795682907104, "learning_rate": 7.376006782534972e-06, "loss": 0.5232, "mean_token_accuracy": 0.8397793769836426, "num_tokens": 13258267.0, "step": 349 }, { "epoch": 0.04452359750667854, "ewc_loss": 0.0181884765625, "ewc_loss_diag": 1.71661376953125e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.6015069484710693, "learning_rate": 7.3972022043238665e-06, "loss": 0.5351, "mean_token_accuracy": 0.8360491991043091, "num_tokens": 13294681.0, "step": 350 }, { "epoch": 0.04465080778526905, "ewc_loss": 0.0181884765625, "ewc_loss_diag": 1.728534698486328e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.574199914932251, "learning_rate": 7.418397626112759e-06, "loss": 0.4745, "mean_token_accuracy": 0.8501217365264893, "num_tokens": 13335469.0, "step": 351 }, { "epoch": 0.04477801806385956, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 1.728534698486328e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.600601315498352, "learning_rate": 7.439593047901654e-06, "loss": 0.4573, "mean_token_accuracy": 0.8535846471786499, "num_tokens": 13372465.0, "step": 352 }, { "epoch": 0.04490522834245007, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 1.728534698486328e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.5914828777313232, "learning_rate": 7.460788469690546e-06, "loss": 0.5065, "mean_token_accuracy": 0.8423025012016296, "num_tokens": 13409218.0, "step": 353 }, { "epoch": 0.04503243862104058, "ewc_loss": 0.0185546875, "ewc_loss_diag": 1.7404556274414062e-05, "ewc_loss_parallel": 1.6808509826660156e-05, "grad_norm": 1.5919338464736938, "learning_rate": 7.481983891479441e-06, "loss": 0.4778, "mean_token_accuracy": 0.8515503406524658, "num_tokens": 13452548.0, "step": 354 }, { "epoch": 0.04515964889963109, "ewc_loss": 0.0185546875, "ewc_loss_diag": 1.7404556274414062e-05, "ewc_loss_parallel": 1.6808509826660156e-05, "grad_norm": 1.623898983001709, "learning_rate": 7.503179313268334e-06, "loss": 0.5613, "mean_token_accuracy": 0.8283206224441528, "num_tokens": 13488325.0, "step": 355 }, { "epoch": 0.0452868591782216, "ewc_loss": 0.018310546875, "ewc_loss_diag": 1.7404556274414062e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.5997241735458374, "learning_rate": 7.524374735057228e-06, "loss": 0.5602, "mean_token_accuracy": 0.8279379606246948, "num_tokens": 13528203.0, "step": 356 }, { "epoch": 0.04541406945681211, "ewc_loss": 0.018798828125, "ewc_loss_diag": 1.7404556274414062e-05, "ewc_loss_parallel": 1.704692840576172e-05, "grad_norm": 1.6059132814407349, "learning_rate": 7.5455701568461215e-06, "loss": 0.4912, "mean_token_accuracy": 0.8462685346603394, "num_tokens": 13567242.0, "step": 357 }, { "epoch": 0.04554127973540262, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 1.7523765563964844e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.6120476722717285, "learning_rate": 7.566765578635016e-06, "loss": 0.54, "mean_token_accuracy": 0.8298465609550476, "num_tokens": 13604521.0, "step": 358 }, { "epoch": 0.04566849001399313, "ewc_loss": 0.0185546875, "ewc_loss_diag": 1.7523765563964844e-05, "ewc_loss_parallel": 1.6808509826660156e-05, "grad_norm": 1.6097619533538818, "learning_rate": 7.587961000423909e-06, "loss": 0.5019, "mean_token_accuracy": 0.8440157771110535, "num_tokens": 13642787.0, "step": 359 }, { "epoch": 0.04579570029258364, "ewc_loss": 0.0186767578125, "ewc_loss_diag": 1.7523765563964844e-05, "ewc_loss_parallel": 1.6927719116210938e-05, "grad_norm": 1.6047227382659912, "learning_rate": 7.609156422212803e-06, "loss": 0.5296, "mean_token_accuracy": 0.8305858373641968, "num_tokens": 13684113.0, "step": 360 }, { "epoch": 0.045922910571174154, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 1.7762184143066406e-05, "ewc_loss_parallel": 1.704692840576172e-05, "grad_norm": 1.6216087341308594, "learning_rate": 7.630351844001696e-06, "loss": 0.4597, "mean_token_accuracy": 0.8530616760253906, "num_tokens": 13719208.0, "step": 361 }, { "epoch": 0.04605012084976466, "ewc_loss": 0.018798828125, "ewc_loss_diag": 1.7762184143066406e-05, "ewc_loss_parallel": 1.6927719116210938e-05, "grad_norm": 1.5986863374710083, "learning_rate": 7.65154726579059e-06, "loss": 0.4791, "mean_token_accuracy": 0.8517345190048218, "num_tokens": 13760159.0, "step": 362 }, { "epoch": 0.04617733112835517, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.728534698486328e-05, "grad_norm": 1.6473315954208374, "learning_rate": 7.672742687579483e-06, "loss": 0.567, "mean_token_accuracy": 0.8248640298843384, "num_tokens": 13798375.0, "step": 363 }, { "epoch": 0.04630454140694568, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.704692840576172e-05, "grad_norm": 1.5857967138290405, "learning_rate": 7.693938109368377e-06, "loss": 0.4489, "mean_token_accuracy": 0.8606141805648804, "num_tokens": 13841924.0, "step": 364 }, { "epoch": 0.04643175168553619, "ewc_loss": 0.01953125, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.7642974853515625e-05, "grad_norm": 1.6682207584381104, "learning_rate": 7.71513353115727e-06, "loss": 0.5476, "mean_token_accuracy": 0.8314643502235413, "num_tokens": 13881436.0, "step": 365 }, { "epoch": 0.0465589619641267, "ewc_loss": 0.018798828125, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.6927719116210938e-05, "grad_norm": 1.6218136548995972, "learning_rate": 7.736328952946165e-06, "loss": 0.5653, "mean_token_accuracy": 0.8255352973937988, "num_tokens": 13916295.0, "step": 366 }, { "epoch": 0.04668617224271721, "ewc_loss": 0.0196533203125, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.7762184143066406e-05, "grad_norm": 1.6779506206512451, "learning_rate": 7.757524374735057e-06, "loss": 0.531, "mean_token_accuracy": 0.8329443335533142, "num_tokens": 13950189.0, "step": 367 }, { "epoch": 0.04681338252130772, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.704692840576172e-05, "grad_norm": 1.6120851039886475, "learning_rate": 7.77871979652395e-06, "loss": 0.5387, "mean_token_accuracy": 0.8335861563682556, "num_tokens": 13988422.0, "step": 368 }, { "epoch": 0.046940592799898234, "ewc_loss": 0.01953125, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.7642974853515625e-05, "grad_norm": 1.66026771068573, "learning_rate": 7.799915218312844e-06, "loss": 0.5363, "mean_token_accuracy": 0.8343406915664673, "num_tokens": 14029199.0, "step": 369 }, { "epoch": 0.04706780307848874, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 1.823902130126953e-05, "ewc_loss_parallel": 1.728534698486328e-05, "grad_norm": 1.621835470199585, "learning_rate": 7.821110640101737e-06, "loss": 0.5289, "mean_token_accuracy": 0.8366931676864624, "num_tokens": 14067976.0, "step": 370 }, { "epoch": 0.04719501335707925, "ewc_loss": 0.0196533203125, "ewc_loss_diag": 1.8477439880371094e-05, "ewc_loss_parallel": 1.7881393432617188e-05, "grad_norm": 1.668890357017517, "learning_rate": 7.842306061890632e-06, "loss": 0.4602, "mean_token_accuracy": 0.8577443361282349, "num_tokens": 14102858.0, "step": 371 }, { "epoch": 0.04732222363566976, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 1.8477439880371094e-05, "ewc_loss_parallel": 1.728534698486328e-05, "grad_norm": 1.6395902633666992, "learning_rate": 7.863501483679524e-06, "loss": 0.5368, "mean_token_accuracy": 0.8377188444137573, "num_tokens": 14138560.0, "step": 372 }, { "epoch": 0.04744943391426027, "ewc_loss": 0.0196533203125, "ewc_loss_diag": 1.8596649169921875e-05, "ewc_loss_parallel": 1.7881393432617188e-05, "grad_norm": 1.703538179397583, "learning_rate": 7.884696905468419e-06, "loss": 0.526, "mean_token_accuracy": 0.8359857797622681, "num_tokens": 14168007.0, "step": 373 }, { "epoch": 0.047576644192850785, "ewc_loss": 0.0194091796875, "ewc_loss_diag": 1.8596649169921875e-05, "ewc_loss_parallel": 1.7523765563964844e-05, "grad_norm": 1.6357085704803467, "learning_rate": 7.905892327257313e-06, "loss": 0.4614, "mean_token_accuracy": 0.8556482791900635, "num_tokens": 14201779.0, "step": 374 }, { "epoch": 0.04770385447144129, "ewc_loss": 0.019775390625, "ewc_loss_diag": 1.8715858459472656e-05, "ewc_loss_parallel": 1.800060272216797e-05, "grad_norm": 1.6723253726959229, "learning_rate": 7.927087749046206e-06, "loss": 0.5855, "mean_token_accuracy": 0.8180193305015564, "num_tokens": 14245917.0, "step": 375 }, { "epoch": 0.0478310647500318, "ewc_loss": 0.0196533203125, "ewc_loss_diag": 1.8715858459472656e-05, "ewc_loss_parallel": 1.7762184143066406e-05, "grad_norm": 1.643310785293579, "learning_rate": 7.9482831708351e-06, "loss": 0.5428, "mean_token_accuracy": 0.8302203416824341, "num_tokens": 14284032.0, "step": 376 }, { "epoch": 0.047958275028622314, "ewc_loss": 0.019775390625, "ewc_loss_diag": 1.8715858459472656e-05, "ewc_loss_parallel": 1.800060272216797e-05, "grad_norm": 1.7162084579467773, "learning_rate": 7.969478592623995e-06, "loss": 0.5926, "mean_token_accuracy": 0.815678060054779, "num_tokens": 14322077.0, "step": 377 }, { "epoch": 0.04808548530721282, "ewc_loss": 0.0196533203125, "ewc_loss_diag": 1.8835067749023438e-05, "ewc_loss_parallel": 1.7762184143066406e-05, "grad_norm": 1.6430552005767822, "learning_rate": 7.990674014412887e-06, "loss": 0.5438, "mean_token_accuracy": 0.8319531679153442, "num_tokens": 14358031.0, "step": 378 }, { "epoch": 0.048212695585803336, "ewc_loss": 0.02001953125, "ewc_loss_diag": 1.8835067749023438e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.7456450462341309, "learning_rate": 8.011869436201782e-06, "loss": 0.6271, "mean_token_accuracy": 0.805083155632019, "num_tokens": 14390230.0, "step": 379 }, { "epoch": 0.048339905864393844, "ewc_loss": 0.019775390625, "ewc_loss_diag": 1.895427703857422e-05, "ewc_loss_parallel": 1.7762184143066406e-05, "grad_norm": 1.627242922782898, "learning_rate": 8.033064857990675e-06, "loss": 0.4794, "mean_token_accuracy": 0.8507604002952576, "num_tokens": 14427394.0, "step": 380 }, { "epoch": 0.04846711614298435, "ewc_loss": 0.0205078125, "ewc_loss_diag": 1.895427703857422e-05, "ewc_loss_parallel": 1.8596649169921875e-05, "grad_norm": 1.7403501272201538, "learning_rate": 8.054260279779569e-06, "loss": 0.4855, "mean_token_accuracy": 0.8492889404296875, "num_tokens": 14466606.0, "step": 381 }, { "epoch": 0.048594326421574865, "ewc_loss": 0.01953125, "ewc_loss_diag": 1.895427703857422e-05, "ewc_loss_parallel": 1.7642974853515625e-05, "grad_norm": 1.6701748371124268, "learning_rate": 8.075455701568462e-06, "loss": 0.5406, "mean_token_accuracy": 0.8343910574913025, "num_tokens": 14494795.0, "step": 382 }, { "epoch": 0.04872153670016537, "ewc_loss": 0.0205078125, "ewc_loss_diag": 1.919269561767578e-05, "ewc_loss_parallel": 1.8596649169921875e-05, "grad_norm": 1.7242509126663208, "learning_rate": 8.096651123357356e-06, "loss": 0.5492, "mean_token_accuracy": 0.8304646611213684, "num_tokens": 14537400.0, "step": 383 }, { "epoch": 0.04884874697875588, "ewc_loss": 0.0198974609375, "ewc_loss_diag": 1.9311904907226562e-05, "ewc_loss_parallel": 1.800060272216797e-05, "grad_norm": 1.6726510524749756, "learning_rate": 8.117846545146249e-06, "loss": 0.5626, "mean_token_accuracy": 0.8338693380355835, "num_tokens": 14575215.0, "step": 384 }, { "epoch": 0.048975957257346395, "ewc_loss": 0.0205078125, "ewc_loss_diag": 1.919269561767578e-05, "ewc_loss_parallel": 1.8596649169921875e-05, "grad_norm": 1.7612786293029785, "learning_rate": 8.139041966935143e-06, "loss": 0.5875, "mean_token_accuracy": 0.8168537020683289, "num_tokens": 14608486.0, "step": 385 }, { "epoch": 0.0491031675359369, "ewc_loss": 0.02001953125, "ewc_loss_diag": 1.9550323486328125e-05, "ewc_loss_parallel": 1.811981201171875e-05, "grad_norm": 1.6561615467071533, "learning_rate": 8.160237388724036e-06, "loss": 0.5189, "mean_token_accuracy": 0.8396537899971008, "num_tokens": 14644583.0, "step": 386 }, { "epoch": 0.049230377814527417, "ewc_loss": 0.0206298828125, "ewc_loss_diag": 1.9669532775878906e-05, "ewc_loss_parallel": 1.8715858459472656e-05, "grad_norm": 1.7442805767059326, "learning_rate": 8.181432810512929e-06, "loss": 0.5385, "mean_token_accuracy": 0.8330336809158325, "num_tokens": 14682976.0, "step": 387 }, { "epoch": 0.049357588093117924, "ewc_loss": 0.0201416015625, "ewc_loss_diag": 1.9788742065429688e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.6787030696868896, "learning_rate": 8.202628232301823e-06, "loss": 0.5132, "mean_token_accuracy": 0.838452935218811, "num_tokens": 14717926.0, "step": 388 }, { "epoch": 0.04948479837170843, "ewc_loss": 0.0206298828125, "ewc_loss_diag": 1.9788742065429688e-05, "ewc_loss_parallel": 1.8715858459472656e-05, "grad_norm": 1.7788416147232056, "learning_rate": 8.223823654090716e-06, "loss": 0.5505, "mean_token_accuracy": 0.8288678526878357, "num_tokens": 14750484.0, "step": 389 }, { "epoch": 0.049612008650298946, "ewc_loss": 0.020263671875, "ewc_loss_diag": 1.9788742065429688e-05, "ewc_loss_parallel": 1.8358230590820312e-05, "grad_norm": 1.6613880395889282, "learning_rate": 8.24501907587961e-06, "loss": 0.572, "mean_token_accuracy": 0.8234426975250244, "num_tokens": 14790044.0, "step": 390 }, { "epoch": 0.04973921892888945, "ewc_loss": 0.0208740234375, "ewc_loss_diag": 1.990795135498047e-05, "ewc_loss_parallel": 1.895427703857422e-05, "grad_norm": 1.7660341262817383, "learning_rate": 8.266214497668503e-06, "loss": 0.4968, "mean_token_accuracy": 0.8428616523742676, "num_tokens": 14828768.0, "step": 391 }, { "epoch": 0.04986642920747997, "ewc_loss": 0.0203857421875, "ewc_loss_diag": 1.990795135498047e-05, "ewc_loss_parallel": 1.8477439880371094e-05, "grad_norm": 1.6557948589324951, "learning_rate": 8.287409919457398e-06, "loss": 0.4863, "mean_token_accuracy": 0.8519707918167114, "num_tokens": 14865848.0, "step": 392 }, { "epoch": 0.049993639486070475, "ewc_loss": 0.021484375, "ewc_loss_diag": 2.002716064453125e-05, "ewc_loss_parallel": 1.9550323486328125e-05, "grad_norm": 1.7948013544082642, "learning_rate": 8.30860534124629e-06, "loss": 0.5232, "mean_token_accuracy": 0.8353581428527832, "num_tokens": 14903973.0, "step": 393 }, { "epoch": 0.05012084976466098, "ewc_loss": 0.020751953125, "ewc_loss_diag": 2.002716064453125e-05, "ewc_loss_parallel": 1.8835067749023438e-05, "grad_norm": 1.6734634637832642, "learning_rate": 8.329800763035185e-06, "loss": 0.5551, "mean_token_accuracy": 0.8299886584281921, "num_tokens": 14944958.0, "step": 394 }, { "epoch": 0.0502480600432515, "ewc_loss": 0.021728515625, "ewc_loss_diag": 2.0265579223632812e-05, "ewc_loss_parallel": 1.9669532775878906e-05, "grad_norm": 1.8046529293060303, "learning_rate": 8.350996184824077e-06, "loss": 0.5478, "mean_token_accuracy": 0.829400360584259, "num_tokens": 14985924.0, "step": 395 }, { "epoch": 0.050375270321842004, "ewc_loss": 0.0208740234375, "ewc_loss_diag": 2.0265579223632812e-05, "ewc_loss_parallel": 1.8835067749023438e-05, "grad_norm": 1.6852518320083618, "learning_rate": 8.372191606612972e-06, "loss": 0.5136, "mean_token_accuracy": 0.8363881707191467, "num_tokens": 15022346.0, "step": 396 }, { "epoch": 0.05050248060043251, "ewc_loss": 0.0218505859375, "ewc_loss_diag": 2.0384788513183594e-05, "ewc_loss_parallel": 1.9788742065429688e-05, "grad_norm": 1.794618844985962, "learning_rate": 8.393387028401865e-06, "loss": 0.465, "mean_token_accuracy": 0.8535103797912598, "num_tokens": 15058656.0, "step": 397 }, { "epoch": 0.050629690879023026, "ewc_loss": 0.021240234375, "ewc_loss_diag": 2.0384788513183594e-05, "ewc_loss_parallel": 1.919269561767578e-05, "grad_norm": 1.6814284324645996, "learning_rate": 8.414582450190759e-06, "loss": 0.498, "mean_token_accuracy": 0.8421691060066223, "num_tokens": 15094165.0, "step": 398 }, { "epoch": 0.05075690115761353, "ewc_loss": 0.0220947265625, "ewc_loss_diag": 2.0384788513183594e-05, "ewc_loss_parallel": 2.002716064453125e-05, "grad_norm": 1.7812024354934692, "learning_rate": 8.435777871979653e-06, "loss": 0.4819, "mean_token_accuracy": 0.8482393622398376, "num_tokens": 15136979.0, "step": 399 }, { "epoch": 0.05088411143620405, "ewc_loss": 0.0213623046875, "ewc_loss_diag": 2.0384788513183594e-05, "ewc_loss_parallel": 1.9311904907226562e-05, "grad_norm": 1.6812047958374023, "learning_rate": 8.456973293768548e-06, "loss": 0.5096, "mean_token_accuracy": 0.8433780670166016, "num_tokens": 15178765.0, "step": 400 }, { "epoch": 0.051011321714794555, "ewc_loss": 0.022216796875, "ewc_loss_diag": 2.0623207092285156e-05, "ewc_loss_parallel": 2.014636993408203e-05, "grad_norm": 1.8037630319595337, "learning_rate": 8.47816871555744e-06, "loss": 0.5084, "mean_token_accuracy": 0.8434333205223083, "num_tokens": 15215892.0, "step": 401 }, { "epoch": 0.05113853199338506, "ewc_loss": 0.021484375, "ewc_loss_diag": 2.0623207092285156e-05, "ewc_loss_parallel": 1.9431114196777344e-05, "grad_norm": 1.7100310325622559, "learning_rate": 8.499364137346335e-06, "loss": 0.4842, "mean_token_accuracy": 0.8516408205032349, "num_tokens": 15248693.0, "step": 402 }, { "epoch": 0.05126574227197558, "ewc_loss": 0.02197265625, "ewc_loss_diag": 2.086162567138672e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 1.8081318140029907, "learning_rate": 8.520559559135228e-06, "loss": 0.5103, "mean_token_accuracy": 0.8432937264442444, "num_tokens": 15289476.0, "step": 403 }, { "epoch": 0.051392952550566084, "ewc_loss": 0.0213623046875, "ewc_loss_diag": 2.086162567138672e-05, "ewc_loss_parallel": 1.9311904907226562e-05, "grad_norm": 1.7138365507125854, "learning_rate": 8.54175498092412e-06, "loss": 0.5395, "mean_token_accuracy": 0.8332952857017517, "num_tokens": 15323032.0, "step": 404 }, { "epoch": 0.0515201628291566, "ewc_loss": 0.022216796875, "ewc_loss_diag": 2.086162567138672e-05, "ewc_loss_parallel": 2.014636993408203e-05, "grad_norm": 1.8261168003082275, "learning_rate": 8.562950402713015e-06, "loss": 0.5708, "mean_token_accuracy": 0.8234922885894775, "num_tokens": 15369155.0, "step": 405 }, { "epoch": 0.051647373107747106, "ewc_loss": 0.021484375, "ewc_loss_diag": 2.09808349609375e-05, "ewc_loss_parallel": 1.9431114196777344e-05, "grad_norm": 1.718764305114746, "learning_rate": 8.584145824501908e-06, "loss": 0.4888, "mean_token_accuracy": 0.8438616394996643, "num_tokens": 15402623.0, "step": 406 }, { "epoch": 0.051774583386337614, "ewc_loss": 0.0225830078125, "ewc_loss_diag": 2.09808349609375e-05, "ewc_loss_parallel": 2.0503997802734375e-05, "grad_norm": 1.8223001956939697, "learning_rate": 8.605341246290802e-06, "loss": 0.5321, "mean_token_accuracy": 0.8376482725143433, "num_tokens": 15440485.0, "step": 407 }, { "epoch": 0.05190179366492813, "ewc_loss": 0.0216064453125, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.9550323486328125e-05, "grad_norm": 1.7030171155929565, "learning_rate": 8.626536668079695e-06, "loss": 0.5011, "mean_token_accuracy": 0.8424277305603027, "num_tokens": 15479641.0, "step": 408 }, { "epoch": 0.052029003943518635, "ewc_loss": 0.0228271484375, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 2.0742416381835938e-05, "grad_norm": 1.8338943719863892, "learning_rate": 8.647732089868589e-06, "loss": 0.5427, "mean_token_accuracy": 0.8339185118675232, "num_tokens": 15519085.0, "step": 409 }, { "epoch": 0.05215621422210915, "ewc_loss": 0.02197265625, "ewc_loss_diag": 2.1338462829589844e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 1.6997541189193726, "learning_rate": 8.668927511657482e-06, "loss": 0.4751, "mean_token_accuracy": 0.8542343974113464, "num_tokens": 15557642.0, "step": 410 }, { "epoch": 0.05228342450069966, "ewc_loss": 0.023193359375, "ewc_loss_diag": 2.1338462829589844e-05, "ewc_loss_parallel": 2.110004425048828e-05, "grad_norm": 1.845490574836731, "learning_rate": 8.690122933446376e-06, "loss": 0.475, "mean_token_accuracy": 0.8514590263366699, "num_tokens": 15593178.0, "step": 411 }, { "epoch": 0.052410634779290165, "ewc_loss": 0.021728515625, "ewc_loss_diag": 2.1338462829589844e-05, "ewc_loss_parallel": 1.9669532775878906e-05, "grad_norm": 1.7116937637329102, "learning_rate": 8.711318355235269e-06, "loss": 0.531, "mean_token_accuracy": 0.8350885510444641, "num_tokens": 15630167.0, "step": 412 }, { "epoch": 0.05253784505788068, "ewc_loss": 0.023681640625, "ewc_loss_diag": 2.1338462829589844e-05, "ewc_loss_parallel": 2.1576881408691406e-05, "grad_norm": 1.8849252462387085, "learning_rate": 8.732513777024163e-06, "loss": 0.5587, "mean_token_accuracy": 0.8288928270339966, "num_tokens": 15669380.0, "step": 413 }, { "epoch": 0.05266505533647119, "ewc_loss": 0.022216796875, "ewc_loss_diag": 2.1576881408691406e-05, "ewc_loss_parallel": 2.002716064453125e-05, "grad_norm": 1.7385174036026, "learning_rate": 8.753709198813056e-06, "loss": 0.5031, "mean_token_accuracy": 0.846381664276123, "num_tokens": 15701791.0, "step": 414 }, { "epoch": 0.052792265615061694, "ewc_loss": 0.0238037109375, "ewc_loss_diag": 2.1696090698242188e-05, "ewc_loss_parallel": 2.1576881408691406e-05, "grad_norm": 1.859445571899414, "learning_rate": 8.77490462060195e-06, "loss": 0.4784, "mean_token_accuracy": 0.8507269620895386, "num_tokens": 15738223.0, "step": 415 }, { "epoch": 0.05291947589365221, "ewc_loss": 0.0220947265625, "ewc_loss_diag": 2.193450927734375e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 1.731885313987732, "learning_rate": 8.796100042390843e-06, "loss": 0.5612, "mean_token_accuracy": 0.8282546997070312, "num_tokens": 15773159.0, "step": 416 }, { "epoch": 0.053046686172242716, "ewc_loss": 0.02392578125, "ewc_loss_diag": 2.193450927734375e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 1.8639193773269653, "learning_rate": 8.817295464179738e-06, "loss": 0.5213, "mean_token_accuracy": 0.839400053024292, "num_tokens": 15813132.0, "step": 417 }, { "epoch": 0.05317389645083323, "ewc_loss": 0.0220947265625, "ewc_loss_diag": 2.193450927734375e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 1.728979229927063, "learning_rate": 8.83849088596863e-06, "loss": 0.5057, "mean_token_accuracy": 0.8409569263458252, "num_tokens": 15849752.0, "step": 418 }, { "epoch": 0.05330110672942374, "ewc_loss": 0.0242919921875, "ewc_loss_diag": 2.205371856689453e-05, "ewc_loss_parallel": 2.205371856689453e-05, "grad_norm": 1.885399580001831, "learning_rate": 8.859686307757525e-06, "loss": 0.4935, "mean_token_accuracy": 0.850073516368866, "num_tokens": 15885331.0, "step": 419 }, { "epoch": 0.053428317008014245, "ewc_loss": 0.0225830078125, "ewc_loss_diag": 2.205371856689453e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 1.7450670003890991, "learning_rate": 8.880881729546418e-06, "loss": 0.5268, "mean_token_accuracy": 0.8405262231826782, "num_tokens": 15924649.0, "step": 420 }, { "epoch": 0.05355552728660476, "ewc_loss": 0.0244140625, "ewc_loss_diag": 2.205371856689453e-05, "ewc_loss_parallel": 2.2172927856445312e-05, "grad_norm": 1.8890609741210938, "learning_rate": 8.902077151335312e-06, "loss": 0.4994, "mean_token_accuracy": 0.8465214371681213, "num_tokens": 15959342.0, "step": 421 }, { "epoch": 0.05368273756519527, "ewc_loss": 0.022705078125, "ewc_loss_diag": 2.2172927856445312e-05, "ewc_loss_parallel": 2.0503997802734375e-05, "grad_norm": 1.7166165113449097, "learning_rate": 8.923272573124205e-06, "loss": 0.4845, "mean_token_accuracy": 0.8487380743026733, "num_tokens": 15998479.0, "step": 422 }, { "epoch": 0.05380994784378578, "ewc_loss": 0.02490234375, "ewc_loss_diag": 2.2172927856445312e-05, "ewc_loss_parallel": 2.2649765014648438e-05, "grad_norm": 1.921600341796875, "learning_rate": 8.9444679949131e-06, "loss": 0.5334, "mean_token_accuracy": 0.8337988257408142, "num_tokens": 16034197.0, "step": 423 }, { "epoch": 0.05393715812237629, "ewc_loss": 0.02294921875, "ewc_loss_diag": 2.2411346435546875e-05, "ewc_loss_parallel": 2.0742416381835938e-05, "grad_norm": 1.745436429977417, "learning_rate": 8.965663416701994e-06, "loss": 0.5445, "mean_token_accuracy": 0.8303889036178589, "num_tokens": 16071448.0, "step": 424 }, { "epoch": 0.054064368400966796, "ewc_loss": 0.025390625, "ewc_loss_diag": 2.2530555725097656e-05, "ewc_loss_parallel": 2.3126602172851562e-05, "grad_norm": 1.9044971466064453, "learning_rate": 8.986858838490886e-06, "loss": 0.4759, "mean_token_accuracy": 0.8549270629882812, "num_tokens": 16116007.0, "step": 425 }, { "epoch": 0.05419157867955731, "ewc_loss": 0.023193359375, "ewc_loss_diag": 2.2530555725097656e-05, "ewc_loss_parallel": 2.086162567138672e-05, "grad_norm": 1.7283740043640137, "learning_rate": 9.00805426027978e-06, "loss": 0.5013, "mean_token_accuracy": 0.8464007377624512, "num_tokens": 16158070.0, "step": 426 }, { "epoch": 0.05431878895814782, "ewc_loss": 0.025634765625, "ewc_loss_diag": 2.2530555725097656e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 1.9355285167694092, "learning_rate": 9.029249682068674e-06, "loss": 0.5423, "mean_token_accuracy": 0.8372950553894043, "num_tokens": 16196310.0, "step": 427 }, { "epoch": 0.054445999236738325, "ewc_loss": 0.0235595703125, "ewc_loss_diag": 2.276897430419922e-05, "ewc_loss_parallel": 2.1219253540039062e-05, "grad_norm": 1.7407201528549194, "learning_rate": 9.050445103857568e-06, "loss": 0.5868, "mean_token_accuracy": 0.8150193691253662, "num_tokens": 16237927.0, "step": 428 }, { "epoch": 0.05457320951532884, "ewc_loss": 0.0260009765625, "ewc_loss_diag": 2.288818359375e-05, "ewc_loss_parallel": 2.372264862060547e-05, "grad_norm": 1.9628701210021973, "learning_rate": 9.07164052564646e-06, "loss": 0.5444, "mean_token_accuracy": 0.8307845592498779, "num_tokens": 16274865.0, "step": 429 }, { "epoch": 0.05470041979391935, "ewc_loss": 0.024169921875, "ewc_loss_diag": 2.288818359375e-05, "ewc_loss_parallel": 2.181529998779297e-05, "grad_norm": 1.7949525117874146, "learning_rate": 9.092835947435355e-06, "loss": 0.5217, "mean_token_accuracy": 0.8364741802215576, "num_tokens": 16302503.0, "step": 430 }, { "epoch": 0.05482763007250986, "ewc_loss": 0.02587890625, "ewc_loss_diag": 2.288818359375e-05, "ewc_loss_parallel": 2.3603439331054688e-05, "grad_norm": 1.949476718902588, "learning_rate": 9.114031369224248e-06, "loss": 0.575, "mean_token_accuracy": 0.8240178823471069, "num_tokens": 16342472.0, "step": 431 }, { "epoch": 0.05495484035110037, "ewc_loss": 0.0238037109375, "ewc_loss_diag": 2.288818359375e-05, "ewc_loss_parallel": 2.1457672119140625e-05, "grad_norm": 1.76775062084198, "learning_rate": 9.135226791013142e-06, "loss": 0.5383, "mean_token_accuracy": 0.8334835767745972, "num_tokens": 16378423.0, "step": 432 }, { "epoch": 0.055082050629690876, "ewc_loss": 0.02587890625, "ewc_loss_diag": 2.288818359375e-05, "ewc_loss_parallel": 2.3603439331054688e-05, "grad_norm": 1.954864740371704, "learning_rate": 9.156422212802035e-06, "loss": 0.5093, "mean_token_accuracy": 0.8438065052032471, "num_tokens": 16418430.0, "step": 433 }, { "epoch": 0.05520926090828139, "ewc_loss": 0.0242919921875, "ewc_loss_diag": 2.300739288330078e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 1.7552695274353027, "learning_rate": 9.17761763459093e-06, "loss": 0.517, "mean_token_accuracy": 0.8415848612785339, "num_tokens": 16460984.0, "step": 434 }, { "epoch": 0.0553364711868719, "ewc_loss": 0.02685546875, "ewc_loss_diag": 2.300739288330078e-05, "ewc_loss_parallel": 2.4557113647460938e-05, "grad_norm": 1.9906651973724365, "learning_rate": 9.198813056379822e-06, "loss": 0.5011, "mean_token_accuracy": 0.8449164628982544, "num_tokens": 16493460.0, "step": 435 }, { "epoch": 0.05546368146546241, "ewc_loss": 0.0245361328125, "ewc_loss_diag": 2.3245811462402344e-05, "ewc_loss_parallel": 2.2172927856445312e-05, "grad_norm": 1.772255301475525, "learning_rate": 9.220008478168717e-06, "loss": 0.4769, "mean_token_accuracy": 0.8514935970306396, "num_tokens": 16528998.0, "step": 436 }, { "epoch": 0.05559089174405292, "ewc_loss": 0.0269775390625, "ewc_loss_diag": 2.3245811462402344e-05, "ewc_loss_parallel": 2.467632293701172e-05, "grad_norm": 2.017198324203491, "learning_rate": 9.24120389995761e-06, "loss": 0.5025, "mean_token_accuracy": 0.8457447290420532, "num_tokens": 16564041.0, "step": 437 }, { "epoch": 0.05571810202264343, "ewc_loss": 0.0245361328125, "ewc_loss_diag": 2.3245811462402344e-05, "ewc_loss_parallel": 2.2172927856445312e-05, "grad_norm": 1.7999860048294067, "learning_rate": 9.262399321746504e-06, "loss": 0.5701, "mean_token_accuracy": 0.8224037289619446, "num_tokens": 16597967.0, "step": 438 }, { "epoch": 0.05584531230123394, "ewc_loss": 0.0272216796875, "ewc_loss_diag": 2.3245811462402344e-05, "ewc_loss_parallel": 2.491474151611328e-05, "grad_norm": 2.009222984313965, "learning_rate": 9.283594743535396e-06, "loss": 0.5624, "mean_token_accuracy": 0.8282592296600342, "num_tokens": 16636756.0, "step": 439 }, { "epoch": 0.05597252257982445, "ewc_loss": 0.02490234375, "ewc_loss_diag": 2.3245811462402344e-05, "ewc_loss_parallel": 2.2530555725097656e-05, "grad_norm": 1.7980287075042725, "learning_rate": 9.30479016532429e-06, "loss": 0.4741, "mean_token_accuracy": 0.8533744812011719, "num_tokens": 16675811.0, "step": 440 }, { "epoch": 0.05609973285841496, "ewc_loss": 0.02734375, "ewc_loss_diag": 2.3365020751953125e-05, "ewc_loss_parallel": 2.5033950805664062e-05, "grad_norm": 2.0144639015197754, "learning_rate": 9.325985587113184e-06, "loss": 0.5579, "mean_token_accuracy": 0.8302900195121765, "num_tokens": 16713714.0, "step": 441 }, { "epoch": 0.05622694313700547, "ewc_loss": 0.025146484375, "ewc_loss_diag": 2.3365020751953125e-05, "ewc_loss_parallel": 2.276897430419922e-05, "grad_norm": 1.784327507019043, "learning_rate": 9.347181008902076e-06, "loss": 0.5141, "mean_token_accuracy": 0.8404204249382019, "num_tokens": 16754317.0, "step": 442 }, { "epoch": 0.05635415341559598, "ewc_loss": 0.0277099609375, "ewc_loss_diag": 2.3603439331054688e-05, "ewc_loss_parallel": 2.5391578674316406e-05, "grad_norm": 2.01420521736145, "learning_rate": 9.36837643069097e-06, "loss": 0.4758, "mean_token_accuracy": 0.853722333908081, "num_tokens": 16793189.0, "step": 443 }, { "epoch": 0.05648136369418649, "ewc_loss": 0.0252685546875, "ewc_loss_diag": 2.3603439331054688e-05, "ewc_loss_parallel": 2.300739288330078e-05, "grad_norm": 1.809677243232727, "learning_rate": 9.389571852479863e-06, "loss": 0.4836, "mean_token_accuracy": 0.8485230803489685, "num_tokens": 16825046.0, "step": 444 }, { "epoch": 0.056608573972777, "ewc_loss": 0.0279541015625, "ewc_loss_diag": 2.3603439331054688e-05, "ewc_loss_parallel": 2.562999725341797e-05, "grad_norm": 2.0406320095062256, "learning_rate": 9.410767274268758e-06, "loss": 0.5449, "mean_token_accuracy": 0.8327182531356812, "num_tokens": 16857769.0, "step": 445 }, { "epoch": 0.05673578425136751, "ewc_loss": 0.025390625, "ewc_loss_diag": 2.3603439331054688e-05, "ewc_loss_parallel": 2.3126602172851562e-05, "grad_norm": 1.799063801765442, "learning_rate": 9.431962696057652e-06, "loss": 0.5002, "mean_token_accuracy": 0.8469045162200928, "num_tokens": 16900878.0, "step": 446 }, { "epoch": 0.05686299452995802, "ewc_loss": 0.0284423828125, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 2.6106834411621094e-05, "grad_norm": 2.032893657684326, "learning_rate": 9.453158117846545e-06, "loss": 0.519, "mean_token_accuracy": 0.8395184278488159, "num_tokens": 16942832.0, "step": 447 }, { "epoch": 0.05699020480854853, "ewc_loss": 0.025634765625, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 1.8327971696853638, "learning_rate": 9.47435353963544e-06, "loss": 0.5211, "mean_token_accuracy": 0.8318698406219482, "num_tokens": 16975609.0, "step": 448 }, { "epoch": 0.057117415087139044, "ewc_loss": 0.028564453125, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 2.6226043701171875e-05, "grad_norm": 2.0300910472869873, "learning_rate": 9.495548961424334e-06, "loss": 0.4948, "mean_token_accuracy": 0.8490779399871826, "num_tokens": 17012620.0, "step": 449 }, { "epoch": 0.05724462536572955, "ewc_loss": 0.02587890625, "ewc_loss_diag": 2.384185791015625e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 1.8299691677093506, "learning_rate": 9.516744383213227e-06, "loss": 0.5341, "mean_token_accuracy": 0.8326208591461182, "num_tokens": 17053820.0, "step": 450 }, { "epoch": 0.05737183564432006, "ewc_loss": 0.028564453125, "ewc_loss_diag": 2.384185791015625e-05, "ewc_loss_parallel": 2.6106834411621094e-05, "grad_norm": 2.034999370574951, "learning_rate": 9.537939805002121e-06, "loss": 0.491, "mean_token_accuracy": 0.8516643047332764, "num_tokens": 17091885.0, "step": 451 }, { "epoch": 0.05749904592291057, "ewc_loss": 0.0260009765625, "ewc_loss_diag": 2.4080276489257812e-05, "ewc_loss_parallel": 2.3603439331054688e-05, "grad_norm": 1.8324575424194336, "learning_rate": 9.559135226791014e-06, "loss": 0.5373, "mean_token_accuracy": 0.8362767100334167, "num_tokens": 17133432.0, "step": 452 }, { "epoch": 0.05762625620150108, "ewc_loss": 0.02880859375, "ewc_loss_diag": 2.4318695068359375e-05, "ewc_loss_parallel": 2.6345252990722656e-05, "grad_norm": 2.035501480102539, "learning_rate": 9.580330648579908e-06, "loss": 0.5607, "mean_token_accuracy": 0.8303277492523193, "num_tokens": 17171845.0, "step": 453 }, { "epoch": 0.057753466480091595, "ewc_loss": 0.0260009765625, "ewc_loss_diag": 2.4318695068359375e-05, "ewc_loss_parallel": 2.3603439331054688e-05, "grad_norm": 1.8509858846664429, "learning_rate": 9.601526070368801e-06, "loss": 0.5492, "mean_token_accuracy": 0.8285738825798035, "num_tokens": 17206137.0, "step": 454 }, { "epoch": 0.0578806767586821, "ewc_loss": 0.0286865234375, "ewc_loss_diag": 2.4318695068359375e-05, "ewc_loss_parallel": 2.6226043701171875e-05, "grad_norm": 2.019186019897461, "learning_rate": 9.622721492157695e-06, "loss": 0.5318, "mean_token_accuracy": 0.8399428129196167, "num_tokens": 17247619.0, "step": 455 }, { "epoch": 0.05800788703727261, "ewc_loss": 0.02587890625, "ewc_loss_diag": 2.4557113647460938e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 1.8677321672439575, "learning_rate": 9.643916913946588e-06, "loss": 0.5516, "mean_token_accuracy": 0.8289105296134949, "num_tokens": 17285053.0, "step": 456 }, { "epoch": 0.058135097315863124, "ewc_loss": 0.0283203125, "ewc_loss_diag": 2.4557113647460938e-05, "ewc_loss_parallel": 2.586841583251953e-05, "grad_norm": 2.0178303718566895, "learning_rate": 9.665112335735482e-06, "loss": 0.5287, "mean_token_accuracy": 0.8374806642532349, "num_tokens": 17320303.0, "step": 457 }, { "epoch": 0.05826230759445363, "ewc_loss": 0.02587890625, "ewc_loss_diag": 2.467632293701172e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 1.8738956451416016, "learning_rate": 9.686307757524375e-06, "loss": 0.5689, "mean_token_accuracy": 0.8301599025726318, "num_tokens": 17354102.0, "step": 458 }, { "epoch": 0.05838951787304414, "ewc_loss": 0.0284423828125, "ewc_loss_diag": 2.47955322265625e-05, "ewc_loss_parallel": 2.5987625122070312e-05, "grad_norm": 2.0035560131073, "learning_rate": 9.707503179313268e-06, "loss": 0.5177, "mean_token_accuracy": 0.8414419293403625, "num_tokens": 17389958.0, "step": 459 }, { "epoch": 0.05851672815163465, "ewc_loss": 0.0257568359375, "ewc_loss_diag": 2.47955322265625e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 1.879686713218689, "learning_rate": 9.728698601102162e-06, "loss": 0.5053, "mean_token_accuracy": 0.8426831960678101, "num_tokens": 17421488.0, "step": 460 }, { "epoch": 0.05864393843022516, "ewc_loss": 0.028076171875, "ewc_loss_diag": 2.47955322265625e-05, "ewc_loss_parallel": 2.562999725341797e-05, "grad_norm": 1.9832261800765991, "learning_rate": 9.749894022891055e-06, "loss": 0.4677, "mean_token_accuracy": 0.8558942675590515, "num_tokens": 17455172.0, "step": 461 }, { "epoch": 0.058771148708815675, "ewc_loss": 0.0263671875, "ewc_loss_diag": 2.5033950805664062e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 1.8662816286087036, "learning_rate": 9.77108944467995e-06, "loss": 0.5633, "mean_token_accuracy": 0.8262333273887634, "num_tokens": 17494798.0, "step": 462 }, { "epoch": 0.05889835898740618, "ewc_loss": 0.028076171875, "ewc_loss_diag": 2.5033950805664062e-05, "ewc_loss_parallel": 2.562999725341797e-05, "grad_norm": 2.002530574798584, "learning_rate": 9.792284866468842e-06, "loss": 0.5275, "mean_token_accuracy": 0.8419137001037598, "num_tokens": 17527619.0, "step": 463 }, { "epoch": 0.05902556926599669, "ewc_loss": 0.026123046875, "ewc_loss_diag": 2.5033950805664062e-05, "ewc_loss_parallel": 2.372264862060547e-05, "grad_norm": 1.8515164852142334, "learning_rate": 9.813480288257737e-06, "loss": 0.4494, "mean_token_accuracy": 0.8614307641983032, "num_tokens": 17564299.0, "step": 464 }, { "epoch": 0.059152779544587204, "ewc_loss": 0.028564453125, "ewc_loss_diag": 2.5033950805664062e-05, "ewc_loss_parallel": 2.5987625122070312e-05, "grad_norm": 2.038792371749878, "learning_rate": 9.83467571004663e-06, "loss": 0.5632, "mean_token_accuracy": 0.8288201689720154, "num_tokens": 17598333.0, "step": 465 }, { "epoch": 0.05927998982317771, "ewc_loss": 0.02587890625, "ewc_loss_diag": 2.5033950805664062e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 1.8383575677871704, "learning_rate": 9.855871131835524e-06, "loss": 0.5484, "mean_token_accuracy": 0.8361140489578247, "num_tokens": 17642106.0, "step": 466 }, { "epoch": 0.059407200101768226, "ewc_loss": 0.0291748046875, "ewc_loss_diag": 2.5153160095214844e-05, "ewc_loss_parallel": 2.658367156982422e-05, "grad_norm": 2.0627243518829346, "learning_rate": 9.877066553624417e-06, "loss": 0.4883, "mean_token_accuracy": 0.8469018340110779, "num_tokens": 17676412.0, "step": 467 }, { "epoch": 0.059534410380358734, "ewc_loss": 0.0262451171875, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 2.372264862060547e-05, "grad_norm": 1.8534444570541382, "learning_rate": 9.898261975413311e-06, "loss": 0.5691, "mean_token_accuracy": 0.8284367918968201, "num_tokens": 17717079.0, "step": 468 }, { "epoch": 0.05966162065894924, "ewc_loss": 0.030029296875, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 2.7418136596679688e-05, "grad_norm": 2.080479145050049, "learning_rate": 9.919457397202204e-06, "loss": 0.468, "mean_token_accuracy": 0.856981635093689, "num_tokens": 17753603.0, "step": 469 }, { "epoch": 0.059788830937539755, "ewc_loss": 0.0263671875, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 1.8314390182495117, "learning_rate": 9.940652818991098e-06, "loss": 0.5249, "mean_token_accuracy": 0.8405670523643494, "num_tokens": 17795960.0, "step": 470 }, { "epoch": 0.05991604121613026, "ewc_loss": 0.0306396484375, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 2.8133392333984375e-05, "grad_norm": 2.1363236904144287, "learning_rate": 9.961848240779993e-06, "loss": 0.5485, "mean_token_accuracy": 0.8359884023666382, "num_tokens": 17829298.0, "step": 471 }, { "epoch": 0.06004325149472077, "ewc_loss": 0.0269775390625, "ewc_loss_diag": 2.5510787963867188e-05, "ewc_loss_parallel": 2.4437904357910156e-05, "grad_norm": 1.841876745223999, "learning_rate": 9.983043662568885e-06, "loss": 0.5832, "mean_token_accuracy": 0.8211824893951416, "num_tokens": 17872027.0, "step": 472 }, { "epoch": 0.060170461773311285, "ewc_loss": 0.031494140625, "ewc_loss_diag": 2.574920654296875e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.1533379554748535, "learning_rate": 1.000423908435778e-05, "loss": 0.4773, "mean_token_accuracy": 0.8537380695343018, "num_tokens": 17911846.0, "step": 473 }, { "epoch": 0.06029767205190179, "ewc_loss": 0.0274658203125, "ewc_loss_diag": 2.574920654296875e-05, "ewc_loss_parallel": 2.491474151611328e-05, "grad_norm": 1.8645625114440918, "learning_rate": 1.0025434506146674e-05, "loss": 0.4598, "mean_token_accuracy": 0.8584028482437134, "num_tokens": 17946095.0, "step": 474 }, { "epoch": 0.060424882330492306, "ewc_loss": 0.032470703125, "ewc_loss_diag": 2.574920654296875e-05, "ewc_loss_parallel": 2.9802322387695312e-05, "grad_norm": 2.1774206161499023, "learning_rate": 1.0046629927935567e-05, "loss": 0.4963, "mean_token_accuracy": 0.845894992351532, "num_tokens": 17983232.0, "step": 475 }, { "epoch": 0.060552092609082814, "ewc_loss": 0.0279541015625, "ewc_loss_diag": 2.586841583251953e-05, "ewc_loss_parallel": 2.5391578674316406e-05, "grad_norm": 1.8810698986053467, "learning_rate": 1.0067825349724461e-05, "loss": 0.5822, "mean_token_accuracy": 0.8192480802536011, "num_tokens": 18022455.0, "step": 476 }, { "epoch": 0.06067930288767332, "ewc_loss": 0.032470703125, "ewc_loss_diag": 2.6226043701171875e-05, "ewc_loss_parallel": 2.9802322387695312e-05, "grad_norm": 2.205782413482666, "learning_rate": 1.0089020771513354e-05, "loss": 0.5524, "mean_token_accuracy": 0.831179141998291, "num_tokens": 18059844.0, "step": 477 }, { "epoch": 0.060806513166263836, "ewc_loss": 0.02880859375, "ewc_loss_diag": 2.6226043701171875e-05, "ewc_loss_parallel": 2.6106834411621094e-05, "grad_norm": 1.9037156105041504, "learning_rate": 1.0110216193302247e-05, "loss": 0.5986, "mean_token_accuracy": 0.8178823590278625, "num_tokens": 18099569.0, "step": 478 }, { "epoch": 0.06093372344485434, "ewc_loss": 0.033447265625, "ewc_loss_diag": 2.6226043701171875e-05, "ewc_loss_parallel": 3.075599670410156e-05, "grad_norm": 2.223850727081299, "learning_rate": 1.0131411615091141e-05, "loss": 0.5616, "mean_token_accuracy": 0.8289022445678711, "num_tokens": 18134747.0, "step": 479 }, { "epoch": 0.06106093372344486, "ewc_loss": 0.02978515625, "ewc_loss_diag": 2.6345252990722656e-05, "ewc_loss_parallel": 2.7060508728027344e-05, "grad_norm": 1.9151506423950195, "learning_rate": 1.0152607036880034e-05, "loss": 0.5672, "mean_token_accuracy": 0.8298017382621765, "num_tokens": 18173989.0, "step": 480 }, { "epoch": 0.061188144002035365, "ewc_loss": 0.03369140625, "ewc_loss_diag": 2.6345252990722656e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.2421066761016846, "learning_rate": 1.0173802458668928e-05, "loss": 0.5443, "mean_token_accuracy": 0.8336683511734009, "num_tokens": 18205065.0, "step": 481 }, { "epoch": 0.06131535428062587, "ewc_loss": 0.03076171875, "ewc_loss_diag": 2.6345252990722656e-05, "ewc_loss_parallel": 2.8133392333984375e-05, "grad_norm": 1.9367009401321411, "learning_rate": 1.0194997880457821e-05, "loss": 0.492, "mean_token_accuracy": 0.8505300283432007, "num_tokens": 18242999.0, "step": 482 }, { "epoch": 0.06144256455921639, "ewc_loss": 0.033935546875, "ewc_loss_diag": 2.6345252990722656e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.233734130859375, "learning_rate": 1.0216193302246715e-05, "loss": 0.5786, "mean_token_accuracy": 0.8247002363204956, "num_tokens": 18280642.0, "step": 483 }, { "epoch": 0.061569774837806894, "ewc_loss": 0.0311279296875, "ewc_loss_diag": 2.6464462280273438e-05, "ewc_loss_parallel": 2.849102020263672e-05, "grad_norm": 1.9531207084655762, "learning_rate": 1.0237388724035608e-05, "loss": 0.4988, "mean_token_accuracy": 0.8489952087402344, "num_tokens": 18317894.0, "step": 484 }, { "epoch": 0.0616969851163974, "ewc_loss": 0.033935546875, "ewc_loss_diag": 2.6464462280273438e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.2271313667297363, "learning_rate": 1.0258584145824503e-05, "loss": 0.5306, "mean_token_accuracy": 0.8393517136573792, "num_tokens": 18350481.0, "step": 485 }, { "epoch": 0.061824195394987916, "ewc_loss": 0.03125, "ewc_loss_diag": 2.6702880859375e-05, "ewc_loss_parallel": 2.872943878173828e-05, "grad_norm": 1.991673469543457, "learning_rate": 1.0279779567613395e-05, "loss": 0.5305, "mean_token_accuracy": 0.8331838250160217, "num_tokens": 18384390.0, "step": 486 }, { "epoch": 0.06195140567357842, "ewc_loss": 0.032958984375, "ewc_loss_diag": 2.6702880859375e-05, "ewc_loss_parallel": 3.0279159545898438e-05, "grad_norm": 2.1734328269958496, "learning_rate": 1.030097498940229e-05, "loss": 0.4964, "mean_token_accuracy": 0.8442829847335815, "num_tokens": 18423958.0, "step": 487 }, { "epoch": 0.06207861595216894, "ewc_loss": 0.03125, "ewc_loss_diag": 2.6941299438476562e-05, "ewc_loss_parallel": 2.86102294921875e-05, "grad_norm": 1.9812042713165283, "learning_rate": 1.0322170411191182e-05, "loss": 0.46, "mean_token_accuracy": 0.8590432405471802, "num_tokens": 18459622.0, "step": 488 }, { "epoch": 0.062205826230759445, "ewc_loss": 0.0322265625, "ewc_loss_diag": 2.6941299438476562e-05, "ewc_loss_parallel": 2.956390380859375e-05, "grad_norm": 2.1502468585968018, "learning_rate": 1.0343365832980077e-05, "loss": 0.52, "mean_token_accuracy": 0.845144510269165, "num_tokens": 18495568.0, "step": 489 }, { "epoch": 0.06233303650934995, "ewc_loss": 0.0306396484375, "ewc_loss_diag": 2.7060508728027344e-05, "ewc_loss_parallel": 2.8014183044433594e-05, "grad_norm": 1.9831115007400513, "learning_rate": 1.036456125476897e-05, "loss": 0.4817, "mean_token_accuracy": 0.8560184240341187, "num_tokens": 18534440.0, "step": 490 }, { "epoch": 0.06246024678794047, "ewc_loss": 0.03173828125, "ewc_loss_diag": 2.7060508728027344e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.1269664764404297, "learning_rate": 1.0385756676557864e-05, "loss": 0.5201, "mean_token_accuracy": 0.8416761159896851, "num_tokens": 18570583.0, "step": 491 }, { "epoch": 0.06258745706653097, "ewc_loss": 0.030029296875, "ewc_loss_diag": 2.7060508728027344e-05, "ewc_loss_parallel": 2.7298927307128906e-05, "grad_norm": 1.9660133123397827, "learning_rate": 1.0406952098346757e-05, "loss": 0.4939, "mean_token_accuracy": 0.8476524949073792, "num_tokens": 18611801.0, "step": 492 }, { "epoch": 0.06271466734512149, "ewc_loss": 0.03173828125, "ewc_loss_diag": 2.7179718017578125e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.133072853088379, "learning_rate": 1.0428147520135651e-05, "loss": 0.5552, "mean_token_accuracy": 0.829418420791626, "num_tokens": 18646985.0, "step": 493 }, { "epoch": 0.06284187762371199, "ewc_loss": 0.0299072265625, "ewc_loss_diag": 2.7298927307128906e-05, "ewc_loss_parallel": 2.7179718017578125e-05, "grad_norm": 1.9629135131835938, "learning_rate": 1.0449342941924544e-05, "loss": 0.5387, "mean_token_accuracy": 0.8401229977607727, "num_tokens": 18688818.0, "step": 494 }, { "epoch": 0.0629690879023025, "ewc_loss": 0.03173828125, "ewc_loss_diag": 2.7298927307128906e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.130094528198242, "learning_rate": 1.0470538363713438e-05, "loss": 0.5172, "mean_token_accuracy": 0.845131516456604, "num_tokens": 18727208.0, "step": 495 }, { "epoch": 0.06309629818089302, "ewc_loss": 0.029541015625, "ewc_loss_diag": 2.7418136596679688e-05, "ewc_loss_parallel": 2.682209014892578e-05, "grad_norm": 1.9414032697677612, "learning_rate": 1.0491733785502333e-05, "loss": 0.4973, "mean_token_accuracy": 0.8476210832595825, "num_tokens": 18772174.0, "step": 496 }, { "epoch": 0.06322350845948353, "ewc_loss": 0.03173828125, "ewc_loss_diag": 2.753734588623047e-05, "ewc_loss_parallel": 2.9087066650390625e-05, "grad_norm": 2.1405797004699707, "learning_rate": 1.0512929207291226e-05, "loss": 0.539, "mean_token_accuracy": 0.8359580039978027, "num_tokens": 18808111.0, "step": 497 }, { "epoch": 0.06335071873807403, "ewc_loss": 0.02978515625, "ewc_loss_diag": 2.753734588623047e-05, "ewc_loss_parallel": 2.6941299438476562e-05, "grad_norm": 1.9554747343063354, "learning_rate": 1.053412462908012e-05, "loss": 0.4884, "mean_token_accuracy": 0.8464604616165161, "num_tokens": 18841912.0, "step": 498 }, { "epoch": 0.06347792901666455, "ewc_loss": 0.031982421875, "ewc_loss_diag": 2.753734588623047e-05, "ewc_loss_parallel": 2.9206275939941406e-05, "grad_norm": 2.1526618003845215, "learning_rate": 1.0555320050869013e-05, "loss": 0.5022, "mean_token_accuracy": 0.8487404584884644, "num_tokens": 18877694.0, "step": 499 }, { "epoch": 0.06360513929525506, "ewc_loss": 0.0296630859375, "ewc_loss_diag": 2.765655517578125e-05, "ewc_loss_parallel": 2.682209014892578e-05, "grad_norm": 1.9497205018997192, "learning_rate": 1.0576515472657907e-05, "loss": 0.5048, "mean_token_accuracy": 0.845861554145813, "num_tokens": 18912253.0, "step": 500 }, { "epoch": 0.06373234957384556, "ewc_loss": 0.0322265625, "ewc_loss_diag": 2.765655517578125e-05, "ewc_loss_parallel": 2.956390380859375e-05, "grad_norm": 2.17453932762146, "learning_rate": 1.05977108944468e-05, "loss": 0.5095, "mean_token_accuracy": 0.848048210144043, "num_tokens": 18945700.0, "step": 501 }, { "epoch": 0.06385955985243608, "ewc_loss": 0.02978515625, "ewc_loss_diag": 2.777576446533203e-05, "ewc_loss_parallel": 2.6941299438476562e-05, "grad_norm": 1.9395928382873535, "learning_rate": 1.0618906316235694e-05, "loss": 0.528, "mean_token_accuracy": 0.8374031782150269, "num_tokens": 18985260.0, "step": 502 }, { "epoch": 0.06398677013102659, "ewc_loss": 0.03271484375, "ewc_loss_diag": 2.777576446533203e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.187800645828247, "learning_rate": 1.0640101738024587e-05, "loss": 0.601, "mean_token_accuracy": 0.8160148859024048, "num_tokens": 19025364.0, "step": 503 }, { "epoch": 0.06411398040961709, "ewc_loss": 0.0296630859375, "ewc_loss_diag": 2.777576446533203e-05, "ewc_loss_parallel": 2.682209014892578e-05, "grad_norm": 1.942785382270813, "learning_rate": 1.0661297159813481e-05, "loss": 0.5395, "mean_token_accuracy": 0.8349008560180664, "num_tokens": 19064262.0, "step": 504 }, { "epoch": 0.0642411906882076, "ewc_loss": 0.033203125, "ewc_loss_diag": 2.777576446533203e-05, "ewc_loss_parallel": 3.039836883544922e-05, "grad_norm": 2.239248752593994, "learning_rate": 1.0682492581602374e-05, "loss": 0.5464, "mean_token_accuracy": 0.8357274532318115, "num_tokens": 19094780.0, "step": 505 }, { "epoch": 0.06436840096679812, "ewc_loss": 0.0302734375, "ewc_loss_diag": 2.7894973754882812e-05, "ewc_loss_parallel": 2.7418136596679688e-05, "grad_norm": 1.9555903673171997, "learning_rate": 1.0703688003391269e-05, "loss": 0.4836, "mean_token_accuracy": 0.8524792194366455, "num_tokens": 19131166.0, "step": 506 }, { "epoch": 0.06449561124538863, "ewc_loss": 0.0341796875, "ewc_loss_diag": 2.8014183044433594e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.2436211109161377, "learning_rate": 1.0724883425180161e-05, "loss": 0.4986, "mean_token_accuracy": 0.8480857610702515, "num_tokens": 19167371.0, "step": 507 }, { "epoch": 0.06462282152397913, "ewc_loss": 0.0306396484375, "ewc_loss_diag": 2.8014183044433594e-05, "ewc_loss_parallel": 2.777576446533203e-05, "grad_norm": 1.944691777229309, "learning_rate": 1.0746078846969056e-05, "loss": 0.5437, "mean_token_accuracy": 0.8347325325012207, "num_tokens": 19206751.0, "step": 508 }, { "epoch": 0.06475003180256965, "ewc_loss": 0.03515625, "ewc_loss_diag": 2.8014183044433594e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.2743120193481445, "learning_rate": 1.0767274268757948e-05, "loss": 0.5035, "mean_token_accuracy": 0.8466055989265442, "num_tokens": 19243227.0, "step": 509 }, { "epoch": 0.06487724208116016, "ewc_loss": 0.03125, "ewc_loss_diag": 2.8371810913085938e-05, "ewc_loss_parallel": 2.849102020263672e-05, "grad_norm": 1.9508590698242188, "learning_rate": 1.0788469690546843e-05, "loss": 0.555, "mean_token_accuracy": 0.8341418504714966, "num_tokens": 19285050.0, "step": 510 }, { "epoch": 0.06500445235975066, "ewc_loss": 0.036376953125, "ewc_loss_diag": 2.8371810913085938e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.2971928119659424, "learning_rate": 1.0809665112335736e-05, "loss": 0.5411, "mean_token_accuracy": 0.8360474109649658, "num_tokens": 19328780.0, "step": 511 }, { "epoch": 0.06513166263834118, "ewc_loss": 0.031982421875, "ewc_loss_diag": 2.8371810913085938e-05, "ewc_loss_parallel": 2.9206275939941406e-05, "grad_norm": 1.970819354057312, "learning_rate": 1.083086053412463e-05, "loss": 0.5314, "mean_token_accuracy": 0.840888261795044, "num_tokens": 19373011.0, "step": 512 }, { "epoch": 0.06525887291693169, "ewc_loss": 0.03662109375, "ewc_loss_diag": 2.8371810913085938e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.333463668823242, "learning_rate": 1.0852055955913523e-05, "loss": 0.5302, "mean_token_accuracy": 0.8392408490180969, "num_tokens": 19409641.0, "step": 513 }, { "epoch": 0.0653860831955222, "ewc_loss": 0.03271484375, "ewc_loss_diag": 2.8371810913085938e-05, "ewc_loss_parallel": 2.9921531677246094e-05, "grad_norm": 2.0055155754089355, "learning_rate": 1.0873251377702415e-05, "loss": 0.5627, "mean_token_accuracy": 0.8311371207237244, "num_tokens": 19451285.0, "step": 514 }, { "epoch": 0.06551329347411271, "ewc_loss": 0.03662109375, "ewc_loss_diag": 2.849102020263672e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.3420474529266357, "learning_rate": 1.089444679949131e-05, "loss": 0.5648, "mean_token_accuracy": 0.8335732221603394, "num_tokens": 19487584.0, "step": 515 }, { "epoch": 0.06564050375270322, "ewc_loss": 0.033447265625, "ewc_loss_diag": 2.849102020263672e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.0152699947357178, "learning_rate": 1.0915642221280203e-05, "loss": 0.5655, "mean_token_accuracy": 0.8333685398101807, "num_tokens": 19531200.0, "step": 516 }, { "epoch": 0.06576771403129372, "ewc_loss": 0.03662109375, "ewc_loss_diag": 2.86102294921875e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.3347339630126953, "learning_rate": 1.0936837643069097e-05, "loss": 0.5758, "mean_token_accuracy": 0.8244662284851074, "num_tokens": 19573126.0, "step": 517 }, { "epoch": 0.06589492430988424, "ewc_loss": 0.0341796875, "ewc_loss_diag": 2.86102294921875e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.0395214557647705, "learning_rate": 1.0958033064857991e-05, "loss": 0.5851, "mean_token_accuracy": 0.8239247798919678, "num_tokens": 19613499.0, "step": 518 }, { "epoch": 0.06602213458847475, "ewc_loss": 0.036376953125, "ewc_loss_diag": 2.86102294921875e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.3180484771728516, "learning_rate": 1.0979228486646884e-05, "loss": 0.5742, "mean_token_accuracy": 0.8214236497879028, "num_tokens": 19652459.0, "step": 519 }, { "epoch": 0.06614934486706527, "ewc_loss": 0.033935546875, "ewc_loss_diag": 2.86102294921875e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.11409854888916, "learning_rate": 1.1000423908435779e-05, "loss": 0.4557, "mean_token_accuracy": 0.8597627878189087, "num_tokens": 19691874.0, "step": 520 }, { "epoch": 0.06627655514565577, "ewc_loss": 0.034912109375, "ewc_loss_diag": 2.8848648071289062e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.242542266845703, "learning_rate": 1.1021619330224673e-05, "loss": 0.5083, "mean_token_accuracy": 0.8485474586486816, "num_tokens": 19731424.0, "step": 521 }, { "epoch": 0.06640376542424628, "ewc_loss": 0.033935546875, "ewc_loss_diag": 2.8848648071289062e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.071178436279297, "learning_rate": 1.1042814752013566e-05, "loss": 0.4975, "mean_token_accuracy": 0.8486303091049194, "num_tokens": 19768530.0, "step": 522 }, { "epoch": 0.0665309757028368, "ewc_loss": 0.0341796875, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.208592653274536, "learning_rate": 1.106401017380246e-05, "loss": 0.5372, "mean_token_accuracy": 0.8371326923370361, "num_tokens": 19807836.0, "step": 523 }, { "epoch": 0.0666581859814273, "ewc_loss": 0.032958984375, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.062619209289551, "learning_rate": 1.1085205595591353e-05, "loss": 0.5144, "mean_token_accuracy": 0.8431609272956848, "num_tokens": 19854338.0, "step": 524 }, { "epoch": 0.06678539626001781, "ewc_loss": 0.03369140625, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 3.075599670410156e-05, "grad_norm": 2.2156758308410645, "learning_rate": 1.1106401017380247e-05, "loss": 0.53, "mean_token_accuracy": 0.8360084295272827, "num_tokens": 19888469.0, "step": 525 }, { "epoch": 0.06691260653860832, "ewc_loss": 0.03173828125, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.0834107398986816, "learning_rate": 1.112759643916914e-05, "loss": 0.5764, "mean_token_accuracy": 0.8256044387817383, "num_tokens": 19925137.0, "step": 526 }, { "epoch": 0.06703981681719882, "ewc_loss": 0.033203125, "ewc_loss_diag": 2.9325485229492188e-05, "ewc_loss_parallel": 3.0279159545898438e-05, "grad_norm": 2.1913843154907227, "learning_rate": 1.1148791860958034e-05, "loss": 0.539, "mean_token_accuracy": 0.8344999551773071, "num_tokens": 19959658.0, "step": 527 }, { "epoch": 0.06716702709578934, "ewc_loss": 0.03173828125, "ewc_loss_diag": 2.9325485229492188e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.065183401107788, "learning_rate": 1.1169987282746927e-05, "loss": 0.4965, "mean_token_accuracy": 0.8477400541305542, "num_tokens": 19998702.0, "step": 528 }, { "epoch": 0.06729423737437985, "ewc_loss": 0.032958984375, "ewc_loss_diag": 2.9325485229492188e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.1794581413269043, "learning_rate": 1.1191182704535822e-05, "loss": 0.5662, "mean_token_accuracy": 0.8306353688240051, "num_tokens": 20038406.0, "step": 529 }, { "epoch": 0.06742144765297035, "ewc_loss": 0.03173828125, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 2.872943878173828e-05, "grad_norm": 2.0691378116607666, "learning_rate": 1.1212378126324714e-05, "loss": 0.517, "mean_token_accuracy": 0.8411952257156372, "num_tokens": 20073185.0, "step": 530 }, { "epoch": 0.06754865793156087, "ewc_loss": 0.032958984375, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.177983522415161, "learning_rate": 1.1233573548113609e-05, "loss": 0.5019, "mean_token_accuracy": 0.8476642966270447, "num_tokens": 20106828.0, "step": 531 }, { "epoch": 0.06767586821015138, "ewc_loss": 0.031494140625, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 2.86102294921875e-05, "grad_norm": 2.0767669677734375, "learning_rate": 1.1254768969902502e-05, "loss": 0.5017, "mean_token_accuracy": 0.8451932668685913, "num_tokens": 20137541.0, "step": 532 }, { "epoch": 0.0678030784887419, "ewc_loss": 0.032958984375, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 2.9921531677246094e-05, "grad_norm": 2.1687521934509277, "learning_rate": 1.1275964391691394e-05, "loss": 0.5224, "mean_token_accuracy": 0.839024543762207, "num_tokens": 20174452.0, "step": 533 }, { "epoch": 0.0679302887673324, "ewc_loss": 0.0322265625, "ewc_loss_diag": 2.968311309814453e-05, "ewc_loss_parallel": 2.9206275939941406e-05, "grad_norm": 2.0735831260681152, "learning_rate": 1.1297159813480289e-05, "loss": 0.5387, "mean_token_accuracy": 0.8367279767990112, "num_tokens": 20212195.0, "step": 534 }, { "epoch": 0.06805749904592291, "ewc_loss": 0.033203125, "ewc_loss_diag": 2.9921531677246094e-05, "ewc_loss_parallel": 3.0279159545898438e-05, "grad_norm": 2.1608760356903076, "learning_rate": 1.1318355235269181e-05, "loss": 0.4521, "mean_token_accuracy": 0.8561413288116455, "num_tokens": 20248344.0, "step": 535 }, { "epoch": 0.06818470932451343, "ewc_loss": 0.031982421875, "ewc_loss_diag": 3.0040740966796875e-05, "ewc_loss_parallel": 2.9087066650390625e-05, "grad_norm": 2.058755397796631, "learning_rate": 1.1339550657058076e-05, "loss": 0.4926, "mean_token_accuracy": 0.8512734174728394, "num_tokens": 20289874.0, "step": 536 }, { "epoch": 0.06831191960310393, "ewc_loss": 0.033447265625, "ewc_loss_diag": 3.0040740966796875e-05, "ewc_loss_parallel": 3.039836883544922e-05, "grad_norm": 2.210740566253662, "learning_rate": 1.1360746078846969e-05, "loss": 0.5064, "mean_token_accuracy": 0.8430044651031494, "num_tokens": 20319582.0, "step": 537 }, { "epoch": 0.06843912988169444, "ewc_loss": 0.031494140625, "ewc_loss_diag": 3.0040740966796875e-05, "ewc_loss_parallel": 2.86102294921875e-05, "grad_norm": 2.0427348613739014, "learning_rate": 1.1381941500635863e-05, "loss": 0.4886, "mean_token_accuracy": 0.8474187850952148, "num_tokens": 20356944.0, "step": 538 }, { "epoch": 0.06856634016028496, "ewc_loss": 0.0341796875, "ewc_loss_diag": 3.0040740966796875e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.22422194480896, "learning_rate": 1.1403136922424756e-05, "loss": 0.5406, "mean_token_accuracy": 0.8332337141036987, "num_tokens": 20398099.0, "step": 539 }, { "epoch": 0.06869355043887546, "ewc_loss": 0.03125, "ewc_loss_diag": 3.0159950256347656e-05, "ewc_loss_parallel": 2.8371810913085938e-05, "grad_norm": 2.0363893508911133, "learning_rate": 1.142433234421365e-05, "loss": 0.5129, "mean_token_accuracy": 0.8418415784835815, "num_tokens": 20435155.0, "step": 540 }, { "epoch": 0.06882076071746597, "ewc_loss": 0.035400390625, "ewc_loss_diag": 3.0279159545898438e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.295393705368042, "learning_rate": 1.1445527766002543e-05, "loss": 0.5729, "mean_token_accuracy": 0.828812837600708, "num_tokens": 20468369.0, "step": 541 }, { "epoch": 0.06894797099605648, "ewc_loss": 0.0311279296875, "ewc_loss_diag": 3.0279159545898438e-05, "ewc_loss_parallel": 2.8133392333984375e-05, "grad_norm": 2.055126905441284, "learning_rate": 1.1466723187791437e-05, "loss": 0.5413, "mean_token_accuracy": 0.8395452499389648, "num_tokens": 20500303.0, "step": 542 }, { "epoch": 0.06907518127464699, "ewc_loss": 0.035888671875, "ewc_loss_diag": 3.0279159545898438e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.294194221496582, "learning_rate": 1.1487918609580332e-05, "loss": 0.5006, "mean_token_accuracy": 0.8492550849914551, "num_tokens": 20541290.0, "step": 543 }, { "epoch": 0.0692023915532375, "ewc_loss": 0.031494140625, "ewc_loss_diag": 3.0517578125e-05, "ewc_loss_parallel": 2.849102020263672e-05, "grad_norm": 2.0040745735168457, "learning_rate": 1.1509114031369224e-05, "loss": 0.5184, "mean_token_accuracy": 0.8438029289245605, "num_tokens": 20583486.0, "step": 544 }, { "epoch": 0.06932960183182801, "ewc_loss": 0.037109375, "ewc_loss_diag": 3.0517578125e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.376676082611084, "learning_rate": 1.1530309453158119e-05, "loss": 0.572, "mean_token_accuracy": 0.8313784599304199, "num_tokens": 20618655.0, "step": 545 }, { "epoch": 0.06945681211041853, "ewc_loss": 0.031982421875, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.8848648071289062e-05, "grad_norm": 1.9913690090179443, "learning_rate": 1.1551504874947013e-05, "loss": 0.524, "mean_token_accuracy": 0.8406670093536377, "num_tokens": 20655733.0, "step": 546 }, { "epoch": 0.06958402238900903, "ewc_loss": 0.038818359375, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.4507598876953125, "learning_rate": 1.1572700296735906e-05, "loss": 0.5567, "mean_token_accuracy": 0.8323933482170105, "num_tokens": 20699338.0, "step": 547 }, { "epoch": 0.06971123266759954, "ewc_loss": 0.033203125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.019631862640381, "learning_rate": 1.15938957185248e-05, "loss": 0.5692, "mean_token_accuracy": 0.8277633786201477, "num_tokens": 20731841.0, "step": 548 }, { "epoch": 0.06983844294619006, "ewc_loss": 0.040283203125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.5399351119995117, "learning_rate": 1.1615091140313693e-05, "loss": 0.4994, "mean_token_accuracy": 0.8512778878211975, "num_tokens": 20761097.0, "step": 549 }, { "epoch": 0.06996565322478056, "ewc_loss": 0.035400390625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.0428309440612793, "learning_rate": 1.1636286562102586e-05, "loss": 0.4758, "mean_token_accuracy": 0.8546554446220398, "num_tokens": 20797013.0, "step": 550 }, { "epoch": 0.07009286350337107, "ewc_loss": 0.042236328125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.6029651165008545, "learning_rate": 1.165748198389148e-05, "loss": 0.5192, "mean_token_accuracy": 0.8419691324234009, "num_tokens": 20833727.0, "step": 551 }, { "epoch": 0.07022007378196159, "ewc_loss": 0.037109375, "ewc_loss_diag": 3.0994415283203125e-05, "ewc_loss_parallel": 3.3855438232421875e-05, "grad_norm": 2.09517765045166, "learning_rate": 1.1678677405680373e-05, "loss": 0.5404, "mean_token_accuracy": 0.8386675119400024, "num_tokens": 20877587.0, "step": 552 }, { "epoch": 0.07034728406055209, "ewc_loss": 0.04345703125, "ewc_loss_diag": 3.0994415283203125e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.6851537227630615, "learning_rate": 1.1699872827469267e-05, "loss": 0.5585, "mean_token_accuracy": 0.8336935639381409, "num_tokens": 20909623.0, "step": 553 }, { "epoch": 0.0704744943391426, "ewc_loss": 0.039306640625, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.221472978591919, "learning_rate": 1.172106824925816e-05, "loss": 0.5351, "mean_token_accuracy": 0.8406530022621155, "num_tokens": 20946348.0, "step": 554 }, { "epoch": 0.07060170461773312, "ewc_loss": 0.043212890625, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.643796443939209, "learning_rate": 1.1742263671047055e-05, "loss": 0.62, "mean_token_accuracy": 0.8163448572158813, "num_tokens": 20989831.0, "step": 555 }, { "epoch": 0.07072891489632362, "ewc_loss": 0.041259765625, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.284055471420288, "learning_rate": 1.1763459092835947e-05, "loss": 0.5644, "mean_token_accuracy": 0.8290364742279053, "num_tokens": 21027007.0, "step": 556 }, { "epoch": 0.07085612517491413, "ewc_loss": 0.0419921875, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.5553908348083496, "learning_rate": 1.1784654514624842e-05, "loss": 0.5114, "mean_token_accuracy": 0.8453701138496399, "num_tokens": 21066488.0, "step": 557 }, { "epoch": 0.07098333545350465, "ewc_loss": 0.040283203125, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.2863540649414062, "learning_rate": 1.1805849936413735e-05, "loss": 0.6006, "mean_token_accuracy": 0.8217946290969849, "num_tokens": 21109049.0, "step": 558 }, { "epoch": 0.07111054573209516, "ewc_loss": 0.039794921875, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.478778123855591, "learning_rate": 1.1827045358202629e-05, "loss": 0.5452, "mean_token_accuracy": 0.8356708884239197, "num_tokens": 21146661.0, "step": 559 }, { "epoch": 0.07123775601068566, "ewc_loss": 0.0390625, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.2857351303100586, "learning_rate": 1.1848240779991522e-05, "loss": 0.5577, "mean_token_accuracy": 0.8307315111160278, "num_tokens": 21179568.0, "step": 560 }, { "epoch": 0.07136496628927617, "ewc_loss": 0.038818359375, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.4052300453186035, "learning_rate": 1.1869436201780416e-05, "loss": 0.5625, "mean_token_accuracy": 0.8305209279060364, "num_tokens": 21216535.0, "step": 561 }, { "epoch": 0.07149217656786669, "ewc_loss": 0.037353515625, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.248544692993164, "learning_rate": 1.1890631623569309e-05, "loss": 0.5214, "mean_token_accuracy": 0.8428173065185547, "num_tokens": 21253887.0, "step": 562 }, { "epoch": 0.07161938684645719, "ewc_loss": 0.03759765625, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.3623204231262207, "learning_rate": 1.1911827045358203e-05, "loss": 0.5306, "mean_token_accuracy": 0.8397001028060913, "num_tokens": 21289813.0, "step": 563 }, { "epoch": 0.0717465971250477, "ewc_loss": 0.03662109375, "ewc_loss_diag": 3.170967102050781e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.2074573040008545, "learning_rate": 1.1933022467147096e-05, "loss": 0.4803, "mean_token_accuracy": 0.8541833162307739, "num_tokens": 21326121.0, "step": 564 }, { "epoch": 0.07187380740363822, "ewc_loss": 0.03759765625, "ewc_loss_diag": 3.170967102050781e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.3443286418914795, "learning_rate": 1.195421788893599e-05, "loss": 0.5747, "mean_token_accuracy": 0.8250327110290527, "num_tokens": 21363644.0, "step": 565 }, { "epoch": 0.07200101768222872, "ewc_loss": 0.0361328125, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.160801649093628, "learning_rate": 1.1975413310724883e-05, "loss": 0.4606, "mean_token_accuracy": 0.8607352375984192, "num_tokens": 21407119.0, "step": 566 }, { "epoch": 0.07212822796081923, "ewc_loss": 0.03759765625, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.347682237625122, "learning_rate": 1.1996608732513778e-05, "loss": 0.5339, "mean_token_accuracy": 0.8407422304153442, "num_tokens": 21447648.0, "step": 567 }, { "epoch": 0.07225543823940975, "ewc_loss": 0.03515625, "ewc_loss_diag": 3.170967102050781e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.154684543609619, "learning_rate": 1.2017804154302672e-05, "loss": 0.5671, "mean_token_accuracy": 0.8345521688461304, "num_tokens": 21488335.0, "step": 568 }, { "epoch": 0.07238264851800025, "ewc_loss": 0.037841796875, "ewc_loss_diag": 3.170967102050781e-05, "ewc_loss_parallel": 3.457069396972656e-05, "grad_norm": 2.3826262950897217, "learning_rate": 1.2038999576091565e-05, "loss": 0.5763, "mean_token_accuracy": 0.824881911277771, "num_tokens": 21525987.0, "step": 569 }, { "epoch": 0.07250985879659076, "ewc_loss": 0.034912109375, "ewc_loss_diag": 3.170967102050781e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.1235787868499756, "learning_rate": 1.2060194997880459e-05, "loss": 0.4699, "mean_token_accuracy": 0.861481249332428, "num_tokens": 21564671.0, "step": 570 }, { "epoch": 0.07263706907518128, "ewc_loss": 0.03857421875, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.4021642208099365, "learning_rate": 1.2081390419669352e-05, "loss": 0.517, "mean_token_accuracy": 0.8412991166114807, "num_tokens": 21608590.0, "step": 571 }, { "epoch": 0.07276427935377179, "ewc_loss": 0.034912109375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.136509656906128, "learning_rate": 1.2102585841458246e-05, "loss": 0.5328, "mean_token_accuracy": 0.841796875, "num_tokens": 21641684.0, "step": 572 }, { "epoch": 0.07289148963236229, "ewc_loss": 0.038818359375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.422344207763672, "learning_rate": 1.2123781263247139e-05, "loss": 0.4732, "mean_token_accuracy": 0.8570296764373779, "num_tokens": 21674923.0, "step": 573 }, { "epoch": 0.0730186999109528, "ewc_loss": 0.034912109375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.124760866165161, "learning_rate": 1.2144976685036033e-05, "loss": 0.5591, "mean_token_accuracy": 0.8321812748908997, "num_tokens": 21714006.0, "step": 574 }, { "epoch": 0.07314591018954332, "ewc_loss": 0.039306640625, "ewc_loss_diag": 3.24249267578125e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.4646003246307373, "learning_rate": 1.2166172106824926e-05, "loss": 0.5172, "mean_token_accuracy": 0.8488669395446777, "num_tokens": 21750580.0, "step": 575 }, { "epoch": 0.07327312046813382, "ewc_loss": 0.03515625, "ewc_loss_diag": 3.24249267578125e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.1248817443847656, "learning_rate": 1.218736752861382e-05, "loss": 0.4927, "mean_token_accuracy": 0.8490829467773438, "num_tokens": 21787400.0, "step": 576 }, { "epoch": 0.07340033074672433, "ewc_loss": 0.040283203125, "ewc_loss_diag": 3.24249267578125e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.5048789978027344, "learning_rate": 1.2208562950402713e-05, "loss": 0.5095, "mean_token_accuracy": 0.8415437936782837, "num_tokens": 21823159.0, "step": 577 }, { "epoch": 0.07352754102531485, "ewc_loss": 0.035888671875, "ewc_loss_diag": 3.24249267578125e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.1433236598968506, "learning_rate": 1.2229758372191608e-05, "loss": 0.5436, "mean_token_accuracy": 0.8360049724578857, "num_tokens": 21858067.0, "step": 578 }, { "epoch": 0.07365475130390535, "ewc_loss": 0.040771484375, "ewc_loss_diag": 3.24249267578125e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.5329997539520264, "learning_rate": 1.22509537939805e-05, "loss": 0.5086, "mean_token_accuracy": 0.8451226949691772, "num_tokens": 21893332.0, "step": 579 }, { "epoch": 0.07378196158249586, "ewc_loss": 0.03662109375, "ewc_loss_diag": 3.24249267578125e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.1693360805511475, "learning_rate": 1.2272149215769395e-05, "loss": 0.6161, "mean_token_accuracy": 0.8137019276618958, "num_tokens": 21932696.0, "step": 580 }, { "epoch": 0.07390917186108638, "ewc_loss": 0.041259765625, "ewc_loss_diag": 3.24249267578125e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.555377244949341, "learning_rate": 1.2293344637558288e-05, "loss": 0.5327, "mean_token_accuracy": 0.832385778427124, "num_tokens": 21964825.0, "step": 581 }, { "epoch": 0.07403638213967688, "ewc_loss": 0.037109375, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 3.3855438232421875e-05, "grad_norm": 2.18182373046875, "learning_rate": 1.2314540059347182e-05, "loss": 0.4885, "mean_token_accuracy": 0.8496143817901611, "num_tokens": 22003396.0, "step": 582 }, { "epoch": 0.0741635924182674, "ewc_loss": 0.04150390625, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.543947696685791, "learning_rate": 1.2335735481136075e-05, "loss": 0.5028, "mean_token_accuracy": 0.8504449725151062, "num_tokens": 22045107.0, "step": 583 }, { "epoch": 0.07429080269685791, "ewc_loss": 0.037841796875, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.2266438007354736, "learning_rate": 1.235693090292497e-05, "loss": 0.4746, "mean_token_accuracy": 0.855263352394104, "num_tokens": 22082754.0, "step": 584 }, { "epoch": 0.07441801297544842, "ewc_loss": 0.041015625, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.5268540382385254, "learning_rate": 1.2378126324713862e-05, "loss": 0.497, "mean_token_accuracy": 0.8503794074058533, "num_tokens": 22122501.0, "step": 585 }, { "epoch": 0.07454522325403892, "ewc_loss": 0.038330078125, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.239492177963257, "learning_rate": 1.2399321746502756e-05, "loss": 0.5309, "mean_token_accuracy": 0.8421027660369873, "num_tokens": 22158110.0, "step": 586 }, { "epoch": 0.07467243353262944, "ewc_loss": 0.041015625, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.489556312561035, "learning_rate": 1.2420517168291649e-05, "loss": 0.4883, "mean_token_accuracy": 0.8486771583557129, "num_tokens": 22190097.0, "step": 587 }, { "epoch": 0.07479964381121995, "ewc_loss": 0.0380859375, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 3.457069396972656e-05, "grad_norm": 2.262974262237549, "learning_rate": 1.2441712590080542e-05, "loss": 0.5535, "mean_token_accuracy": 0.8323662877082825, "num_tokens": 22227568.0, "step": 588 }, { "epoch": 0.07492685408981045, "ewc_loss": 0.0400390625, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.4444079399108887, "learning_rate": 1.2462908011869436e-05, "loss": 0.5193, "mean_token_accuracy": 0.8454967141151428, "num_tokens": 22265376.0, "step": 589 }, { "epoch": 0.07505406436840097, "ewc_loss": 0.037841796875, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.2473342418670654, "learning_rate": 1.248410343365833e-05, "loss": 0.5431, "mean_token_accuracy": 0.8378312587738037, "num_tokens": 22310398.0, "step": 590 }, { "epoch": 0.07518127464699148, "ewc_loss": 0.03955078125, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.4150588512420654, "learning_rate": 1.2505298855447223e-05, "loss": 0.4538, "mean_token_accuracy": 0.86212158203125, "num_tokens": 22341413.0, "step": 591 }, { "epoch": 0.07530848492558198, "ewc_loss": 0.03759765625, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.2501380443573, "learning_rate": 1.2526494277236118e-05, "loss": 0.5216, "mean_token_accuracy": 0.841524064540863, "num_tokens": 22383228.0, "step": 592 }, { "epoch": 0.0754356952041725, "ewc_loss": 0.038818359375, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.3960535526275635, "learning_rate": 1.2547689699025012e-05, "loss": 0.5416, "mean_token_accuracy": 0.8378242254257202, "num_tokens": 22423750.0, "step": 593 }, { "epoch": 0.07556290548276301, "ewc_loss": 0.03759765625, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.276078224182129, "learning_rate": 1.2568885120813907e-05, "loss": 0.5022, "mean_token_accuracy": 0.8478959798812866, "num_tokens": 22454717.0, "step": 594 }, { "epoch": 0.07569011576135352, "ewc_loss": 0.03857421875, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 3.504753112792969e-05, "grad_norm": 2.3546905517578125, "learning_rate": 1.2590080542602798e-05, "loss": 0.5019, "mean_token_accuracy": 0.8490794897079468, "num_tokens": 22492798.0, "step": 595 }, { "epoch": 0.07581732603994402, "ewc_loss": 0.037353515625, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 3.3855438232421875e-05, "grad_norm": 2.2815864086151123, "learning_rate": 1.2611275964391692e-05, "loss": 0.541, "mean_token_accuracy": 0.8371703028678894, "num_tokens": 22530774.0, "step": 596 }, { "epoch": 0.07594453631853454, "ewc_loss": 0.037841796875, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.3331925868988037, "learning_rate": 1.2632471386180587e-05, "loss": 0.5213, "mean_token_accuracy": 0.8424196243286133, "num_tokens": 22566355.0, "step": 597 }, { "epoch": 0.07607174659712505, "ewc_loss": 0.037353515625, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 3.3855438232421875e-05, "grad_norm": 2.277688980102539, "learning_rate": 1.2653666807969481e-05, "loss": 0.5058, "mean_token_accuracy": 0.848125696182251, "num_tokens": 22603633.0, "step": 598 }, { "epoch": 0.07619895687571555, "ewc_loss": 0.037841796875, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.3378138542175293, "learning_rate": 1.2674862229758372e-05, "loss": 0.5378, "mean_token_accuracy": 0.8398703932762146, "num_tokens": 22641481.0, "step": 599 }, { "epoch": 0.07632616715430607, "ewc_loss": 0.037353515625, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 3.3855438232421875e-05, "grad_norm": 2.27626371383667, "learning_rate": 1.2696057651547266e-05, "loss": 0.4801, "mean_token_accuracy": 0.8550744652748108, "num_tokens": 22672853.0, "step": 600 }, { "epoch": 0.07645337743289658, "ewc_loss": 0.037841796875, "ewc_loss_diag": 3.409385681152344e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.3501691818237305, "learning_rate": 1.271725307333616e-05, "loss": 0.5494, "mean_token_accuracy": 0.8320965766906738, "num_tokens": 22704851.0, "step": 601 }, { "epoch": 0.07658058771148708, "ewc_loss": 0.036865234375, "ewc_loss_diag": 3.409385681152344e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.270622491836548, "learning_rate": 1.2738448495125055e-05, "loss": 0.5703, "mean_token_accuracy": 0.8277177810668945, "num_tokens": 22742578.0, "step": 602 }, { "epoch": 0.0767077979900776, "ewc_loss": 0.037841796875, "ewc_loss_diag": 3.4332275390625e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.3399298191070557, "learning_rate": 1.2759643916913946e-05, "loss": 0.5142, "mean_token_accuracy": 0.8424193859100342, "num_tokens": 22783960.0, "step": 603 }, { "epoch": 0.07683500826866811, "ewc_loss": 0.036865234375, "ewc_loss_diag": 3.4332275390625e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.2812461853027344, "learning_rate": 1.278083933870284e-05, "loss": 0.4733, "mean_token_accuracy": 0.8543994426727295, "num_tokens": 22813643.0, "step": 604 }, { "epoch": 0.07696221854725861, "ewc_loss": 0.037841796875, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.349759340286255, "learning_rate": 1.2802034760491735e-05, "loss": 0.5639, "mean_token_accuracy": 0.8278980255126953, "num_tokens": 22850290.0, "step": 605 }, { "epoch": 0.07708942882584913, "ewc_loss": 0.037109375, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.265669822692871, "learning_rate": 1.2823230182280626e-05, "loss": 0.5202, "mean_token_accuracy": 0.8404455184936523, "num_tokens": 22894009.0, "step": 606 }, { "epoch": 0.07721663910443964, "ewc_loss": 0.037841796875, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.3734211921691895, "learning_rate": 1.284442560406952e-05, "loss": 0.5129, "mean_token_accuracy": 0.8460656404495239, "num_tokens": 22931418.0, "step": 607 }, { "epoch": 0.07734384938303016, "ewc_loss": 0.03662109375, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.2525758743286133, "learning_rate": 1.2865621025858415e-05, "loss": 0.5084, "mean_token_accuracy": 0.8496602773666382, "num_tokens": 22970074.0, "step": 608 }, { "epoch": 0.07747105966162066, "ewc_loss": 0.038818359375, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.4150876998901367, "learning_rate": 1.288681644764731e-05, "loss": 0.5425, "mean_token_accuracy": 0.8382280468940735, "num_tokens": 23008185.0, "step": 609 }, { "epoch": 0.07759826994021117, "ewc_loss": 0.03662109375, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.2224550247192383, "learning_rate": 1.29080118694362e-05, "loss": 0.4595, "mean_token_accuracy": 0.8591793775558472, "num_tokens": 23039667.0, "step": 610 }, { "epoch": 0.07772548021880168, "ewc_loss": 0.0390625, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.479665517807007, "learning_rate": 1.2929207291225095e-05, "loss": 0.5273, "mean_token_accuracy": 0.842164158821106, "num_tokens": 23075012.0, "step": 611 }, { "epoch": 0.07785269049739219, "ewc_loss": 0.03564453125, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.1850743293762207, "learning_rate": 1.295040271301399e-05, "loss": 0.459, "mean_token_accuracy": 0.8588827848434448, "num_tokens": 23108397.0, "step": 612 }, { "epoch": 0.0779799007759827, "ewc_loss": 0.041015625, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.5589406490325928, "learning_rate": 1.2971598134802884e-05, "loss": 0.5395, "mean_token_accuracy": 0.8428221344947815, "num_tokens": 23146419.0, "step": 613 }, { "epoch": 0.07810711105457321, "ewc_loss": 0.03564453125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.122175693511963, "learning_rate": 1.2992793556591776e-05, "loss": 0.5349, "mean_token_accuracy": 0.8402777910232544, "num_tokens": 23191750.0, "step": 614 }, { "epoch": 0.07823432133316371, "ewc_loss": 0.045166015625, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.7152554988861084, "learning_rate": 1.3013988978380671e-05, "loss": 0.5498, "mean_token_accuracy": 0.8352099657058716, "num_tokens": 23230199.0, "step": 615 }, { "epoch": 0.07836153161175423, "ewc_loss": 0.036376953125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.1105337142944336, "learning_rate": 1.3035184400169564e-05, "loss": 0.5274, "mean_token_accuracy": 0.8339582681655884, "num_tokens": 23264184.0, "step": 616 }, { "epoch": 0.07848874189034474, "ewc_loss": 0.051025390625, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 2.9215896129608154, "learning_rate": 1.3056379821958458e-05, "loss": 0.4969, "mean_token_accuracy": 0.8527042865753174, "num_tokens": 23295543.0, "step": 617 }, { "epoch": 0.07861595216893524, "ewc_loss": 0.0390625, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.188917636871338, "learning_rate": 1.307757524374735e-05, "loss": 0.5089, "mean_token_accuracy": 0.8467577695846558, "num_tokens": 23328148.0, "step": 618 }, { "epoch": 0.07874316244752576, "ewc_loss": 0.059326171875, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.093113422393799, "learning_rate": 1.3098770665536245e-05, "loss": 0.5241, "mean_token_accuracy": 0.8476230502128601, "num_tokens": 23369441.0, "step": 619 }, { "epoch": 0.07887037272611627, "ewc_loss": 0.0498046875, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 2.4159371852874756, "learning_rate": 1.311996608732514e-05, "loss": 0.5702, "mean_token_accuracy": 0.8294765949249268, "num_tokens": 23402977.0, "step": 620 }, { "epoch": 0.07899758300470679, "ewc_loss": 0.060302734375, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.1421520709991455, "learning_rate": 1.3141161509114034e-05, "loss": 0.5538, "mean_token_accuracy": 0.8416270017623901, "num_tokens": 23441988.0, "step": 621 }, { "epoch": 0.07912479328329729, "ewc_loss": 0.05615234375, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 2.611116409301758, "learning_rate": 1.3162356930902925e-05, "loss": 0.5817, "mean_token_accuracy": 0.8300352692604065, "num_tokens": 23480481.0, "step": 622 }, { "epoch": 0.0792520035618878, "ewc_loss": 0.056640625, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.0562667846679688, "learning_rate": 1.318355235269182e-05, "loss": 0.5807, "mean_token_accuracy": 0.828583300113678, "num_tokens": 23515188.0, "step": 623 }, { "epoch": 0.07937921384047832, "ewc_loss": 0.05615234375, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 2.6749041080474854, "learning_rate": 1.3204747774480714e-05, "loss": 0.5645, "mean_token_accuracy": 0.8369776606559753, "num_tokens": 23557777.0, "step": 624 }, { "epoch": 0.07950642411906882, "ewc_loss": 0.05224609375, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 2.920295476913452, "learning_rate": 1.3225943196269605e-05, "loss": 0.5483, "mean_token_accuracy": 0.8420222401618958, "num_tokens": 23588357.0, "step": 625 }, { "epoch": 0.07963363439765933, "ewc_loss": 0.0517578125, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 2.62300968170166, "learning_rate": 1.32471386180585e-05, "loss": 0.5137, "mean_token_accuracy": 0.8449300527572632, "num_tokens": 23628656.0, "step": 626 }, { "epoch": 0.07976084467624985, "ewc_loss": 0.05029296875, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 2.8056421279907227, "learning_rate": 1.3268334039847394e-05, "loss": 0.6093, "mean_token_accuracy": 0.8199599981307983, "num_tokens": 23666390.0, "step": 627 }, { "epoch": 0.07988805495484035, "ewc_loss": 0.0478515625, "ewc_loss_diag": 3.5762786865234375e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.558379888534546, "learning_rate": 1.3289529461636288e-05, "loss": 0.4698, "mean_token_accuracy": 0.8637461066246033, "num_tokens": 23699620.0, "step": 628 }, { "epoch": 0.08001526523343086, "ewc_loss": 0.04736328125, "ewc_loss_diag": 3.5762786865234375e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.713660717010498, "learning_rate": 1.331072488342518e-05, "loss": 0.5493, "mean_token_accuracy": 0.837895393371582, "num_tokens": 23735883.0, "step": 629 }, { "epoch": 0.08014247551202137, "ewc_loss": 0.04443359375, "ewc_loss_diag": 3.5762786865234375e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.4846787452697754, "learning_rate": 1.3331920305214074e-05, "loss": 0.5045, "mean_token_accuracy": 0.8497719764709473, "num_tokens": 23769918.0, "step": 630 }, { "epoch": 0.08026968579061187, "ewc_loss": 0.046142578125, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.6494691371917725, "learning_rate": 1.3353115727002968e-05, "loss": 0.5135, "mean_token_accuracy": 0.8473055362701416, "num_tokens": 23815402.0, "step": 631 }, { "epoch": 0.08039689606920239, "ewc_loss": 0.0419921875, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.4086203575134277, "learning_rate": 1.3374311148791863e-05, "loss": 0.5106, "mean_token_accuracy": 0.8481953144073486, "num_tokens": 23858205.0, "step": 632 }, { "epoch": 0.0805241063477929, "ewc_loss": 0.044921875, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.660618305206299, "learning_rate": 1.3395506570580754e-05, "loss": 0.5249, "mean_token_accuracy": 0.8463462591171265, "num_tokens": 23893506.0, "step": 633 }, { "epoch": 0.08065131662638342, "ewc_loss": 0.040283203125, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.369662284851074, "learning_rate": 1.3416701992369648e-05, "loss": 0.5129, "mean_token_accuracy": 0.8431970477104187, "num_tokens": 23929457.0, "step": 634 }, { "epoch": 0.08077852690497392, "ewc_loss": 0.04541015625, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.668776273727417, "learning_rate": 1.3437897414158542e-05, "loss": 0.5109, "mean_token_accuracy": 0.8504970073699951, "num_tokens": 23974577.0, "step": 635 }, { "epoch": 0.08090573718356443, "ewc_loss": 0.039794921875, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.352700710296631, "learning_rate": 1.3459092835947437e-05, "loss": 0.5801, "mean_token_accuracy": 0.8288741111755371, "num_tokens": 24007490.0, "step": 636 }, { "epoch": 0.08103294746215495, "ewc_loss": 0.045654296875, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.710803985595703, "learning_rate": 1.348028825773633e-05, "loss": 0.6242, "mean_token_accuracy": 0.8133798837661743, "num_tokens": 24048058.0, "step": 637 }, { "epoch": 0.08116015774074545, "ewc_loss": 0.039794921875, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.305246353149414, "learning_rate": 1.3501483679525222e-05, "loss": 0.4846, "mean_token_accuracy": 0.8531073331832886, "num_tokens": 24089776.0, "step": 638 }, { "epoch": 0.08128736801933596, "ewc_loss": 0.049072265625, "ewc_loss_diag": 3.647804260253906e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 2.761059522628784, "learning_rate": 1.3522679101314117e-05, "loss": 0.5444, "mean_token_accuracy": 0.8345108032226562, "num_tokens": 24135730.0, "step": 639 }, { "epoch": 0.08141457829792648, "ewc_loss": 0.04052734375, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 2.3419041633605957, "learning_rate": 1.3543874523103011e-05, "loss": 0.4937, "mean_token_accuracy": 0.8522714376449585, "num_tokens": 24168513.0, "step": 640 }, { "epoch": 0.08154178857651698, "ewc_loss": 0.049560546875, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 2.803696632385254, "learning_rate": 1.3565069944891904e-05, "loss": 0.5683, "mean_token_accuracy": 0.830880880355835, "num_tokens": 24206555.0, "step": 641 }, { "epoch": 0.08166899885510749, "ewc_loss": 0.042236328125, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.3626608848571777, "learning_rate": 1.3586265366680798e-05, "loss": 0.467, "mean_token_accuracy": 0.8598446846008301, "num_tokens": 24248919.0, "step": 642 }, { "epoch": 0.081796209133698, "ewc_loss": 0.05029296875, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 2.797001361846924, "learning_rate": 1.3607460788469693e-05, "loss": 0.4699, "mean_token_accuracy": 0.8619645833969116, "num_tokens": 24290493.0, "step": 643 }, { "epoch": 0.0819234194122885, "ewc_loss": 0.0439453125, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.4280898571014404, "learning_rate": 1.3628656210258584e-05, "loss": 0.5163, "mean_token_accuracy": 0.8479725122451782, "num_tokens": 24327968.0, "step": 644 }, { "epoch": 0.08205062969087902, "ewc_loss": 0.049560546875, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 2.7494781017303467, "learning_rate": 1.3649851632047478e-05, "loss": 0.5558, "mean_token_accuracy": 0.8381307721138, "num_tokens": 24370019.0, "step": 645 }, { "epoch": 0.08217783996946953, "ewc_loss": 0.045166015625, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.4844398498535156, "learning_rate": 1.3671047053836373e-05, "loss": 0.5222, "mean_token_accuracy": 0.8482580780982971, "num_tokens": 24403911.0, "step": 646 }, { "epoch": 0.08230505024806005, "ewc_loss": 0.045654296875, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.6518054008483887, "learning_rate": 1.3692242475625267e-05, "loss": 0.5404, "mean_token_accuracy": 0.8399877548217773, "num_tokens": 24445039.0, "step": 647 }, { "epoch": 0.08243226052665055, "ewc_loss": 0.0439453125, "ewc_loss_diag": 3.695487976074219e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.5257232189178467, "learning_rate": 1.3713437897414158e-05, "loss": 0.5369, "mean_token_accuracy": 0.836794912815094, "num_tokens": 24479543.0, "step": 648 }, { "epoch": 0.08255947080524106, "ewc_loss": 0.042724609375, "ewc_loss_diag": 3.695487976074219e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.5145864486694336, "learning_rate": 1.3734633319203052e-05, "loss": 0.4976, "mean_token_accuracy": 0.8500096797943115, "num_tokens": 24517483.0, "step": 649 }, { "epoch": 0.08268668108383158, "ewc_loss": 0.043212890625, "ewc_loss_diag": 3.695487976074219e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.5410397052764893, "learning_rate": 1.3755828740991947e-05, "loss": 0.5327, "mean_token_accuracy": 0.8391494750976562, "num_tokens": 24557702.0, "step": 650 }, { "epoch": 0.08281389136242208, "ewc_loss": 0.041015625, "ewc_loss_diag": 3.695487976074219e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.430201292037964, "learning_rate": 1.3777024162780841e-05, "loss": 0.531, "mean_token_accuracy": 0.84104323387146, "num_tokens": 24595145.0, "step": 651 }, { "epoch": 0.0829411016410126, "ewc_loss": 0.042724609375, "ewc_loss_diag": 3.719329833984375e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.5682618618011475, "learning_rate": 1.3798219584569732e-05, "loss": 0.5217, "mean_token_accuracy": 0.8424680829048157, "num_tokens": 24631564.0, "step": 652 }, { "epoch": 0.08306831191960311, "ewc_loss": 0.040283203125, "ewc_loss_diag": 3.719329833984375e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.385678291320801, "learning_rate": 1.3819415006358627e-05, "loss": 0.5758, "mean_token_accuracy": 0.8269063830375671, "num_tokens": 24668563.0, "step": 653 }, { "epoch": 0.08319552219819361, "ewc_loss": 0.04296875, "ewc_loss_diag": 3.719329833984375e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.607215642929077, "learning_rate": 1.3840610428147521e-05, "loss": 0.563, "mean_token_accuracy": 0.8288559913635254, "num_tokens": 24703662.0, "step": 654 }, { "epoch": 0.08332273247678412, "ewc_loss": 0.03955078125, "ewc_loss_diag": 3.719329833984375e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.3382115364074707, "learning_rate": 1.3861805849936416e-05, "loss": 0.5154, "mean_token_accuracy": 0.8480405807495117, "num_tokens": 24739504.0, "step": 655 }, { "epoch": 0.08344994275537464, "ewc_loss": 0.0439453125, "ewc_loss_diag": 3.719329833984375e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.648756980895996, "learning_rate": 1.3883001271725307e-05, "loss": 0.528, "mean_token_accuracy": 0.8399295806884766, "num_tokens": 24781884.0, "step": 656 }, { "epoch": 0.08357715303396514, "ewc_loss": 0.03955078125, "ewc_loss_diag": 3.743171691894531e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.313037157058716, "learning_rate": 1.3904196693514201e-05, "loss": 0.5763, "mean_token_accuracy": 0.826179563999176, "num_tokens": 24823964.0, "step": 657 }, { "epoch": 0.08370436331255565, "ewc_loss": 0.04638671875, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 2.7291884422302246, "learning_rate": 1.3925392115303095e-05, "loss": 0.5255, "mean_token_accuracy": 0.8435892462730408, "num_tokens": 24859552.0, "step": 658 }, { "epoch": 0.08383157359114617, "ewc_loss": 0.039306640625, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.2960376739501953, "learning_rate": 1.394658753709199e-05, "loss": 0.5363, "mean_token_accuracy": 0.8376787900924683, "num_tokens": 24897580.0, "step": 659 }, { "epoch": 0.08395878386973668, "ewc_loss": 0.04931640625, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 2.8111572265625, "learning_rate": 1.3967782958880881e-05, "loss": 0.4946, "mean_token_accuracy": 0.8507101535797119, "num_tokens": 24934616.0, "step": 660 }, { "epoch": 0.08408599414832718, "ewc_loss": 0.039794921875, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.3046154975891113, "learning_rate": 1.3988978380669775e-05, "loss": 0.5554, "mean_token_accuracy": 0.8319264650344849, "num_tokens": 24972149.0, "step": 661 }, { "epoch": 0.0842132044269177, "ewc_loss": 0.0517578125, "ewc_loss_diag": 3.790855407714844e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 2.895179271697998, "learning_rate": 1.401017380245867e-05, "loss": 0.455, "mean_token_accuracy": 0.8672078847885132, "num_tokens": 25008682.0, "step": 662 }, { "epoch": 0.08434041470550821, "ewc_loss": 0.0419921875, "ewc_loss_diag": 3.790855407714844e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.3184213638305664, "learning_rate": 1.4031369224247563e-05, "loss": 0.4662, "mean_token_accuracy": 0.8582110404968262, "num_tokens": 25048454.0, "step": 663 }, { "epoch": 0.08446762498409871, "ewc_loss": 0.056640625, "ewc_loss_diag": 3.790855407714844e-05, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 2.9976227283477783, "learning_rate": 1.4052564646036457e-05, "loss": 0.5589, "mean_token_accuracy": 0.8303146958351135, "num_tokens": 25086772.0, "step": 664 }, { "epoch": 0.08459483526268922, "ewc_loss": 0.04736328125, "ewc_loss_diag": 3.814697265625e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.404578924179077, "learning_rate": 1.4073760067825351e-05, "loss": 0.5034, "mean_token_accuracy": 0.8471320867538452, "num_tokens": 25122215.0, "step": 665 }, { "epoch": 0.08472204554127974, "ewc_loss": 0.05859375, "ewc_loss_diag": 3.814697265625e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.0328540802001953, "learning_rate": 1.4094955489614246e-05, "loss": 0.5575, "mean_token_accuracy": 0.8412795662879944, "num_tokens": 25159966.0, "step": 666 }, { "epoch": 0.08484925581987024, "ewc_loss": 0.05029296875, "ewc_loss_diag": 3.814697265625e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 2.5165889263153076, "learning_rate": 1.4116150911403137e-05, "loss": 0.5004, "mean_token_accuracy": 0.8532058000564575, "num_tokens": 25191929.0, "step": 667 }, { "epoch": 0.08497646609846075, "ewc_loss": 0.056884765625, "ewc_loss_diag": 3.814697265625e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 2.9687016010284424, "learning_rate": 1.4137346333192031e-05, "loss": 0.5269, "mean_token_accuracy": 0.8439103364944458, "num_tokens": 25233385.0, "step": 668 }, { "epoch": 0.08510367637705127, "ewc_loss": 0.052490234375, "ewc_loss_diag": 3.814697265625e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 2.5688445568084717, "learning_rate": 1.4158541754980926e-05, "loss": 0.5403, "mean_token_accuracy": 0.841446042060852, "num_tokens": 25276943.0, "step": 669 }, { "epoch": 0.08523088665564178, "ewc_loss": 0.054443359375, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 2.864393472671509, "learning_rate": 1.417973717676982e-05, "loss": 0.5263, "mean_token_accuracy": 0.8460016250610352, "num_tokens": 25316552.0, "step": 670 }, { "epoch": 0.08535809693423228, "ewc_loss": 0.05078125, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 2.6121928691864014, "learning_rate": 1.4200932598558711e-05, "loss": 0.5721, "mean_token_accuracy": 0.8358231782913208, "num_tokens": 25357517.0, "step": 671 }, { "epoch": 0.0854853072128228, "ewc_loss": 0.04931640625, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 2.7510530948638916, "learning_rate": 1.4222128020347606e-05, "loss": 0.5635, "mean_token_accuracy": 0.8336212635040283, "num_tokens": 25392693.0, "step": 672 }, { "epoch": 0.08561251749141331, "ewc_loss": 0.0498046875, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 2.6117968559265137, "learning_rate": 1.42433234421365e-05, "loss": 0.5259, "mean_token_accuracy": 0.8435800075531006, "num_tokens": 25430623.0, "step": 673 }, { "epoch": 0.08573972777000381, "ewc_loss": 0.04638671875, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.621229410171509, "learning_rate": 1.4264518863925394e-05, "loss": 0.4625, "mean_token_accuracy": 0.8618945479393005, "num_tokens": 25465043.0, "step": 674 }, { "epoch": 0.08586693804859433, "ewc_loss": 0.047607421875, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.604665517807007, "learning_rate": 1.4285714285714285e-05, "loss": 0.4857, "mean_token_accuracy": 0.8565373420715332, "num_tokens": 25503483.0, "step": 675 }, { "epoch": 0.08599414832718484, "ewc_loss": 0.04443359375, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.533289670944214, "learning_rate": 1.430690970750318e-05, "loss": 0.532, "mean_token_accuracy": 0.8404629230499268, "num_tokens": 25547079.0, "step": 676 }, { "epoch": 0.08612135860577534, "ewc_loss": 0.04541015625, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.6370859146118164, "learning_rate": 1.4328105129292074e-05, "loss": 0.5721, "mean_token_accuracy": 0.8265453577041626, "num_tokens": 25586494.0, "step": 677 }, { "epoch": 0.08624856888436586, "ewc_loss": 0.04248046875, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.496647596359253, "learning_rate": 1.4349300551080965e-05, "loss": 0.5709, "mean_token_accuracy": 0.8267104625701904, "num_tokens": 25618998.0, "step": 678 }, { "epoch": 0.08637577916295637, "ewc_loss": 0.044677734375, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.622826099395752, "learning_rate": 1.437049597286986e-05, "loss": 0.5115, "mean_token_accuracy": 0.8468815088272095, "num_tokens": 25663147.0, "step": 679 }, { "epoch": 0.08650298944154687, "ewc_loss": 0.0419921875, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.4657578468322754, "learning_rate": 1.4391691394658754e-05, "loss": 0.5553, "mean_token_accuracy": 0.8381397128105164, "num_tokens": 25696830.0, "step": 680 }, { "epoch": 0.08663019972013739, "ewc_loss": 0.044921875, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.642110824584961, "learning_rate": 1.4412886816447649e-05, "loss": 0.4896, "mean_token_accuracy": 0.8576260805130005, "num_tokens": 25732828.0, "step": 681 }, { "epoch": 0.0867574099987279, "ewc_loss": 0.0419921875, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.4288129806518555, "learning_rate": 1.443408223823654e-05, "loss": 0.475, "mean_token_accuracy": 0.855957567691803, "num_tokens": 25773545.0, "step": 682 }, { "epoch": 0.08688462027731841, "ewc_loss": 0.046142578125, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.6792378425598145, "learning_rate": 1.4455277660025434e-05, "loss": 0.5221, "mean_token_accuracy": 0.8445709347724915, "num_tokens": 25814792.0, "step": 683 }, { "epoch": 0.08701183055590891, "ewc_loss": 0.040771484375, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 2.424814462661743, "learning_rate": 1.4476473081814328e-05, "loss": 0.527, "mean_token_accuracy": 0.8388819098472595, "num_tokens": 25850713.0, "step": 684 }, { "epoch": 0.08713904083449943, "ewc_loss": 0.04736328125, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.726168394088745, "learning_rate": 1.4497668503603223e-05, "loss": 0.5065, "mean_token_accuracy": 0.8455377221107483, "num_tokens": 25888320.0, "step": 685 }, { "epoch": 0.08726625111308994, "ewc_loss": 0.04052734375, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.386143207550049, "learning_rate": 1.4518863925392116e-05, "loss": 0.5418, "mean_token_accuracy": 0.8352312445640564, "num_tokens": 25924941.0, "step": 686 }, { "epoch": 0.08739346139168044, "ewc_loss": 0.0498046875, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 2.8028011322021484, "learning_rate": 1.454005934718101e-05, "loss": 0.5168, "mean_token_accuracy": 0.8459609746932983, "num_tokens": 25964051.0, "step": 687 }, { "epoch": 0.08752067167027096, "ewc_loss": 0.04052734375, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.3341948986053467, "learning_rate": 1.4561254768969903e-05, "loss": 0.5082, "mean_token_accuracy": 0.8472280502319336, "num_tokens": 26007300.0, "step": 688 }, { "epoch": 0.08764788194886147, "ewc_loss": 0.053955078125, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 2.94407057762146, "learning_rate": 1.4582450190758797e-05, "loss": 0.579, "mean_token_accuracy": 0.8259358406066895, "num_tokens": 26045833.0, "step": 689 }, { "epoch": 0.08777509222745197, "ewc_loss": 0.041748046875, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.3110525608062744, "learning_rate": 1.460364561254769e-05, "loss": 0.5607, "mean_token_accuracy": 0.8314614295959473, "num_tokens": 26092277.0, "step": 690 }, { "epoch": 0.08790230250604249, "ewc_loss": 0.061279296875, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.1252808570861816, "learning_rate": 1.4624841034336584e-05, "loss": 0.5357, "mean_token_accuracy": 0.8431162238121033, "num_tokens": 26130082.0, "step": 691 }, { "epoch": 0.088029512784633, "ewc_loss": 0.047607421875, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.369198799133301, "learning_rate": 1.4646036456125479e-05, "loss": 0.4956, "mean_token_accuracy": 0.8549835085868835, "num_tokens": 26170087.0, "step": 692 }, { "epoch": 0.0881567230632235, "ewc_loss": 0.06884765625, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.301003932952881, "learning_rate": 1.4667231877914373e-05, "loss": 0.5955, "mean_token_accuracy": 0.8281407356262207, "num_tokens": 26207645.0, "step": 693 }, { "epoch": 0.08828393334181402, "ewc_loss": 0.056640625, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 2.572355031967163, "learning_rate": 1.4688427299703264e-05, "loss": 0.5108, "mean_token_accuracy": 0.8476517796516418, "num_tokens": 26242417.0, "step": 694 }, { "epoch": 0.08841114362040453, "ewc_loss": 0.0693359375, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.344162940979004, "learning_rate": 1.4709622721492159e-05, "loss": 0.587, "mean_token_accuracy": 0.8322681188583374, "num_tokens": 26275427.0, "step": 695 }, { "epoch": 0.08853835389899505, "ewc_loss": 0.0654296875, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 2.7645812034606934, "learning_rate": 1.4730818143281053e-05, "loss": 0.5326, "mean_token_accuracy": 0.8461273908615112, "num_tokens": 26308692.0, "step": 696 }, { "epoch": 0.08866556417758555, "ewc_loss": 0.06494140625, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.2013206481933594, "learning_rate": 1.4752013565069944e-05, "loss": 0.5283, "mean_token_accuracy": 0.848081111907959, "num_tokens": 26352155.0, "step": 697 }, { "epoch": 0.08879277445617606, "ewc_loss": 0.0654296875, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 2.8147125244140625, "learning_rate": 1.4773208986858839e-05, "loss": 0.5735, "mean_token_accuracy": 0.8350479006767273, "num_tokens": 26392410.0, "step": 698 }, { "epoch": 0.08891998473476657, "ewc_loss": 0.0595703125, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.0415115356445312, "learning_rate": 1.4794404408647733e-05, "loss": 0.4947, "mean_token_accuracy": 0.8557920455932617, "num_tokens": 26430120.0, "step": 699 }, { "epoch": 0.08904719501335707, "ewc_loss": 0.060546875, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 2.791055679321289, "learning_rate": 1.4815599830436627e-05, "loss": 0.5564, "mean_token_accuracy": 0.8365797996520996, "num_tokens": 26466159.0, "step": 700 }, { "epoch": 0.08917440529194759, "ewc_loss": 0.05615234375, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 2.883171796798706, "learning_rate": 1.4836795252225518e-05, "loss": 0.474, "mean_token_accuracy": 0.8629308938980103, "num_tokens": 26502382.0, "step": 701 }, { "epoch": 0.0893016155705381, "ewc_loss": 0.055908203125, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 2.7401978969573975, "learning_rate": 1.4857990674014413e-05, "loss": 0.5131, "mean_token_accuracy": 0.8502917885780334, "num_tokens": 26540911.0, "step": 702 }, { "epoch": 0.0894288258491286, "ewc_loss": 0.05126953125, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 2.7723562717437744, "learning_rate": 1.4879186095803307e-05, "loss": 0.5691, "mean_token_accuracy": 0.8295748829841614, "num_tokens": 26580081.0, "step": 703 }, { "epoch": 0.08955603612771912, "ewc_loss": 0.051025390625, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 2.6938626766204834, "learning_rate": 1.4900381517592202e-05, "loss": 0.5356, "mean_token_accuracy": 0.8409260511398315, "num_tokens": 26622033.0, "step": 704 }, { "epoch": 0.08968324640630963, "ewc_loss": 0.048583984375, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 2.7115204334259033, "learning_rate": 1.4921576939381093e-05, "loss": 0.5332, "mean_token_accuracy": 0.8402666449546814, "num_tokens": 26657042.0, "step": 705 }, { "epoch": 0.08981045668490013, "ewc_loss": 0.048583984375, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 2.657155752182007, "learning_rate": 1.4942772361169987e-05, "loss": 0.5457, "mean_token_accuracy": 0.8396548628807068, "num_tokens": 26696668.0, "step": 706 }, { "epoch": 0.08993766696349065, "ewc_loss": 0.046875, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.670305013656616, "learning_rate": 1.4963967782958882e-05, "loss": 0.5406, "mean_token_accuracy": 0.8396863341331482, "num_tokens": 26736550.0, "step": 707 }, { "epoch": 0.09006487724208116, "ewc_loss": 0.046875, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.630131244659424, "learning_rate": 1.4985163204747776e-05, "loss": 0.5483, "mean_token_accuracy": 0.8367756605148315, "num_tokens": 26776311.0, "step": 708 }, { "epoch": 0.09019208752067168, "ewc_loss": 0.0458984375, "ewc_loss_diag": 3.9577484130859375e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.7089016437530518, "learning_rate": 1.5006358626536669e-05, "loss": 0.4903, "mean_token_accuracy": 0.8536071181297302, "num_tokens": 26812094.0, "step": 709 }, { "epoch": 0.09031929779926218, "ewc_loss": 0.044189453125, "ewc_loss_diag": 3.981590270996094e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.5767998695373535, "learning_rate": 1.5027554048325561e-05, "loss": 0.5228, "mean_token_accuracy": 0.8447955846786499, "num_tokens": 26846317.0, "step": 710 }, { "epoch": 0.09044650807785269, "ewc_loss": 0.0478515625, "ewc_loss_diag": 3.981590270996094e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 2.699028491973877, "learning_rate": 1.5048749470114456e-05, "loss": 0.4709, "mean_token_accuracy": 0.8598427772521973, "num_tokens": 26885247.0, "step": 711 }, { "epoch": 0.0905737183564432, "ewc_loss": 0.043212890625, "ewc_loss_diag": 3.981590270996094e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.5509393215179443, "learning_rate": 1.506994489190335e-05, "loss": 0.5276, "mean_token_accuracy": 0.8428975939750671, "num_tokens": 26925578.0, "step": 712 }, { "epoch": 0.0907009286350337, "ewc_loss": 0.04833984375, "ewc_loss_diag": 3.981590270996094e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 2.7382826805114746, "learning_rate": 1.5091140313692243e-05, "loss": 0.5577, "mean_token_accuracy": 0.8380361795425415, "num_tokens": 26961924.0, "step": 713 }, { "epoch": 0.09082813891362422, "ewc_loss": 0.04248046875, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.5009043216705322, "learning_rate": 1.5112335735481137e-05, "loss": 0.4811, "mean_token_accuracy": 0.8538221716880798, "num_tokens": 26999433.0, "step": 714 }, { "epoch": 0.09095534919221474, "ewc_loss": 0.050048828125, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 2.7905263900756836, "learning_rate": 1.5133531157270032e-05, "loss": 0.5102, "mean_token_accuracy": 0.8496412038803101, "num_tokens": 27044427.0, "step": 715 }, { "epoch": 0.09108255947080524, "ewc_loss": 0.0419921875, "ewc_loss_diag": 4.029273986816406e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.4582417011260986, "learning_rate": 1.5154726579058923e-05, "loss": 0.5084, "mean_token_accuracy": 0.8462958335876465, "num_tokens": 27080645.0, "step": 716 }, { "epoch": 0.09120976974939575, "ewc_loss": 0.05224609375, "ewc_loss_diag": 4.029273986816406e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.1872572898864746, "learning_rate": 1.5175922000847817e-05, "loss": 0.5814, "mean_token_accuracy": 0.8298933506011963, "num_tokens": 27118213.0, "step": 717 }, { "epoch": 0.09133698002798626, "ewc_loss": 0.04248046875, "ewc_loss_diag": 4.029273986816406e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.3335113525390625, "learning_rate": 1.5197117422636712e-05, "loss": 0.4922, "mean_token_accuracy": 0.8514389395713806, "num_tokens": 27156827.0, "step": 718 }, { "epoch": 0.09146419030657676, "ewc_loss": 0.06982421875, "ewc_loss_diag": 4.029273986816406e-05, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.4139180183410645, "learning_rate": 1.5218312844425606e-05, "loss": 0.5549, "mean_token_accuracy": 0.8375048041343689, "num_tokens": 27192016.0, "step": 719 }, { "epoch": 0.09159140058516728, "ewc_loss": 0.0517578125, "ewc_loss_diag": 4.076957702636719e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 2.4666731357574463, "learning_rate": 1.5239508266214497e-05, "loss": 0.5761, "mean_token_accuracy": 0.8269884586334229, "num_tokens": 27226037.0, "step": 720 }, { "epoch": 0.0917186108637578, "ewc_loss": 0.07763671875, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 7.390975952148438e-05, "grad_norm": 3.6913881301879883, "learning_rate": 1.526070368800339e-05, "loss": 0.5243, "mean_token_accuracy": 0.8512665033340454, "num_tokens": 27260399.0, "step": 721 }, { "epoch": 0.09184582114234831, "ewc_loss": 0.06982421875, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 2.8579049110412598, "learning_rate": 1.5281899109792286e-05, "loss": 0.5267, "mean_token_accuracy": 0.8515604734420776, "num_tokens": 27297293.0, "step": 722 }, { "epoch": 0.09197303142093881, "ewc_loss": 0.076171875, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 7.2479248046875e-05, "grad_norm": 3.676038980484009, "learning_rate": 1.530309453158118e-05, "loss": 0.5237, "mean_token_accuracy": 0.8525338172912598, "num_tokens": 27333045.0, "step": 723 }, { "epoch": 0.09210024169952932, "ewc_loss": 0.0751953125, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 7.152557373046875e-05, "grad_norm": 3.1317548751831055, "learning_rate": 1.532428995337007e-05, "loss": 0.5226, "mean_token_accuracy": 0.8489104509353638, "num_tokens": 27369044.0, "step": 724 }, { "epoch": 0.09222745197811984, "ewc_loss": 0.07080078125, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.4704270362854004, "learning_rate": 1.5345485375158966e-05, "loss": 0.5746, "mean_token_accuracy": 0.8357261419296265, "num_tokens": 27407486.0, "step": 725 }, { "epoch": 0.09235466225671034, "ewc_loss": 0.07177734375, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 3.104098320007324, "learning_rate": 1.536668079694786e-05, "loss": 0.5365, "mean_token_accuracy": 0.8475252389907837, "num_tokens": 27441835.0, "step": 726 }, { "epoch": 0.09248187253530085, "ewc_loss": 0.06640625, "ewc_loss_diag": 4.124641418457031e-05, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.2792303562164307, "learning_rate": 1.5387876218736755e-05, "loss": 0.5489, "mean_token_accuracy": 0.83992999792099, "num_tokens": 27483105.0, "step": 727 }, { "epoch": 0.09260908281389137, "ewc_loss": 0.06640625, "ewc_loss_diag": 4.124641418457031e-05, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 2.968876600265503, "learning_rate": 1.5409071640525646e-05, "loss": 0.565, "mean_token_accuracy": 0.8394113779067993, "num_tokens": 27528888.0, "step": 728 }, { "epoch": 0.09273629309248187, "ewc_loss": 0.06298828125, "ewc_loss_diag": 4.124641418457031e-05, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.1661720275878906, "learning_rate": 1.543026706231454e-05, "loss": 0.6046, "mean_token_accuracy": 0.8241778016090393, "num_tokens": 27571763.0, "step": 729 }, { "epoch": 0.09286350337107238, "ewc_loss": 0.059326171875, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 2.852774143218994, "learning_rate": 1.5451462484103435e-05, "loss": 0.5263, "mean_token_accuracy": 0.8484849333763123, "num_tokens": 27606980.0, "step": 730 }, { "epoch": 0.0929907136496629, "ewc_loss": 0.06103515625, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.1169662475585938, "learning_rate": 1.547265790589233e-05, "loss": 0.5258, "mean_token_accuracy": 0.8455625772476196, "num_tokens": 27640969.0, "step": 731 }, { "epoch": 0.0931179239282534, "ewc_loss": 0.05419921875, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 2.7416653633117676, "learning_rate": 1.549385332768122e-05, "loss": 0.6478, "mean_token_accuracy": 0.806260347366333, "num_tokens": 27684360.0, "step": 732 }, { "epoch": 0.09324513420684391, "ewc_loss": 0.060791015625, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.091123580932617, "learning_rate": 1.5515048749470115e-05, "loss": 0.4983, "mean_token_accuracy": 0.8534492254257202, "num_tokens": 27723596.0, "step": 733 }, { "epoch": 0.09337234448543442, "ewc_loss": 0.05078125, "ewc_loss_diag": 4.1961669921875e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 2.6615965366363525, "learning_rate": 1.553624417125901e-05, "loss": 0.5237, "mean_token_accuracy": 0.8400006294250488, "num_tokens": 27759469.0, "step": 734 }, { "epoch": 0.09349955476402494, "ewc_loss": 0.0625, "ewc_loss_diag": 4.2438507080078125e-05, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.1622836589813232, "learning_rate": 1.55574395930479e-05, "loss": 0.5287, "mean_token_accuracy": 0.8471478223800659, "num_tokens": 27795783.0, "step": 735 }, { "epoch": 0.09362676504261544, "ewc_loss": 0.050537109375, "ewc_loss_diag": 4.2438507080078125e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 2.6404075622558594, "learning_rate": 1.5578635014836794e-05, "loss": 0.5271, "mean_token_accuracy": 0.8423431515693665, "num_tokens": 27826146.0, "step": 736 }, { "epoch": 0.09375397532120595, "ewc_loss": 0.06640625, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.228243589401245, "learning_rate": 1.559983043662569e-05, "loss": 0.5421, "mean_token_accuracy": 0.8453007340431213, "num_tokens": 27865581.0, "step": 737 }, { "epoch": 0.09388118559979647, "ewc_loss": 0.0537109375, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 2.6604855060577393, "learning_rate": 1.5621025858414583e-05, "loss": 0.578, "mean_token_accuracy": 0.8309593200683594, "num_tokens": 27911209.0, "step": 738 }, { "epoch": 0.09400839587838697, "ewc_loss": 0.06982421875, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.2831125259399414, "learning_rate": 1.5642221280203474e-05, "loss": 0.5543, "mean_token_accuracy": 0.8410792946815491, "num_tokens": 27951639.0, "step": 739 }, { "epoch": 0.09413560615697748, "ewc_loss": 0.05908203125, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 2.754920482635498, "learning_rate": 1.566341670199237e-05, "loss": 0.4779, "mean_token_accuracy": 0.8590798377990723, "num_tokens": 27990808.0, "step": 740 }, { "epoch": 0.094262816435568, "ewc_loss": 0.06787109375, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.243163824081421, "learning_rate": 1.5684612123781263e-05, "loss": 0.5106, "mean_token_accuracy": 0.8548281192779541, "num_tokens": 28025421.0, "step": 741 }, { "epoch": 0.0943900267141585, "ewc_loss": 0.0615234375, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 2.838144540786743, "learning_rate": 1.5705807545570158e-05, "loss": 0.4756, "mean_token_accuracy": 0.8638238310813904, "num_tokens": 28068652.0, "step": 742 }, { "epoch": 0.09451723699274901, "ewc_loss": 0.06298828125, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.1140007972717285, "learning_rate": 1.572700296735905e-05, "loss": 0.5217, "mean_token_accuracy": 0.8471208214759827, "num_tokens": 28104852.0, "step": 743 }, { "epoch": 0.09464444727133953, "ewc_loss": 0.06103515625, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 2.885779619216919, "learning_rate": 1.5748198389147943e-05, "loss": 0.4932, "mean_token_accuracy": 0.8558344841003418, "num_tokens": 28141646.0, "step": 744 }, { "epoch": 0.09477165754993004, "ewc_loss": 0.058349609375, "ewc_loss_diag": 4.291534423828125e-05, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 2.9295198917388916, "learning_rate": 1.5769393810936837e-05, "loss": 0.5872, "mean_token_accuracy": 0.8334776163101196, "num_tokens": 28178951.0, "step": 745 }, { "epoch": 0.09489886782852054, "ewc_loss": 0.0576171875, "ewc_loss_diag": 4.291534423828125e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 2.8972606658935547, "learning_rate": 1.5790589232725732e-05, "loss": 0.5987, "mean_token_accuracy": 0.8265334367752075, "num_tokens": 28221911.0, "step": 746 }, { "epoch": 0.09502607810711106, "ewc_loss": 0.05224609375, "ewc_loss_diag": 4.291534423828125e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 2.786713123321533, "learning_rate": 1.5811784654514626e-05, "loss": 0.5178, "mean_token_accuracy": 0.8474351167678833, "num_tokens": 28252828.0, "step": 747 }, { "epoch": 0.09515328838570157, "ewc_loss": 0.0556640625, "ewc_loss_diag": 4.291534423828125e-05, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 2.8993771076202393, "learning_rate": 1.583298007630352e-05, "loss": 0.494, "mean_token_accuracy": 0.8553130626678467, "num_tokens": 28293459.0, "step": 748 }, { "epoch": 0.09528049866429207, "ewc_loss": 0.048095703125, "ewc_loss_diag": 4.291534423828125e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.6751387119293213, "learning_rate": 1.5854175498092412e-05, "loss": 0.5218, "mean_token_accuracy": 0.8442427515983582, "num_tokens": 28332775.0, "step": 749 }, { "epoch": 0.09540770894288259, "ewc_loss": 0.054443359375, "ewc_loss_diag": 4.291534423828125e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 2.909174680709839, "learning_rate": 1.5875370919881306e-05, "loss": 0.503, "mean_token_accuracy": 0.8515924215316772, "num_tokens": 28370367.0, "step": 750 }, { "epoch": 0.0955349192214731, "ewc_loss": 0.047119140625, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 2.628643751144409, "learning_rate": 1.58965663416702e-05, "loss": 0.5259, "mean_token_accuracy": 0.844692587852478, "num_tokens": 28404913.0, "step": 751 }, { "epoch": 0.0956621295000636, "ewc_loss": 0.0556640625, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 2.941262722015381, "learning_rate": 1.5917761763459095e-05, "loss": 0.4761, "mean_token_accuracy": 0.8590389490127563, "num_tokens": 28441057.0, "step": 752 }, { "epoch": 0.09578933977865411, "ewc_loss": 0.046630859375, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.6007285118103027, "learning_rate": 1.593895718524799e-05, "loss": 0.5582, "mean_token_accuracy": 0.8325269222259521, "num_tokens": 28479691.0, "step": 753 }, { "epoch": 0.09591655005724463, "ewc_loss": 0.057861328125, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.002058982849121, "learning_rate": 1.596015260703688e-05, "loss": 0.565, "mean_token_accuracy": 0.8353273272514343, "num_tokens": 28520469.0, "step": 754 }, { "epoch": 0.09604376033583513, "ewc_loss": 0.046630859375, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.5694942474365234, "learning_rate": 1.5981348028825775e-05, "loss": 0.4937, "mean_token_accuracy": 0.8516199588775635, "num_tokens": 28560622.0, "step": 755 }, { "epoch": 0.09617097061442564, "ewc_loss": 0.06005859375, "ewc_loss_diag": 4.3392181396484375e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.0872669219970703, "learning_rate": 1.600254345061467e-05, "loss": 0.5467, "mean_token_accuracy": 0.8389726281166077, "num_tokens": 28594890.0, "step": 756 }, { "epoch": 0.09629818089301616, "ewc_loss": 0.0478515625, "ewc_loss_diag": 4.3392181396484375e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.5551846027374268, "learning_rate": 1.6023738872403564e-05, "loss": 0.5283, "mean_token_accuracy": 0.8427119255065918, "num_tokens": 28633112.0, "step": 757 }, { "epoch": 0.09642539117160667, "ewc_loss": 0.06494140625, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.1903719902038574, "learning_rate": 1.6044934294192455e-05, "loss": 0.5701, "mean_token_accuracy": 0.8323949575424194, "num_tokens": 28672195.0, "step": 758 }, { "epoch": 0.09655260145019717, "ewc_loss": 0.049560546875, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 2.564877510070801, "learning_rate": 1.606612971598135e-05, "loss": 0.5012, "mean_token_accuracy": 0.8533588647842407, "num_tokens": 28710315.0, "step": 759 }, { "epoch": 0.09667981172878769, "ewc_loss": 0.06982421875, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.2818398475646973, "learning_rate": 1.6087325137770244e-05, "loss": 0.5579, "mean_token_accuracy": 0.8445246815681458, "num_tokens": 28748713.0, "step": 760 }, { "epoch": 0.0968070220073782, "ewc_loss": 0.054931640625, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 2.633052349090576, "learning_rate": 1.6108520559559138e-05, "loss": 0.5067, "mean_token_accuracy": 0.8500050902366638, "num_tokens": 28789832.0, "step": 761 }, { "epoch": 0.0969342322859687, "ewc_loss": 0.07177734375, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.343315601348877, "learning_rate": 1.612971598134803e-05, "loss": 0.5515, "mean_token_accuracy": 0.8439156413078308, "num_tokens": 28822781.0, "step": 762 }, { "epoch": 0.09706144256455922, "ewc_loss": 0.059814453125, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 2.749009609222412, "learning_rate": 1.6150911403136924e-05, "loss": 0.523, "mean_token_accuracy": 0.843209981918335, "num_tokens": 28857734.0, "step": 763 }, { "epoch": 0.09718865284314973, "ewc_loss": 0.07080078125, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.2672312259674072, "learning_rate": 1.6172106824925818e-05, "loss": 0.5628, "mean_token_accuracy": 0.8388646841049194, "num_tokens": 28901296.0, "step": 764 }, { "epoch": 0.09731586312174023, "ewc_loss": 0.0634765625, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 2.8344056606292725, "learning_rate": 1.6193302246714712e-05, "loss": 0.5242, "mean_token_accuracy": 0.8467978239059448, "num_tokens": 28942053.0, "step": 765 }, { "epoch": 0.09744307340033075, "ewc_loss": 0.06494140625, "ewc_loss_diag": 4.410743713378906e-05, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.1394636631011963, "learning_rate": 1.6214497668503603e-05, "loss": 0.4947, "mean_token_accuracy": 0.854182779788971, "num_tokens": 28975320.0, "step": 766 }, { "epoch": 0.09757028367892126, "ewc_loss": 0.06396484375, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 2.8904759883880615, "learning_rate": 1.6235693090292498e-05, "loss": 0.5447, "mean_token_accuracy": 0.8360317945480347, "num_tokens": 29019941.0, "step": 767 }, { "epoch": 0.09769749395751176, "ewc_loss": 0.060302734375, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 2.946037769317627, "learning_rate": 1.6256888512081392e-05, "loss": 0.5685, "mean_token_accuracy": 0.8365609645843506, "num_tokens": 29062608.0, "step": 768 }, { "epoch": 0.09782470423610228, "ewc_loss": 0.061767578125, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 2.9331164360046387, "learning_rate": 1.6278083933870287e-05, "loss": 0.5167, "mean_token_accuracy": 0.849842369556427, "num_tokens": 29106423.0, "step": 769 }, { "epoch": 0.09795191451469279, "ewc_loss": 0.0546875, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 2.78481125831604, "learning_rate": 1.6299279355659178e-05, "loss": 0.4577, "mean_token_accuracy": 0.8658385276794434, "num_tokens": 29148717.0, "step": 770 }, { "epoch": 0.0980791247932833, "ewc_loss": 0.06005859375, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 2.9914915561676025, "learning_rate": 1.6320474777448072e-05, "loss": 0.5081, "mean_token_accuracy": 0.8526571393013, "num_tokens": 29189464.0, "step": 771 }, { "epoch": 0.0982063350718738, "ewc_loss": 0.050537109375, "ewc_loss_diag": 4.458427429199219e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 2.677877187728882, "learning_rate": 1.6341670199236967e-05, "loss": 0.5263, "mean_token_accuracy": 0.843346893787384, "num_tokens": 29228384.0, "step": 772 }, { "epoch": 0.09833354535046432, "ewc_loss": 0.061279296875, "ewc_loss_diag": 4.482269287109375e-05, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.072141408920288, "learning_rate": 1.6362865621025858e-05, "loss": 0.5238, "mean_token_accuracy": 0.8512974977493286, "num_tokens": 29270942.0, "step": 773 }, { "epoch": 0.09846075562905483, "ewc_loss": 0.050048828125, "ewc_loss_diag": 4.482269287109375e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 2.60825252532959, "learning_rate": 1.6384061042814752e-05, "loss": 0.4934, "mean_token_accuracy": 0.8534579873085022, "num_tokens": 29313770.0, "step": 774 }, { "epoch": 0.09858796590764533, "ewc_loss": 0.064453125, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.190641403198242, "learning_rate": 1.6405256464603646e-05, "loss": 0.5081, "mean_token_accuracy": 0.8510174751281738, "num_tokens": 29347924.0, "step": 775 }, { "epoch": 0.09871517618623585, "ewc_loss": 0.05029296875, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 2.579331636428833, "learning_rate": 1.642645188639254e-05, "loss": 0.5385, "mean_token_accuracy": 0.8379127979278564, "num_tokens": 29386899.0, "step": 776 }, { "epoch": 0.09884238646482636, "ewc_loss": 0.0703125, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.3204803466796875, "learning_rate": 1.6447647308181432e-05, "loss": 0.5066, "mean_token_accuracy": 0.8575031757354736, "num_tokens": 29423349.0, "step": 777 }, { "epoch": 0.09896959674341686, "ewc_loss": 0.05517578125, "ewc_loss_diag": 4.5299530029296875e-05, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 2.608083963394165, "learning_rate": 1.6468842729970326e-05, "loss": 0.5004, "mean_token_accuracy": 0.8507357835769653, "num_tokens": 29461127.0, "step": 778 }, { "epoch": 0.09909680702200738, "ewc_loss": 0.07421875, "ewc_loss_diag": 4.5299530029296875e-05, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 3.450385332107544, "learning_rate": 1.649003815175922e-05, "loss": 0.6009, "mean_token_accuracy": 0.8329490423202515, "num_tokens": 29498335.0, "step": 779 }, { "epoch": 0.09922401730059789, "ewc_loss": 0.061279296875, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 2.717097043991089, "learning_rate": 1.6511233573548115e-05, "loss": 0.5167, "mean_token_accuracy": 0.8494990468025208, "num_tokens": 29541174.0, "step": 780 }, { "epoch": 0.09935122757918839, "ewc_loss": 0.07568359375, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 7.152557373046875e-05, "grad_norm": 3.4823455810546875, "learning_rate": 1.6532428995337006e-05, "loss": 0.5555, "mean_token_accuracy": 0.8444879055023193, "num_tokens": 29578684.0, "step": 781 }, { "epoch": 0.0994784378577789, "ewc_loss": 0.06787109375, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 2.8702149391174316, "learning_rate": 1.65536244171259e-05, "loss": 0.4977, "mean_token_accuracy": 0.8571469783782959, "num_tokens": 29611236.0, "step": 782 }, { "epoch": 0.09960564813636942, "ewc_loss": 0.07373046875, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 3.3726916313171387, "learning_rate": 1.6574819838914795e-05, "loss": 0.5199, "mean_token_accuracy": 0.8524918556213379, "num_tokens": 29653769.0, "step": 783 }, { "epoch": 0.09973285841495994, "ewc_loss": 0.0693359375, "ewc_loss_diag": 4.57763671875e-05, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 2.950610637664795, "learning_rate": 1.659601526070369e-05, "loss": 0.5276, "mean_token_accuracy": 0.8507329225540161, "num_tokens": 29695231.0, "step": 784 }, { "epoch": 0.09986006869355044, "ewc_loss": 0.0693359375, "ewc_loss_diag": 4.57763671875e-05, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.2100675106048584, "learning_rate": 1.661721068249258e-05, "loss": 0.5092, "mean_token_accuracy": 0.855201244354248, "num_tokens": 29725755.0, "step": 785 }, { "epoch": 0.09998727897214095, "ewc_loss": 0.068359375, "ewc_loss_diag": 4.57763671875e-05, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 2.9917824268341064, "learning_rate": 1.6638406104281475e-05, "loss": 0.5275, "mean_token_accuracy": 0.8482282161712646, "num_tokens": 29760270.0, "step": 786 }, { "epoch": 0.10011448925073146, "ewc_loss": 0.0634765625, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.019584894180298, "learning_rate": 1.665960152607037e-05, "loss": 0.5723, "mean_token_accuracy": 0.8398780226707458, "num_tokens": 29797292.0, "step": 787 }, { "epoch": 0.10024169952932196, "ewc_loss": 0.0654296875, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 2.997982978820801, "learning_rate": 1.668079694785926e-05, "loss": 0.4982, "mean_token_accuracy": 0.8598827123641968, "num_tokens": 29833265.0, "step": 788 }, { "epoch": 0.10036890980791248, "ewc_loss": 0.0595703125, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 2.8718769550323486, "learning_rate": 1.6701992369648155e-05, "loss": 0.5711, "mean_token_accuracy": 0.834496259689331, "num_tokens": 29876399.0, "step": 789 }, { "epoch": 0.100496120086503, "ewc_loss": 0.0625, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.025597333908081, "learning_rate": 1.672318779143705e-05, "loss": 0.5353, "mean_token_accuracy": 0.846926748752594, "num_tokens": 29913883.0, "step": 790 }, { "epoch": 0.1006233303650935, "ewc_loss": 0.055419921875, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 2.8047380447387695, "learning_rate": 1.6744383213225944e-05, "loss": 0.5813, "mean_token_accuracy": 0.8362371325492859, "num_tokens": 29951691.0, "step": 791 }, { "epoch": 0.10075054064368401, "ewc_loss": 0.060791015625, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.017801523208618, "learning_rate": 1.6765578635014838e-05, "loss": 0.55, "mean_token_accuracy": 0.8396881222724915, "num_tokens": 29992978.0, "step": 792 }, { "epoch": 0.10087775092227452, "ewc_loss": 0.05419921875, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 2.745272159576416, "learning_rate": 1.678677405680373e-05, "loss": 0.5319, "mean_token_accuracy": 0.8411128520965576, "num_tokens": 30035602.0, "step": 793 }, { "epoch": 0.10100496120086502, "ewc_loss": 0.06103515625, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.0669047832489014, "learning_rate": 1.6807969478592624e-05, "loss": 0.5385, "mean_token_accuracy": 0.8416136503219604, "num_tokens": 30070063.0, "step": 794 }, { "epoch": 0.10113217147945554, "ewc_loss": 0.052734375, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 2.7242112159729004, "learning_rate": 1.6829164900381518e-05, "loss": 0.5659, "mean_token_accuracy": 0.8285248279571533, "num_tokens": 30101636.0, "step": 795 }, { "epoch": 0.10125938175804605, "ewc_loss": 0.0625, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.1144466400146484, "learning_rate": 1.6850360322170412e-05, "loss": 0.541, "mean_token_accuracy": 0.8349435925483704, "num_tokens": 30137429.0, "step": 796 }, { "epoch": 0.10138659203663657, "ewc_loss": 0.05224609375, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 2.677852153778076, "learning_rate": 1.6871555743959307e-05, "loss": 0.507, "mean_token_accuracy": 0.8508915305137634, "num_tokens": 30176173.0, "step": 797 }, { "epoch": 0.10151380231522707, "ewc_loss": 0.0654296875, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.205456018447876, "learning_rate": 1.68927511657482e-05, "loss": 0.5154, "mean_token_accuracy": 0.8519598841667175, "num_tokens": 30206588.0, "step": 798 }, { "epoch": 0.10164101259381758, "ewc_loss": 0.052734375, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 2.661649703979492, "learning_rate": 1.6913946587537096e-05, "loss": 0.5303, "mean_token_accuracy": 0.8424934148788452, "num_tokens": 30237345.0, "step": 799 }, { "epoch": 0.1017682228724081, "ewc_loss": 0.0703125, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.313527822494507, "learning_rate": 1.6935142009325987e-05, "loss": 0.5639, "mean_token_accuracy": 0.8419581055641174, "num_tokens": 30274543.0, "step": 800 }, { "epoch": 0.1018954331509986, "ewc_loss": 0.055419921875, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 2.7976467609405518, "learning_rate": 1.695633743111488e-05, "loss": 0.5205, "mean_token_accuracy": 0.8439579010009766, "num_tokens": 30312687.0, "step": 801 }, { "epoch": 0.10202264342958911, "ewc_loss": 0.0703125, "ewc_loss_diag": 4.649162292480469e-05, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.276043176651001, "learning_rate": 1.6977532852903776e-05, "loss": 0.5689, "mean_token_accuracy": 0.8347763419151306, "num_tokens": 30343993.0, "step": 802 }, { "epoch": 0.10214985370817962, "ewc_loss": 0.05810546875, "ewc_loss_diag": 4.649162292480469e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 2.7375688552856445, "learning_rate": 1.699872827469267e-05, "loss": 0.5101, "mean_token_accuracy": 0.851651668548584, "num_tokens": 30378998.0, "step": 803 }, { "epoch": 0.10227706398677013, "ewc_loss": 0.072265625, "ewc_loss_diag": 4.7206878662109375e-05, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.2926199436187744, "learning_rate": 1.701992369648156e-05, "loss": 0.5836, "mean_token_accuracy": 0.8309481143951416, "num_tokens": 30412205.0, "step": 804 }, { "epoch": 0.10240427426536064, "ewc_loss": 0.0595703125, "ewc_loss_diag": 4.7206878662109375e-05, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 2.7423360347747803, "learning_rate": 1.7041119118270455e-05, "loss": 0.4973, "mean_token_accuracy": 0.8532888889312744, "num_tokens": 30448489.0, "step": 805 }, { "epoch": 0.10253148454395115, "ewc_loss": 0.0732421875, "ewc_loss_diag": 4.7206878662109375e-05, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.3314051628112793, "learning_rate": 1.706231454005935e-05, "loss": 0.5646, "mean_token_accuracy": 0.8453174829483032, "num_tokens": 30488741.0, "step": 806 }, { "epoch": 0.10265869482254165, "ewc_loss": 0.0615234375, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 2.7780916690826416, "learning_rate": 1.708350996184824e-05, "loss": 0.5521, "mean_token_accuracy": 0.8460255861282349, "num_tokens": 30535715.0, "step": 807 }, { "epoch": 0.10278590510113217, "ewc_loss": 0.0732421875, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.3280420303344727, "learning_rate": 1.7104705383637135e-05, "loss": 0.5188, "mean_token_accuracy": 0.8549765944480896, "num_tokens": 30574992.0, "step": 808 }, { "epoch": 0.10291311537972268, "ewc_loss": 0.06494140625, "ewc_loss_diag": 4.8160552978515625e-05, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 2.829577922821045, "learning_rate": 1.712590080542603e-05, "loss": 0.5799, "mean_token_accuracy": 0.8278021216392517, "num_tokens": 30615384.0, "step": 809 }, { "epoch": 0.1030403256583132, "ewc_loss": 0.0732421875, "ewc_loss_diag": 4.8160552978515625e-05, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.3041858673095703, "learning_rate": 1.7147096227214924e-05, "loss": 0.5346, "mean_token_accuracy": 0.8488693237304688, "num_tokens": 30651011.0, "step": 810 }, { "epoch": 0.1031675359369037, "ewc_loss": 0.0654296875, "ewc_loss_diag": 4.863739013671875e-05, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 2.875241756439209, "learning_rate": 1.7168291649003815e-05, "loss": 0.5158, "mean_token_accuracy": 0.8528064489364624, "num_tokens": 30687213.0, "step": 811 }, { "epoch": 0.10329474621549421, "ewc_loss": 0.0712890625, "ewc_loss_diag": 4.863739013671875e-05, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.209237813949585, "learning_rate": 1.718948707079271e-05, "loss": 0.5582, "mean_token_accuracy": 0.8416442275047302, "num_tokens": 30726950.0, "step": 812 }, { "epoch": 0.10342195649408473, "ewc_loss": 0.06689453125, "ewc_loss_diag": 4.863739013671875e-05, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 2.9153542518615723, "learning_rate": 1.7210682492581604e-05, "loss": 0.4926, "mean_token_accuracy": 0.8582298159599304, "num_tokens": 30769359.0, "step": 813 }, { "epoch": 0.10354916677267523, "ewc_loss": 0.0673828125, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.0993757247924805, "learning_rate": 1.72318779143705e-05, "loss": 0.5586, "mean_token_accuracy": 0.8399108052253723, "num_tokens": 30812480.0, "step": 814 }, { "epoch": 0.10367637705126574, "ewc_loss": 0.06494140625, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 2.9684104919433594, "learning_rate": 1.725307333615939e-05, "loss": 0.499, "mean_token_accuracy": 0.8536349534988403, "num_tokens": 30847287.0, "step": 815 }, { "epoch": 0.10380358732985626, "ewc_loss": 0.0625, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 2.9868175983428955, "learning_rate": 1.7274268757948284e-05, "loss": 0.539, "mean_token_accuracy": 0.8422769904136658, "num_tokens": 30887050.0, "step": 816 }, { "epoch": 0.10393079760844676, "ewc_loss": 0.06396484375, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.0084824562072754, "learning_rate": 1.7295464179737178e-05, "loss": 0.5443, "mean_token_accuracy": 0.842679500579834, "num_tokens": 30926543.0, "step": 817 }, { "epoch": 0.10405800788703727, "ewc_loss": 0.05908203125, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 2.884833335876465, "learning_rate": 1.7316659601526073e-05, "loss": 0.554, "mean_token_accuracy": 0.8369044661521912, "num_tokens": 30966439.0, "step": 818 }, { "epoch": 0.10418521816562779, "ewc_loss": 0.06396484375, "ewc_loss_diag": 4.935264587402344e-05, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.0797648429870605, "learning_rate": 1.7337855023314964e-05, "loss": 0.5051, "mean_token_accuracy": 0.8543649911880493, "num_tokens": 30999860.0, "step": 819 }, { "epoch": 0.1043124284442183, "ewc_loss": 0.056640625, "ewc_loss_diag": 4.9591064453125e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 2.7961604595184326, "learning_rate": 1.7359050445103858e-05, "loss": 0.5396, "mean_token_accuracy": 0.8427385091781616, "num_tokens": 31037040.0, "step": 820 }, { "epoch": 0.1044396387228088, "ewc_loss": 0.06494140625, "ewc_loss_diag": 4.9591064453125e-05, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.176752805709839, "learning_rate": 1.7380245866892753e-05, "loss": 0.545, "mean_token_accuracy": 0.8381450176239014, "num_tokens": 31071508.0, "step": 821 }, { "epoch": 0.10456684900139931, "ewc_loss": 0.0546875, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 2.7096259593963623, "learning_rate": 1.7401441288681647e-05, "loss": 0.5224, "mean_token_accuracy": 0.8419257402420044, "num_tokens": 31107358.0, "step": 822 }, { "epoch": 0.10469405927998983, "ewc_loss": 0.06982421875, "ewc_loss_diag": 5.0067901611328125e-05, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.325589179992676, "learning_rate": 1.7422636710470538e-05, "loss": 0.5461, "mean_token_accuracy": 0.8423001766204834, "num_tokens": 31144986.0, "step": 823 }, { "epoch": 0.10482126955858033, "ewc_loss": 0.05419921875, "ewc_loss_diag": 5.0067901611328125e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 2.6494100093841553, "learning_rate": 1.7443832132259432e-05, "loss": 0.52, "mean_token_accuracy": 0.8413232564926147, "num_tokens": 31179709.0, "step": 824 }, { "epoch": 0.10494847983717084, "ewc_loss": 0.07568359375, "ewc_loss_diag": 5.0067901611328125e-05, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 3.505603790283203, "learning_rate": 1.7465027554048327e-05, "loss": 0.5584, "mean_token_accuracy": 0.8421941995620728, "num_tokens": 31222242.0, "step": 825 }, { "epoch": 0.10507569011576136, "ewc_loss": 0.05810546875, "ewc_loss_diag": 5.0067901611328125e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 2.638443946838379, "learning_rate": 1.7486222975837218e-05, "loss": 0.511, "mean_token_accuracy": 0.8490354418754578, "num_tokens": 31264793.0, "step": 826 }, { "epoch": 0.10520290039435186, "ewc_loss": 0.08349609375, "ewc_loss_diag": 5.0067901611328125e-05, "ewc_loss_parallel": 7.867813110351562e-05, "grad_norm": 3.7545554637908936, "learning_rate": 1.7507418397626112e-05, "loss": 0.5924, "mean_token_accuracy": 0.8345365524291992, "num_tokens": 31300254.0, "step": 827 }, { "epoch": 0.10533011067294237, "ewc_loss": 0.068359375, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 2.7990810871124268, "learning_rate": 1.7528613819415007e-05, "loss": 0.5193, "mean_token_accuracy": 0.8489126563072205, "num_tokens": 31339786.0, "step": 828 }, { "epoch": 0.10545732095153289, "ewc_loss": 0.09033203125, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 8.535385131835938e-05, "grad_norm": 3.8943026065826416, "learning_rate": 1.75498092412039e-05, "loss": 0.5767, "mean_token_accuracy": 0.8373540043830872, "num_tokens": 31375100.0, "step": 829 }, { "epoch": 0.10558453123012339, "ewc_loss": 0.0810546875, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 7.62939453125e-05, "grad_norm": 3.0783276557922363, "learning_rate": 1.7571004662992792e-05, "loss": 0.6216, "mean_token_accuracy": 0.8249964118003845, "num_tokens": 31419000.0, "step": 830 }, { "epoch": 0.1057117415087139, "ewc_loss": 0.08642578125, "ewc_loss_diag": 5.078315734863281e-05, "ewc_loss_parallel": 8.153915405273438e-05, "grad_norm": 3.8054089546203613, "learning_rate": 1.7592200084781687e-05, "loss": 0.5712, "mean_token_accuracy": 0.8416949510574341, "num_tokens": 31459287.0, "step": 831 }, { "epoch": 0.10583895178730442, "ewc_loss": 0.087890625, "ewc_loss_diag": 5.078315734863281e-05, "ewc_loss_parallel": 8.296966552734375e-05, "grad_norm": 3.2457668781280518, "learning_rate": 1.761339550657058e-05, "loss": 0.5785, "mean_token_accuracy": 0.846047580242157, "num_tokens": 31502617.0, "step": 832 }, { "epoch": 0.10596616206589493, "ewc_loss": 0.080078125, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 7.534027099609375e-05, "grad_norm": 3.5672850608825684, "learning_rate": 1.7634590928359476e-05, "loss": 0.5831, "mean_token_accuracy": 0.834334135055542, "num_tokens": 31545159.0, "step": 833 }, { "epoch": 0.10609337234448543, "ewc_loss": 0.0849609375, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 8.0108642578125e-05, "grad_norm": 3.269075393676758, "learning_rate": 1.7655786350148367e-05, "loss": 0.5178, "mean_token_accuracy": 0.8599711656570435, "num_tokens": 31579193.0, "step": 834 }, { "epoch": 0.10622058262307595, "ewc_loss": 0.07373046875, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 3.3059515953063965, "learning_rate": 1.767698177193726e-05, "loss": 0.5269, "mean_token_accuracy": 0.8521440625190735, "num_tokens": 31618995.0, "step": 835 }, { "epoch": 0.10634779290166646, "ewc_loss": 0.07666015625, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 3.252655506134033, "learning_rate": 1.7698177193726155e-05, "loss": 0.551, "mean_token_accuracy": 0.8427788019180298, "num_tokens": 31657024.0, "step": 836 }, { "epoch": 0.10647500318025696, "ewc_loss": 0.06884765625, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.078505516052246, "learning_rate": 1.771937261551505e-05, "loss": 0.5382, "mean_token_accuracy": 0.8464884757995605, "num_tokens": 31700521.0, "step": 837 }, { "epoch": 0.10660221345884748, "ewc_loss": 0.07421875, "ewc_loss_diag": 5.125999450683594e-05, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 3.2725372314453125, "learning_rate": 1.774056803730394e-05, "loss": 0.5125, "mean_token_accuracy": 0.8581817150115967, "num_tokens": 31735942.0, "step": 838 }, { "epoch": 0.10672942373743799, "ewc_loss": 0.064453125, "ewc_loss_diag": 5.125999450683594e-05, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 2.9305405616760254, "learning_rate": 1.7761763459092835e-05, "loss": 0.5299, "mean_token_accuracy": 0.848980188369751, "num_tokens": 31774092.0, "step": 839 }, { "epoch": 0.10685663401602849, "ewc_loss": 0.0732421875, "ewc_loss_diag": 5.125999450683594e-05, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.317985773086548, "learning_rate": 1.778295888088173e-05, "loss": 0.5179, "mean_token_accuracy": 0.8527030944824219, "num_tokens": 31816292.0, "step": 840 }, { "epoch": 0.106983844294619, "ewc_loss": 0.060791015625, "ewc_loss_diag": 5.125999450683594e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 2.8578577041625977, "learning_rate": 1.7804154302670624e-05, "loss": 0.5226, "mean_token_accuracy": 0.8490591049194336, "num_tokens": 31848641.0, "step": 841 }, { "epoch": 0.10711105457320952, "ewc_loss": 0.07470703125, "ewc_loss_diag": 5.14984130859375e-05, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 3.4038610458374023, "learning_rate": 1.782534972445952e-05, "loss": 0.5903, "mean_token_accuracy": 0.830942690372467, "num_tokens": 31881292.0, "step": 842 }, { "epoch": 0.10723826485180002, "ewc_loss": 0.060546875, "ewc_loss_diag": 5.14984130859375e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 2.806908130645752, "learning_rate": 1.784654514624841e-05, "loss": 0.5332, "mean_token_accuracy": 0.8458518981933594, "num_tokens": 31915810.0, "step": 843 }, { "epoch": 0.10736547513039053, "ewc_loss": 0.07861328125, "ewc_loss_diag": 5.14984130859375e-05, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 3.5104873180389404, "learning_rate": 1.7867740568037304e-05, "loss": 0.5372, "mean_token_accuracy": 0.8496618866920471, "num_tokens": 31953621.0, "step": 844 }, { "epoch": 0.10749268540898105, "ewc_loss": 0.0634765625, "ewc_loss_diag": 5.14984130859375e-05, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 2.819589853286743, "learning_rate": 1.78889359898262e-05, "loss": 0.5581, "mean_token_accuracy": 0.8378165364265442, "num_tokens": 31988738.0, "step": 845 }, { "epoch": 0.10761989568757156, "ewc_loss": 0.08154296875, "ewc_loss_diag": 5.173683166503906e-05, "ewc_loss_parallel": 7.62939453125e-05, "grad_norm": 3.6103506088256836, "learning_rate": 1.7910131411615093e-05, "loss": 0.5235, "mean_token_accuracy": 0.8585561513900757, "num_tokens": 32029184.0, "step": 846 }, { "epoch": 0.10774710596616206, "ewc_loss": 0.06884765625, "ewc_loss_diag": 5.173683166503906e-05, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 2.8964388370513916, "learning_rate": 1.7931326833403987e-05, "loss": 0.5902, "mean_token_accuracy": 0.8294264078140259, "num_tokens": 32064413.0, "step": 847 }, { "epoch": 0.10787431624475258, "ewc_loss": 0.08349609375, "ewc_loss_diag": 5.173683166503906e-05, "ewc_loss_parallel": 7.82012939453125e-05, "grad_norm": 3.6439743041992188, "learning_rate": 1.795252225519288e-05, "loss": 0.5722, "mean_token_accuracy": 0.8430355787277222, "num_tokens": 32104043.0, "step": 848 }, { "epoch": 0.10800152652334309, "ewc_loss": 0.0732421875, "ewc_loss_diag": 5.173683166503906e-05, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 3.0022480487823486, "learning_rate": 1.7973717676981773e-05, "loss": 0.5485, "mean_token_accuracy": 0.84376460313797, "num_tokens": 32140956.0, "step": 849 }, { "epoch": 0.10812873680193359, "ewc_loss": 0.0810546875, "ewc_loss_diag": 5.173683166503906e-05, "ewc_loss_parallel": 7.581710815429688e-05, "grad_norm": 3.579167604446411, "learning_rate": 1.7994913098770667e-05, "loss": 0.604, "mean_token_accuracy": 0.8313108682632446, "num_tokens": 32181368.0, "step": 850 }, { "epoch": 0.1082559470805241, "ewc_loss": 0.07666015625, "ewc_loss_diag": 5.173683166503906e-05, "ewc_loss_parallel": 7.152557373046875e-05, "grad_norm": 3.0906741619110107, "learning_rate": 1.801610852055956e-05, "loss": 0.5496, "mean_token_accuracy": 0.8411011695861816, "num_tokens": 32222714.0, "step": 851 }, { "epoch": 0.10838315735911462, "ewc_loss": 0.07861328125, "ewc_loss_diag": 5.173683166503906e-05, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 3.4128599166870117, "learning_rate": 1.8037303942348456e-05, "loss": 0.5162, "mean_token_accuracy": 0.8574994206428528, "num_tokens": 32267870.0, "step": 852 }, { "epoch": 0.10851036763770512, "ewc_loss": 0.0751953125, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 3.130371570587158, "learning_rate": 1.8058499364137347e-05, "loss": 0.4846, "mean_token_accuracy": 0.8639719486236572, "num_tokens": 32307408.0, "step": 853 }, { "epoch": 0.10863757791629564, "ewc_loss": 0.07373046875, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.249359607696533, "learning_rate": 1.807969478592624e-05, "loss": 0.6169, "mean_token_accuracy": 0.8265618085861206, "num_tokens": 32343503.0, "step": 854 }, { "epoch": 0.10876478819488615, "ewc_loss": 0.07275390625, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.145160675048828, "learning_rate": 1.8100890207715136e-05, "loss": 0.562, "mean_token_accuracy": 0.8412653207778931, "num_tokens": 32386375.0, "step": 855 }, { "epoch": 0.10889199847347665, "ewc_loss": 0.0703125, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.123817205429077, "learning_rate": 1.812208562950403e-05, "loss": 0.5677, "mean_token_accuracy": 0.8391838669776917, "num_tokens": 32424370.0, "step": 856 }, { "epoch": 0.10901920875206716, "ewc_loss": 0.07080078125, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.1454315185546875, "learning_rate": 1.814328105129292e-05, "loss": 0.5794, "mean_token_accuracy": 0.8344347476959229, "num_tokens": 32461998.0, "step": 857 }, { "epoch": 0.10914641903065768, "ewc_loss": 0.06591796875, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.034785747528076, "learning_rate": 1.8164476473081816e-05, "loss": 0.5114, "mean_token_accuracy": 0.8521512746810913, "num_tokens": 32498124.0, "step": 858 }, { "epoch": 0.1092736293092482, "ewc_loss": 0.06787109375, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.1499385833740234, "learning_rate": 1.818567189487071e-05, "loss": 0.5083, "mean_token_accuracy": 0.8540397882461548, "num_tokens": 32537735.0, "step": 859 }, { "epoch": 0.1094008395878387, "ewc_loss": 0.06298828125, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 2.9903311729431152, "learning_rate": 1.8206867316659605e-05, "loss": 0.5441, "mean_token_accuracy": 0.8427412509918213, "num_tokens": 32576219.0, "step": 860 }, { "epoch": 0.10952804986642921, "ewc_loss": 0.0673828125, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.1726419925689697, "learning_rate": 1.8228062738448496e-05, "loss": 0.5144, "mean_token_accuracy": 0.8469990491867065, "num_tokens": 32612868.0, "step": 861 }, { "epoch": 0.10965526014501972, "ewc_loss": 0.06103515625, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 2.9332971572875977, "learning_rate": 1.824925816023739e-05, "loss": 0.5006, "mean_token_accuracy": 0.8567113876342773, "num_tokens": 32655412.0, "step": 862 }, { "epoch": 0.10978247042361022, "ewc_loss": 0.06787109375, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.221696376800537, "learning_rate": 1.8270453582026284e-05, "loss": 0.476, "mean_token_accuracy": 0.8649932742118835, "num_tokens": 32692579.0, "step": 863 }, { "epoch": 0.10990968070220074, "ewc_loss": 0.05908203125, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 2.90254807472229, "learning_rate": 1.8291649003815176e-05, "loss": 0.5634, "mean_token_accuracy": 0.8305076956748962, "num_tokens": 32731642.0, "step": 864 }, { "epoch": 0.11003689098079125, "ewc_loss": 0.06884765625, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.298072338104248, "learning_rate": 1.831284442560407e-05, "loss": 0.5384, "mean_token_accuracy": 0.8445377349853516, "num_tokens": 32765707.0, "step": 865 }, { "epoch": 0.11016410125938175, "ewc_loss": 0.058349609375, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 2.841421127319336, "learning_rate": 1.8334039847392964e-05, "loss": 0.5592, "mean_token_accuracy": 0.8327738046646118, "num_tokens": 32804613.0, "step": 866 }, { "epoch": 0.11029131153797227, "ewc_loss": 0.07275390625, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.4131739139556885, "learning_rate": 1.835523526918186e-05, "loss": 0.5584, "mean_token_accuracy": 0.839195966720581, "num_tokens": 32841198.0, "step": 867 }, { "epoch": 0.11041852181656278, "ewc_loss": 0.05859375, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 2.7665810585021973, "learning_rate": 1.837643069097075e-05, "loss": 0.5143, "mean_token_accuracy": 0.8491228818893433, "num_tokens": 32878241.0, "step": 868 }, { "epoch": 0.11054573209515328, "ewc_loss": 0.078125, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 3.6305458545684814, "learning_rate": 1.8397626112759644e-05, "loss": 0.5909, "mean_token_accuracy": 0.8340777158737183, "num_tokens": 32915510.0, "step": 869 }, { "epoch": 0.1106729423737438, "ewc_loss": 0.060302734375, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 2.7195351123809814, "learning_rate": 1.841882153454854e-05, "loss": 0.4742, "mean_token_accuracy": 0.8618418574333191, "num_tokens": 32953371.0, "step": 870 }, { "epoch": 0.11080015265233431, "ewc_loss": 0.09033203125, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 8.487701416015625e-05, "grad_norm": 3.893651008605957, "learning_rate": 1.8440016956337433e-05, "loss": 0.5659, "mean_token_accuracy": 0.8467184901237488, "num_tokens": 32998167.0, "step": 871 }, { "epoch": 0.11092736293092482, "ewc_loss": 0.07080078125, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 2.842963218688965, "learning_rate": 1.8461212378126324e-05, "loss": 0.5138, "mean_token_accuracy": 0.851806640625, "num_tokens": 33046660.0, "step": 872 }, { "epoch": 0.11105457320951533, "ewc_loss": 0.10107421875, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 9.584426879882812e-05, "grad_norm": 4.1234331130981445, "learning_rate": 1.848240779991522e-05, "loss": 0.5866, "mean_token_accuracy": 0.8368210196495056, "num_tokens": 33080672.0, "step": 873 }, { "epoch": 0.11118178348810584, "ewc_loss": 0.0888671875, "ewc_loss_diag": 5.340576171875e-05, "ewc_loss_parallel": 8.344650268554688e-05, "grad_norm": 3.214794397354126, "learning_rate": 1.8503603221704113e-05, "loss": 0.5187, "mean_token_accuracy": 0.8560848236083984, "num_tokens": 33117023.0, "step": 874 }, { "epoch": 0.11130899376669635, "ewc_loss": 0.09716796875, "ewc_loss_diag": 5.340576171875e-05, "ewc_loss_parallel": 9.1552734375e-05, "grad_norm": 4.031470775604248, "learning_rate": 1.8524798643493007e-05, "loss": 0.6373, "mean_token_accuracy": 0.8248404264450073, "num_tokens": 33154843.0, "step": 875 }, { "epoch": 0.11143620404528685, "ewc_loss": 0.09912109375, "ewc_loss_diag": 5.340576171875e-05, "ewc_loss_parallel": 9.393692016601562e-05, "grad_norm": 3.440519332885742, "learning_rate": 1.85459940652819e-05, "loss": 0.5966, "mean_token_accuracy": 0.8362990617752075, "num_tokens": 33193500.0, "step": 876 }, { "epoch": 0.11156341432387737, "ewc_loss": 0.08447265625, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 7.915496826171875e-05, "grad_norm": 3.7350692749023438, "learning_rate": 1.8567189487070793e-05, "loss": 0.5729, "mean_token_accuracy": 0.8411200046539307, "num_tokens": 33240528.0, "step": 877 }, { "epoch": 0.11169062460246788, "ewc_loss": 0.09375, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 8.821487426757812e-05, "grad_norm": 3.453221321105957, "learning_rate": 1.8588384908859687e-05, "loss": 0.5867, "mean_token_accuracy": 0.8383762836456299, "num_tokens": 33276959.0, "step": 878 }, { "epoch": 0.11181783488105838, "ewc_loss": 0.0791015625, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 7.390975952148438e-05, "grad_norm": 3.432143211364746, "learning_rate": 1.860958033064858e-05, "loss": 0.4852, "mean_token_accuracy": 0.8648617267608643, "num_tokens": 33313573.0, "step": 879 }, { "epoch": 0.1119450451596489, "ewc_loss": 0.08544921875, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 8.0108642578125e-05, "grad_norm": 3.4482409954071045, "learning_rate": 1.8630775752437473e-05, "loss": 0.5638, "mean_token_accuracy": 0.8467634320259094, "num_tokens": 33347408.0, "step": 880 }, { "epoch": 0.11207225543823941, "ewc_loss": 0.07177734375, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.1851539611816406, "learning_rate": 1.8651971174226367e-05, "loss": 0.4827, "mean_token_accuracy": 0.8634591698646545, "num_tokens": 33381894.0, "step": 881 }, { "epoch": 0.11219946571682991, "ewc_loss": 0.0810546875, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 7.581710815429688e-05, "grad_norm": 3.4774246215820312, "learning_rate": 1.867316659601526e-05, "loss": 0.5124, "mean_token_accuracy": 0.8597762584686279, "num_tokens": 33421746.0, "step": 882 }, { "epoch": 0.11232667599542043, "ewc_loss": 0.0693359375, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.0238513946533203, "learning_rate": 1.8694362017804153e-05, "loss": 0.5197, "mean_token_accuracy": 0.8556731939315796, "num_tokens": 33465203.0, "step": 883 }, { "epoch": 0.11245388627401094, "ewc_loss": 0.0810546875, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 7.581710815429688e-05, "grad_norm": 3.563018321990967, "learning_rate": 1.8715557439593047e-05, "loss": 0.5699, "mean_token_accuracy": 0.8391474485397339, "num_tokens": 33500333.0, "step": 884 }, { "epoch": 0.11258109655260146, "ewc_loss": 0.06787109375, "ewc_loss_diag": 5.435943603515625e-05, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 2.9521896839141846, "learning_rate": 1.873675286138194e-05, "loss": 0.5752, "mean_token_accuracy": 0.8332265615463257, "num_tokens": 33535180.0, "step": 885 }, { "epoch": 0.11270830683119196, "ewc_loss": 0.08349609375, "ewc_loss_diag": 5.435943603515625e-05, "ewc_loss_parallel": 7.82012939453125e-05, "grad_norm": 3.6786515712738037, "learning_rate": 1.8757948283170836e-05, "loss": 0.5411, "mean_token_accuracy": 0.8490207195281982, "num_tokens": 33571298.0, "step": 886 }, { "epoch": 0.11283551710978247, "ewc_loss": 0.0693359375, "ewc_loss_diag": 5.435943603515625e-05, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 2.939626455307007, "learning_rate": 1.8779143704959727e-05, "loss": 0.5533, "mean_token_accuracy": 0.8405442833900452, "num_tokens": 33613693.0, "step": 887 }, { "epoch": 0.11296272738837299, "ewc_loss": 0.08642578125, "ewc_loss_diag": 5.435943603515625e-05, "ewc_loss_parallel": 8.106231689453125e-05, "grad_norm": 3.7926905155181885, "learning_rate": 1.880033912674862e-05, "loss": 0.5334, "mean_token_accuracy": 0.8555693626403809, "num_tokens": 33649760.0, "step": 888 }, { "epoch": 0.11308993766696349, "ewc_loss": 0.07470703125, "ewc_loss_diag": 5.435943603515625e-05, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 3.015620470046997, "learning_rate": 1.8821534548537516e-05, "loss": 0.4864, "mean_token_accuracy": 0.8636969327926636, "num_tokens": 33692441.0, "step": 889 }, { "epoch": 0.113217147945554, "ewc_loss": 0.08935546875, "ewc_loss_diag": 5.435943603515625e-05, "ewc_loss_parallel": 8.392333984375e-05, "grad_norm": 3.844085216522217, "learning_rate": 1.884272997032641e-05, "loss": 0.5476, "mean_token_accuracy": 0.8477718830108643, "num_tokens": 33723034.0, "step": 890 }, { "epoch": 0.11334435822414451, "ewc_loss": 0.07958984375, "ewc_loss_diag": 5.435943603515625e-05, "ewc_loss_parallel": 7.43865966796875e-05, "grad_norm": 3.1521759033203125, "learning_rate": 1.8863925392115305e-05, "loss": 0.5553, "mean_token_accuracy": 0.8451359272003174, "num_tokens": 33764102.0, "step": 891 }, { "epoch": 0.11347156850273502, "ewc_loss": 0.0869140625, "ewc_loss_diag": 5.4836273193359375e-05, "ewc_loss_parallel": 8.153915405273438e-05, "grad_norm": 3.7538418769836426, "learning_rate": 1.88851208139042e-05, "loss": 0.5272, "mean_token_accuracy": 0.8550388813018799, "num_tokens": 33803293.0, "step": 892 }, { "epoch": 0.11359877878132553, "ewc_loss": 0.0830078125, "ewc_loss_diag": 5.4836273193359375e-05, "ewc_loss_parallel": 7.772445678710938e-05, "grad_norm": 3.253023386001587, "learning_rate": 1.890631623569309e-05, "loss": 0.5194, "mean_token_accuracy": 0.8556472063064575, "num_tokens": 33841681.0, "step": 893 }, { "epoch": 0.11372598905991604, "ewc_loss": 0.08251953125, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 7.724761962890625e-05, "grad_norm": 3.5728979110717773, "learning_rate": 1.8927511657481984e-05, "loss": 0.543, "mean_token_accuracy": 0.8427401185035706, "num_tokens": 33877207.0, "step": 894 }, { "epoch": 0.11385319933850654, "ewc_loss": 0.080078125, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 7.486343383789062e-05, "grad_norm": 3.3203253746032715, "learning_rate": 1.894870707927088e-05, "loss": 0.5515, "mean_token_accuracy": 0.8411920070648193, "num_tokens": 33917703.0, "step": 895 }, { "epoch": 0.11398040961709706, "ewc_loss": 0.07666015625, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 3.3598227500915527, "learning_rate": 1.8969902501059773e-05, "loss": 0.5423, "mean_token_accuracy": 0.8472939729690552, "num_tokens": 33957007.0, "step": 896 }, { "epoch": 0.11410761989568757, "ewc_loss": 0.0771484375, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 3.2983813285827637, "learning_rate": 1.8991097922848668e-05, "loss": 0.4743, "mean_token_accuracy": 0.8664300441741943, "num_tokens": 33992615.0, "step": 897 }, { "epoch": 0.11423483017427809, "ewc_loss": 0.0732421875, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 3.2399392127990723, "learning_rate": 1.901229334463756e-05, "loss": 0.5887, "mean_token_accuracy": 0.8338226675987244, "num_tokens": 34032049.0, "step": 898 }, { "epoch": 0.11436204045286859, "ewc_loss": 0.0751953125, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 3.273191213607788, "learning_rate": 1.9033488766426453e-05, "loss": 0.5211, "mean_token_accuracy": 0.8508915305137634, "num_tokens": 34065978.0, "step": 899 }, { "epoch": 0.1144892507314591, "ewc_loss": 0.0712890625, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.1813080310821533, "learning_rate": 1.9054684188215348e-05, "loss": 0.5002, "mean_token_accuracy": 0.8581111431121826, "num_tokens": 34103454.0, "step": 900 }, { "epoch": 0.11461646101004962, "ewc_loss": 0.072265625, "ewc_loss_diag": 5.5789947509765625e-05, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.2370994091033936, "learning_rate": 1.9075879610004242e-05, "loss": 0.4731, "mean_token_accuracy": 0.8632517457008362, "num_tokens": 34137970.0, "step": 901 }, { "epoch": 0.11474367128864012, "ewc_loss": 0.0693359375, "ewc_loss_diag": 5.5789947509765625e-05, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.1561970710754395, "learning_rate": 1.9097075031793133e-05, "loss": 0.5631, "mean_token_accuracy": 0.8383868932723999, "num_tokens": 34176997.0, "step": 902 }, { "epoch": 0.11487088156723063, "ewc_loss": 0.0712890625, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.2228002548217773, "learning_rate": 1.9118270453582028e-05, "loss": 0.5423, "mean_token_accuracy": 0.8437257409095764, "num_tokens": 34215790.0, "step": 903 }, { "epoch": 0.11499809184582115, "ewc_loss": 0.0673828125, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.1397786140441895, "learning_rate": 1.9139465875370922e-05, "loss": 0.5507, "mean_token_accuracy": 0.8474394679069519, "num_tokens": 34255212.0, "step": 904 }, { "epoch": 0.11512530212441165, "ewc_loss": 0.0693359375, "ewc_loss_diag": 5.650520324707031e-05, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.2249300479888916, "learning_rate": 1.9160661297159816e-05, "loss": 0.487, "mean_token_accuracy": 0.8629807233810425, "num_tokens": 34288432.0, "step": 905 }, { "epoch": 0.11525251240300216, "ewc_loss": 0.0673828125, "ewc_loss_diag": 5.650520324707031e-05, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.1244943141937256, "learning_rate": 1.9181856718948707e-05, "loss": 0.5892, "mean_token_accuracy": 0.827714204788208, "num_tokens": 34328991.0, "step": 906 }, { "epoch": 0.11537972268159268, "ewc_loss": 0.0693359375, "ewc_loss_diag": 5.650520324707031e-05, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.2392561435699463, "learning_rate": 1.9203052140737602e-05, "loss": 0.5775, "mean_token_accuracy": 0.837368905544281, "num_tokens": 34366712.0, "step": 907 }, { "epoch": 0.11550693296018319, "ewc_loss": 0.0654296875, "ewc_loss_diag": 5.650520324707031e-05, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.0883238315582275, "learning_rate": 1.9224247562526496e-05, "loss": 0.5609, "mean_token_accuracy": 0.8353030681610107, "num_tokens": 34405757.0, "step": 908 }, { "epoch": 0.11563414323877369, "ewc_loss": 0.06982421875, "ewc_loss_diag": 5.6743621826171875e-05, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.2704503536224365, "learning_rate": 1.924544298431539e-05, "loss": 0.5212, "mean_token_accuracy": 0.8525764346122742, "num_tokens": 34448019.0, "step": 909 }, { "epoch": 0.1157613535173642, "ewc_loss": 0.06396484375, "ewc_loss_diag": 5.6743621826171875e-05, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.0457935333251953, "learning_rate": 1.9266638406104282e-05, "loss": 0.535, "mean_token_accuracy": 0.8446016311645508, "num_tokens": 34491237.0, "step": 910 }, { "epoch": 0.11588856379595472, "ewc_loss": 0.0712890625, "ewc_loss_diag": 5.698204040527344e-05, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.348980188369751, "learning_rate": 1.9287833827893176e-05, "loss": 0.5105, "mean_token_accuracy": 0.8543746471405029, "num_tokens": 34522842.0, "step": 911 }, { "epoch": 0.11601577407454522, "ewc_loss": 0.061279296875, "ewc_loss_diag": 5.698204040527344e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 2.9777445793151855, "learning_rate": 1.930902924968207e-05, "loss": 0.5232, "mean_token_accuracy": 0.8521987199783325, "num_tokens": 34560551.0, "step": 912 }, { "epoch": 0.11614298435313573, "ewc_loss": 0.07421875, "ewc_loss_diag": 5.698204040527344e-05, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.473722219467163, "learning_rate": 1.9330224671470965e-05, "loss": 0.5811, "mean_token_accuracy": 0.8410736322402954, "num_tokens": 34603064.0, "step": 913 }, { "epoch": 0.11627019463172625, "ewc_loss": 0.060302734375, "ewc_loss_diag": 5.698204040527344e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 2.8792731761932373, "learning_rate": 1.9351420093259856e-05, "loss": 0.5446, "mean_token_accuracy": 0.8408017158508301, "num_tokens": 34640283.0, "step": 914 }, { "epoch": 0.11639740491031675, "ewc_loss": 0.0810546875, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 7.534027099609375e-05, "grad_norm": 3.697331666946411, "learning_rate": 1.937261551504875e-05, "loss": 0.5157, "mean_token_accuracy": 0.8577510118484497, "num_tokens": 34679840.0, "step": 915 }, { "epoch": 0.11652461518890726, "ewc_loss": 0.060546875, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 2.7574172019958496, "learning_rate": 1.9393810936837645e-05, "loss": 0.4822, "mean_token_accuracy": 0.8599483966827393, "num_tokens": 34717045.0, "step": 916 }, { "epoch": 0.11665182546749778, "ewc_loss": 0.09619140625, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 9.012222290039062e-05, "grad_norm": 4.065230369567871, "learning_rate": 1.9415006358626536e-05, "loss": 0.527, "mean_token_accuracy": 0.8576870560646057, "num_tokens": 34755849.0, "step": 917 }, { "epoch": 0.11677903574608828, "ewc_loss": 0.06884765625, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 2.798319101333618, "learning_rate": 1.943620178041543e-05, "loss": 0.4708, "mean_token_accuracy": 0.8631058931350708, "num_tokens": 34795202.0, "step": 918 }, { "epoch": 0.11690624602467879, "ewc_loss": 0.12255859375, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 0.00011682510375976562, "grad_norm": 4.550017833709717, "learning_rate": 1.9457397202204325e-05, "loss": 0.5616, "mean_token_accuracy": 0.8549008965492249, "num_tokens": 34824279.0, "step": 919 }, { "epoch": 0.1170334563032693, "ewc_loss": 0.0986328125, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 9.298324584960938e-05, "grad_norm": 3.3442342281341553, "learning_rate": 1.947859262399322e-05, "loss": 0.5784, "mean_token_accuracy": 0.8407462239265442, "num_tokens": 34857631.0, "step": 920 }, { "epoch": 0.11716066658185982, "ewc_loss": 0.1240234375, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 0.000118255615234375, "grad_norm": 4.618134021759033, "learning_rate": 1.949978804578211e-05, "loss": 0.5756, "mean_token_accuracy": 0.8508378863334656, "num_tokens": 34896969.0, "step": 921 }, { "epoch": 0.11728787686045032, "ewc_loss": 0.1240234375, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 0.000118255615234375, "grad_norm": 3.8052170276641846, "learning_rate": 1.9520983467571005e-05, "loss": 0.5934, "mean_token_accuracy": 0.8446198105812073, "num_tokens": 34942177.0, "step": 922 }, { "epoch": 0.11741508713904084, "ewc_loss": 0.1083984375, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 0.00010251998901367188, "grad_norm": 4.355931282043457, "learning_rate": 1.95421788893599e-05, "loss": 0.6472, "mean_token_accuracy": 0.8257744312286377, "num_tokens": 34980440.0, "step": 923 }, { "epoch": 0.11754229741763135, "ewc_loss": 0.11572265625, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 0.00010967254638671875, "grad_norm": 3.8314931392669678, "learning_rate": 1.9563374311148793e-05, "loss": 0.6052, "mean_token_accuracy": 0.8382624983787537, "num_tokens": 35015343.0, "step": 924 }, { "epoch": 0.11766950769622185, "ewc_loss": 0.09912109375, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 9.34600830078125e-05, "grad_norm": 4.0929179191589355, "learning_rate": 1.9584569732937684e-05, "loss": 0.579, "mean_token_accuracy": 0.8425002098083496, "num_tokens": 35050265.0, "step": 925 }, { "epoch": 0.11779671797481236, "ewc_loss": 0.09912109375, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 9.34600830078125e-05, "grad_norm": 3.6522629261016846, "learning_rate": 1.960576515472658e-05, "loss": 0.6002, "mean_token_accuracy": 0.8362021446228027, "num_tokens": 35089235.0, "step": 926 }, { "epoch": 0.11792392825340288, "ewc_loss": 0.09326171875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 8.726119995117188e-05, "grad_norm": 3.893341302871704, "learning_rate": 1.9626960576515473e-05, "loss": 0.5479, "mean_token_accuracy": 0.8486620187759399, "num_tokens": 35122887.0, "step": 927 }, { "epoch": 0.11805113853199338, "ewc_loss": 0.0859375, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 8.0108642578125e-05, "grad_norm": 3.4774298667907715, "learning_rate": 1.9648155998304368e-05, "loss": 0.5409, "mean_token_accuracy": 0.8475556373596191, "num_tokens": 35158686.0, "step": 928 }, { "epoch": 0.1181783488105839, "ewc_loss": 0.08837890625, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 8.249282836914062e-05, "grad_norm": 3.776332139968872, "learning_rate": 1.966935142009326e-05, "loss": 0.5439, "mean_token_accuracy": 0.8475997447967529, "num_tokens": 35195215.0, "step": 929 }, { "epoch": 0.11830555908917441, "ewc_loss": 0.080078125, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 7.43865966796875e-05, "grad_norm": 3.3242955207824707, "learning_rate": 1.9690546841882153e-05, "loss": 0.5378, "mean_token_accuracy": 0.8497377634048462, "num_tokens": 35236499.0, "step": 930 }, { "epoch": 0.11843276936776491, "ewc_loss": 0.0869140625, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 8.106231689453125e-05, "grad_norm": 3.759122848510742, "learning_rate": 1.9711742263671048e-05, "loss": 0.5964, "mean_token_accuracy": 0.8331712484359741, "num_tokens": 35278809.0, "step": 931 }, { "epoch": 0.11855997964635542, "ewc_loss": 0.076171875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 3.2113208770751953, "learning_rate": 1.9732937685459942e-05, "loss": 0.5327, "mean_token_accuracy": 0.8473609685897827, "num_tokens": 35318342.0, "step": 932 }, { "epoch": 0.11868718992494594, "ewc_loss": 0.0888671875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 8.296966552734375e-05, "grad_norm": 3.8206725120544434, "learning_rate": 1.9754133107248833e-05, "loss": 0.5446, "mean_token_accuracy": 0.8500896692276001, "num_tokens": 35361802.0, "step": 933 }, { "epoch": 0.11881440020353645, "ewc_loss": 0.07421875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.1392054557800293, "learning_rate": 1.9775328529037728e-05, "loss": 0.5029, "mean_token_accuracy": 0.8582211136817932, "num_tokens": 35401244.0, "step": 934 }, { "epoch": 0.11894161048212695, "ewc_loss": 0.0966796875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 9.059906005859375e-05, "grad_norm": 3.975756883621216, "learning_rate": 1.9796523950826622e-05, "loss": 0.5626, "mean_token_accuracy": 0.84882652759552, "num_tokens": 35438557.0, "step": 935 }, { "epoch": 0.11906882076071747, "ewc_loss": 0.0771484375, "ewc_loss_diag": 5.793571472167969e-05, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 3.1916065216064453, "learning_rate": 1.9817719372615516e-05, "loss": 0.5323, "mean_token_accuracy": 0.8513188362121582, "num_tokens": 35471686.0, "step": 936 }, { "epoch": 0.11919603103930798, "ewc_loss": 0.1015625, "ewc_loss_diag": 5.793571472167969e-05, "ewc_loss_parallel": 9.584426879882812e-05, "grad_norm": 4.055111408233643, "learning_rate": 1.9838914794404407e-05, "loss": 0.5823, "mean_token_accuracy": 0.8401091694831848, "num_tokens": 35514288.0, "step": 937 }, { "epoch": 0.11932324131789848, "ewc_loss": 0.08544921875, "ewc_loss_diag": 5.793571472167969e-05, "ewc_loss_parallel": 7.963180541992188e-05, "grad_norm": 3.363355875015259, "learning_rate": 1.9860110216193302e-05, "loss": 0.5741, "mean_token_accuracy": 0.8407545685768127, "num_tokens": 35547588.0, "step": 938 }, { "epoch": 0.119450451596489, "ewc_loss": 0.09814453125, "ewc_loss_diag": 5.793571472167969e-05, "ewc_loss_parallel": 9.250640869140625e-05, "grad_norm": 3.95227313041687, "learning_rate": 1.9881305637982196e-05, "loss": 0.5226, "mean_token_accuracy": 0.8590285181999207, "num_tokens": 35582099.0, "step": 939 }, { "epoch": 0.11957766187507951, "ewc_loss": 0.08935546875, "ewc_loss_diag": 5.793571472167969e-05, "ewc_loss_parallel": 8.344650268554688e-05, "grad_norm": 3.4915049076080322, "learning_rate": 1.990250105977109e-05, "loss": 0.4974, "mean_token_accuracy": 0.8623108863830566, "num_tokens": 35616173.0, "step": 940 }, { "epoch": 0.11970487215367001, "ewc_loss": 0.08740234375, "ewc_loss_diag": 5.793571472167969e-05, "ewc_loss_parallel": 8.153915405273438e-05, "grad_norm": 3.691337823867798, "learning_rate": 1.9923696481559985e-05, "loss": 0.5931, "mean_token_accuracy": 0.8383296132087708, "num_tokens": 35654801.0, "step": 941 }, { "epoch": 0.11983208243226053, "ewc_loss": 0.087890625, "ewc_loss_diag": 5.793571472167969e-05, "ewc_loss_parallel": 8.20159912109375e-05, "grad_norm": 3.545520067214966, "learning_rate": 1.994489190334888e-05, "loss": 0.538, "mean_token_accuracy": 0.8524585962295532, "num_tokens": 35692219.0, "step": 942 }, { "epoch": 0.11995929271085104, "ewc_loss": 0.0791015625, "ewc_loss_diag": 5.841255187988281e-05, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 3.3970444202423096, "learning_rate": 1.996608732513777e-05, "loss": 0.5909, "mean_token_accuracy": 0.8356618881225586, "num_tokens": 35732886.0, "step": 943 }, { "epoch": 0.12008650298944154, "ewc_loss": 0.08544921875, "ewc_loss_diag": 5.841255187988281e-05, "ewc_loss_parallel": 7.963180541992188e-05, "grad_norm": 3.602849245071411, "learning_rate": 1.9987282746926665e-05, "loss": 0.4871, "mean_token_accuracy": 0.8669781684875488, "num_tokens": 35766796.0, "step": 944 }, { "epoch": 0.12021371326803205, "ewc_loss": 0.07373046875, "ewc_loss_diag": 5.841255187988281e-05, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 3.18856143951416, "learning_rate": 2.000847816871556e-05, "loss": 0.5392, "mean_token_accuracy": 0.8461982011795044, "num_tokens": 35803065.0, "step": 945 }, { "epoch": 0.12034092354662257, "ewc_loss": 0.0849609375, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 7.915496826171875e-05, "grad_norm": 3.671849250793457, "learning_rate": 2.0029673590504454e-05, "loss": 0.5429, "mean_token_accuracy": 0.8483717441558838, "num_tokens": 35836490.0, "step": 946 }, { "epoch": 0.12046813382521308, "ewc_loss": 0.07080078125, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.0691332817077637, "learning_rate": 2.0050869012293348e-05, "loss": 0.4861, "mean_token_accuracy": 0.8648491501808167, "num_tokens": 35877239.0, "step": 947 }, { "epoch": 0.12059534410380358, "ewc_loss": 0.0869140625, "ewc_loss_diag": 5.91278076171875e-05, "ewc_loss_parallel": 8.106231689453125e-05, "grad_norm": 3.782219171524048, "learning_rate": 2.007206443408224e-05, "loss": 0.6064, "mean_token_accuracy": 0.8314369916915894, "num_tokens": 35917779.0, "step": 948 }, { "epoch": 0.1207225543823941, "ewc_loss": 0.06982421875, "ewc_loss_diag": 5.91278076171875e-05, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.0459938049316406, "learning_rate": 2.0093259855871134e-05, "loss": 0.5211, "mean_token_accuracy": 0.8497948050498962, "num_tokens": 35955083.0, "step": 949 }, { "epoch": 0.12084976466098461, "ewc_loss": 0.09228515625, "ewc_loss_diag": 5.91278076171875e-05, "ewc_loss_parallel": 8.630752563476562e-05, "grad_norm": 3.897862434387207, "learning_rate": 2.0114455277660028e-05, "loss": 0.5697, "mean_token_accuracy": 0.8442561626434326, "num_tokens": 35999581.0, "step": 950 }, { "epoch": 0.12097697493957511, "ewc_loss": 0.0732421875, "ewc_loss_diag": 5.91278076171875e-05, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.112938642501831, "learning_rate": 2.0135650699448923e-05, "loss": 0.5954, "mean_token_accuracy": 0.8338304758071899, "num_tokens": 36040891.0, "step": 951 }, { "epoch": 0.12110418521816563, "ewc_loss": 0.095703125, "ewc_loss_diag": 5.91278076171875e-05, "ewc_loss_parallel": 8.96453857421875e-05, "grad_norm": 3.984428644180298, "learning_rate": 2.0156846121237814e-05, "loss": 0.5381, "mean_token_accuracy": 0.8496502637863159, "num_tokens": 36074958.0, "step": 952 }, { "epoch": 0.12123139549675614, "ewc_loss": 0.0791015625, "ewc_loss_diag": 5.936622619628906e-05, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 3.2220568656921387, "learning_rate": 2.0178041543026708e-05, "loss": 0.557, "mean_token_accuracy": 0.8406095504760742, "num_tokens": 36115211.0, "step": 953 }, { "epoch": 0.12135860577534664, "ewc_loss": 0.09814453125, "ewc_loss_diag": 5.936622619628906e-05, "ewc_loss_parallel": 9.250640869140625e-05, "grad_norm": 3.956500768661499, "learning_rate": 2.0199236964815602e-05, "loss": 0.5731, "mean_token_accuracy": 0.8463497757911682, "num_tokens": 36156876.0, "step": 954 }, { "epoch": 0.12148581605393716, "ewc_loss": 0.0849609375, "ewc_loss_diag": 5.936622619628906e-05, "ewc_loss_parallel": 7.915496826171875e-05, "grad_norm": 3.3508315086364746, "learning_rate": 2.0220432386604493e-05, "loss": 0.5411, "mean_token_accuracy": 0.8471247553825378, "num_tokens": 36197277.0, "step": 955 }, { "epoch": 0.12161302633252767, "ewc_loss": 0.09130859375, "ewc_loss_diag": 5.9604644775390625e-05, "ewc_loss_parallel": 8.535385131835938e-05, "grad_norm": 3.8125717639923096, "learning_rate": 2.0241627808393388e-05, "loss": 0.577, "mean_token_accuracy": 0.8406962156295776, "num_tokens": 36232839.0, "step": 956 }, { "epoch": 0.12174023661111817, "ewc_loss": 0.0859375, "ewc_loss_diag": 5.9604644775390625e-05, "ewc_loss_parallel": 8.0108642578125e-05, "grad_norm": 3.4252586364746094, "learning_rate": 2.0262823230182282e-05, "loss": 0.5642, "mean_token_accuracy": 0.8425164222717285, "num_tokens": 36273531.0, "step": 957 }, { "epoch": 0.12186744688970869, "ewc_loss": 0.0859375, "ewc_loss_diag": 5.9604644775390625e-05, "ewc_loss_parallel": 8.0108642578125e-05, "grad_norm": 3.6288630962371826, "learning_rate": 2.0284018651971177e-05, "loss": 0.5777, "mean_token_accuracy": 0.8363302946090698, "num_tokens": 36309325.0, "step": 958 }, { "epoch": 0.1219946571682992, "ewc_loss": 0.083984375, "ewc_loss_diag": 5.9604644775390625e-05, "ewc_loss_parallel": 7.82012939453125e-05, "grad_norm": 3.451416015625, "learning_rate": 2.0305214073760068e-05, "loss": 0.5495, "mean_token_accuracy": 0.8471516966819763, "num_tokens": 36344457.0, "step": 959 }, { "epoch": 0.12212186744688971, "ewc_loss": 0.080078125, "ewc_loss_diag": 5.9604644775390625e-05, "ewc_loss_parallel": 7.43865966796875e-05, "grad_norm": 3.482353448867798, "learning_rate": 2.0326409495548962e-05, "loss": 0.5471, "mean_token_accuracy": 0.8431199193000793, "num_tokens": 36376731.0, "step": 960 }, { "epoch": 0.12224907772548022, "ewc_loss": 0.0810546875, "ewc_loss_diag": 5.9604644775390625e-05, "ewc_loss_parallel": 7.534027099609375e-05, "grad_norm": 3.436251163482666, "learning_rate": 2.0347604917337857e-05, "loss": 0.5075, "mean_token_accuracy": 0.8586581349372864, "num_tokens": 36415529.0, "step": 961 }, { "epoch": 0.12237628800407073, "ewc_loss": 0.078125, "ewc_loss_diag": 5.984306335449219e-05, "ewc_loss_parallel": 7.2479248046875e-05, "grad_norm": 3.399393320083618, "learning_rate": 2.036880033912675e-05, "loss": 0.5321, "mean_token_accuracy": 0.8486937880516052, "num_tokens": 36449522.0, "step": 962 }, { "epoch": 0.12250349828266124, "ewc_loss": 0.078125, "ewc_loss_diag": 5.984306335449219e-05, "ewc_loss_parallel": 7.2479248046875e-05, "grad_norm": 3.4073498249053955, "learning_rate": 2.0389995760915642e-05, "loss": 0.4811, "mean_token_accuracy": 0.8636050224304199, "num_tokens": 36483010.0, "step": 963 }, { "epoch": 0.12263070856125174, "ewc_loss": 0.07666015625, "ewc_loss_diag": 5.984306335449219e-05, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 3.365560531616211, "learning_rate": 2.0411191182704537e-05, "loss": 0.5527, "mean_token_accuracy": 0.8468123078346252, "num_tokens": 36527993.0, "step": 964 }, { "epoch": 0.12275791883984226, "ewc_loss": 0.07568359375, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 3.3623874187469482, "learning_rate": 2.043238660449343e-05, "loss": 0.5219, "mean_token_accuracy": 0.8505059480667114, "num_tokens": 36568479.0, "step": 965 }, { "epoch": 0.12288512911843277, "ewc_loss": 0.0751953125, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 3.3845040798187256, "learning_rate": 2.0453582026282325e-05, "loss": 0.5585, "mean_token_accuracy": 0.8412636518478394, "num_tokens": 36607196.0, "step": 966 }, { "epoch": 0.12301233939702327, "ewc_loss": 0.0732421875, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.324443817138672, "learning_rate": 2.0474777448071216e-05, "loss": 0.5234, "mean_token_accuracy": 0.8503016233444214, "num_tokens": 36640845.0, "step": 967 }, { "epoch": 0.12313954967561379, "ewc_loss": 0.07421875, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.3898794651031494, "learning_rate": 2.049597286986011e-05, "loss": 0.524, "mean_token_accuracy": 0.8518255949020386, "num_tokens": 36682182.0, "step": 968 }, { "epoch": 0.1232667599542043, "ewc_loss": 0.07275390625, "ewc_loss_diag": 6.0558319091796875e-05, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.2993102073669434, "learning_rate": 2.0517168291649005e-05, "loss": 0.5641, "mean_token_accuracy": 0.8401720523834229, "num_tokens": 36723629.0, "step": 969 }, { "epoch": 0.1233939702327948, "ewc_loss": 0.07421875, "ewc_loss_diag": 6.0558319091796875e-05, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.4115352630615234, "learning_rate": 2.05383637134379e-05, "loss": 0.5497, "mean_token_accuracy": 0.8437009453773499, "num_tokens": 36763560.0, "step": 970 }, { "epoch": 0.12352118051138532, "ewc_loss": 0.07177734375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.2832298278808594, "learning_rate": 2.055955913522679e-05, "loss": 0.6135, "mean_token_accuracy": 0.8194952011108398, "num_tokens": 36802267.0, "step": 971 }, { "epoch": 0.12364839078997583, "ewc_loss": 0.0751953125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 3.437836170196533, "learning_rate": 2.0580754557015685e-05, "loss": 0.5065, "mean_token_accuracy": 0.8589348793029785, "num_tokens": 36840047.0, "step": 972 }, { "epoch": 0.12377560106856635, "ewc_loss": 0.0703125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.2633652687072754, "learning_rate": 2.060194997880458e-05, "loss": 0.5612, "mean_token_accuracy": 0.8409978151321411, "num_tokens": 36876788.0, "step": 973 }, { "epoch": 0.12390281134715685, "ewc_loss": 0.07568359375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 3.4770305156707764, "learning_rate": 2.062314540059347e-05, "loss": 0.5189, "mean_token_accuracy": 0.8513386845588684, "num_tokens": 36909954.0, "step": 974 }, { "epoch": 0.12403002162574736, "ewc_loss": 0.06884765625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.2122714519500732, "learning_rate": 2.0644340822382365e-05, "loss": 0.5476, "mean_token_accuracy": 0.8413248062133789, "num_tokens": 36948371.0, "step": 975 }, { "epoch": 0.12415723190433788, "ewc_loss": 0.0771484375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 3.559865951538086, "learning_rate": 2.066553624417126e-05, "loss": 0.5645, "mean_token_accuracy": 0.838474452495575, "num_tokens": 36983304.0, "step": 976 }, { "epoch": 0.12428444218292838, "ewc_loss": 0.0673828125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.1199607849121094, "learning_rate": 2.0686731665960154e-05, "loss": 0.4788, "mean_token_accuracy": 0.8614468574523926, "num_tokens": 37019644.0, "step": 977 }, { "epoch": 0.12441165246151889, "ewc_loss": 0.083984375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 7.772445678710938e-05, "grad_norm": 3.746734380722046, "learning_rate": 2.0707927087749045e-05, "loss": 0.5653, "mean_token_accuracy": 0.8404420614242554, "num_tokens": 37056906.0, "step": 978 }, { "epoch": 0.1245388627401094, "ewc_loss": 0.06494140625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 2.976414918899536, "learning_rate": 2.072912250953794e-05, "loss": 0.5024, "mean_token_accuracy": 0.8535906672477722, "num_tokens": 37090396.0, "step": 979 }, { "epoch": 0.1246660730186999, "ewc_loss": 0.0947265625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 8.821487426757812e-05, "grad_norm": 4.066446781158447, "learning_rate": 2.0750317931326834e-05, "loss": 0.5714, "mean_token_accuracy": 0.8442787528038025, "num_tokens": 37129246.0, "step": 980 }, { "epoch": 0.12479328329729042, "ewc_loss": 0.06591796875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 2.844954490661621, "learning_rate": 2.0771513353115728e-05, "loss": 0.4911, "mean_token_accuracy": 0.8574191927909851, "num_tokens": 37166344.0, "step": 981 }, { "epoch": 0.12492049357588093, "ewc_loss": 0.1220703125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 0.00011587142944335938, "grad_norm": 4.577681064605713, "learning_rate": 2.079270877490462e-05, "loss": 0.5306, "mean_token_accuracy": 0.8653706312179565, "num_tokens": 37202827.0, "step": 982 }, { "epoch": 0.12504770385447145, "ewc_loss": 0.08349609375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 7.724761962890625e-05, "grad_norm": 3.090794324874878, "learning_rate": 2.0813904196693514e-05, "loss": 0.5572, "mean_token_accuracy": 0.8414137959480286, "num_tokens": 37239606.0, "step": 983 }, { "epoch": 0.12517491413306195, "ewc_loss": 0.1474609375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 0.000141143798828125, "grad_norm": 5.0492448806762695, "learning_rate": 2.0835099618482408e-05, "loss": 0.616, "mean_token_accuracy": 0.8426276445388794, "num_tokens": 37275154.0, "step": 984 }, { "epoch": 0.12530212441165245, "ewc_loss": 0.1328125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 0.00012683868408203125, "grad_norm": 3.889575958251953, "learning_rate": 2.0856295040271302e-05, "loss": 0.558, "mean_token_accuracy": 0.8594170808792114, "num_tokens": 37311849.0, "step": 985 }, { "epoch": 0.12542933469024298, "ewc_loss": 0.1357421875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 0.00012969970703125, "grad_norm": 4.877871513366699, "learning_rate": 2.0877490462060197e-05, "loss": 0.6149, "mean_token_accuracy": 0.84339439868927, "num_tokens": 37353174.0, "step": 986 }, { "epoch": 0.12555654496883348, "ewc_loss": 0.142578125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.17293119430542, "learning_rate": 2.0898685883849088e-05, "loss": 0.5753, "mean_token_accuracy": 0.8571951985359192, "num_tokens": 37392081.0, "step": 987 }, { "epoch": 0.12568375524742398, "ewc_loss": 0.11865234375, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 0.0001125335693359375, "grad_norm": 4.571597576141357, "learning_rate": 2.0919881305637982e-05, "loss": 0.5487, "mean_token_accuracy": 0.8580989837646484, "num_tokens": 37430639.0, "step": 988 }, { "epoch": 0.1258109655260145, "ewc_loss": 0.1279296875, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 0.00012159347534179688, "grad_norm": 4.016204357147217, "learning_rate": 2.0941076727426877e-05, "loss": 0.6423, "mean_token_accuracy": 0.8295568227767944, "num_tokens": 37467126.0, "step": 989 }, { "epoch": 0.125938175804605, "ewc_loss": 0.1142578125, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 0.00010776519775390625, "grad_norm": 4.35176944732666, "learning_rate": 2.096227214921577e-05, "loss": 0.5666, "mean_token_accuracy": 0.8508691787719727, "num_tokens": 37505657.0, "step": 990 }, { "epoch": 0.12606538608319554, "ewc_loss": 0.10546875, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 9.918212890625e-05, "grad_norm": 3.786024332046509, "learning_rate": 2.0983467571004666e-05, "loss": 0.5628, "mean_token_accuracy": 0.8503637313842773, "num_tokens": 37549262.0, "step": 991 }, { "epoch": 0.12619259636178604, "ewc_loss": 0.10791015625, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 0.00010156631469726562, "grad_norm": 4.216907501220703, "learning_rate": 2.100466299279356e-05, "loss": 0.5718, "mean_token_accuracy": 0.8436853289604187, "num_tokens": 37589967.0, "step": 992 }, { "epoch": 0.12631980664037654, "ewc_loss": 0.09619140625, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 8.96453857421875e-05, "grad_norm": 3.598102569580078, "learning_rate": 2.102585841458245e-05, "loss": 0.5655, "mean_token_accuracy": 0.8464293479919434, "num_tokens": 37635686.0, "step": 993 }, { "epoch": 0.12644701691896706, "ewc_loss": 0.1083984375, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 0.00010204315185546875, "grad_norm": 4.200845718383789, "learning_rate": 2.1047053836371345e-05, "loss": 0.619, "mean_token_accuracy": 0.8372375965118408, "num_tokens": 37673753.0, "step": 994 }, { "epoch": 0.12657422719755757, "ewc_loss": 0.08935546875, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 8.296966552734375e-05, "grad_norm": 3.501068353652954, "learning_rate": 2.106824925816024e-05, "loss": 0.5045, "mean_token_accuracy": 0.8615376949310303, "num_tokens": 37706345.0, "step": 995 }, { "epoch": 0.12670143747614807, "ewc_loss": 0.11181640625, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 0.00010538101196289062, "grad_norm": 4.248007297515869, "learning_rate": 2.1089444679949134e-05, "loss": 0.6013, "mean_token_accuracy": 0.8370252847671509, "num_tokens": 37750356.0, "step": 996 }, { "epoch": 0.1268286477547386, "ewc_loss": 0.08984375, "ewc_loss_diag": 6.246566772460938e-05, "ewc_loss_parallel": 8.344650268554688e-05, "grad_norm": 3.476245641708374, "learning_rate": 2.1110640101738025e-05, "loss": 0.5717, "mean_token_accuracy": 0.8427751064300537, "num_tokens": 37791466.0, "step": 997 }, { "epoch": 0.1269558580333291, "ewc_loss": 0.1171875, "ewc_loss_diag": 6.246566772460938e-05, "ewc_loss_parallel": 0.000110626220703125, "grad_norm": 4.322146415710449, "learning_rate": 2.113183552352692e-05, "loss": 0.6059, "mean_token_accuracy": 0.8385998010635376, "num_tokens": 37837257.0, "step": 998 }, { "epoch": 0.1270830683119196, "ewc_loss": 0.09521484375, "ewc_loss_diag": 6.246566772460938e-05, "ewc_loss_parallel": 8.869171142578125e-05, "grad_norm": 3.56276798248291, "learning_rate": 2.1153030945315814e-05, "loss": 0.5719, "mean_token_accuracy": 0.8440828919410706, "num_tokens": 37880223.0, "step": 999 }, { "epoch": 0.12721027859051012, "ewc_loss": 0.1181640625, "ewc_loss_diag": 6.246566772460938e-05, "ewc_loss_parallel": 0.00011205673217773438, "grad_norm": 4.318793773651123, "learning_rate": 2.117422636710471e-05, "loss": 0.6695, "mean_token_accuracy": 0.826901912689209, "num_tokens": 37917662.0, "step": 1000 }, { "epoch": 0.12733748886910062, "ewc_loss": 0.1005859375, "ewc_loss_diag": 6.246566772460938e-05, "ewc_loss_parallel": 9.441375732421875e-05, "grad_norm": 3.685142755508423, "learning_rate": 2.11954217888936e-05, "loss": 0.5822, "mean_token_accuracy": 0.8390526175498962, "num_tokens": 37961985.0, "step": 1001 }, { "epoch": 0.12746469914769112, "ewc_loss": 0.1083984375, "ewc_loss_diag": 6.246566772460938e-05, "ewc_loss_parallel": 0.00010204315185546875, "grad_norm": 4.143968105316162, "learning_rate": 2.1216617210682494e-05, "loss": 0.5766, "mean_token_accuracy": 0.8481607437133789, "num_tokens": 37998960.0, "step": 1002 }, { "epoch": 0.12759190942628165, "ewc_loss": 0.1005859375, "ewc_loss_diag": 6.246566772460938e-05, "ewc_loss_parallel": 9.441375732421875e-05, "grad_norm": 3.7491579055786133, "learning_rate": 2.123781263247139e-05, "loss": 0.5453, "mean_token_accuracy": 0.8551092147827148, "num_tokens": 38040920.0, "step": 1003 }, { "epoch": 0.12771911970487215, "ewc_loss": 0.09814453125, "ewc_loss_diag": 6.246566772460938e-05, "ewc_loss_parallel": 9.202957153320312e-05, "grad_norm": 3.9078242778778076, "learning_rate": 2.1259008054260283e-05, "loss": 0.5855, "mean_token_accuracy": 0.839291512966156, "num_tokens": 38071814.0, "step": 1004 }, { "epoch": 0.12784632998346265, "ewc_loss": 0.0966796875, "ewc_loss_diag": 6.29425048828125e-05, "ewc_loss_parallel": 9.012222290039062e-05, "grad_norm": 3.7115724086761475, "learning_rate": 2.1280203476049174e-05, "loss": 0.6027, "mean_token_accuracy": 0.8365273475646973, "num_tokens": 38112784.0, "step": 1005 }, { "epoch": 0.12797354026205318, "ewc_loss": 0.0888671875, "ewc_loss_diag": 6.29425048828125e-05, "ewc_loss_parallel": 8.249282836914062e-05, "grad_norm": 3.7037839889526367, "learning_rate": 2.130139889783807e-05, "loss": 0.6089, "mean_token_accuracy": 0.8327439427375793, "num_tokens": 38152140.0, "step": 1006 }, { "epoch": 0.12810075054064368, "ewc_loss": 0.08740234375, "ewc_loss_diag": 6.29425048828125e-05, "ewc_loss_parallel": 8.106231689453125e-05, "grad_norm": 3.6344246864318848, "learning_rate": 2.1322594319626963e-05, "loss": 0.5185, "mean_token_accuracy": 0.8583405017852783, "num_tokens": 38188293.0, "step": 1007 }, { "epoch": 0.12822796081923418, "ewc_loss": 0.0859375, "ewc_loss_diag": 6.29425048828125e-05, "ewc_loss_parallel": 7.963180541992188e-05, "grad_norm": 3.603315591812134, "learning_rate": 2.1343789741415854e-05, "loss": 0.566, "mean_token_accuracy": 0.8413940668106079, "num_tokens": 38229136.0, "step": 1008 }, { "epoch": 0.1283551710978247, "ewc_loss": 0.0830078125, "ewc_loss_diag": 6.29425048828125e-05, "ewc_loss_parallel": 7.677078247070312e-05, "grad_norm": 3.549492120742798, "learning_rate": 2.1364985163204748e-05, "loss": 0.5813, "mean_token_accuracy": 0.8390242457389832, "num_tokens": 38273330.0, "step": 1009 }, { "epoch": 0.1284823813764152, "ewc_loss": 0.08349609375, "ewc_loss_diag": 6.341934204101562e-05, "ewc_loss_parallel": 7.724761962890625e-05, "grad_norm": 3.5836362838745117, "learning_rate": 2.1386180584993643e-05, "loss": 0.5436, "mean_token_accuracy": 0.8471120595932007, "num_tokens": 38315066.0, "step": 1010 }, { "epoch": 0.1286095916550057, "ewc_loss": 0.0791015625, "ewc_loss_diag": 6.341934204101562e-05, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 3.448396682739258, "learning_rate": 2.1407376006782537e-05, "loss": 0.5449, "mean_token_accuracy": 0.8453952074050903, "num_tokens": 38353430.0, "step": 1011 }, { "epoch": 0.12873680193359624, "ewc_loss": 0.08251953125, "ewc_loss_diag": 6.341934204101562e-05, "ewc_loss_parallel": 7.62939453125e-05, "grad_norm": 3.6224124431610107, "learning_rate": 2.1428571428571428e-05, "loss": 0.5781, "mean_token_accuracy": 0.8372169137001038, "num_tokens": 38391919.0, "step": 1012 }, { "epoch": 0.12886401221218674, "ewc_loss": 0.076171875, "ewc_loss_diag": 6.341934204101562e-05, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 3.3595733642578125, "learning_rate": 2.1449766850360323e-05, "loss": 0.5213, "mean_token_accuracy": 0.854300856590271, "num_tokens": 38430104.0, "step": 1013 }, { "epoch": 0.12899122249077727, "ewc_loss": 0.08447265625, "ewc_loss_diag": 6.341934204101562e-05, "ewc_loss_parallel": 7.82012939453125e-05, "grad_norm": 3.7158379554748535, "learning_rate": 2.1470962272149217e-05, "loss": 0.5721, "mean_token_accuracy": 0.8402059674263, "num_tokens": 38468775.0, "step": 1014 }, { "epoch": 0.12911843276936777, "ewc_loss": 0.07373046875, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.2413551807403564, "learning_rate": 2.149215769393811e-05, "loss": 0.4485, "mean_token_accuracy": 0.8757029175758362, "num_tokens": 38510330.0, "step": 1015 }, { "epoch": 0.12924564304795827, "ewc_loss": 0.09130859375, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 8.487701416015625e-05, "grad_norm": 3.8896031379699707, "learning_rate": 2.1513353115727002e-05, "loss": 0.6021, "mean_token_accuracy": 0.8344254493713379, "num_tokens": 38543840.0, "step": 1016 }, { "epoch": 0.1293728533265488, "ewc_loss": 0.07275390625, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.1500461101531982, "learning_rate": 2.1534548537515897e-05, "loss": 0.5404, "mean_token_accuracy": 0.8493101000785828, "num_tokens": 38581670.0, "step": 1017 }, { "epoch": 0.1295000636051393, "ewc_loss": 0.1005859375, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 9.441375732421875e-05, "grad_norm": 4.1231560707092285, "learning_rate": 2.155574395930479e-05, "loss": 0.5581, "mean_token_accuracy": 0.8471074104309082, "num_tokens": 38612642.0, "step": 1018 }, { "epoch": 0.1296272738837298, "ewc_loss": 0.07421875, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 3.130242347717285, "learning_rate": 2.1576939381093686e-05, "loss": 0.5691, "mean_token_accuracy": 0.8426134586334229, "num_tokens": 38647579.0, "step": 1019 }, { "epoch": 0.12975448416232033, "ewc_loss": 0.1162109375, "ewc_loss_diag": 6.4849853515625e-05, "ewc_loss_parallel": 0.00010967254638671875, "grad_norm": 4.376181125640869, "learning_rate": 2.1598134802882577e-05, "loss": 0.6122, "mean_token_accuracy": 0.8338617086410522, "num_tokens": 38683114.0, "step": 1020 }, { "epoch": 0.12988169444091083, "ewc_loss": 0.0849609375, "ewc_loss_diag": 6.4849853515625e-05, "ewc_loss_parallel": 7.867813110351562e-05, "grad_norm": 3.3146164417266846, "learning_rate": 2.161933022467147e-05, "loss": 0.5251, "mean_token_accuracy": 0.8543521761894226, "num_tokens": 38722990.0, "step": 1021 }, { "epoch": 0.13000890471950133, "ewc_loss": 0.12451171875, "ewc_loss_diag": 6.4849853515625e-05, "ewc_loss_parallel": 0.000118255615234375, "grad_norm": 4.437989234924316, "learning_rate": 2.1640525646460366e-05, "loss": 0.5905, "mean_token_accuracy": 0.8452957272529602, "num_tokens": 38765041.0, "step": 1022 }, { "epoch": 0.13013611499809186, "ewc_loss": 0.10302734375, "ewc_loss_diag": 6.4849853515625e-05, "ewc_loss_parallel": 9.679794311523438e-05, "grad_norm": 3.65769624710083, "learning_rate": 2.166172106824926e-05, "loss": 0.5324, "mean_token_accuracy": 0.8569649457931519, "num_tokens": 38802241.0, "step": 1023 }, { "epoch": 0.13026332527668236, "ewc_loss": 0.1083984375, "ewc_loss_diag": 6.4849853515625e-05, "ewc_loss_parallel": 0.00010204315185546875, "grad_norm": 4.140275001525879, "learning_rate": 2.168291649003815e-05, "loss": 0.5651, "mean_token_accuracy": 0.8491906523704529, "num_tokens": 38839444.0, "step": 1024 }, { "epoch": 0.13039053555527286, "ewc_loss": 0.10986328125, "ewc_loss_diag": 6.4849853515625e-05, "ewc_loss_parallel": 0.00010347366333007812, "grad_norm": 3.8703691959381104, "learning_rate": 2.1704111911827045e-05, "loss": 0.5485, "mean_token_accuracy": 0.8543947339057922, "num_tokens": 38882102.0, "step": 1025 }, { "epoch": 0.13051774583386339, "ewc_loss": 0.08837890625, "ewc_loss_diag": 6.4849853515625e-05, "ewc_loss_parallel": 8.20159912109375e-05, "grad_norm": 3.643221378326416, "learning_rate": 2.172530733361594e-05, "loss": 0.5181, "mean_token_accuracy": 0.8593300580978394, "num_tokens": 38925855.0, "step": 1026 }, { "epoch": 0.13064495611245389, "ewc_loss": 0.111328125, "ewc_loss_diag": 6.4849853515625e-05, "ewc_loss_parallel": 0.0001049041748046875, "grad_norm": 4.088473796844482, "learning_rate": 2.174650275540483e-05, "loss": 0.5971, "mean_token_accuracy": 0.8400928974151611, "num_tokens": 38962293.0, "step": 1027 }, { "epoch": 0.1307721663910444, "ewc_loss": 0.078125, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 3.276155948638916, "learning_rate": 2.1767698177193725e-05, "loss": 0.5299, "mean_token_accuracy": 0.8492878675460815, "num_tokens": 38999000.0, "step": 1028 }, { "epoch": 0.13089937666963491, "ewc_loss": 0.119140625, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 0.00011301040649414062, "grad_norm": 4.411340236663818, "learning_rate": 2.178889359898262e-05, "loss": 0.6344, "mean_token_accuracy": 0.8295608758926392, "num_tokens": 39032819.0, "step": 1029 }, { "epoch": 0.13102658694822542, "ewc_loss": 0.08251953125, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 7.62939453125e-05, "grad_norm": 3.2795114517211914, "learning_rate": 2.1810089020771514e-05, "loss": 0.5221, "mean_token_accuracy": 0.8550988435745239, "num_tokens": 39070247.0, "step": 1030 }, { "epoch": 0.13115379722681592, "ewc_loss": 0.1328125, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 0.000125885009765625, "grad_norm": 4.635194301605225, "learning_rate": 2.1831284442560405e-05, "loss": 0.5662, "mean_token_accuracy": 0.8573951721191406, "num_tokens": 39105310.0, "step": 1031 }, { "epoch": 0.13128100750540644, "ewc_loss": 0.10400390625, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 9.775161743164062e-05, "grad_norm": 3.634416103363037, "learning_rate": 2.18524798643493e-05, "loss": 0.5691, "mean_token_accuracy": 0.8456183671951294, "num_tokens": 39144218.0, "step": 1032 }, { "epoch": 0.13140821778399694, "ewc_loss": 0.125, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 0.00011873245239257812, "grad_norm": 4.50261116027832, "learning_rate": 2.1873675286138194e-05, "loss": 0.58, "mean_token_accuracy": 0.8472083806991577, "num_tokens": 39177248.0, "step": 1033 }, { "epoch": 0.13153542806258745, "ewc_loss": 0.11572265625, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 0.00010919570922851562, "grad_norm": 3.865834951400757, "learning_rate": 2.189487070792709e-05, "loss": 0.5767, "mean_token_accuracy": 0.8472819328308105, "num_tokens": 39214230.0, "step": 1034 }, { "epoch": 0.13166263834117797, "ewc_loss": 0.11083984375, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 0.00010442733764648438, "grad_norm": 4.174887657165527, "learning_rate": 2.1916066129715983e-05, "loss": 0.5684, "mean_token_accuracy": 0.8471686840057373, "num_tokens": 39252189.0, "step": 1035 }, { "epoch": 0.13178984861976847, "ewc_loss": 0.10888671875, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 0.00010251998901367188, "grad_norm": 3.8542730808258057, "learning_rate": 2.1937261551504877e-05, "loss": 0.5152, "mean_token_accuracy": 0.8650339245796204, "num_tokens": 39288058.0, "step": 1036 }, { "epoch": 0.13191705889835897, "ewc_loss": 0.09912109375, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 9.298324584960938e-05, "grad_norm": 3.9240827560424805, "learning_rate": 2.195845697329377e-05, "loss": 0.5152, "mean_token_accuracy": 0.8579907417297363, "num_tokens": 39322157.0, "step": 1037 }, { "epoch": 0.1320442691769495, "ewc_loss": 0.0986328125, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 9.1552734375e-05, "grad_norm": 3.739769697189331, "learning_rate": 2.1979652395082663e-05, "loss": 0.53, "mean_token_accuracy": 0.8511033058166504, "num_tokens": 39357678.0, "step": 1038 }, { "epoch": 0.13217147945554, "ewc_loss": 0.095703125, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 8.916854858398438e-05, "grad_norm": 3.822299003601074, "learning_rate": 2.2000847816871557e-05, "loss": 0.5485, "mean_token_accuracy": 0.8497304320335388, "num_tokens": 39398909.0, "step": 1039 }, { "epoch": 0.13229868973413053, "ewc_loss": 0.0869140625, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 8.058547973632812e-05, "grad_norm": 3.587639570236206, "learning_rate": 2.202204323866045e-05, "loss": 0.6128, "mean_token_accuracy": 0.8297585844993591, "num_tokens": 39437934.0, "step": 1040 }, { "epoch": 0.13242590001272103, "ewc_loss": 0.0927734375, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 8.58306884765625e-05, "grad_norm": 3.859351396560669, "learning_rate": 2.2043238660449346e-05, "loss": 0.5403, "mean_token_accuracy": 0.8489798307418823, "num_tokens": 39471772.0, "step": 1041 }, { "epoch": 0.13255311029131153, "ewc_loss": 0.0810546875, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 7.43865966796875e-05, "grad_norm": 3.4148378372192383, "learning_rate": 2.206443408223824e-05, "loss": 0.5443, "mean_token_accuracy": 0.8520676493644714, "num_tokens": 39512398.0, "step": 1042 }, { "epoch": 0.13268032056990206, "ewc_loss": 0.09765625, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 9.059906005859375e-05, "grad_norm": 3.9767675399780273, "learning_rate": 2.208562950402713e-05, "loss": 0.5547, "mean_token_accuracy": 0.8498935699462891, "num_tokens": 39554735.0, "step": 1043 }, { "epoch": 0.13280753084849256, "ewc_loss": 0.078125, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 3.307075262069702, "learning_rate": 2.2106824925816026e-05, "loss": 0.5831, "mean_token_accuracy": 0.8390913605690002, "num_tokens": 39595397.0, "step": 1044 }, { "epoch": 0.13293474112708306, "ewc_loss": 0.103515625, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 9.679794311523438e-05, "grad_norm": 4.144688606262207, "learning_rate": 2.212802034760492e-05, "loss": 0.5949, "mean_token_accuracy": 0.8398325443267822, "num_tokens": 39632474.0, "step": 1045 }, { "epoch": 0.1330619514056736, "ewc_loss": 0.0791015625, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 7.2479248046875e-05, "grad_norm": 3.2736897468566895, "learning_rate": 2.214921576939381e-05, "loss": 0.5539, "mean_token_accuracy": 0.8412595987319946, "num_tokens": 39666408.0, "step": 1046 }, { "epoch": 0.1331891616842641, "ewc_loss": 0.115234375, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 0.0001087188720703125, "grad_norm": 4.3263444900512695, "learning_rate": 2.2170411191182706e-05, "loss": 0.6245, "mean_token_accuracy": 0.8326270580291748, "num_tokens": 39709033.0, "step": 1047 }, { "epoch": 0.1333163719628546, "ewc_loss": 0.0849609375, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 7.82012939453125e-05, "grad_norm": 3.3538525104522705, "learning_rate": 2.21916066129716e-05, "loss": 0.5682, "mean_token_accuracy": 0.8437012434005737, "num_tokens": 39749732.0, "step": 1048 }, { "epoch": 0.13344358224144512, "ewc_loss": 0.125, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 0.00011873245239257812, "grad_norm": 4.418773174285889, "learning_rate": 2.2212802034760495e-05, "loss": 0.6169, "mean_token_accuracy": 0.835890531539917, "num_tokens": 39788125.0, "step": 1049 }, { "epoch": 0.13357079252003562, "ewc_loss": 0.099609375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 9.298324584960938e-05, "grad_norm": 3.5614840984344482, "learning_rate": 2.2233997456549386e-05, "loss": 0.5325, "mean_token_accuracy": 0.8581333756446838, "num_tokens": 39830125.0, "step": 1050 }, { "epoch": 0.13369800279862612, "ewc_loss": 0.11962890625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 0.00011301040649414062, "grad_norm": 4.294439315795898, "learning_rate": 2.225519287833828e-05, "loss": 0.6082, "mean_token_accuracy": 0.8368252515792847, "num_tokens": 39868236.0, "step": 1051 }, { "epoch": 0.13382521307721665, "ewc_loss": 0.109375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 0.00010251998901367188, "grad_norm": 3.7676281929016113, "learning_rate": 2.2276388300127175e-05, "loss": 0.5653, "mean_token_accuracy": 0.8481662273406982, "num_tokens": 39913254.0, "step": 1052 }, { "epoch": 0.13395242335580715, "ewc_loss": 0.1025390625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 9.584426879882812e-05, "grad_norm": 3.95399808883667, "learning_rate": 2.229758372191607e-05, "loss": 0.568, "mean_token_accuracy": 0.8441431522369385, "num_tokens": 39948242.0, "step": 1053 }, { "epoch": 0.13407963363439765, "ewc_loss": 0.11083984375, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 0.00010395050048828125, "grad_norm": 3.914060592651367, "learning_rate": 2.231877914370496e-05, "loss": 0.5966, "mean_token_accuracy": 0.8407009840011597, "num_tokens": 39989196.0, "step": 1054 }, { "epoch": 0.13420684391298818, "ewc_loss": 0.08837890625, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 8.153915405273438e-05, "grad_norm": 3.573676109313965, "learning_rate": 2.2339974565493854e-05, "loss": 0.5753, "mean_token_accuracy": 0.8483258485794067, "num_tokens": 40028662.0, "step": 1055 }, { "epoch": 0.13433405419157868, "ewc_loss": 0.10888671875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 0.00010204315185546875, "grad_norm": 4.078902721405029, "learning_rate": 2.236116998728275e-05, "loss": 0.5916, "mean_token_accuracy": 0.8422700762748718, "num_tokens": 40066178.0, "step": 1056 }, { "epoch": 0.13446126447016918, "ewc_loss": 0.07958984375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 3.3297154903411865, "learning_rate": 2.2382365409071643e-05, "loss": 0.5576, "mean_token_accuracy": 0.8417879939079285, "num_tokens": 40106908.0, "step": 1057 }, { "epoch": 0.1345884747487597, "ewc_loss": 0.11328125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 0.00010633468627929688, "grad_norm": 4.276823997497559, "learning_rate": 2.2403560830860534e-05, "loss": 0.5523, "mean_token_accuracy": 0.8542645573616028, "num_tokens": 40146872.0, "step": 1058 }, { "epoch": 0.1347156850273502, "ewc_loss": 0.08154296875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 7.486343383789062e-05, "grad_norm": 3.2899537086486816, "learning_rate": 2.242475625264943e-05, "loss": 0.5496, "mean_token_accuracy": 0.8448150157928467, "num_tokens": 40186717.0, "step": 1059 }, { "epoch": 0.1348428953059407, "ewc_loss": 0.1259765625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 0.00011968612670898438, "grad_norm": 4.463010311126709, "learning_rate": 2.2445951674438323e-05, "loss": 0.6325, "mean_token_accuracy": 0.8382346630096436, "num_tokens": 40227119.0, "step": 1060 }, { "epoch": 0.13497010558453124, "ewc_loss": 0.0947265625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 8.7738037109375e-05, "grad_norm": 3.4651403427124023, "learning_rate": 2.2467147096227218e-05, "loss": 0.5789, "mean_token_accuracy": 0.8436671495437622, "num_tokens": 40262302.0, "step": 1061 }, { "epoch": 0.13509731586312174, "ewc_loss": 0.1279296875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 0.00012111663818359375, "grad_norm": 4.463179111480713, "learning_rate": 2.248834251801611e-05, "loss": 0.5585, "mean_token_accuracy": 0.8562229871749878, "num_tokens": 40296783.0, "step": 1062 }, { "epoch": 0.13522452614171224, "ewc_loss": 0.1083984375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 0.00010156631469726562, "grad_norm": 3.7074739933013916, "learning_rate": 2.2509537939805003e-05, "loss": 0.6022, "mean_token_accuracy": 0.8355343341827393, "num_tokens": 40331323.0, "step": 1063 }, { "epoch": 0.13535173642030277, "ewc_loss": 0.1181640625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 0.00011110305786132812, "grad_norm": 4.249721527099609, "learning_rate": 2.2530733361593897e-05, "loss": 0.584, "mean_token_accuracy": 0.847200334072113, "num_tokens": 40369965.0, "step": 1064 }, { "epoch": 0.13547894669889327, "ewc_loss": 0.10986328125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 0.000102996826171875, "grad_norm": 3.8061459064483643, "learning_rate": 2.255192878338279e-05, "loss": 0.6246, "mean_token_accuracy": 0.8324896693229675, "num_tokens": 40410820.0, "step": 1065 }, { "epoch": 0.1356061569774838, "ewc_loss": 0.10546875, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 9.870529174804688e-05, "grad_norm": 3.981837272644043, "learning_rate": 2.2573124205171683e-05, "loss": 0.5226, "mean_token_accuracy": 0.8606096506118774, "num_tokens": 40446942.0, "step": 1066 }, { "epoch": 0.1357333672560743, "ewc_loss": 0.10693359375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 0.00010013580322265625, "grad_norm": 3.8164350986480713, "learning_rate": 2.2594319626960577e-05, "loss": 0.5458, "mean_token_accuracy": 0.8527380228042603, "num_tokens": 40484684.0, "step": 1067 }, { "epoch": 0.1358605775346648, "ewc_loss": 0.09814453125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 9.107589721679688e-05, "grad_norm": 3.8023877143859863, "learning_rate": 2.2615515048749472e-05, "loss": 0.5149, "mean_token_accuracy": 0.8582010865211487, "num_tokens": 40521164.0, "step": 1068 }, { "epoch": 0.13598778781325532, "ewc_loss": 0.0966796875, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 8.96453857421875e-05, "grad_norm": 3.752056121826172, "learning_rate": 2.2636710470538363e-05, "loss": 0.5461, "mean_token_accuracy": 0.8470326066017151, "num_tokens": 40566192.0, "step": 1069 }, { "epoch": 0.13611499809184582, "ewc_loss": 0.09326171875, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 8.630752563476562e-05, "grad_norm": 3.781273365020752, "learning_rate": 2.2657905892327257e-05, "loss": 0.5799, "mean_token_accuracy": 0.8397825956344604, "num_tokens": 40605534.0, "step": 1070 }, { "epoch": 0.13624220837043632, "ewc_loss": 0.087890625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 8.106231689453125e-05, "grad_norm": 3.621782064437866, "learning_rate": 2.267910131411615e-05, "loss": 0.5271, "mean_token_accuracy": 0.8539295196533203, "num_tokens": 40645478.0, "step": 1071 }, { "epoch": 0.13636941864902685, "ewc_loss": 0.09521484375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 8.821487426757812e-05, "grad_norm": 3.853166341781616, "learning_rate": 2.2700296735905046e-05, "loss": 0.6205, "mean_token_accuracy": 0.8275495767593384, "num_tokens": 40684591.0, "step": 1072 }, { "epoch": 0.13649662892761735, "ewc_loss": 0.08203125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 7.534027099609375e-05, "grad_norm": 3.4954845905303955, "learning_rate": 2.2721492157693937e-05, "loss": 0.5429, "mean_token_accuracy": 0.8476041555404663, "num_tokens": 40724437.0, "step": 1073 }, { "epoch": 0.13662383920620785, "ewc_loss": 0.0986328125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 9.1552734375e-05, "grad_norm": 3.9725382328033447, "learning_rate": 2.274268757948283e-05, "loss": 0.6025, "mean_token_accuracy": 0.833191454410553, "num_tokens": 40758324.0, "step": 1074 }, { "epoch": 0.13675104948479838, "ewc_loss": 0.07958984375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 3.392157554626465, "learning_rate": 2.2763883001271726e-05, "loss": 0.5655, "mean_token_accuracy": 0.8390625715255737, "num_tokens": 40794182.0, "step": 1075 }, { "epoch": 0.13687825976338888, "ewc_loss": 0.10400390625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 9.72747802734375e-05, "grad_norm": 4.112053871154785, "learning_rate": 2.278507842306062e-05, "loss": 0.54, "mean_token_accuracy": 0.8563764691352844, "num_tokens": 40832293.0, "step": 1076 }, { "epoch": 0.13700547004197938, "ewc_loss": 0.07861328125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 3.309988021850586, "learning_rate": 2.280627384484951e-05, "loss": 0.5112, "mean_token_accuracy": 0.8547835946083069, "num_tokens": 40871078.0, "step": 1077 }, { "epoch": 0.1371326803205699, "ewc_loss": 0.11474609375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 0.00010776519775390625, "grad_norm": 4.314992427825928, "learning_rate": 2.2827469266638406e-05, "loss": 0.5593, "mean_token_accuracy": 0.852592408657074, "num_tokens": 40909061.0, "step": 1078 }, { "epoch": 0.1372598905991604, "ewc_loss": 0.0810546875, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 7.43865966796875e-05, "grad_norm": 3.309701919555664, "learning_rate": 2.28486646884273e-05, "loss": 0.563, "mean_token_accuracy": 0.8467074632644653, "num_tokens": 40945623.0, "step": 1079 }, { "epoch": 0.1373871008777509, "ewc_loss": 0.1259765625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 0.00011968612670898438, "grad_norm": 4.501095294952393, "learning_rate": 2.2869860110216195e-05, "loss": 0.671, "mean_token_accuracy": 0.8222167491912842, "num_tokens": 40985009.0, "step": 1080 }, { "epoch": 0.13751431115634144, "ewc_loss": 0.09521484375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 8.821487426757812e-05, "grad_norm": 3.45392107963562, "learning_rate": 2.2891055532005086e-05, "loss": 0.5932, "mean_token_accuracy": 0.8376331329345703, "num_tokens": 41027377.0, "step": 1081 }, { "epoch": 0.13764152143493194, "ewc_loss": 0.1328125, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 0.000125885009765625, "grad_norm": 4.524701118469238, "learning_rate": 2.291225095379398e-05, "loss": 0.6314, "mean_token_accuracy": 0.8359335660934448, "num_tokens": 41068816.0, "step": 1082 }, { "epoch": 0.13776873171352244, "ewc_loss": 0.11279296875, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 0.00010585784912109375, "grad_norm": 3.716792583465576, "learning_rate": 2.2933446375582875e-05, "loss": 0.5336, "mean_token_accuracy": 0.8584039211273193, "num_tokens": 41106603.0, "step": 1083 }, { "epoch": 0.13789594199211297, "ewc_loss": 0.1201171875, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 0.00011348724365234375, "grad_norm": 4.2644267082214355, "learning_rate": 2.295464179737177e-05, "loss": 0.624, "mean_token_accuracy": 0.8343534469604492, "num_tokens": 41142862.0, "step": 1084 }, { "epoch": 0.13802315227070347, "ewc_loss": 0.11865234375, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 0.00011157989501953125, "grad_norm": 3.9414281845092773, "learning_rate": 2.2975837219160663e-05, "loss": 0.577, "mean_token_accuracy": 0.8494909405708313, "num_tokens": 41182761.0, "step": 1085 }, { "epoch": 0.13815036254929397, "ewc_loss": 0.10009765625, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 9.34600830078125e-05, "grad_norm": 3.8145179748535156, "learning_rate": 2.2997032640949558e-05, "loss": 0.509, "mean_token_accuracy": 0.8631932735443115, "num_tokens": 41221318.0, "step": 1086 }, { "epoch": 0.1382775728278845, "ewc_loss": 0.119140625, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 0.0001125335693359375, "grad_norm": 4.16218376159668, "learning_rate": 2.301822806273845e-05, "loss": 0.5658, "mean_token_accuracy": 0.8493618369102478, "num_tokens": 41257748.0, "step": 1087 }, { "epoch": 0.138404783106475, "ewc_loss": 0.083984375, "ewc_loss_diag": 6.866455078125e-05, "ewc_loss_parallel": 7.724761962890625e-05, "grad_norm": 3.4286043643951416, "learning_rate": 2.3039423484527343e-05, "loss": 0.5397, "mean_token_accuracy": 0.8490263223648071, "num_tokens": 41299287.0, "step": 1088 }, { "epoch": 0.1385319933850655, "ewc_loss": 0.126953125, "ewc_loss_diag": 6.866455078125e-05, "ewc_loss_parallel": 0.0001201629638671875, "grad_norm": 4.477506160736084, "learning_rate": 2.3060618906316238e-05, "loss": 0.6252, "mean_token_accuracy": 0.836673378944397, "num_tokens": 41344600.0, "step": 1089 }, { "epoch": 0.13865920366365603, "ewc_loss": 0.0849609375, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 7.82012939453125e-05, "grad_norm": 3.3207528591156006, "learning_rate": 2.3081814328105132e-05, "loss": 0.4733, "mean_token_accuracy": 0.8670504689216614, "num_tokens": 41380681.0, "step": 1090 }, { "epoch": 0.13878641394224653, "ewc_loss": 0.14453125, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.773292064666748, "learning_rate": 2.3103009749894027e-05, "loss": 0.6255, "mean_token_accuracy": 0.8448124527931213, "num_tokens": 41418751.0, "step": 1091 }, { "epoch": 0.13891362422083706, "ewc_loss": 0.1083984375, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.00010156631469726562, "grad_norm": 3.5946531295776367, "learning_rate": 2.3124205171682918e-05, "loss": 0.6371, "mean_token_accuracy": 0.8269087672233582, "num_tokens": 41455617.0, "step": 1092 }, { "epoch": 0.13904083449942756, "ewc_loss": 0.1435546875, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.786061763763428, "learning_rate": 2.3145400593471812e-05, "loss": 0.5831, "mean_token_accuracy": 0.8541838526725769, "num_tokens": 41494480.0, "step": 1093 }, { "epoch": 0.13916804477801806, "ewc_loss": 0.12890625, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.0001220703125, "grad_norm": 3.939164876937866, "learning_rate": 2.3166596015260706e-05, "loss": 0.5843, "mean_token_accuracy": 0.8504437208175659, "num_tokens": 41529294.0, "step": 1094 }, { "epoch": 0.13929525505660859, "ewc_loss": 0.1328125, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.000125885009765625, "grad_norm": 4.483667850494385, "learning_rate": 2.31877914370496e-05, "loss": 0.5672, "mean_token_accuracy": 0.8545145988464355, "num_tokens": 41570787.0, "step": 1095 }, { "epoch": 0.1394224653351991, "ewc_loss": 0.12890625, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.0001220703125, "grad_norm": 4.052098751068115, "learning_rate": 2.3208986858838492e-05, "loss": 0.5741, "mean_token_accuracy": 0.8556580543518066, "num_tokens": 41606596.0, "step": 1096 }, { "epoch": 0.1395496756137896, "ewc_loss": 0.1171875, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.00011014938354492188, "grad_norm": 4.128309726715088, "learning_rate": 2.3230182280627386e-05, "loss": 0.5828, "mean_token_accuracy": 0.8429770469665527, "num_tokens": 41645317.0, "step": 1097 }, { "epoch": 0.13967688589238011, "ewc_loss": 0.1171875, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.00011014938354492188, "grad_norm": 4.036610126495361, "learning_rate": 2.325137770241628e-05, "loss": 0.5507, "mean_token_accuracy": 0.8513824939727783, "num_tokens": 41680424.0, "step": 1098 }, { "epoch": 0.13980409617097062, "ewc_loss": 0.1044921875, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 9.775161743164062e-05, "grad_norm": 3.905858039855957, "learning_rate": 2.3272573124205172e-05, "loss": 0.5664, "mean_token_accuracy": 0.849166214466095, "num_tokens": 41712492.0, "step": 1099 }, { "epoch": 0.13993130644956112, "ewc_loss": 0.10693359375, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.00010013580322265625, "grad_norm": 3.9265644550323486, "learning_rate": 2.3293768545994066e-05, "loss": 0.5676, "mean_token_accuracy": 0.8494676947593689, "num_tokens": 41747666.0, "step": 1100 }, { "epoch": 0.14005851672815164, "ewc_loss": 0.1025390625, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 9.584426879882812e-05, "grad_norm": 3.87544584274292, "learning_rate": 2.331496396778296e-05, "loss": 0.5436, "mean_token_accuracy": 0.8540051579475403, "num_tokens": 41784638.0, "step": 1101 }, { "epoch": 0.14018572700674214, "ewc_loss": 0.09375, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 8.678436279296875e-05, "grad_norm": 3.7542476654052734, "learning_rate": 2.3336159389571855e-05, "loss": 0.6201, "mean_token_accuracy": 0.8282457590103149, "num_tokens": 41826547.0, "step": 1102 }, { "epoch": 0.14031293728533265, "ewc_loss": 0.10107421875, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 9.441375732421875e-05, "grad_norm": 3.955794095993042, "learning_rate": 2.3357354811360746e-05, "loss": 0.5764, "mean_token_accuracy": 0.8415464162826538, "num_tokens": 41859526.0, "step": 1103 }, { "epoch": 0.14044014756392317, "ewc_loss": 0.08740234375, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 8.058547973632812e-05, "grad_norm": 3.587789297103882, "learning_rate": 2.337855023314964e-05, "loss": 0.5628, "mean_token_accuracy": 0.8448249101638794, "num_tokens": 41901177.0, "step": 1104 }, { "epoch": 0.14056735784251367, "ewc_loss": 0.103515625, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 9.679794311523438e-05, "grad_norm": 4.078505039215088, "learning_rate": 2.3399745654938535e-05, "loss": 0.5908, "mean_token_accuracy": 0.8374631404876709, "num_tokens": 41941579.0, "step": 1105 }, { "epoch": 0.14069456812110417, "ewc_loss": 0.0810546875, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 7.43865966796875e-05, "grad_norm": 3.458683729171753, "learning_rate": 2.342094107672743e-05, "loss": 0.4974, "mean_token_accuracy": 0.8604127168655396, "num_tokens": 41975315.0, "step": 1106 }, { "epoch": 0.1408217783996947, "ewc_loss": 0.11083984375, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.00010395050048828125, "grad_norm": 4.264317512512207, "learning_rate": 2.344213649851632e-05, "loss": 0.5209, "mean_token_accuracy": 0.8595463037490845, "num_tokens": 42011950.0, "step": 1107 }, { "epoch": 0.1409489886782852, "ewc_loss": 0.08056640625, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 7.390975952148438e-05, "grad_norm": 3.356358766555786, "learning_rate": 2.3463331920305215e-05, "loss": 0.551, "mean_token_accuracy": 0.8450739979743958, "num_tokens": 42051764.0, "step": 1108 }, { "epoch": 0.1410761989568757, "ewc_loss": 0.1259765625, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.00011873245239257812, "grad_norm": 4.518805503845215, "learning_rate": 2.348452734209411e-05, "loss": 0.6097, "mean_token_accuracy": 0.8362559080123901, "num_tokens": 42094218.0, "step": 1109 }, { "epoch": 0.14120340923546623, "ewc_loss": 0.0869140625, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 8.0108642578125e-05, "grad_norm": 3.373347043991089, "learning_rate": 2.3505722763883004e-05, "loss": 0.5493, "mean_token_accuracy": 0.8474363684654236, "num_tokens": 42129120.0, "step": 1110 }, { "epoch": 0.14133061951405673, "ewc_loss": 0.142578125, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.0001354217529296875, "grad_norm": 4.743247032165527, "learning_rate": 2.3526918185671895e-05, "loss": 0.6219, "mean_token_accuracy": 0.8429123163223267, "num_tokens": 42167437.0, "step": 1111 }, { "epoch": 0.14145782979264723, "ewc_loss": 0.10791015625, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.0001010894775390625, "grad_norm": 3.602022171020508, "learning_rate": 2.354811360746079e-05, "loss": 0.5543, "mean_token_accuracy": 0.8545757532119751, "num_tokens": 42203143.0, "step": 1112 }, { "epoch": 0.14158504007123776, "ewc_loss": 0.1455078125, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 0.00013828277587890625, "grad_norm": 4.76061487197876, "learning_rate": 2.3569309029249684e-05, "loss": 0.5929, "mean_token_accuracy": 0.8529024124145508, "num_tokens": 42237024.0, "step": 1113 }, { "epoch": 0.14171225034982826, "ewc_loss": 0.12890625, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 0.0001220703125, "grad_norm": 3.93178391456604, "learning_rate": 2.3590504451038578e-05, "loss": 0.6487, "mean_token_accuracy": 0.8285754919052124, "num_tokens": 42272847.0, "step": 1114 }, { "epoch": 0.1418394606284188, "ewc_loss": 0.1298828125, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 0.00012302398681640625, "grad_norm": 4.446120738983154, "learning_rate": 2.361169987282747e-05, "loss": 0.6033, "mean_token_accuracy": 0.8420110940933228, "num_tokens": 42309335.0, "step": 1115 }, { "epoch": 0.1419666709070093, "ewc_loss": 0.1337890625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 0.00012683868408203125, "grad_norm": 4.129509449005127, "learning_rate": 2.3632895294616363e-05, "loss": 0.605, "mean_token_accuracy": 0.8476431369781494, "num_tokens": 42346349.0, "step": 1116 }, { "epoch": 0.1420938811855998, "ewc_loss": 0.1064453125, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 9.965896606445312e-05, "grad_norm": 3.966118335723877, "learning_rate": 2.3654090716405258e-05, "loss": 0.6075, "mean_token_accuracy": 0.8397848010063171, "num_tokens": 42389310.0, "step": 1117 }, { "epoch": 0.14222109146419032, "ewc_loss": 0.12890625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 0.0001220703125, "grad_norm": 4.289444923400879, "learning_rate": 2.367528613819415e-05, "loss": 0.5687, "mean_token_accuracy": 0.8545023202896118, "num_tokens": 42427808.0, "step": 1118 }, { "epoch": 0.14234830174278082, "ewc_loss": 0.09326171875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 8.630752563476562e-05, "grad_norm": 3.6036856174468994, "learning_rate": 2.3696481559983043e-05, "loss": 0.5563, "mean_token_accuracy": 0.8457215428352356, "num_tokens": 42465531.0, "step": 1119 }, { "epoch": 0.14247551202137132, "ewc_loss": 0.1328125, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 0.000125885009765625, "grad_norm": 4.509393692016602, "learning_rate": 2.3717676981771938e-05, "loss": 0.5917, "mean_token_accuracy": 0.8522049784660339, "num_tokens": 42507613.0, "step": 1120 }, { "epoch": 0.14260272229996185, "ewc_loss": 0.09326171875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 8.630752563476562e-05, "grad_norm": 3.4931583404541016, "learning_rate": 2.3738872403560832e-05, "loss": 0.5736, "mean_token_accuracy": 0.8456175327301025, "num_tokens": 42542876.0, "step": 1121 }, { "epoch": 0.14272993257855235, "ewc_loss": 0.1416015625, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 0.00013446807861328125, "grad_norm": 4.696010589599609, "learning_rate": 2.3760067825349723e-05, "loss": 0.641, "mean_token_accuracy": 0.8375185132026672, "num_tokens": 42577587.0, "step": 1122 }, { "epoch": 0.14285714285714285, "ewc_loss": 0.10693359375, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 0.00010013580322265625, "grad_norm": 3.636167049407959, "learning_rate": 2.3781263247138618e-05, "loss": 0.5726, "mean_token_accuracy": 0.8463541269302368, "num_tokens": 42612733.0, "step": 1123 }, { "epoch": 0.14298435313573338, "ewc_loss": 0.142578125, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 0.0001354217529296875, "grad_norm": 4.708673477172852, "learning_rate": 2.3802458668927512e-05, "loss": 0.5546, "mean_token_accuracy": 0.8619510531425476, "num_tokens": 42646831.0, "step": 1124 }, { "epoch": 0.14311156341432388, "ewc_loss": 0.12060546875, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 0.00011396408081054688, "grad_norm": 3.8828375339508057, "learning_rate": 2.3823654090716406e-05, "loss": 0.5814, "mean_token_accuracy": 0.8413184285163879, "num_tokens": 42681351.0, "step": 1125 }, { "epoch": 0.14323877369291438, "ewc_loss": 0.1357421875, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 0.00012874603271484375, "grad_norm": 4.487020015716553, "learning_rate": 2.3844849512505297e-05, "loss": 0.6465, "mean_token_accuracy": 0.8329607248306274, "num_tokens": 42723491.0, "step": 1126 }, { "epoch": 0.1433659839715049, "ewc_loss": 0.1259765625, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 0.00011920928955078125, "grad_norm": 3.9937191009521484, "learning_rate": 2.3866044934294192e-05, "loss": 0.5978, "mean_token_accuracy": 0.8441750407218933, "num_tokens": 42763349.0, "step": 1127 }, { "epoch": 0.1434931942500954, "ewc_loss": 0.1220703125, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 0.00011539459228515625, "grad_norm": 4.2130303382873535, "learning_rate": 2.3887240356083086e-05, "loss": 0.6094, "mean_token_accuracy": 0.8401673436164856, "num_tokens": 42803977.0, "step": 1128 }, { "epoch": 0.1436204045286859, "ewc_loss": 0.11767578125, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 0.000110626220703125, "grad_norm": 3.9894638061523438, "learning_rate": 2.390843577787198e-05, "loss": 0.6046, "mean_token_accuracy": 0.8375319838523865, "num_tokens": 42847313.0, "step": 1129 }, { "epoch": 0.14374761480727644, "ewc_loss": 0.11181640625, "ewc_loss_diag": 7.104873657226562e-05, "ewc_loss_parallel": 0.00010442733764648438, "grad_norm": 4.034872531890869, "learning_rate": 2.3929631199660875e-05, "loss": 0.548, "mean_token_accuracy": 0.8535161018371582, "num_tokens": 42887468.0, "step": 1130 }, { "epoch": 0.14387482508586694, "ewc_loss": 0.11083984375, "ewc_loss_diag": 7.104873657226562e-05, "ewc_loss_parallel": 0.00010347366333007812, "grad_norm": 3.906937599182129, "learning_rate": 2.3950826621449766e-05, "loss": 0.575, "mean_token_accuracy": 0.8469533920288086, "num_tokens": 42925502.0, "step": 1131 }, { "epoch": 0.14400203536445744, "ewc_loss": 0.1083984375, "ewc_loss_diag": 7.104873657226562e-05, "ewc_loss_parallel": 0.0001010894775390625, "grad_norm": 4.024871826171875, "learning_rate": 2.397202204323866e-05, "loss": 0.5733, "mean_token_accuracy": 0.8490815758705139, "num_tokens": 42961930.0, "step": 1132 }, { "epoch": 0.14412924564304797, "ewc_loss": 0.0986328125, "ewc_loss_diag": 7.104873657226562e-05, "ewc_loss_parallel": 9.107589721679688e-05, "grad_norm": 3.7577757835388184, "learning_rate": 2.3993217465027555e-05, "loss": 0.5102, "mean_token_accuracy": 0.8615937829017639, "num_tokens": 43002121.0, "step": 1133 }, { "epoch": 0.14425645592163847, "ewc_loss": 0.10986328125, "ewc_loss_diag": 7.104873657226562e-05, "ewc_loss_parallel": 0.00010251998901367188, "grad_norm": 4.122555732727051, "learning_rate": 2.401441288681645e-05, "loss": 0.6177, "mean_token_accuracy": 0.8354413509368896, "num_tokens": 43039343.0, "step": 1134 }, { "epoch": 0.14438366620022897, "ewc_loss": 0.08935546875, "ewc_loss_diag": 7.104873657226562e-05, "ewc_loss_parallel": 8.20159912109375e-05, "grad_norm": 3.6098597049713135, "learning_rate": 2.4035608308605344e-05, "loss": 0.5011, "mean_token_accuracy": 0.8639576435089111, "num_tokens": 43075035.0, "step": 1135 }, { "epoch": 0.1445108764788195, "ewc_loss": 0.11474609375, "ewc_loss_diag": 7.104873657226562e-05, "ewc_loss_parallel": 0.00010728836059570312, "grad_norm": 4.265214443206787, "learning_rate": 2.405680373039424e-05, "loss": 0.5766, "mean_token_accuracy": 0.8499171733856201, "num_tokens": 43114350.0, "step": 1136 }, { "epoch": 0.14463808675741, "ewc_loss": 0.08642578125, "ewc_loss_diag": 7.104873657226562e-05, "ewc_loss_parallel": 7.915496826171875e-05, "grad_norm": 3.5189454555511475, "learning_rate": 2.407799915218313e-05, "loss": 0.549, "mean_token_accuracy": 0.8468220829963684, "num_tokens": 43146445.0, "step": 1137 }, { "epoch": 0.1447652970360005, "ewc_loss": 0.1220703125, "ewc_loss_diag": 7.104873657226562e-05, "ewc_loss_parallel": 0.00011491775512695312, "grad_norm": 4.423152446746826, "learning_rate": 2.4099194573972024e-05, "loss": 0.5707, "mean_token_accuracy": 0.8513845801353455, "num_tokens": 43182216.0, "step": 1138 }, { "epoch": 0.14489250731459102, "ewc_loss": 0.0888671875, "ewc_loss_diag": 7.104873657226562e-05, "ewc_loss_parallel": 8.153915405273438e-05, "grad_norm": 3.488563060760498, "learning_rate": 2.4120389995760918e-05, "loss": 0.543, "mean_token_accuracy": 0.8470466136932373, "num_tokens": 43216014.0, "step": 1139 }, { "epoch": 0.14501971759318152, "ewc_loss": 0.1337890625, "ewc_loss_diag": 7.152557373046875e-05, "ewc_loss_parallel": 0.00012683868408203125, "grad_norm": 4.5896897315979, "learning_rate": 2.4141585417549813e-05, "loss": 0.5928, "mean_token_accuracy": 0.8468544483184814, "num_tokens": 43252909.0, "step": 1140 }, { "epoch": 0.14514692787177205, "ewc_loss": 0.09912109375, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 9.1552734375e-05, "grad_norm": 3.561471939086914, "learning_rate": 2.4162780839338704e-05, "loss": 0.5379, "mean_token_accuracy": 0.8579154014587402, "num_tokens": 43285653.0, "step": 1141 }, { "epoch": 0.14527413815036255, "ewc_loss": 0.140625, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 0.000133514404296875, "grad_norm": 4.673516750335693, "learning_rate": 2.4183976261127598e-05, "loss": 0.5773, "mean_token_accuracy": 0.8581802845001221, "num_tokens": 43322087.0, "step": 1142 }, { "epoch": 0.14540134842895305, "ewc_loss": 0.11328125, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 0.00010585784912109375, "grad_norm": 3.761929512023926, "learning_rate": 2.4205171682916493e-05, "loss": 0.5628, "mean_token_accuracy": 0.8542385697364807, "num_tokens": 43360526.0, "step": 1143 }, { "epoch": 0.14552855870754358, "ewc_loss": 0.134765625, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 0.0001277923583984375, "grad_norm": 4.521472454071045, "learning_rate": 2.4226367104705387e-05, "loss": 0.5665, "mean_token_accuracy": 0.8576491475105286, "num_tokens": 43400267.0, "step": 1144 }, { "epoch": 0.14565576898613408, "ewc_loss": 0.1240234375, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 0.00011682510375976562, "grad_norm": 4.018107891082764, "learning_rate": 2.4247562526494278e-05, "loss": 0.5897, "mean_token_accuracy": 0.8434078693389893, "num_tokens": 43430411.0, "step": 1145 }, { "epoch": 0.14578297926472458, "ewc_loss": 0.1142578125, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 0.0001068115234375, "grad_norm": 4.100822925567627, "learning_rate": 2.4268757948283172e-05, "loss": 0.567, "mean_token_accuracy": 0.8504449129104614, "num_tokens": 43472199.0, "step": 1146 }, { "epoch": 0.1459101895433151, "ewc_loss": 0.1279296875, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 0.00012111663818359375, "grad_norm": 4.222083568572998, "learning_rate": 2.4289953370072067e-05, "loss": 0.6026, "mean_token_accuracy": 0.8438090085983276, "num_tokens": 43515896.0, "step": 1147 }, { "epoch": 0.1460373998219056, "ewc_loss": 0.09521484375, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 8.7738037109375e-05, "grad_norm": 3.7016000747680664, "learning_rate": 2.431114879186096e-05, "loss": 0.5896, "mean_token_accuracy": 0.8358258605003357, "num_tokens": 43556088.0, "step": 1148 }, { "epoch": 0.1461646101004961, "ewc_loss": 0.1318359375, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 0.00012493133544921875, "grad_norm": 4.4723358154296875, "learning_rate": 2.4332344213649852e-05, "loss": 0.633, "mean_token_accuracy": 0.8340812921524048, "num_tokens": 43596623.0, "step": 1149 }, { "epoch": 0.14629182037908664, "ewc_loss": 0.0888671875, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 8.153915405273438e-05, "grad_norm": 3.4692485332489014, "learning_rate": 2.4353539635438747e-05, "loss": 0.5612, "mean_token_accuracy": 0.8458280563354492, "num_tokens": 43634757.0, "step": 1150 }, { "epoch": 0.14641903065767714, "ewc_loss": 0.14453125, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.765594959259033, "learning_rate": 2.437473505722764e-05, "loss": 0.6239, "mean_token_accuracy": 0.8436551690101624, "num_tokens": 43671979.0, "step": 1151 }, { "epoch": 0.14654624093626764, "ewc_loss": 0.09912109375, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 9.1552734375e-05, "grad_norm": 3.502065896987915, "learning_rate": 2.4395930479016536e-05, "loss": 0.504, "mean_token_accuracy": 0.8648018836975098, "num_tokens": 43711898.0, "step": 1152 }, { "epoch": 0.14667345121485817, "ewc_loss": 0.1552734375, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 0.00014781951904296875, "grad_norm": 4.984156608581543, "learning_rate": 2.4417125900805427e-05, "loss": 0.5905, "mean_token_accuracy": 0.8554890155792236, "num_tokens": 43748394.0, "step": 1153 }, { "epoch": 0.14680066149344867, "ewc_loss": 0.125, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 0.00011777877807617188, "grad_norm": 3.8546183109283447, "learning_rate": 2.443832132259432e-05, "loss": 0.6107, "mean_token_accuracy": 0.8452693819999695, "num_tokens": 43784730.0, "step": 1154 }, { "epoch": 0.14692787177203917, "ewc_loss": 0.1513671875, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 0.00014400482177734375, "grad_norm": 4.865743160247803, "learning_rate": 2.4459516744383215e-05, "loss": 0.629, "mean_token_accuracy": 0.8471897840499878, "num_tokens": 43823993.0, "step": 1155 }, { "epoch": 0.1470550820506297, "ewc_loss": 0.14453125, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.158027172088623, "learning_rate": 2.4480712166172106e-05, "loss": 0.6579, "mean_token_accuracy": 0.8387080430984497, "num_tokens": 43861673.0, "step": 1156 }, { "epoch": 0.1471822923292202, "ewc_loss": 0.1376953125, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 0.00012969970703125, "grad_norm": 4.469301223754883, "learning_rate": 2.4501907587961e-05, "loss": 0.6453, "mean_token_accuracy": 0.8388367891311646, "num_tokens": 43900123.0, "step": 1157 }, { "epoch": 0.1473095026078107, "ewc_loss": 0.138671875, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 0.00013065338134765625, "grad_norm": 4.252363681793213, "learning_rate": 2.4523103009749895e-05, "loss": 0.6019, "mean_token_accuracy": 0.8486233949661255, "num_tokens": 43939371.0, "step": 1158 }, { "epoch": 0.14743671288640123, "ewc_loss": 0.119140625, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 0.00011157989501953125, "grad_norm": 4.1109747886657715, "learning_rate": 2.454429843153879e-05, "loss": 0.6091, "mean_token_accuracy": 0.8428750038146973, "num_tokens": 43972960.0, "step": 1159 }, { "epoch": 0.14756392316499173, "ewc_loss": 0.126953125, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 0.00011968612670898438, "grad_norm": 4.2413458824157715, "learning_rate": 2.456549385332768e-05, "loss": 0.6253, "mean_token_accuracy": 0.8409469723701477, "num_tokens": 44009919.0, "step": 1160 }, { "epoch": 0.14769113344358223, "ewc_loss": 0.1123046875, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 0.0001049041748046875, "grad_norm": 3.9456841945648193, "learning_rate": 2.4586689275116575e-05, "loss": 0.5576, "mean_token_accuracy": 0.8488034009933472, "num_tokens": 44050491.0, "step": 1161 }, { "epoch": 0.14781834372217276, "ewc_loss": 0.1162109375, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 0.0001087188720703125, "grad_norm": 4.134776592254639, "learning_rate": 2.460788469690547e-05, "loss": 0.5801, "mean_token_accuracy": 0.8526799082756042, "num_tokens": 44087999.0, "step": 1162 }, { "epoch": 0.14794555400076326, "ewc_loss": 0.109375, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 0.00010204315185546875, "grad_norm": 3.965742349624634, "learning_rate": 2.4629080118694364e-05, "loss": 0.6209, "mean_token_accuracy": 0.8336604237556458, "num_tokens": 44126014.0, "step": 1163 }, { "epoch": 0.14807276427935376, "ewc_loss": 0.10595703125, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 9.870529174804688e-05, "grad_norm": 3.98152494430542, "learning_rate": 2.4650275540483255e-05, "loss": 0.5079, "mean_token_accuracy": 0.8665589690208435, "num_tokens": 44154713.0, "step": 1164 }, { "epoch": 0.1481999745579443, "ewc_loss": 0.10791015625, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 0.00010061264038085938, "grad_norm": 4.039422988891602, "learning_rate": 2.467147096227215e-05, "loss": 0.6353, "mean_token_accuracy": 0.8284573554992676, "num_tokens": 44189624.0, "step": 1165 }, { "epoch": 0.1483271848365348, "ewc_loss": 0.09765625, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 9.012222290039062e-05, "grad_norm": 3.837207555770874, "learning_rate": 2.4692666384061044e-05, "loss": 0.5678, "mean_token_accuracy": 0.8458174467086792, "num_tokens": 44225010.0, "step": 1166 }, { "epoch": 0.14845439511512531, "ewc_loss": 0.1064453125, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 9.918212890625e-05, "grad_norm": 4.108272552490234, "learning_rate": 2.471386180584994e-05, "loss": 0.4893, "mean_token_accuracy": 0.867146372795105, "num_tokens": 44264987.0, "step": 1167 }, { "epoch": 0.14858160539371582, "ewc_loss": 0.0927734375, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 8.535385131835938e-05, "grad_norm": 3.727135419845581, "learning_rate": 2.473505722763883e-05, "loss": 0.5475, "mean_token_accuracy": 0.8518454432487488, "num_tokens": 44306705.0, "step": 1168 }, { "epoch": 0.14870881567230632, "ewc_loss": 0.1103515625, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 0.000102996826171875, "grad_norm": 4.221113681793213, "learning_rate": 2.4756252649427724e-05, "loss": 0.5264, "mean_token_accuracy": 0.8616818785667419, "num_tokens": 44347227.0, "step": 1169 }, { "epoch": 0.14883602595089684, "ewc_loss": 0.08740234375, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 8.0108642578125e-05, "grad_norm": 3.624948501586914, "learning_rate": 2.4777448071216618e-05, "loss": 0.6373, "mean_token_accuracy": 0.8177849054336548, "num_tokens": 44385273.0, "step": 1170 }, { "epoch": 0.14896323622948734, "ewc_loss": 0.11865234375, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 0.00011110305786132812, "grad_norm": 4.383383274078369, "learning_rate": 2.4798643493005513e-05, "loss": 0.5875, "mean_token_accuracy": 0.8408674001693726, "num_tokens": 44426177.0, "step": 1171 }, { "epoch": 0.14909044650807785, "ewc_loss": 0.083984375, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 7.677078247070312e-05, "grad_norm": 3.501763343811035, "learning_rate": 2.4819838914794404e-05, "loss": 0.5316, "mean_token_accuracy": 0.8534268140792847, "num_tokens": 44468609.0, "step": 1172 }, { "epoch": 0.14921765678666837, "ewc_loss": 0.1337890625, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 0.000125885009765625, "grad_norm": 4.651960372924805, "learning_rate": 2.4841034336583298e-05, "loss": 0.5781, "mean_token_accuracy": 0.8516520261764526, "num_tokens": 44507212.0, "step": 1173 }, { "epoch": 0.14934486706525887, "ewc_loss": 0.08642578125, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 7.915496826171875e-05, "grad_norm": 3.412308931350708, "learning_rate": 2.4862229758372193e-05, "loss": 0.5174, "mean_token_accuracy": 0.8553698658943176, "num_tokens": 44538551.0, "step": 1174 }, { "epoch": 0.14947207734384937, "ewc_loss": 0.150390625, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 0.00014209747314453125, "grad_norm": 5.016982555389404, "learning_rate": 2.4883425180161084e-05, "loss": 0.6687, "mean_token_accuracy": 0.8287251591682434, "num_tokens": 44577105.0, "step": 1175 }, { "epoch": 0.1495992876224399, "ewc_loss": 0.10791015625, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 0.00010061264038085938, "grad_norm": 3.5537796020507812, "learning_rate": 2.4904620601949978e-05, "loss": 0.5092, "mean_token_accuracy": 0.866486132144928, "num_tokens": 44615548.0, "step": 1176 }, { "epoch": 0.1497264979010304, "ewc_loss": 0.16796875, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 0.00016021728515625, "grad_norm": 5.27305793762207, "learning_rate": 2.4925816023738872e-05, "loss": 0.5763, "mean_token_accuracy": 0.8628993034362793, "num_tokens": 44654447.0, "step": 1177 }, { "epoch": 0.1498537081796209, "ewc_loss": 0.1455078125, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.046855449676514, "learning_rate": 2.4947011445527767e-05, "loss": 0.6081, "mean_token_accuracy": 0.8510928153991699, "num_tokens": 44697995.0, "step": 1178 }, { "epoch": 0.14998091845821143, "ewc_loss": 0.16015625, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 0.000152587890625, "grad_norm": 5.091984272003174, "learning_rate": 2.496820686731666e-05, "loss": 0.6804, "mean_token_accuracy": 0.8296793699264526, "num_tokens": 44734376.0, "step": 1179 }, { "epoch": 0.15010812873680193, "ewc_loss": 0.16015625, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 0.000152587890625, "grad_norm": 4.418853282928467, "learning_rate": 2.4989402289105556e-05, "loss": 0.629, "mean_token_accuracy": 0.8428375124931335, "num_tokens": 44772745.0, "step": 1180 }, { "epoch": 0.15023533901539243, "ewc_loss": 0.1376953125, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 0.00012969970703125, "grad_norm": 4.524810314178467, "learning_rate": 2.5010597710894447e-05, "loss": 0.5947, "mean_token_accuracy": 0.8496695756912231, "num_tokens": 44816530.0, "step": 1181 }, { "epoch": 0.15036254929398296, "ewc_loss": 0.15625, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 0.000148773193359375, "grad_norm": 4.58204984664917, "learning_rate": 2.503179313268334e-05, "loss": 0.5762, "mean_token_accuracy": 0.8617138862609863, "num_tokens": 44853428.0, "step": 1182 }, { "epoch": 0.15048975957257346, "ewc_loss": 0.11181640625, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 0.00010442733764648438, "grad_norm": 3.961681842803955, "learning_rate": 2.5052988554472236e-05, "loss": 0.5518, "mean_token_accuracy": 0.8530746698379517, "num_tokens": 44891118.0, "step": 1183 }, { "epoch": 0.15061696985116396, "ewc_loss": 0.15234375, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 0.00014495849609375, "grad_norm": 4.835756301879883, "learning_rate": 2.507418397626113e-05, "loss": 0.5845, "mean_token_accuracy": 0.8535181283950806, "num_tokens": 44929143.0, "step": 1184 }, { "epoch": 0.1507441801297545, "ewc_loss": 0.109375, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 0.00010204315185546875, "grad_norm": 3.7361934185028076, "learning_rate": 2.5095379398050024e-05, "loss": 0.5448, "mean_token_accuracy": 0.8588156700134277, "num_tokens": 44966361.0, "step": 1185 }, { "epoch": 0.150871390408345, "ewc_loss": 0.1611328125, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 0.00015354156494140625, "grad_norm": 5.085335731506348, "learning_rate": 2.511657481983892e-05, "loss": 0.6352, "mean_token_accuracy": 0.8420952558517456, "num_tokens": 45007993.0, "step": 1186 }, { "epoch": 0.1509986006869355, "ewc_loss": 0.1279296875, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 0.00012063980102539062, "grad_norm": 3.914362907409668, "learning_rate": 2.5137770241627813e-05, "loss": 0.6358, "mean_token_accuracy": 0.8348126411437988, "num_tokens": 45040337.0, "step": 1187 }, { "epoch": 0.15112581096552602, "ewc_loss": 0.16015625, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.000152587890625, "grad_norm": 5.104899883270264, "learning_rate": 2.5158965663416704e-05, "loss": 0.6776, "mean_token_accuracy": 0.8309611082077026, "num_tokens": 45080730.0, "step": 1188 }, { "epoch": 0.15125302124411652, "ewc_loss": 0.1474609375, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.0001392364501953125, "grad_norm": 4.199041366577148, "learning_rate": 2.5180161085205595e-05, "loss": 0.6262, "mean_token_accuracy": 0.841435432434082, "num_tokens": 45116748.0, "step": 1189 }, { "epoch": 0.15138023152270705, "ewc_loss": 0.1474609375, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 0.0001392364501953125, "grad_norm": 4.838438034057617, "learning_rate": 2.520135650699449e-05, "loss": 0.6135, "mean_token_accuracy": 0.8434897661209106, "num_tokens": 45158596.0, "step": 1190 }, { "epoch": 0.15150744180129755, "ewc_loss": 0.1484375, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.00014019012451171875, "grad_norm": 4.30460786819458, "learning_rate": 2.5222551928783384e-05, "loss": 0.5911, "mean_token_accuracy": 0.8510095477104187, "num_tokens": 45202433.0, "step": 1191 }, { "epoch": 0.15163465207988805, "ewc_loss": 0.13671875, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.00012874603271484375, "grad_norm": 4.504400253295898, "learning_rate": 2.524374735057228e-05, "loss": 0.5867, "mean_token_accuracy": 0.8515304923057556, "num_tokens": 45239202.0, "step": 1192 }, { "epoch": 0.15176186235847858, "ewc_loss": 0.1357421875, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.0001277923583984375, "grad_norm": 4.264827728271484, "learning_rate": 2.5264942772361173e-05, "loss": 0.6018, "mean_token_accuracy": 0.8440684080123901, "num_tokens": 45270150.0, "step": 1193 }, { "epoch": 0.15188907263706908, "ewc_loss": 0.125, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.00011777877807617188, "grad_norm": 4.283732891082764, "learning_rate": 2.5286138194150067e-05, "loss": 0.5939, "mean_token_accuracy": 0.8461421728134155, "num_tokens": 45306394.0, "step": 1194 }, { "epoch": 0.15201628291565958, "ewc_loss": 0.123046875, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.00011539459228515625, "grad_norm": 4.149649620056152, "learning_rate": 2.5307333615938962e-05, "loss": 0.6074, "mean_token_accuracy": 0.8426707983016968, "num_tokens": 45344945.0, "step": 1195 }, { "epoch": 0.1521434931942501, "ewc_loss": 0.11962890625, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.00011157989501953125, "grad_norm": 4.219902992248535, "learning_rate": 2.532852903772785e-05, "loss": 0.557, "mean_token_accuracy": 0.8576908111572266, "num_tokens": 45378179.0, "step": 1196 }, { "epoch": 0.1522707034728406, "ewc_loss": 0.1123046875, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.00010442733764648438, "grad_norm": 3.9983630180358887, "learning_rate": 2.5349724459516744e-05, "loss": 0.5623, "mean_token_accuracy": 0.8493841886520386, "num_tokens": 45421194.0, "step": 1197 }, { "epoch": 0.1523979137514311, "ewc_loss": 0.11767578125, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.00010967254638671875, "grad_norm": 4.240426540374756, "learning_rate": 2.537091988130564e-05, "loss": 0.5773, "mean_token_accuracy": 0.8463953733444214, "num_tokens": 45458226.0, "step": 1198 }, { "epoch": 0.15252512403002164, "ewc_loss": 0.10546875, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 9.775161743164062e-05, "grad_norm": 3.882413148880005, "learning_rate": 2.5392115303094533e-05, "loss": 0.5691, "mean_token_accuracy": 0.846996545791626, "num_tokens": 45500007.0, "step": 1199 }, { "epoch": 0.15265233430861214, "ewc_loss": 0.1181640625, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.00011014938354492188, "grad_norm": 4.316103458404541, "learning_rate": 2.5413310724883427e-05, "loss": 0.5732, "mean_token_accuracy": 0.8462282419204712, "num_tokens": 45536605.0, "step": 1200 }, { "epoch": 0.15277954458720264, "ewc_loss": 0.09716796875, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 8.916854858398438e-05, "grad_norm": 3.7714548110961914, "learning_rate": 2.543450614667232e-05, "loss": 0.5503, "mean_token_accuracy": 0.8532115817070007, "num_tokens": 45572137.0, "step": 1201 }, { "epoch": 0.15290675486579317, "ewc_loss": 0.12451171875, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 0.00011682510375976562, "grad_norm": 4.444441795349121, "learning_rate": 2.5455701568461216e-05, "loss": 0.5795, "mean_token_accuracy": 0.8484108448028564, "num_tokens": 45608659.0, "step": 1202 }, { "epoch": 0.15303396514438367, "ewc_loss": 0.0947265625, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 8.678436279296875e-05, "grad_norm": 3.6585776805877686, "learning_rate": 2.547689699025011e-05, "loss": 0.5381, "mean_token_accuracy": 0.8511643409729004, "num_tokens": 45647002.0, "step": 1203 }, { "epoch": 0.15316117542297417, "ewc_loss": 0.134765625, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 0.00012683868408203125, "grad_norm": 4.654012203216553, "learning_rate": 2.5498092412038998e-05, "loss": 0.5998, "mean_token_accuracy": 0.8441087007522583, "num_tokens": 45680194.0, "step": 1204 }, { "epoch": 0.1532883857015647, "ewc_loss": 0.09375, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 8.58306884765625e-05, "grad_norm": 3.5802204608917236, "learning_rate": 2.5519287833827893e-05, "loss": 0.5316, "mean_token_accuracy": 0.8536548614501953, "num_tokens": 45719629.0, "step": 1205 }, { "epoch": 0.1534155959801552, "ewc_loss": 0.1474609375, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 0.0001392364501953125, "grad_norm": 4.914204120635986, "learning_rate": 2.5540483255616787e-05, "loss": 0.5596, "mean_token_accuracy": 0.8627991080284119, "num_tokens": 45754630.0, "step": 1206 }, { "epoch": 0.1535428062587457, "ewc_loss": 0.1083984375, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 0.00010061264038085938, "grad_norm": 3.660710573196411, "learning_rate": 2.556167867740568e-05, "loss": 0.5548, "mean_token_accuracy": 0.851414680480957, "num_tokens": 45791904.0, "step": 1207 }, { "epoch": 0.15367001653733622, "ewc_loss": 0.16015625, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 0.000152587890625, "grad_norm": 5.060652732849121, "learning_rate": 2.5582874099194576e-05, "loss": 0.5295, "mean_token_accuracy": 0.8771469593048096, "num_tokens": 45830553.0, "step": 1208 }, { "epoch": 0.15379722681592672, "ewc_loss": 0.1337890625, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 0.000125885009765625, "grad_norm": 3.975139856338501, "learning_rate": 2.560406952098347e-05, "loss": 0.5791, "mean_token_accuracy": 0.8546274900436401, "num_tokens": 45865957.0, "step": 1209 }, { "epoch": 0.15392443709451722, "ewc_loss": 0.154296875, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 0.0001468658447265625, "grad_norm": 4.882810115814209, "learning_rate": 2.5625264942772365e-05, "loss": 0.6556, "mean_token_accuracy": 0.8369214534759521, "num_tokens": 45898993.0, "step": 1210 }, { "epoch": 0.15405164737310775, "ewc_loss": 0.146484375, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 0.00013828277587890625, "grad_norm": 4.2863898277282715, "learning_rate": 2.5646460364561252e-05, "loss": 0.5785, "mean_token_accuracy": 0.8568417429924011, "num_tokens": 45932843.0, "step": 1211 }, { "epoch": 0.15417885765169825, "ewc_loss": 0.1318359375, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 0.0001239776611328125, "grad_norm": 4.381089210510254, "learning_rate": 2.5667655786350147e-05, "loss": 0.5633, "mean_token_accuracy": 0.8588820099830627, "num_tokens": 45973247.0, "step": 1212 }, { "epoch": 0.15430606793028875, "ewc_loss": 0.146484375, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 0.00013828277587890625, "grad_norm": 4.514984607696533, "learning_rate": 2.568885120813904e-05, "loss": 0.5413, "mean_token_accuracy": 0.8684811592102051, "num_tokens": 46008641.0, "step": 1213 }, { "epoch": 0.15443327820887928, "ewc_loss": 0.1083984375, "ewc_loss_diag": 7.724761962890625e-05, "ewc_loss_parallel": 0.00010061264038085938, "grad_norm": 3.8872649669647217, "learning_rate": 2.5710046629927936e-05, "loss": 0.5313, "mean_token_accuracy": 0.8615031242370605, "num_tokens": 46046789.0, "step": 1214 }, { "epoch": 0.15456048848746978, "ewc_loss": 0.1484375, "ewc_loss_diag": 7.724761962890625e-05, "ewc_loss_parallel": 0.00014019012451171875, "grad_norm": 4.782269477844238, "learning_rate": 2.573124205171683e-05, "loss": 0.6095, "mean_token_accuracy": 0.8487985730171204, "num_tokens": 46086904.0, "step": 1215 }, { "epoch": 0.1546876987660603, "ewc_loss": 0.10302734375, "ewc_loss_diag": 7.724761962890625e-05, "ewc_loss_parallel": 9.5367431640625e-05, "grad_norm": 3.6669111251831055, "learning_rate": 2.5752437473505724e-05, "loss": 0.5172, "mean_token_accuracy": 0.8639712929725647, "num_tokens": 46129743.0, "step": 1216 }, { "epoch": 0.1548149090446508, "ewc_loss": 0.16015625, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 0.000152587890625, "grad_norm": 5.062917232513428, "learning_rate": 2.577363289529462e-05, "loss": 0.6147, "mean_token_accuracy": 0.849599301815033, "num_tokens": 46164902.0, "step": 1217 }, { "epoch": 0.1549421193232413, "ewc_loss": 0.119140625, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 0.00011110305786132812, "grad_norm": 3.809705972671509, "learning_rate": 2.5794828317083513e-05, "loss": 0.5729, "mean_token_accuracy": 0.8525259494781494, "num_tokens": 46209689.0, "step": 1218 }, { "epoch": 0.15506932960183184, "ewc_loss": 0.1650390625, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 0.00015735626220703125, "grad_norm": 5.132357597351074, "learning_rate": 2.58160237388724e-05, "loss": 0.5705, "mean_token_accuracy": 0.8632466793060303, "num_tokens": 46251365.0, "step": 1219 }, { "epoch": 0.15519653988042234, "ewc_loss": 0.14453125, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.148782730102539, "learning_rate": 2.5837219160661295e-05, "loss": 0.598, "mean_token_accuracy": 0.8468639850616455, "num_tokens": 46290961.0, "step": 1220 }, { "epoch": 0.15532375015901284, "ewc_loss": 0.154296875, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 0.0001468658447265625, "grad_norm": 4.884791374206543, "learning_rate": 2.585841458245019e-05, "loss": 0.5879, "mean_token_accuracy": 0.8569327592849731, "num_tokens": 46326317.0, "step": 1221 }, { "epoch": 0.15545096043760337, "ewc_loss": 0.1484375, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 0.00014019012451171875, "grad_norm": 4.333355903625488, "learning_rate": 2.5879610004239084e-05, "loss": 0.5526, "mean_token_accuracy": 0.863752007484436, "num_tokens": 46361018.0, "step": 1222 }, { "epoch": 0.15557817071619387, "ewc_loss": 0.1396484375, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 0.0001316070556640625, "grad_norm": 4.529273509979248, "learning_rate": 2.590080542602798e-05, "loss": 0.5748, "mean_token_accuracy": 0.8546181917190552, "num_tokens": 46397344.0, "step": 1223 }, { "epoch": 0.15570538099478437, "ewc_loss": 0.138671875, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 0.00013065338134765625, "grad_norm": 4.33671236038208, "learning_rate": 2.5922000847816873e-05, "loss": 0.6287, "mean_token_accuracy": 0.8393871188163757, "num_tokens": 46437257.0, "step": 1224 }, { "epoch": 0.1558325912733749, "ewc_loss": 0.130859375, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 0.00012302398681640625, "grad_norm": 4.322168350219727, "learning_rate": 2.5943196269605767e-05, "loss": 0.5975, "mean_token_accuracy": 0.8450251221656799, "num_tokens": 46477844.0, "step": 1225 }, { "epoch": 0.1559598015519654, "ewc_loss": 0.126953125, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 0.00011920928955078125, "grad_norm": 4.203363418579102, "learning_rate": 2.5964391691394662e-05, "loss": 0.5584, "mean_token_accuracy": 0.8538087010383606, "num_tokens": 46514556.0, "step": 1226 }, { "epoch": 0.1560870118305559, "ewc_loss": 0.1298828125, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 0.00012159347534179688, "grad_norm": 4.32546854019165, "learning_rate": 2.5985587113183553e-05, "loss": 0.6365, "mean_token_accuracy": 0.8345473408699036, "num_tokens": 46559274.0, "step": 1227 }, { "epoch": 0.15621422210914643, "ewc_loss": 0.11328125, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 0.00010538101196289062, "grad_norm": 4.008408546447754, "learning_rate": 2.6006782534972447e-05, "loss": 0.5508, "mean_token_accuracy": 0.8521414995193481, "num_tokens": 46591104.0, "step": 1228 }, { "epoch": 0.15634143238773693, "ewc_loss": 0.12890625, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 0.00012111663818359375, "grad_norm": 4.495946884155273, "learning_rate": 2.6027977956761342e-05, "loss": 0.6491, "mean_token_accuracy": 0.8278399109840393, "num_tokens": 46631296.0, "step": 1229 }, { "epoch": 0.15646864266632743, "ewc_loss": 0.10205078125, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 9.441375732421875e-05, "grad_norm": 3.793647050857544, "learning_rate": 2.6049173378550236e-05, "loss": 0.5638, "mean_token_accuracy": 0.848281741142273, "num_tokens": 46673079.0, "step": 1230 }, { "epoch": 0.15659585294491796, "ewc_loss": 0.142578125, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 0.00013446807861328125, "grad_norm": 4.7031426429748535, "learning_rate": 2.6070368800339127e-05, "loss": 0.6643, "mean_token_accuracy": 0.8318109512329102, "num_tokens": 46706820.0, "step": 1231 }, { "epoch": 0.15672306322350846, "ewc_loss": 0.099609375, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 9.1552734375e-05, "grad_norm": 3.6771609783172607, "learning_rate": 2.609156422212802e-05, "loss": 0.5514, "mean_token_accuracy": 0.8514443635940552, "num_tokens": 46742466.0, "step": 1232 }, { "epoch": 0.15685027350209896, "ewc_loss": 0.150390625, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 0.0001430511474609375, "grad_norm": 4.9458441734313965, "learning_rate": 2.6112759643916916e-05, "loss": 0.5822, "mean_token_accuracy": 0.8540924787521362, "num_tokens": 46778882.0, "step": 1233 }, { "epoch": 0.1569774837806895, "ewc_loss": 0.11181640625, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 0.00010395050048828125, "grad_norm": 3.7260401248931885, "learning_rate": 2.6133955065705807e-05, "loss": 0.6032, "mean_token_accuracy": 0.8417444825172424, "num_tokens": 46821509.0, "step": 1234 }, { "epoch": 0.15710469405928, "ewc_loss": 0.1611328125, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 0.00015354156494140625, "grad_norm": 5.107822895050049, "learning_rate": 2.61551504874947e-05, "loss": 0.5951, "mean_token_accuracy": 0.8549568057060242, "num_tokens": 46862247.0, "step": 1235 }, { "epoch": 0.1572319043378705, "ewc_loss": 0.1337890625, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.000125885009765625, "grad_norm": 4.000215530395508, "learning_rate": 2.6176345909283596e-05, "loss": 0.592, "mean_token_accuracy": 0.8496259450912476, "num_tokens": 46906645.0, "step": 1236 }, { "epoch": 0.15735911461646102, "ewc_loss": 0.1591796875, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.00015163421630859375, "grad_norm": 4.997152328491211, "learning_rate": 2.619754133107249e-05, "loss": 0.6637, "mean_token_accuracy": 0.8313657641410828, "num_tokens": 46940492.0, "step": 1237 }, { "epoch": 0.15748632489505152, "ewc_loss": 0.1494140625, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.000141143798828125, "grad_norm": 4.317049026489258, "learning_rate": 2.6218736752861385e-05, "loss": 0.6139, "mean_token_accuracy": 0.8487423658370972, "num_tokens": 46976871.0, "step": 1238 }, { "epoch": 0.15761353517364202, "ewc_loss": 0.140625, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.00013256072998046875, "grad_norm": 4.605550289154053, "learning_rate": 2.623993217465028e-05, "loss": 0.5984, "mean_token_accuracy": 0.8502823710441589, "num_tokens": 47013689.0, "step": 1239 }, { "epoch": 0.15774074545223254, "ewc_loss": 0.1484375, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.00014019012451171875, "grad_norm": 4.513757705688477, "learning_rate": 2.6261127596439174e-05, "loss": 0.6644, "mean_token_accuracy": 0.8297789692878723, "num_tokens": 47050995.0, "step": 1240 }, { "epoch": 0.15786795573082305, "ewc_loss": 0.11767578125, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.00010967254638671875, "grad_norm": 4.0777363777160645, "learning_rate": 2.6282323018228068e-05, "loss": 0.5644, "mean_token_accuracy": 0.8552363514900208, "num_tokens": 47088571.0, "step": 1241 }, { "epoch": 0.15799516600941357, "ewc_loss": 0.1494140625, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.000141143798828125, "grad_norm": 4.739895343780518, "learning_rate": 2.6303518440016956e-05, "loss": 0.6133, "mean_token_accuracy": 0.847713053226471, "num_tokens": 47130547.0, "step": 1242 }, { "epoch": 0.15812237628800407, "ewc_loss": 0.10791015625, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.00010013580322265625, "grad_norm": 3.7740159034729004, "learning_rate": 2.632471386180585e-05, "loss": 0.5638, "mean_token_accuracy": 0.8493592739105225, "num_tokens": 47162724.0, "step": 1243 }, { "epoch": 0.15824958656659457, "ewc_loss": 0.1572265625, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.00014972686767578125, "grad_norm": 5.020766735076904, "learning_rate": 2.6345909283594745e-05, "loss": 0.6579, "mean_token_accuracy": 0.8349908590316772, "num_tokens": 47200081.0, "step": 1244 }, { "epoch": 0.1583767968451851, "ewc_loss": 0.115234375, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.00010728836059570312, "grad_norm": 3.761456251144409, "learning_rate": 2.636710470538364e-05, "loss": 0.5545, "mean_token_accuracy": 0.8559635877609253, "num_tokens": 47235074.0, "step": 1245 }, { "epoch": 0.1585040071237756, "ewc_loss": 0.1689453125, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.00016117095947265625, "grad_norm": 5.191530704498291, "learning_rate": 2.6388300127172533e-05, "loss": 0.5717, "mean_token_accuracy": 0.8638596534729004, "num_tokens": 47274258.0, "step": 1246 }, { "epoch": 0.1586312174023661, "ewc_loss": 0.1357421875, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.0001277923583984375, "grad_norm": 4.034997463226318, "learning_rate": 2.6409495548961428e-05, "loss": 0.5611, "mean_token_accuracy": 0.8586676120758057, "num_tokens": 47314232.0, "step": 1247 }, { "epoch": 0.15875842768095663, "ewc_loss": 0.1640625, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 0.000156402587890625, "grad_norm": 5.089090824127197, "learning_rate": 2.6430690970750322e-05, "loss": 0.632, "mean_token_accuracy": 0.847176194190979, "num_tokens": 47355245.0, "step": 1248 }, { "epoch": 0.15888563795954713, "ewc_loss": 0.15234375, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 0.00014495849609375, "grad_norm": 4.339539527893066, "learning_rate": 2.645188639253921e-05, "loss": 0.6593, "mean_token_accuracy": 0.8401514291763306, "num_tokens": 47394356.0, "step": 1249 }, { "epoch": 0.15901284823813763, "ewc_loss": 0.150390625, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 0.0001430511474609375, "grad_norm": 4.7395501136779785, "learning_rate": 2.6473081814328104e-05, "loss": 0.5925, "mean_token_accuracy": 0.8544095158576965, "num_tokens": 47435220.0, "step": 1250 }, { "epoch": 0.15914005851672816, "ewc_loss": 0.1484375, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 0.00014019012451171875, "grad_norm": 4.443812847137451, "learning_rate": 2.6494277236117e-05, "loss": 0.6675, "mean_token_accuracy": 0.8376920819282532, "num_tokens": 47477294.0, "step": 1251 }, { "epoch": 0.15926726879531866, "ewc_loss": 0.138671875, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 0.00013065338134765625, "grad_norm": 4.409924507141113, "learning_rate": 2.6515472657905893e-05, "loss": 0.6136, "mean_token_accuracy": 0.8428162336349487, "num_tokens": 47512216.0, "step": 1252 }, { "epoch": 0.15939447907390916, "ewc_loss": 0.13671875, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 0.00012874603271484375, "grad_norm": 4.395164966583252, "learning_rate": 2.6536668079694788e-05, "loss": 0.5306, "mean_token_accuracy": 0.8701155185699463, "num_tokens": 47550075.0, "step": 1253 }, { "epoch": 0.1595216893524997, "ewc_loss": 0.1328125, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 0.00012493133544921875, "grad_norm": 4.335845947265625, "learning_rate": 2.6557863501483682e-05, "loss": 0.6106, "mean_token_accuracy": 0.8419265151023865, "num_tokens": 47581595.0, "step": 1254 }, { "epoch": 0.1596488996310902, "ewc_loss": 0.1240234375, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 0.00011587142944335938, "grad_norm": 4.190892219543457, "learning_rate": 2.6579058923272576e-05, "loss": 0.5076, "mean_token_accuracy": 0.8689231872558594, "num_tokens": 47617336.0, "step": 1255 }, { "epoch": 0.1597761099096807, "ewc_loss": 0.1318359375, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 0.0001239776611328125, "grad_norm": 4.45828914642334, "learning_rate": 2.660025434506147e-05, "loss": 0.59, "mean_token_accuracy": 0.8499395251274109, "num_tokens": 47662701.0, "step": 1256 }, { "epoch": 0.15990332018827122, "ewc_loss": 0.11181640625, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 0.00010347366333007812, "grad_norm": 3.9500691890716553, "learning_rate": 2.662144976685036e-05, "loss": 0.6187, "mean_token_accuracy": 0.8367566466331482, "num_tokens": 47703335.0, "step": 1257 }, { "epoch": 0.16003053046686172, "ewc_loss": 0.140625, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 0.00013256072998046875, "grad_norm": 4.663618087768555, "learning_rate": 2.6642645188639253e-05, "loss": 0.5899, "mean_token_accuracy": 0.8515267372131348, "num_tokens": 47746980.0, "step": 1258 }, { "epoch": 0.16015774074545222, "ewc_loss": 0.10400390625, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 9.584426879882812e-05, "grad_norm": 3.789130210876465, "learning_rate": 2.6663840610428147e-05, "loss": 0.5475, "mean_token_accuracy": 0.8505978584289551, "num_tokens": 47784606.0, "step": 1259 }, { "epoch": 0.16028495102404275, "ewc_loss": 0.146484375, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 0.00013828277587890625, "grad_norm": 4.898208141326904, "learning_rate": 2.6685036032217042e-05, "loss": 0.6251, "mean_token_accuracy": 0.8415358066558838, "num_tokens": 47821298.0, "step": 1260 }, { "epoch": 0.16041216130263325, "ewc_loss": 0.10693359375, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 9.870529174804688e-05, "grad_norm": 3.7286922931671143, "learning_rate": 2.6706231454005936e-05, "loss": 0.5474, "mean_token_accuracy": 0.8566327095031738, "num_tokens": 47863102.0, "step": 1261 }, { "epoch": 0.16053937158122375, "ewc_loss": 0.1591796875, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 0.00015163421630859375, "grad_norm": 5.133675575256348, "learning_rate": 2.672742687579483e-05, "loss": 0.5868, "mean_token_accuracy": 0.8570406436920166, "num_tokens": 47903124.0, "step": 1262 }, { "epoch": 0.16066658185981428, "ewc_loss": 0.12353515625, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 0.00011539459228515625, "grad_norm": 3.896799087524414, "learning_rate": 2.6748622297583725e-05, "loss": 0.5479, "mean_token_accuracy": 0.8567749261856079, "num_tokens": 47939441.0, "step": 1263 }, { "epoch": 0.16079379213840478, "ewc_loss": 0.166015625, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 0.0001583099365234375, "grad_norm": 5.17633581161499, "learning_rate": 2.6769817719372613e-05, "loss": 0.6568, "mean_token_accuracy": 0.8396693468093872, "num_tokens": 47981236.0, "step": 1264 }, { "epoch": 0.1609210024169953, "ewc_loss": 0.1455078125, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.226263046264648, "learning_rate": 2.6791013141161507e-05, "loss": 0.6008, "mean_token_accuracy": 0.8503807783126831, "num_tokens": 48011431.0, "step": 1265 }, { "epoch": 0.1610482126955858, "ewc_loss": 0.154296875, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 0.0001468658447265625, "grad_norm": 4.856306076049805, "learning_rate": 2.68122085629504e-05, "loss": 0.6001, "mean_token_accuracy": 0.8515682220458984, "num_tokens": 48051003.0, "step": 1266 }, { "epoch": 0.1611754229741763, "ewc_loss": 0.150390625, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 0.0001430511474609375, "grad_norm": 4.49617338180542, "learning_rate": 2.6833403984739296e-05, "loss": 0.642, "mean_token_accuracy": 0.8396368026733398, "num_tokens": 48091239.0, "step": 1267 }, { "epoch": 0.16130263325276684, "ewc_loss": 0.1328125, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 0.00012493133544921875, "grad_norm": 4.334333896636963, "learning_rate": 2.685459940652819e-05, "loss": 0.6092, "mean_token_accuracy": 0.8463187217712402, "num_tokens": 48126021.0, "step": 1268 }, { "epoch": 0.16142984353135734, "ewc_loss": 0.15234375, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 0.00014495849609375, "grad_norm": 4.733648777008057, "learning_rate": 2.6875794828317085e-05, "loss": 0.6169, "mean_token_accuracy": 0.8484886884689331, "num_tokens": 48160459.0, "step": 1269 }, { "epoch": 0.16155705380994784, "ewc_loss": 0.11181640625, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 0.00010347366333007812, "grad_norm": 3.9113523960113525, "learning_rate": 2.689699025010598e-05, "loss": 0.5621, "mean_token_accuracy": 0.8515938520431519, "num_tokens": 48200199.0, "step": 1270 }, { "epoch": 0.16168426408853837, "ewc_loss": 0.15625, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 0.000148773193359375, "grad_norm": 5.015183448791504, "learning_rate": 2.6918185671894874e-05, "loss": 0.5915, "mean_token_accuracy": 0.8536736965179443, "num_tokens": 48233926.0, "step": 1271 }, { "epoch": 0.16181147436712887, "ewc_loss": 0.11279296875, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 0.00010442733764648438, "grad_norm": 3.7813665866851807, "learning_rate": 2.6939381093683765e-05, "loss": 0.6365, "mean_token_accuracy": 0.8260443806648254, "num_tokens": 48279667.0, "step": 1272 }, { "epoch": 0.16193868464571937, "ewc_loss": 0.16796875, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 0.00016021728515625, "grad_norm": 5.2458086013793945, "learning_rate": 2.696057651547266e-05, "loss": 0.5815, "mean_token_accuracy": 0.8623104691505432, "num_tokens": 48319185.0, "step": 1273 }, { "epoch": 0.1620658949243099, "ewc_loss": 0.1328125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.0001239776611328125, "grad_norm": 3.978546619415283, "learning_rate": 2.6981771937261554e-05, "loss": 0.5347, "mean_token_accuracy": 0.8675429821014404, "num_tokens": 48357250.0, "step": 1274 }, { "epoch": 0.1621931052029004, "ewc_loss": 0.1728515625, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.000164031982421875, "grad_norm": 5.272563934326172, "learning_rate": 2.7002967359050445e-05, "loss": 0.6587, "mean_token_accuracy": 0.841558575630188, "num_tokens": 48396577.0, "step": 1275 }, { "epoch": 0.1623203154814909, "ewc_loss": 0.15234375, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.00014400482177734375, "grad_norm": 4.3138604164123535, "learning_rate": 2.702416278083934e-05, "loss": 0.6326, "mean_token_accuracy": 0.8408598899841309, "num_tokens": 48431360.0, "step": 1276 }, { "epoch": 0.16244752576008142, "ewc_loss": 0.16015625, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.00015163421630859375, "grad_norm": 5.0005974769592285, "learning_rate": 2.7045358202628233e-05, "loss": 0.6092, "mean_token_accuracy": 0.8531867265701294, "num_tokens": 48471640.0, "step": 1277 }, { "epoch": 0.16257473603867192, "ewc_loss": 0.1533203125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.00014495849609375, "grad_norm": 4.481804370880127, "learning_rate": 2.7066553624417128e-05, "loss": 0.6581, "mean_token_accuracy": 0.8296164274215698, "num_tokens": 48509870.0, "step": 1278 }, { "epoch": 0.16270194631726242, "ewc_loss": 0.1484375, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.0001392364501953125, "grad_norm": 4.62996244430542, "learning_rate": 2.7087749046206022e-05, "loss": 0.6284, "mean_token_accuracy": 0.8468794226646423, "num_tokens": 48554174.0, "step": 1279 }, { "epoch": 0.16282915659585295, "ewc_loss": 0.1455078125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.503655433654785, "learning_rate": 2.7108944467994913e-05, "loss": 0.6317, "mean_token_accuracy": 0.8407034873962402, "num_tokens": 48588590.0, "step": 1280 }, { "epoch": 0.16295636687444345, "ewc_loss": 0.140625, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.0001316070556640625, "grad_norm": 4.4258317947387695, "learning_rate": 2.7130139889783808e-05, "loss": 0.6476, "mean_token_accuracy": 0.8352134823799133, "num_tokens": 48633567.0, "step": 1281 }, { "epoch": 0.16308357715303395, "ewc_loss": 0.13671875, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.0001277923583984375, "grad_norm": 4.3743720054626465, "learning_rate": 2.7151335311572702e-05, "loss": 0.6619, "mean_token_accuracy": 0.8290997743606567, "num_tokens": 48673764.0, "step": 1282 }, { "epoch": 0.16321078743162448, "ewc_loss": 0.13671875, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.0001277923583984375, "grad_norm": 4.453065872192383, "learning_rate": 2.7172530733361597e-05, "loss": 0.6399, "mean_token_accuracy": 0.8354320526123047, "num_tokens": 48722531.0, "step": 1283 }, { "epoch": 0.16333799771021498, "ewc_loss": 0.1220703125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.00011396408081054688, "grad_norm": 4.138572692871094, "learning_rate": 2.719372615515049e-05, "loss": 0.6144, "mean_token_accuracy": 0.8355758190155029, "num_tokens": 48758273.0, "step": 1284 }, { "epoch": 0.16346520798880548, "ewc_loss": 0.1416015625, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.00013256072998046875, "grad_norm": 4.6284098625183105, "learning_rate": 2.7214921576939385e-05, "loss": 0.5271, "mean_token_accuracy": 0.8701706528663635, "num_tokens": 48793016.0, "step": 1285 }, { "epoch": 0.163592418267396, "ewc_loss": 0.11083984375, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.00010251998901367188, "grad_norm": 3.943080186843872, "learning_rate": 2.723611699872828e-05, "loss": 0.5701, "mean_token_accuracy": 0.850949764251709, "num_tokens": 48832204.0, "step": 1286 }, { "epoch": 0.1637196285459865, "ewc_loss": 0.146484375, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.836148738861084, "learning_rate": 2.7257312420517167e-05, "loss": 0.5814, "mean_token_accuracy": 0.8560376167297363, "num_tokens": 48867199.0, "step": 1287 }, { "epoch": 0.163846838824577, "ewc_loss": 0.10888671875, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.00010061264038085938, "grad_norm": 3.8208627700805664, "learning_rate": 2.7278507842306062e-05, "loss": 0.5577, "mean_token_accuracy": 0.8526198267936707, "num_tokens": 48901423.0, "step": 1288 }, { "epoch": 0.16397404910316754, "ewc_loss": 0.1572265625, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 0.000148773193359375, "grad_norm": 5.078003883361816, "learning_rate": 2.7299703264094956e-05, "loss": 0.687, "mean_token_accuracy": 0.8272053003311157, "num_tokens": 48941883.0, "step": 1289 }, { "epoch": 0.16410125938175804, "ewc_loss": 0.11767578125, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 0.00010919570922851562, "grad_norm": 3.845484972000122, "learning_rate": 2.732089868588385e-05, "loss": 0.5792, "mean_token_accuracy": 0.8461820483207703, "num_tokens": 48971764.0, "step": 1290 }, { "epoch": 0.16422846966034857, "ewc_loss": 0.1669921875, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 0.0001583099365234375, "grad_norm": 5.254181861877441, "learning_rate": 2.7342094107672745e-05, "loss": 0.5723, "mean_token_accuracy": 0.8636038303375244, "num_tokens": 49003687.0, "step": 1291 }, { "epoch": 0.16435567993893907, "ewc_loss": 0.13671875, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 0.0001277923583984375, "grad_norm": 4.060852527618408, "learning_rate": 2.736328952946164e-05, "loss": 0.5489, "mean_token_accuracy": 0.8630498647689819, "num_tokens": 49036896.0, "step": 1292 }, { "epoch": 0.16448289021752957, "ewc_loss": 0.169921875, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 0.00016117095947265625, "grad_norm": 5.18467378616333, "learning_rate": 2.7384484951250534e-05, "loss": 0.6978, "mean_token_accuracy": 0.8285889625549316, "num_tokens": 49076549.0, "step": 1293 }, { "epoch": 0.1646101004961201, "ewc_loss": 0.154296875, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 0.00014591217041015625, "grad_norm": 4.389141082763672, "learning_rate": 2.740568037303943e-05, "loss": 0.5676, "mean_token_accuracy": 0.8567305207252502, "num_tokens": 49114507.0, "step": 1294 }, { "epoch": 0.1647373107747106, "ewc_loss": 0.15234375, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 0.00014400482177734375, "grad_norm": 4.761167049407959, "learning_rate": 2.7426875794828316e-05, "loss": 0.5791, "mean_token_accuracy": 0.8594655394554138, "num_tokens": 49159794.0, "step": 1295 }, { "epoch": 0.1648645210533011, "ewc_loss": 0.1572265625, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 0.000148773193359375, "grad_norm": 4.663768768310547, "learning_rate": 2.744807121661721e-05, "loss": 0.623, "mean_token_accuracy": 0.8465652465820312, "num_tokens": 49198766.0, "step": 1296 }, { "epoch": 0.16499173133189163, "ewc_loss": 0.1279296875, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 0.00011920928955078125, "grad_norm": 4.203476428985596, "learning_rate": 2.7469266638406105e-05, "loss": 0.5394, "mean_token_accuracy": 0.8631604909896851, "num_tokens": 49236481.0, "step": 1297 }, { "epoch": 0.16511894161048213, "ewc_loss": 0.1572265625, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 0.000148773193359375, "grad_norm": 4.960524082183838, "learning_rate": 2.7490462060195e-05, "loss": 0.6877, "mean_token_accuracy": 0.8311856985092163, "num_tokens": 49271730.0, "step": 1298 }, { "epoch": 0.16524615188907263, "ewc_loss": 0.11474609375, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 0.00010633468627929688, "grad_norm": 3.8464903831481934, "learning_rate": 2.7511657481983894e-05, "loss": 0.5377, "mean_token_accuracy": 0.8570384979248047, "num_tokens": 49316208.0, "step": 1299 }, { "epoch": 0.16537336216766316, "ewc_loss": 0.1689453125, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 0.00016021728515625, "grad_norm": 5.307467460632324, "learning_rate": 2.7532852903772788e-05, "loss": 0.6306, "mean_token_accuracy": 0.8478962779045105, "num_tokens": 49355343.0, "step": 1300 }, { "epoch": 0.16550057244625366, "ewc_loss": 0.123046875, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 0.00011491775512695312, "grad_norm": 3.8689212799072266, "learning_rate": 2.7554048325561683e-05, "loss": 0.5587, "mean_token_accuracy": 0.8577014207839966, "num_tokens": 49394055.0, "step": 1301 }, { "epoch": 0.16562778272484416, "ewc_loss": 0.1865234375, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 0.0001773834228515625, "grad_norm": 5.506251335144043, "learning_rate": 2.757524374735057e-05, "loss": 0.6512, "mean_token_accuracy": 0.8486669659614563, "num_tokens": 49432466.0, "step": 1302 }, { "epoch": 0.1657549930034347, "ewc_loss": 0.15234375, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 0.00014400482177734375, "grad_norm": 4.267955780029297, "learning_rate": 2.7596439169139465e-05, "loss": 0.5991, "mean_token_accuracy": 0.8509261012077332, "num_tokens": 49470746.0, "step": 1303 }, { "epoch": 0.1658822032820252, "ewc_loss": 0.177734375, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 0.00016880035400390625, "grad_norm": 5.341649055480957, "learning_rate": 2.761763459092836e-05, "loss": 0.6658, "mean_token_accuracy": 0.8374884128570557, "num_tokens": 49502879.0, "step": 1304 }, { "epoch": 0.1660094135606157, "ewc_loss": 0.1689453125, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 0.00016021728515625, "grad_norm": 4.615049362182617, "learning_rate": 2.7638830012717254e-05, "loss": 0.5775, "mean_token_accuracy": 0.8598721027374268, "num_tokens": 49534875.0, "step": 1305 }, { "epoch": 0.16613662383920622, "ewc_loss": 0.1591796875, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 0.0001506805419921875, "grad_norm": 4.868756294250488, "learning_rate": 2.7660025434506148e-05, "loss": 0.6376, "mean_token_accuracy": 0.8450877070426941, "num_tokens": 49573537.0, "step": 1306 }, { "epoch": 0.16626383411779672, "ewc_loss": 0.1611328125, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 0.000152587890625, "grad_norm": 4.710800647735596, "learning_rate": 2.7681220856295042e-05, "loss": 0.6023, "mean_token_accuracy": 0.8540198802947998, "num_tokens": 49615190.0, "step": 1307 }, { "epoch": 0.16639104439638722, "ewc_loss": 0.1455078125, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.51164436340332, "learning_rate": 2.7702416278083937e-05, "loss": 0.6852, "mean_token_accuracy": 0.8251280784606934, "num_tokens": 49649788.0, "step": 1308 }, { "epoch": 0.16651825467497774, "ewc_loss": 0.1494140625, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 0.00014019012451171875, "grad_norm": 4.648669719696045, "learning_rate": 2.772361169987283e-05, "loss": 0.5963, "mean_token_accuracy": 0.8527472019195557, "num_tokens": 49683728.0, "step": 1309 }, { "epoch": 0.16664546495356825, "ewc_loss": 0.140625, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 0.0001316070556640625, "grad_norm": 4.429976940155029, "learning_rate": 2.774480712166172e-05, "loss": 0.5829, "mean_token_accuracy": 0.8486889004707336, "num_tokens": 49719331.0, "step": 1310 }, { "epoch": 0.16677267523215875, "ewc_loss": 0.138671875, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 0.00012969970703125, "grad_norm": 4.41986608505249, "learning_rate": 2.7766002543450613e-05, "loss": 0.5374, "mean_token_accuracy": 0.8661924600601196, "num_tokens": 49760158.0, "step": 1311 }, { "epoch": 0.16689988551074927, "ewc_loss": 0.142578125, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 0.000133514404296875, "grad_norm": 4.56477165222168, "learning_rate": 2.7787197965239508e-05, "loss": 0.6415, "mean_token_accuracy": 0.8350827097892761, "num_tokens": 49795285.0, "step": 1312 }, { "epoch": 0.16702709578933977, "ewc_loss": 0.1201171875, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 0.00011157989501953125, "grad_norm": 4.147981643676758, "learning_rate": 2.7808393387028402e-05, "loss": 0.5764, "mean_token_accuracy": 0.8488473892211914, "num_tokens": 49839071.0, "step": 1313 }, { "epoch": 0.16715430606793028, "ewc_loss": 0.146484375, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.786381721496582, "learning_rate": 2.7829588808817297e-05, "loss": 0.5851, "mean_token_accuracy": 0.8554606437683105, "num_tokens": 49870063.0, "step": 1314 }, { "epoch": 0.1672815163465208, "ewc_loss": 0.111328125, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 0.00010251998901367188, "grad_norm": 3.945950746536255, "learning_rate": 2.785078423060619e-05, "loss": 0.5799, "mean_token_accuracy": 0.8432118892669678, "num_tokens": 49908232.0, "step": 1315 }, { "epoch": 0.1674087266251113, "ewc_loss": 0.154296875, "ewc_loss_diag": 8.58306884765625e-05, "ewc_loss_parallel": 0.00014591217041015625, "grad_norm": 5.027949810028076, "learning_rate": 2.7871979652395085e-05, "loss": 0.6468, "mean_token_accuracy": 0.8382561206817627, "num_tokens": 49944676.0, "step": 1316 }, { "epoch": 0.16753593690370183, "ewc_loss": 0.1142578125, "ewc_loss_diag": 8.58306884765625e-05, "ewc_loss_parallel": 0.00010538101196289062, "grad_norm": 3.8492400646209717, "learning_rate": 2.789317507418398e-05, "loss": 0.5536, "mean_token_accuracy": 0.8569358587265015, "num_tokens": 49978472.0, "step": 1317 }, { "epoch": 0.16766314718229233, "ewc_loss": 0.169921875, "ewc_loss_diag": 8.58306884765625e-05, "ewc_loss_parallel": 0.00016117095947265625, "grad_norm": 5.31126594543457, "learning_rate": 2.7914370495972867e-05, "loss": 0.6537, "mean_token_accuracy": 0.8435118198394775, "num_tokens": 50010564.0, "step": 1318 }, { "epoch": 0.16779035746088283, "ewc_loss": 0.1259765625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 0.00011730194091796875, "grad_norm": 3.9125165939331055, "learning_rate": 2.7935565917761762e-05, "loss": 0.605, "mean_token_accuracy": 0.8437520265579224, "num_tokens": 50049538.0, "step": 1319 }, { "epoch": 0.16791756773947336, "ewc_loss": 0.1865234375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 0.0001773834228515625, "grad_norm": 5.503983497619629, "learning_rate": 2.7956761339550656e-05, "loss": 0.6207, "mean_token_accuracy": 0.856816291809082, "num_tokens": 50090779.0, "step": 1320 }, { "epoch": 0.16804477801806386, "ewc_loss": 0.150390625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 0.000141143798828125, "grad_norm": 4.254049777984619, "learning_rate": 2.797795676133955e-05, "loss": 0.6336, "mean_token_accuracy": 0.8434484601020813, "num_tokens": 50129527.0, "step": 1321 }, { "epoch": 0.16817198829665436, "ewc_loss": 0.181640625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 0.00017261505126953125, "grad_norm": 5.344785213470459, "learning_rate": 2.7999152183128445e-05, "loss": 0.6232, "mean_token_accuracy": 0.8535795211791992, "num_tokens": 50167950.0, "step": 1322 }, { "epoch": 0.1682991985752449, "ewc_loss": 0.1689453125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 0.00016021728515625, "grad_norm": 4.604300022125244, "learning_rate": 2.802034760491734e-05, "loss": 0.6148, "mean_token_accuracy": 0.853628396987915, "num_tokens": 50205079.0, "step": 1323 }, { "epoch": 0.1684264088538354, "ewc_loss": 0.1572265625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 0.000148773193359375, "grad_norm": 4.853799819946289, "learning_rate": 2.8041543026706234e-05, "loss": 0.5942, "mean_token_accuracy": 0.8534350395202637, "num_tokens": 50237909.0, "step": 1324 }, { "epoch": 0.1685536191324259, "ewc_loss": 0.1640625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 0.00015544891357421875, "grad_norm": 4.828513145446777, "learning_rate": 2.8062738448495125e-05, "loss": 0.663, "mean_token_accuracy": 0.8391256928443909, "num_tokens": 50272825.0, "step": 1325 }, { "epoch": 0.16868082941101642, "ewc_loss": 0.13671875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 0.0001277923583984375, "grad_norm": 4.319072723388672, "learning_rate": 2.808393387028402e-05, "loss": 0.621, "mean_token_accuracy": 0.8439856171607971, "num_tokens": 50312964.0, "step": 1326 }, { "epoch": 0.16880803968960692, "ewc_loss": 0.1630859375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 0.0001544952392578125, "grad_norm": 5.0199689865112305, "learning_rate": 2.8105129292072914e-05, "loss": 0.6448, "mean_token_accuracy": 0.8404326438903809, "num_tokens": 50355912.0, "step": 1327 }, { "epoch": 0.16893524996819742, "ewc_loss": 0.12353515625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 0.00011491775512695312, "grad_norm": 4.053516864776611, "learning_rate": 2.812632471386181e-05, "loss": 0.5945, "mean_token_accuracy": 0.8476386070251465, "num_tokens": 50395875.0, "step": 1328 }, { "epoch": 0.16906246024678795, "ewc_loss": 0.1650390625, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 0.000156402587890625, "grad_norm": 5.187395095825195, "learning_rate": 2.8147520135650703e-05, "loss": 0.6842, "mean_token_accuracy": 0.8368303775787354, "num_tokens": 50435628.0, "step": 1329 }, { "epoch": 0.16918967052537845, "ewc_loss": 0.1298828125, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 0.00012111663818359375, "grad_norm": 4.0613274574279785, "learning_rate": 2.8168715557439597e-05, "loss": 0.6524, "mean_token_accuracy": 0.8264066576957703, "num_tokens": 50474812.0, "step": 1330 }, { "epoch": 0.16931688080396895, "ewc_loss": 0.1708984375, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 0.0001621246337890625, "grad_norm": 5.251420021057129, "learning_rate": 2.818991097922849e-05, "loss": 0.5756, "mean_token_accuracy": 0.8666958808898926, "num_tokens": 50510058.0, "step": 1331 }, { "epoch": 0.16944409108255948, "ewc_loss": 0.140625, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 0.0001316070556640625, "grad_norm": 4.208634376525879, "learning_rate": 2.8211106401017383e-05, "loss": 0.5413, "mean_token_accuracy": 0.8672868013381958, "num_tokens": 50546282.0, "step": 1332 }, { "epoch": 0.16957130136114998, "ewc_loss": 0.1708984375, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 0.0001621246337890625, "grad_norm": 5.195985794067383, "learning_rate": 2.8232301822806274e-05, "loss": 0.5897, "mean_token_accuracy": 0.8604587912559509, "num_tokens": 50577457.0, "step": 1333 }, { "epoch": 0.16969851163974048, "ewc_loss": 0.1484375, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 0.0001392364501953125, "grad_norm": 4.37975549697876, "learning_rate": 2.8253497244595168e-05, "loss": 0.6679, "mean_token_accuracy": 0.8311684727668762, "num_tokens": 50619754.0, "step": 1334 }, { "epoch": 0.169825721918331, "ewc_loss": 0.1611328125, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 0.000152587890625, "grad_norm": 4.99675989151001, "learning_rate": 2.8274692666384062e-05, "loss": 0.6144, "mean_token_accuracy": 0.850027322769165, "num_tokens": 50660674.0, "step": 1335 }, { "epoch": 0.1699529321969215, "ewc_loss": 0.150390625, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 0.000141143798828125, "grad_norm": 4.474977970123291, "learning_rate": 2.8295888088172957e-05, "loss": 0.5962, "mean_token_accuracy": 0.8532043099403381, "num_tokens": 50698404.0, "step": 1336 }, { "epoch": 0.170080142475512, "ewc_loss": 0.1513671875, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 0.0001430511474609375, "grad_norm": 4.774916172027588, "learning_rate": 2.831708350996185e-05, "loss": 0.6028, "mean_token_accuracy": 0.8495889902114868, "num_tokens": 50736329.0, "step": 1337 }, { "epoch": 0.17020735275410254, "ewc_loss": 0.14453125, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 0.0001354217529296875, "grad_norm": 4.472385406494141, "learning_rate": 2.8338278931750746e-05, "loss": 0.6541, "mean_token_accuracy": 0.8250513076782227, "num_tokens": 50770065.0, "step": 1338 }, { "epoch": 0.17033456303269304, "ewc_loss": 0.1474609375, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 0.00013828277587890625, "grad_norm": 4.668282508850098, "learning_rate": 2.835947435353964e-05, "loss": 0.5969, "mean_token_accuracy": 0.8542708158493042, "num_tokens": 50811603.0, "step": 1339 }, { "epoch": 0.17046177331128357, "ewc_loss": 0.1376953125, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 0.00012874603271484375, "grad_norm": 4.364517688751221, "learning_rate": 2.8380669775328528e-05, "loss": 0.5293, "mean_token_accuracy": 0.8654620051383972, "num_tokens": 50846141.0, "step": 1340 }, { "epoch": 0.17058898358987407, "ewc_loss": 0.146484375, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.724486827850342, "learning_rate": 2.8401865197117422e-05, "loss": 0.5831, "mean_token_accuracy": 0.8567622303962708, "num_tokens": 50885511.0, "step": 1341 }, { "epoch": 0.17071619386846457, "ewc_loss": 0.126953125, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 0.00011873245239257812, "grad_norm": 4.2086710929870605, "learning_rate": 2.8423060618906317e-05, "loss": 0.6096, "mean_token_accuracy": 0.838454008102417, "num_tokens": 50926374.0, "step": 1342 }, { "epoch": 0.1708434041470551, "ewc_loss": 0.1484375, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 0.0001392364501953125, "grad_norm": 4.85951566696167, "learning_rate": 2.844425604069521e-05, "loss": 0.6393, "mean_token_accuracy": 0.8348333835601807, "num_tokens": 50965453.0, "step": 1343 }, { "epoch": 0.1709706144256456, "ewc_loss": 0.123046875, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 0.00011444091796875, "grad_norm": 4.074211597442627, "learning_rate": 2.8465451462484106e-05, "loss": 0.6079, "mean_token_accuracy": 0.8455837965011597, "num_tokens": 51007921.0, "step": 1344 }, { "epoch": 0.1710978247042361, "ewc_loss": 0.158203125, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 0.00014972686767578125, "grad_norm": 5.057172775268555, "learning_rate": 2.8486646884273e-05, "loss": 0.6228, "mean_token_accuracy": 0.8465960025787354, "num_tokens": 51046982.0, "step": 1345 }, { "epoch": 0.17122503498282662, "ewc_loss": 0.123046875, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 0.00011444091796875, "grad_norm": 3.9864773750305176, "learning_rate": 2.8507842306061894e-05, "loss": 0.5225, "mean_token_accuracy": 0.8667465448379517, "num_tokens": 51083192.0, "step": 1346 }, { "epoch": 0.17135224526141712, "ewc_loss": 0.169921875, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 0.00016117095947265625, "grad_norm": 5.246100425720215, "learning_rate": 2.852903772785079e-05, "loss": 0.5813, "mean_token_accuracy": 0.8622505068778992, "num_tokens": 51125258.0, "step": 1347 }, { "epoch": 0.17147945554000762, "ewc_loss": 0.1318359375, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 0.00012302398681640625, "grad_norm": 4.058828830718994, "learning_rate": 2.8550233149639676e-05, "loss": 0.5756, "mean_token_accuracy": 0.8521649837493896, "num_tokens": 51164036.0, "step": 1348 }, { "epoch": 0.17160666581859815, "ewc_loss": 0.1708984375, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 0.0001621246337890625, "grad_norm": 5.28989315032959, "learning_rate": 2.857142857142857e-05, "loss": 0.5821, "mean_token_accuracy": 0.8615176677703857, "num_tokens": 51201288.0, "step": 1349 }, { "epoch": 0.17173387609718865, "ewc_loss": 0.146484375, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.337402820587158, "learning_rate": 2.8592623993217465e-05, "loss": 0.6254, "mean_token_accuracy": 0.8426388502120972, "num_tokens": 51235931.0, "step": 1350 }, { "epoch": 0.17186108637577915, "ewc_loss": 0.1611328125, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 0.000152587890625, "grad_norm": 4.994572162628174, "learning_rate": 2.861381941500636e-05, "loss": 0.6525, "mean_token_accuracy": 0.8405663371086121, "num_tokens": 51277188.0, "step": 1351 }, { "epoch": 0.17198829665436968, "ewc_loss": 0.1552734375, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 0.0001468658447265625, "grad_norm": 4.662824630737305, "learning_rate": 2.8635014836795254e-05, "loss": 0.6085, "mean_token_accuracy": 0.8495453000068665, "num_tokens": 51313829.0, "step": 1352 }, { "epoch": 0.17211550693296018, "ewc_loss": 0.140625, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 0.0001316070556640625, "grad_norm": 4.453411102294922, "learning_rate": 2.865621025858415e-05, "loss": 0.6589, "mean_token_accuracy": 0.8278090357780457, "num_tokens": 51355678.0, "step": 1353 }, { "epoch": 0.17224271721155068, "ewc_loss": 0.16015625, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 0.00015163421630859375, "grad_norm": 4.994379997253418, "learning_rate": 2.8677405680373043e-05, "loss": 0.5891, "mean_token_accuracy": 0.8557038307189941, "num_tokens": 51390125.0, "step": 1354 }, { "epoch": 0.1723699274901412, "ewc_loss": 0.119140625, "ewc_loss_diag": 8.869171142578125e-05, "ewc_loss_parallel": 0.00011014938354492188, "grad_norm": 3.9958126544952393, "learning_rate": 2.869860110216193e-05, "loss": 0.5467, "mean_token_accuracy": 0.8563305139541626, "num_tokens": 51426384.0, "step": 1355 }, { "epoch": 0.1724971377687317, "ewc_loss": 0.17578125, "ewc_loss_diag": 8.869171142578125e-05, "ewc_loss_parallel": 0.00016689300537109375, "grad_norm": 5.401102542877197, "learning_rate": 2.8719796523950825e-05, "loss": 0.6949, "mean_token_accuracy": 0.8349924087524414, "num_tokens": 51467087.0, "step": 1356 }, { "epoch": 0.1726243480473222, "ewc_loss": 0.12060546875, "ewc_loss_diag": 8.869171142578125e-05, "ewc_loss_parallel": 0.00011205673217773438, "grad_norm": 3.8335235118865967, "learning_rate": 2.874099194573972e-05, "loss": 0.6003, "mean_token_accuracy": 0.8431323766708374, "num_tokens": 51505126.0, "step": 1357 }, { "epoch": 0.17275155832591274, "ewc_loss": 0.20703125, "ewc_loss_diag": 8.869171142578125e-05, "ewc_loss_parallel": 0.0001983642578125, "grad_norm": 5.794839859008789, "learning_rate": 2.8762187367528614e-05, "loss": 0.7043, "mean_token_accuracy": 0.8390055894851685, "num_tokens": 51546623.0, "step": 1358 }, { "epoch": 0.17287876860450324, "ewc_loss": 0.150390625, "ewc_loss_diag": 8.869171142578125e-05, "ewc_loss_parallel": 0.000141143798828125, "grad_norm": 4.196341514587402, "learning_rate": 2.878338278931751e-05, "loss": 0.6202, "mean_token_accuracy": 0.8445707559585571, "num_tokens": 51583175.0, "step": 1359 }, { "epoch": 0.17300597888309374, "ewc_loss": 0.212890625, "ewc_loss_diag": 8.869171142578125e-05, "ewc_loss_parallel": 0.0002040863037109375, "grad_norm": 5.793371200561523, "learning_rate": 2.8804578211106403e-05, "loss": 0.6397, "mean_token_accuracy": 0.8567860722541809, "num_tokens": 51621454.0, "step": 1360 }, { "epoch": 0.17313318916168427, "ewc_loss": 0.193359375, "ewc_loss_diag": 8.869171142578125e-05, "ewc_loss_parallel": 0.0001850128173828125, "grad_norm": 4.75695276260376, "learning_rate": 2.8825773632895297e-05, "loss": 0.6426, "mean_token_accuracy": 0.852653980255127, "num_tokens": 51659482.0, "step": 1361 }, { "epoch": 0.17326039944027477, "ewc_loss": 0.1826171875, "ewc_loss_diag": 8.869171142578125e-05, "ewc_loss_parallel": 0.0001735687255859375, "grad_norm": 5.271962642669678, "learning_rate": 2.884696905468419e-05, "loss": 0.6398, "mean_token_accuracy": 0.849541962146759, "num_tokens": 51695478.0, "step": 1362 }, { "epoch": 0.17338760971886527, "ewc_loss": 0.1923828125, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 0.00018310546875, "grad_norm": 4.993807315826416, "learning_rate": 2.886816447647308e-05, "loss": 0.6907, "mean_token_accuracy": 0.8349325656890869, "num_tokens": 51736436.0, "step": 1363 }, { "epoch": 0.1735148199974558, "ewc_loss": 0.1533203125, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 0.00014495849609375, "grad_norm": 4.6775712966918945, "learning_rate": 2.8889359898261974e-05, "loss": 0.5848, "mean_token_accuracy": 0.8561873435974121, "num_tokens": 51772502.0, "step": 1364 }, { "epoch": 0.1736420302760463, "ewc_loss": 0.1708984375, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 0.0001621246337890625, "grad_norm": 5.087886333465576, "learning_rate": 2.8910555320050868e-05, "loss": 0.6369, "mean_token_accuracy": 0.8455288410186768, "num_tokens": 51816343.0, "step": 1365 }, { "epoch": 0.17376924055463683, "ewc_loss": 0.146484375, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.408919811248779, "learning_rate": 2.8931750741839762e-05, "loss": 0.6075, "mean_token_accuracy": 0.8497114181518555, "num_tokens": 51861317.0, "step": 1366 }, { "epoch": 0.17389645083322733, "ewc_loss": 0.1630859375, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 0.0001544952392578125, "grad_norm": 5.007945537567139, "learning_rate": 2.8952946163628657e-05, "loss": 0.623, "mean_token_accuracy": 0.8468151092529297, "num_tokens": 51902994.0, "step": 1367 }, { "epoch": 0.17402366111181783, "ewc_loss": 0.1455078125, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.44999361038208, "learning_rate": 2.897414158541755e-05, "loss": 0.5907, "mean_token_accuracy": 0.8537968397140503, "num_tokens": 51936374.0, "step": 1368 }, { "epoch": 0.17415087139040836, "ewc_loss": 0.1513671875, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 0.00014209747314453125, "grad_norm": 4.7724385261535645, "learning_rate": 2.8995337007206446e-05, "loss": 0.5959, "mean_token_accuracy": 0.8509138822555542, "num_tokens": 51970235.0, "step": 1369 }, { "epoch": 0.17427808166899886, "ewc_loss": 0.1455078125, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.550088405609131, "learning_rate": 2.901653242899534e-05, "loss": 0.5789, "mean_token_accuracy": 0.8576482534408569, "num_tokens": 52006220.0, "step": 1370 }, { "epoch": 0.17440529194758936, "ewc_loss": 0.1416015625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 0.00013256072998046875, "grad_norm": 4.565794944763184, "learning_rate": 2.903772785078423e-05, "loss": 0.5612, "mean_token_accuracy": 0.8598800897598267, "num_tokens": 52046634.0, "step": 1371 }, { "epoch": 0.1745325022261799, "ewc_loss": 0.1435546875, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 0.00013446807861328125, "grad_norm": 4.631047248840332, "learning_rate": 2.9058923272573126e-05, "loss": 0.6473, "mean_token_accuracy": 0.833184003829956, "num_tokens": 52079768.0, "step": 1372 }, { "epoch": 0.1746597125047704, "ewc_loss": 0.1328125, "ewc_loss_diag": 9.012222290039062e-05, "ewc_loss_parallel": 0.0001239776611328125, "grad_norm": 4.428402900695801, "learning_rate": 2.908011869436202e-05, "loss": 0.5596, "mean_token_accuracy": 0.8575356006622314, "num_tokens": 52118160.0, "step": 1373 }, { "epoch": 0.1747869227833609, "ewc_loss": 0.1416015625, "ewc_loss_diag": 9.012222290039062e-05, "ewc_loss_parallel": 0.00013256072998046875, "grad_norm": 4.689064979553223, "learning_rate": 2.9101314116150914e-05, "loss": 0.6103, "mean_token_accuracy": 0.8459857702255249, "num_tokens": 52160219.0, "step": 1374 }, { "epoch": 0.17491413306195142, "ewc_loss": 0.1259765625, "ewc_loss_diag": 9.012222290039062e-05, "ewc_loss_parallel": 0.00011682510375976562, "grad_norm": 4.344021797180176, "learning_rate": 2.9122509537939806e-05, "loss": 0.5524, "mean_token_accuracy": 0.8562793731689453, "num_tokens": 52194746.0, "step": 1375 }, { "epoch": 0.17504134334054192, "ewc_loss": 0.140625, "ewc_loss_diag": 9.012222290039062e-05, "ewc_loss_parallel": 0.0001316070556640625, "grad_norm": 4.770477294921875, "learning_rate": 2.91437049597287e-05, "loss": 0.6568, "mean_token_accuracy": 0.830150306224823, "num_tokens": 52232625.0, "step": 1376 }, { "epoch": 0.17516855361913242, "ewc_loss": 0.12109375, "ewc_loss_diag": 9.012222290039062e-05, "ewc_loss_parallel": 0.00011205673217773438, "grad_norm": 4.239146709442139, "learning_rate": 2.9164900381517594e-05, "loss": 0.6344, "mean_token_accuracy": 0.8323200941085815, "num_tokens": 52273030.0, "step": 1377 }, { "epoch": 0.17529576389772294, "ewc_loss": 0.1474609375, "ewc_loss_diag": 9.012222290039062e-05, "ewc_loss_parallel": 0.00013828277587890625, "grad_norm": 4.908356189727783, "learning_rate": 2.9186095803306485e-05, "loss": 0.5563, "mean_token_accuracy": 0.8610596060752869, "num_tokens": 52306490.0, "step": 1378 }, { "epoch": 0.17542297417631345, "ewc_loss": 0.1162109375, "ewc_loss_diag": 9.012222290039062e-05, "ewc_loss_parallel": 0.00010728836059570312, "grad_norm": 4.085898399353027, "learning_rate": 2.920729122509538e-05, "loss": 0.5209, "mean_token_accuracy": 0.8637829422950745, "num_tokens": 52346717.0, "step": 1379 }, { "epoch": 0.17555018445490395, "ewc_loss": 0.158203125, "ewc_loss_diag": 9.059906005859375e-05, "ewc_loss_parallel": 0.00014972686767578125, "grad_norm": 5.173569202423096, "learning_rate": 2.9228486646884274e-05, "loss": 0.6125, "mean_token_accuracy": 0.851205050945282, "num_tokens": 52378165.0, "step": 1380 }, { "epoch": 0.17567739473349447, "ewc_loss": 0.11181640625, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 0.00010251998901367188, "grad_norm": 3.893543243408203, "learning_rate": 2.924968206867317e-05, "loss": 0.6299, "mean_token_accuracy": 0.832587718963623, "num_tokens": 52418719.0, "step": 1381 }, { "epoch": 0.17580460501208497, "ewc_loss": 0.181640625, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 0.00017261505126953125, "grad_norm": 5.562623500823975, "learning_rate": 2.9270877490462063e-05, "loss": 0.6087, "mean_token_accuracy": 0.8584965467453003, "num_tokens": 52459094.0, "step": 1382 }, { "epoch": 0.17593181529067548, "ewc_loss": 0.125, "ewc_loss_diag": 9.1552734375e-05, "ewc_loss_parallel": 0.0001163482666015625, "grad_norm": 3.830096483230591, "learning_rate": 2.9292072912250958e-05, "loss": 0.5857, "mean_token_accuracy": 0.8477348685264587, "num_tokens": 52503117.0, "step": 1383 }, { "epoch": 0.176059025569266, "ewc_loss": 0.2216796875, "ewc_loss_diag": 9.1552734375e-05, "ewc_loss_parallel": 0.00021266937255859375, "grad_norm": 6.019906997680664, "learning_rate": 2.9313268334039852e-05, "loss": 0.7055, "mean_token_accuracy": 0.8405052423477173, "num_tokens": 52542643.0, "step": 1384 }, { "epoch": 0.1761862358478565, "ewc_loss": 0.1552734375, "ewc_loss_diag": 9.1552734375e-05, "ewc_loss_parallel": 0.0001468658447265625, "grad_norm": 4.2201361656188965, "learning_rate": 2.9334463755828746e-05, "loss": 0.6156, "mean_token_accuracy": 0.8481325507164001, "num_tokens": 52585885.0, "step": 1385 }, { "epoch": 0.176313446126447, "ewc_loss": 0.23828125, "ewc_loss_diag": 9.1552734375e-05, "ewc_loss_parallel": 0.00022983551025390625, "grad_norm": 6.1731791496276855, "learning_rate": 2.9355659177617634e-05, "loss": 0.7468, "mean_token_accuracy": 0.8332396745681763, "num_tokens": 52621629.0, "step": 1386 }, { "epoch": 0.17644065640503753, "ewc_loss": 0.220703125, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 0.0002117156982421875, "grad_norm": 4.9313645362854, "learning_rate": 2.937685459940653e-05, "loss": 0.68, "mean_token_accuracy": 0.8451716303825378, "num_tokens": 52667056.0, "step": 1387 }, { "epoch": 0.17656786668362803, "ewc_loss": 0.19921875, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 0.00019073486328125, "grad_norm": 5.632881164550781, "learning_rate": 2.9398050021195423e-05, "loss": 0.7174, "mean_token_accuracy": 0.8348717093467712, "num_tokens": 52705857.0, "step": 1388 }, { "epoch": 0.17669507696221853, "ewc_loss": 0.23046875, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 0.00022125244140625, "grad_norm": 5.264941692352295, "learning_rate": 2.9419245442984317e-05, "loss": 0.7354, "mean_token_accuracy": 0.8359406590461731, "num_tokens": 52748908.0, "step": 1389 }, { "epoch": 0.17682228724080906, "ewc_loss": 0.16015625, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 0.0001506805419921875, "grad_norm": 4.8423237800598145, "learning_rate": 2.9440440864773212e-05, "loss": 0.6385, "mean_token_accuracy": 0.8442294597625732, "num_tokens": 52783600.0, "step": 1390 }, { "epoch": 0.17694949751939956, "ewc_loss": 0.216796875, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 0.0002079010009765625, "grad_norm": 5.533773422241211, "learning_rate": 2.9461636286562106e-05, "loss": 0.6215, "mean_token_accuracy": 0.8667023181915283, "num_tokens": 52819873.0, "step": 1391 }, { "epoch": 0.1770767077979901, "ewc_loss": 0.146484375, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.4675092697143555, "learning_rate": 2.9482831708351e-05, "loss": 0.5513, "mean_token_accuracy": 0.8599842190742493, "num_tokens": 52858883.0, "step": 1392 }, { "epoch": 0.1772039180765806, "ewc_loss": 0.203125, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 0.000194549560546875, "grad_norm": 5.713611602783203, "learning_rate": 2.9504027130139888e-05, "loss": 0.6628, "mean_token_accuracy": 0.8487613201141357, "num_tokens": 52895501.0, "step": 1393 }, { "epoch": 0.1773311283551711, "ewc_loss": 0.1552734375, "ewc_loss_diag": 9.34600830078125e-05, "ewc_loss_parallel": 0.00014591217041015625, "grad_norm": 4.3710737228393555, "learning_rate": 2.9525222551928783e-05, "loss": 0.5626, "mean_token_accuracy": 0.8633790016174316, "num_tokens": 52934678.0, "step": 1394 }, { "epoch": 0.17745833863376162, "ewc_loss": 0.2080078125, "ewc_loss_diag": 9.441375732421875e-05, "ewc_loss_parallel": 0.0001983642578125, "grad_norm": 5.806839942932129, "learning_rate": 2.9546417973717677e-05, "loss": 0.7158, "mean_token_accuracy": 0.8322266340255737, "num_tokens": 52972540.0, "step": 1395 }, { "epoch": 0.17758554891235212, "ewc_loss": 0.181640625, "ewc_loss_diag": 9.441375732421875e-05, "ewc_loss_parallel": 0.000171661376953125, "grad_norm": 4.629493713378906, "learning_rate": 2.956761339550657e-05, "loss": 0.6259, "mean_token_accuracy": 0.8481813669204712, "num_tokens": 53010819.0, "step": 1396 }, { "epoch": 0.17771275919094262, "ewc_loss": 0.1962890625, "ewc_loss_diag": 9.441375732421875e-05, "ewc_loss_parallel": 0.000186920166015625, "grad_norm": 5.614298343658447, "learning_rate": 2.9588808817295466e-05, "loss": 0.7066, "mean_token_accuracy": 0.8321428298950195, "num_tokens": 53050142.0, "step": 1397 }, { "epoch": 0.17783996946953315, "ewc_loss": 0.1923828125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 0.00018215179443359375, "grad_norm": 4.841613292694092, "learning_rate": 2.961000423908436e-05, "loss": 0.631, "mean_token_accuracy": 0.8566393256187439, "num_tokens": 53087306.0, "step": 1398 }, { "epoch": 0.17796717974812365, "ewc_loss": 0.173828125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 0.000164031982421875, "grad_norm": 5.201028347015381, "learning_rate": 2.9631199660873255e-05, "loss": 0.6693, "mean_token_accuracy": 0.8378088474273682, "num_tokens": 53131850.0, "step": 1399 }, { "epoch": 0.17809439002671415, "ewc_loss": 0.1767578125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 0.00016689300537109375, "grad_norm": 4.892735958099365, "learning_rate": 2.965239508266215e-05, "loss": 0.6261, "mean_token_accuracy": 0.8491643667221069, "num_tokens": 53168902.0, "step": 1400 }, { "epoch": 0.17822160030530468, "ewc_loss": 0.16015625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 0.0001506805419921875, "grad_norm": 4.859064102172852, "learning_rate": 2.9673590504451037e-05, "loss": 0.5896, "mean_token_accuracy": 0.8561602830886841, "num_tokens": 53208497.0, "step": 1401 }, { "epoch": 0.17834881058389518, "ewc_loss": 0.1630859375, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 0.00015354156494140625, "grad_norm": 4.866100788116455, "learning_rate": 2.969478592623993e-05, "loss": 0.6447, "mean_token_accuracy": 0.8385587930679321, "num_tokens": 53245618.0, "step": 1402 }, { "epoch": 0.17847602086248568, "ewc_loss": 0.15234375, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 0.00014209747314453125, "grad_norm": 4.678650379180908, "learning_rate": 2.9715981348028826e-05, "loss": 0.5971, "mean_token_accuracy": 0.8507922887802124, "num_tokens": 53280315.0, "step": 1403 }, { "epoch": 0.1786032311410762, "ewc_loss": 0.1552734375, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 0.00014591217041015625, "grad_norm": 4.816033363342285, "learning_rate": 2.973717676981772e-05, "loss": 0.6371, "mean_token_accuracy": 0.8434299230575562, "num_tokens": 53313291.0, "step": 1404 }, { "epoch": 0.1787304414196667, "ewc_loss": 0.1455078125, "ewc_loss_diag": 9.5367431640625e-05, "ewc_loss_parallel": 0.0001354217529296875, "grad_norm": 4.5765838623046875, "learning_rate": 2.9758372191606615e-05, "loss": 0.6138, "mean_token_accuracy": 0.846099853515625, "num_tokens": 53347764.0, "step": 1405 }, { "epoch": 0.1788576516982572, "ewc_loss": 0.15234375, "ewc_loss_diag": 9.5367431640625e-05, "ewc_loss_parallel": 0.00014209747314453125, "grad_norm": 4.818143844604492, "learning_rate": 2.977956761339551e-05, "loss": 0.6186, "mean_token_accuracy": 0.8413870334625244, "num_tokens": 53385184.0, "step": 1406 }, { "epoch": 0.17898486197684774, "ewc_loss": 0.1396484375, "ewc_loss_diag": 9.5367431640625e-05, "ewc_loss_parallel": 0.00012969970703125, "grad_norm": 4.469534397125244, "learning_rate": 2.9800763035184403e-05, "loss": 0.6807, "mean_token_accuracy": 0.8219958543777466, "num_tokens": 53425850.0, "step": 1407 }, { "epoch": 0.17911207225543824, "ewc_loss": 0.15234375, "ewc_loss_diag": 9.5367431640625e-05, "ewc_loss_parallel": 0.0001430511474609375, "grad_norm": 4.896149158477783, "learning_rate": 2.9821958456973298e-05, "loss": 0.6297, "mean_token_accuracy": 0.8443878293037415, "num_tokens": 53472444.0, "step": 1408 }, { "epoch": 0.17923928253402874, "ewc_loss": 0.1318359375, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 0.0001220703125, "grad_norm": 4.359963417053223, "learning_rate": 2.9843153878762185e-05, "loss": 0.6436, "mean_token_accuracy": 0.8335531949996948, "num_tokens": 53506996.0, "step": 1409 }, { "epoch": 0.17936649281261927, "ewc_loss": 0.154296875, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 0.00014495849609375, "grad_norm": 4.98947286605835, "learning_rate": 2.986434930055108e-05, "loss": 0.6418, "mean_token_accuracy": 0.8375750780105591, "num_tokens": 53542005.0, "step": 1410 }, { "epoch": 0.17949370309120977, "ewc_loss": 0.126953125, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 0.00011777877807617188, "grad_norm": 4.255631923675537, "learning_rate": 2.9885544722339974e-05, "loss": 0.6521, "mean_token_accuracy": 0.8346593379974365, "num_tokens": 53582130.0, "step": 1411 }, { "epoch": 0.17962091336980027, "ewc_loss": 0.16015625, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 0.0001506805419921875, "grad_norm": 5.149958610534668, "learning_rate": 2.990674014412887e-05, "loss": 0.6049, "mean_token_accuracy": 0.8564196825027466, "num_tokens": 53624528.0, "step": 1412 }, { "epoch": 0.1797481236483908, "ewc_loss": 0.125, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 0.00011587142944335938, "grad_norm": 4.1310906410217285, "learning_rate": 2.9927935565917763e-05, "loss": 0.6172, "mean_token_accuracy": 0.8392924070358276, "num_tokens": 53663701.0, "step": 1413 }, { "epoch": 0.1798753339269813, "ewc_loss": 0.17578125, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 0.0001659393310546875, "grad_norm": 5.399507999420166, "learning_rate": 2.9949130987706658e-05, "loss": 0.6568, "mean_token_accuracy": 0.8461071252822876, "num_tokens": 53700655.0, "step": 1414 }, { "epoch": 0.18000254420557182, "ewc_loss": 0.12890625, "ewc_loss_diag": 9.632110595703125e-05, "ewc_loss_parallel": 0.00011920928955078125, "grad_norm": 4.046360969543457, "learning_rate": 2.9970326409495552e-05, "loss": 0.6277, "mean_token_accuracy": 0.8398826718330383, "num_tokens": 53747387.0, "step": 1415 }, { "epoch": 0.18012975448416232, "ewc_loss": 0.1953125, "ewc_loss_diag": 9.632110595703125e-05, "ewc_loss_parallel": 0.00018596649169921875, "grad_norm": 5.657912731170654, "learning_rate": 2.9991521831284443e-05, "loss": 0.6622, "mean_token_accuracy": 0.8467248678207397, "num_tokens": 53789416.0, "step": 1416 }, { "epoch": 0.18025696476275282, "ewc_loss": 0.142578125, "ewc_loss_diag": 9.632110595703125e-05, "ewc_loss_parallel": 0.00013256072998046875, "grad_norm": 4.188094139099121, "learning_rate": 3.0012717253073337e-05, "loss": 0.5363, "mean_token_accuracy": 0.8703676462173462, "num_tokens": 53824300.0, "step": 1417 }, { "epoch": 0.18038417504134335, "ewc_loss": 0.2080078125, "ewc_loss_diag": 9.632110595703125e-05, "ewc_loss_parallel": 0.0001983642578125, "grad_norm": 5.708250045776367, "learning_rate": 3.0033912674862232e-05, "loss": 0.6739, "mean_token_accuracy": 0.8465465903282166, "num_tokens": 53861366.0, "step": 1418 }, { "epoch": 0.18051138531993385, "ewc_loss": 0.1669921875, "ewc_loss_diag": 9.632110595703125e-05, "ewc_loss_parallel": 0.00015735626220703125, "grad_norm": 4.559264183044434, "learning_rate": 3.0055108096651123e-05, "loss": 0.6537, "mean_token_accuracy": 0.8393086791038513, "num_tokens": 53899655.0, "step": 1419 }, { "epoch": 0.18063859559852435, "ewc_loss": 0.185546875, "ewc_loss_diag": 9.632110595703125e-05, "ewc_loss_parallel": 0.00017547607421875, "grad_norm": 5.365756988525391, "learning_rate": 3.0076303518440017e-05, "loss": 0.6547, "mean_token_accuracy": 0.8416970372200012, "num_tokens": 53935563.0, "step": 1420 }, { "epoch": 0.18076580587711488, "ewc_loss": 0.18359375, "ewc_loss_diag": 9.632110595703125e-05, "ewc_loss_parallel": 0.0001735687255859375, "grad_norm": 4.9297285079956055, "learning_rate": 3.0097498940228912e-05, "loss": 0.6494, "mean_token_accuracy": 0.8472477197647095, "num_tokens": 53970490.0, "step": 1421 }, { "epoch": 0.18089301615570538, "ewc_loss": 0.1533203125, "ewc_loss_diag": 9.632110595703125e-05, "ewc_loss_parallel": 0.00014400482177734375, "grad_norm": 4.7226786613464355, "learning_rate": 3.0118694362017806e-05, "loss": 0.5894, "mean_token_accuracy": 0.8548228740692139, "num_tokens": 54000460.0, "step": 1422 }, { "epoch": 0.18102022643429588, "ewc_loss": 0.1865234375, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 0.00017642974853515625, "grad_norm": 5.293346405029297, "learning_rate": 3.01398897838067e-05, "loss": 0.6447, "mean_token_accuracy": 0.8532617092132568, "num_tokens": 54038007.0, "step": 1423 }, { "epoch": 0.1811474367128864, "ewc_loss": 0.1357421875, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 0.000125885009765625, "grad_norm": 4.2073516845703125, "learning_rate": 3.016108520559559e-05, "loss": 0.6098, "mean_token_accuracy": 0.8407043218612671, "num_tokens": 54074515.0, "step": 1424 }, { "epoch": 0.1812746469914769, "ewc_loss": 0.203125, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 0.00019359588623046875, "grad_norm": 5.704926013946533, "learning_rate": 3.0182280627384486e-05, "loss": 0.7097, "mean_token_accuracy": 0.8360762596130371, "num_tokens": 54109355.0, "step": 1425 }, { "epoch": 0.1814018572700674, "ewc_loss": 0.138671875, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 0.00012874603271484375, "grad_norm": 4.131441116333008, "learning_rate": 3.020347604917338e-05, "loss": 0.5483, "mean_token_accuracy": 0.860302746295929, "num_tokens": 54150966.0, "step": 1426 }, { "epoch": 0.18152906754865794, "ewc_loss": 0.2255859375, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 5.967555522918701, "learning_rate": 3.0224671470962275e-05, "loss": 0.7292, "mean_token_accuracy": 0.8340842723846436, "num_tokens": 54188398.0, "step": 1427 }, { "epoch": 0.18165627782724844, "ewc_loss": 0.177734375, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 0.0001678466796875, "grad_norm": 4.555944442749023, "learning_rate": 3.024586689275117e-05, "loss": 0.6354, "mean_token_accuracy": 0.849108099937439, "num_tokens": 54233669.0, "step": 1428 }, { "epoch": 0.18178348810583894, "ewc_loss": 0.2119140625, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 0.000202178955078125, "grad_norm": 5.751354217529297, "learning_rate": 3.0267062314540064e-05, "loss": 0.6513, "mean_token_accuracy": 0.8533155918121338, "num_tokens": 54272377.0, "step": 1429 }, { "epoch": 0.18191069838442947, "ewc_loss": 0.205078125, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 0.00019550323486328125, "grad_norm": 4.977719306945801, "learning_rate": 3.0288257736328958e-05, "loss": 0.7157, "mean_token_accuracy": 0.8346466422080994, "num_tokens": 54313831.0, "step": 1430 }, { "epoch": 0.18203790866301997, "ewc_loss": 0.173828125, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 0.000164031982421875, "grad_norm": 5.172646999359131, "learning_rate": 3.0309453158117846e-05, "loss": 0.5912, "mean_token_accuracy": 0.8582412004470825, "num_tokens": 54356724.0, "step": 1431 }, { "epoch": 0.18216511894161047, "ewc_loss": 0.193359375, "ewc_loss_diag": 9.822845458984375e-05, "ewc_loss_parallel": 0.00018310546875, "grad_norm": 5.131042957305908, "learning_rate": 3.033064857990674e-05, "loss": 0.6422, "mean_token_accuracy": 0.8486496210098267, "num_tokens": 54398407.0, "step": 1432 }, { "epoch": 0.182292329220201, "ewc_loss": 0.1611328125, "ewc_loss_diag": 9.822845458984375e-05, "ewc_loss_parallel": 0.00015163421630859375, "grad_norm": 4.763276100158691, "learning_rate": 3.0351844001695635e-05, "loss": 0.5678, "mean_token_accuracy": 0.8648420572280884, "num_tokens": 54430023.0, "step": 1433 }, { "epoch": 0.1824195394987915, "ewc_loss": 0.1728515625, "ewc_loss_diag": 9.822845458984375e-05, "ewc_loss_parallel": 0.00016307830810546875, "grad_norm": 5.074102878570557, "learning_rate": 3.037303942348453e-05, "loss": 0.6105, "mean_token_accuracy": 0.8537636995315552, "num_tokens": 54467154.0, "step": 1434 }, { "epoch": 0.182546749777382, "ewc_loss": 0.154296875, "ewc_loss_diag": 9.822845458984375e-05, "ewc_loss_parallel": 0.00014495849609375, "grad_norm": 4.733381748199463, "learning_rate": 3.0394234845273423e-05, "loss": 0.6221, "mean_token_accuracy": 0.846142053604126, "num_tokens": 54501139.0, "step": 1435 }, { "epoch": 0.18267396005597253, "ewc_loss": 0.15234375, "ewc_loss_diag": 9.822845458984375e-05, "ewc_loss_parallel": 0.0001430511474609375, "grad_norm": 4.794090747833252, "learning_rate": 3.0415430267062318e-05, "loss": 0.6033, "mean_token_accuracy": 0.8508325219154358, "num_tokens": 54541084.0, "step": 1436 }, { "epoch": 0.18280117033456303, "ewc_loss": 0.1572265625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00014781951904296875, "grad_norm": 4.92372465133667, "learning_rate": 3.0436625688851212e-05, "loss": 0.5862, "mean_token_accuracy": 0.8563929796218872, "num_tokens": 54576745.0, "step": 1437 }, { "epoch": 0.18292838061315353, "ewc_loss": 0.138671875, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00012874603271484375, "grad_norm": 4.461557865142822, "learning_rate": 3.0457821110640107e-05, "loss": 0.6448, "mean_token_accuracy": 0.8376883268356323, "num_tokens": 54619039.0, "step": 1438 }, { "epoch": 0.18305559089174406, "ewc_loss": 0.1669921875, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00015735626220703125, "grad_norm": 5.167212009429932, "learning_rate": 3.0479016532428994e-05, "loss": 0.6196, "mean_token_accuracy": 0.8476911783218384, "num_tokens": 54660009.0, "step": 1439 }, { "epoch": 0.18318280117033456, "ewc_loss": 0.1298828125, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.0001201629638671875, "grad_norm": 4.231765270233154, "learning_rate": 3.050021195421789e-05, "loss": 0.6107, "mean_token_accuracy": 0.8425481915473938, "num_tokens": 54701253.0, "step": 1440 }, { "epoch": 0.1833100114489251, "ewc_loss": 0.1806640625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00017070770263671875, "grad_norm": 5.4692792892456055, "learning_rate": 3.052140737600678e-05, "loss": 0.6188, "mean_token_accuracy": 0.8556505441665649, "num_tokens": 54738406.0, "step": 1441 }, { "epoch": 0.1834372217275156, "ewc_loss": 0.12890625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00011920928955078125, "grad_norm": 4.0971808433532715, "learning_rate": 3.0542602797795674e-05, "loss": 0.6209, "mean_token_accuracy": 0.8400139212608337, "num_tokens": 54771808.0, "step": 1442 }, { "epoch": 0.1835644320061061, "ewc_loss": 0.205078125, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00019550323486328125, "grad_norm": 5.8035197257995605, "learning_rate": 3.056379821958457e-05, "loss": 0.5963, "mean_token_accuracy": 0.8690350651741028, "num_tokens": 54807337.0, "step": 1443 }, { "epoch": 0.18369164228469662, "ewc_loss": 0.14453125, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00013446807861328125, "grad_norm": 4.177159786224365, "learning_rate": 3.058499364137346e-05, "loss": 0.6212, "mean_token_accuracy": 0.8445181250572205, "num_tokens": 54844932.0, "step": 1444 }, { "epoch": 0.18381885256328712, "ewc_loss": 0.2333984375, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 6.053939342498779, "learning_rate": 3.060618906316236e-05, "loss": 0.7116, "mean_token_accuracy": 0.8380411267280579, "num_tokens": 54884492.0, "step": 1445 }, { "epoch": 0.18394606284187762, "ewc_loss": 0.1787109375, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00016880035400390625, "grad_norm": 4.586553573608398, "learning_rate": 3.062738448495125e-05, "loss": 0.6123, "mean_token_accuracy": 0.8589032888412476, "num_tokens": 54924164.0, "step": 1446 }, { "epoch": 0.18407327312046814, "ewc_loss": 0.228515625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00021839141845703125, "grad_norm": 5.908583164215088, "learning_rate": 3.064857990674014e-05, "loss": 0.6898, "mean_token_accuracy": 0.8507130742073059, "num_tokens": 54961676.0, "step": 1447 }, { "epoch": 0.18420048339905865, "ewc_loss": 0.205078125, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00019550323486328125, "grad_norm": 4.9908223152160645, "learning_rate": 3.066977532852904e-05, "loss": 0.692, "mean_token_accuracy": 0.8365322947502136, "num_tokens": 55000889.0, "step": 1448 }, { "epoch": 0.18432769367764915, "ewc_loss": 0.1923828125, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00018215179443359375, "grad_norm": 5.389123439788818, "learning_rate": 3.069097075031793e-05, "loss": 0.6454, "mean_token_accuracy": 0.8524627685546875, "num_tokens": 55039471.0, "step": 1449 }, { "epoch": 0.18445490395623967, "ewc_loss": 0.203125, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00019359588623046875, "grad_norm": 5.177496433258057, "learning_rate": 3.071216617210683e-05, "loss": 0.6884, "mean_token_accuracy": 0.838913083076477, "num_tokens": 55078408.0, "step": 1450 }, { "epoch": 0.18458211423483017, "ewc_loss": 0.1669921875, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.00015735626220703125, "grad_norm": 4.875053405761719, "learning_rate": 3.073336159389572e-05, "loss": 0.6528, "mean_token_accuracy": 0.8404457569122314, "num_tokens": 55116068.0, "step": 1451 }, { "epoch": 0.18470932451342068, "ewc_loss": 0.181640625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 0.000171661376953125, "grad_norm": 5.219395160675049, "learning_rate": 3.075455701568462e-05, "loss": 0.6525, "mean_token_accuracy": 0.8473069667816162, "num_tokens": 55149933.0, "step": 1452 }, { "epoch": 0.1848365347920112, "ewc_loss": 0.1533203125, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 0.00014400482177734375, "grad_norm": 4.651513576507568, "learning_rate": 3.077575243747351e-05, "loss": 0.5569, "mean_token_accuracy": 0.8657717704772949, "num_tokens": 55184674.0, "step": 1453 }, { "epoch": 0.1849637450706017, "ewc_loss": 0.169921875, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 0.00016021728515625, "grad_norm": 5.1400532722473145, "learning_rate": 3.07969478592624e-05, "loss": 0.607, "mean_token_accuracy": 0.8545138835906982, "num_tokens": 55227918.0, "step": 1454 }, { "epoch": 0.1850909553491922, "ewc_loss": 0.15234375, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 0.00014209747314453125, "grad_norm": 4.609557151794434, "learning_rate": 3.081814328105129e-05, "loss": 0.5977, "mean_token_accuracy": 0.852384090423584, "num_tokens": 55274312.0, "step": 1455 }, { "epoch": 0.18521816562778273, "ewc_loss": 0.1640625, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 0.0001544952392578125, "grad_norm": 5.0800323486328125, "learning_rate": 3.083933870284019e-05, "loss": 0.676, "mean_token_accuracy": 0.8320157527923584, "num_tokens": 55310583.0, "step": 1456 }, { "epoch": 0.18534537590637323, "ewc_loss": 0.1484375, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 0.00013828277587890625, "grad_norm": 4.564909934997559, "learning_rate": 3.086053412462908e-05, "loss": 0.5916, "mean_token_accuracy": 0.8509683012962341, "num_tokens": 55352365.0, "step": 1457 }, { "epoch": 0.18547258618496373, "ewc_loss": 0.1650390625, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 0.00015544891357421875, "grad_norm": 5.1269073486328125, "learning_rate": 3.088172954641798e-05, "loss": 0.6229, "mean_token_accuracy": 0.8470869064331055, "num_tokens": 55391268.0, "step": 1458 }, { "epoch": 0.18559979646355426, "ewc_loss": 0.140625, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 0.00013065338134765625, "grad_norm": 4.440632343292236, "learning_rate": 3.090292496820687e-05, "loss": 0.5752, "mean_token_accuracy": 0.8563125133514404, "num_tokens": 55428842.0, "step": 1459 }, { "epoch": 0.18572700674214476, "ewc_loss": 0.1728515625, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 0.00016307830810546875, "grad_norm": 5.281182289123535, "learning_rate": 3.092412038999577e-05, "loss": 0.5899, "mean_token_accuracy": 0.8599474430084229, "num_tokens": 55464341.0, "step": 1460 }, { "epoch": 0.18585421702073526, "ewc_loss": 0.13671875, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 0.00012683868408203125, "grad_norm": 4.331973552703857, "learning_rate": 3.094531581178466e-05, "loss": 0.6487, "mean_token_accuracy": 0.831285834312439, "num_tokens": 55498985.0, "step": 1461 }, { "epoch": 0.1859814272993258, "ewc_loss": 0.1826171875, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 0.00017261505126953125, "grad_norm": 5.4628424644470215, "learning_rate": 3.096651123357355e-05, "loss": 0.622, "mean_token_accuracy": 0.8536895513534546, "num_tokens": 55537762.0, "step": 1462 }, { "epoch": 0.1861086375779163, "ewc_loss": 0.138671875, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 0.00012874603271484375, "grad_norm": 4.271613597869873, "learning_rate": 3.098770665536244e-05, "loss": 0.6217, "mean_token_accuracy": 0.8394990563392639, "num_tokens": 55572571.0, "step": 1463 }, { "epoch": 0.1862358478565068, "ewc_loss": 0.2001953125, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.00019073486328125, "grad_norm": 5.674976825714111, "learning_rate": 3.100890207715134e-05, "loss": 0.6507, "mean_token_accuracy": 0.8520696759223938, "num_tokens": 55611858.0, "step": 1464 }, { "epoch": 0.18636305813509732, "ewc_loss": 0.1474609375, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.335870265960693, "learning_rate": 3.103009749894023e-05, "loss": 0.6334, "mean_token_accuracy": 0.8411059379577637, "num_tokens": 55649168.0, "step": 1465 }, { "epoch": 0.18649026841368782, "ewc_loss": 0.212890625, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.00020313262939453125, "grad_norm": 5.741753101348877, "learning_rate": 3.105129292072913e-05, "loss": 0.6434, "mean_token_accuracy": 0.8550106883049011, "num_tokens": 55686127.0, "step": 1466 }, { "epoch": 0.18661747869227835, "ewc_loss": 0.166015625, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.000156402587890625, "grad_norm": 4.607513904571533, "learning_rate": 3.107248834251802e-05, "loss": 0.6517, "mean_token_accuracy": 0.8371695280075073, "num_tokens": 55724062.0, "step": 1467 }, { "epoch": 0.18674468897086885, "ewc_loss": 0.19921875, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.00018978118896484375, "grad_norm": 5.495096206665039, "learning_rate": 3.1093683764306916e-05, "loss": 0.6425, "mean_token_accuracy": 0.8542071580886841, "num_tokens": 55752137.0, "step": 1468 }, { "epoch": 0.18687189924945935, "ewc_loss": 0.177734375, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.0001678466796875, "grad_norm": 4.9129180908203125, "learning_rate": 3.11148791860958e-05, "loss": 0.6631, "mean_token_accuracy": 0.8472980260848999, "num_tokens": 55793171.0, "step": 1469 }, { "epoch": 0.18699910952804988, "ewc_loss": 0.1640625, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.0001544952392578125, "grad_norm": 4.985983371734619, "learning_rate": 3.11360746078847e-05, "loss": 0.6709, "mean_token_accuracy": 0.8401369452476501, "num_tokens": 55831118.0, "step": 1470 }, { "epoch": 0.18712631980664038, "ewc_loss": 0.181640625, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.000171661376953125, "grad_norm": 5.165744304656982, "learning_rate": 3.115727002967359e-05, "loss": 0.5891, "mean_token_accuracy": 0.8645516633987427, "num_tokens": 55865568.0, "step": 1471 }, { "epoch": 0.18725353008523088, "ewc_loss": 0.1474609375, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.546192646026611, "learning_rate": 3.117846545146249e-05, "loss": 0.6702, "mean_token_accuracy": 0.8343382477760315, "num_tokens": 55904624.0, "step": 1472 }, { "epoch": 0.1873807403638214, "ewc_loss": 0.1845703125, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.00017452239990234375, "grad_norm": 5.368598937988281, "learning_rate": 3.119966087325138e-05, "loss": 0.5878, "mean_token_accuracy": 0.8661372065544128, "num_tokens": 55944714.0, "step": 1473 }, { "epoch": 0.1875079506424119, "ewc_loss": 0.140625, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.00013065338134765625, "grad_norm": 4.3697590827941895, "learning_rate": 3.1220856295040275e-05, "loss": 0.6274, "mean_token_accuracy": 0.837862491607666, "num_tokens": 55979444.0, "step": 1474 }, { "epoch": 0.1876351609210024, "ewc_loss": 0.1865234375, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.00017642974853515625, "grad_norm": 5.462007522583008, "learning_rate": 3.1242051716829167e-05, "loss": 0.6109, "mean_token_accuracy": 0.8596189022064209, "num_tokens": 56017171.0, "step": 1475 }, { "epoch": 0.18776237119959294, "ewc_loss": 0.146484375, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.39996862411499, "learning_rate": 3.1263247138618064e-05, "loss": 0.5886, "mean_token_accuracy": 0.8560538291931152, "num_tokens": 56060562.0, "step": 1476 }, { "epoch": 0.18788958147818344, "ewc_loss": 0.189453125, "ewc_loss_diag": 0.0001010894775390625, "ewc_loss_parallel": 0.000179290771484375, "grad_norm": 5.452930927276611, "learning_rate": 3.128444256040695e-05, "loss": 0.6282, "mean_token_accuracy": 0.8541263341903687, "num_tokens": 56098366.0, "step": 1477 }, { "epoch": 0.18801679175677394, "ewc_loss": 0.1513671875, "ewc_loss_diag": 0.0001010894775390625, "ewc_loss_parallel": 0.000141143798828125, "grad_norm": 4.5178961753845215, "learning_rate": 3.1305637982195846e-05, "loss": 0.6516, "mean_token_accuracy": 0.8333116769790649, "num_tokens": 56143323.0, "step": 1478 }, { "epoch": 0.18814400203536447, "ewc_loss": 0.1826171875, "ewc_loss_diag": 0.0001010894775390625, "ewc_loss_parallel": 0.00017261505126953125, "grad_norm": 5.352375030517578, "learning_rate": 3.132683340398474e-05, "loss": 0.5818, "mean_token_accuracy": 0.8641175031661987, "num_tokens": 56183846.0, "step": 1479 }, { "epoch": 0.18827121231395497, "ewc_loss": 0.154296875, "ewc_loss_diag": 0.0001010894775390625, "ewc_loss_parallel": 0.00014495849609375, "grad_norm": 4.644114017486572, "learning_rate": 3.1348028825773635e-05, "loss": 0.6086, "mean_token_accuracy": 0.8506661653518677, "num_tokens": 56227412.0, "step": 1480 }, { "epoch": 0.18839842259254547, "ewc_loss": 0.173828125, "ewc_loss_diag": 0.0001010894775390625, "ewc_loss_parallel": 0.000164031982421875, "grad_norm": 5.1972455978393555, "learning_rate": 3.1369224247562526e-05, "loss": 0.6309, "mean_token_accuracy": 0.8483233451843262, "num_tokens": 56267649.0, "step": 1481 }, { "epoch": 0.188525632871136, "ewc_loss": 0.1552734375, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 0.00014591217041015625, "grad_norm": 4.723962306976318, "learning_rate": 3.1390419669351424e-05, "loss": 0.5398, "mean_token_accuracy": 0.8706583976745605, "num_tokens": 56310802.0, "step": 1482 }, { "epoch": 0.1886528431497265, "ewc_loss": 0.1640625, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 0.0001544952392578125, "grad_norm": 5.064182758331299, "learning_rate": 3.1411615091140315e-05, "loss": 0.5865, "mean_token_accuracy": 0.8622627854347229, "num_tokens": 56345972.0, "step": 1483 }, { "epoch": 0.188780053428317, "ewc_loss": 0.1533203125, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 0.00014400482177734375, "grad_norm": 4.731993198394775, "learning_rate": 3.1432810512929206e-05, "loss": 0.658, "mean_token_accuracy": 0.8395031690597534, "num_tokens": 56391778.0, "step": 1484 }, { "epoch": 0.18890726370690752, "ewc_loss": 0.16015625, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 0.0001506805419921875, "grad_norm": 5.029357433319092, "learning_rate": 3.14540059347181e-05, "loss": 0.5506, "mean_token_accuracy": 0.866809070110321, "num_tokens": 56426306.0, "step": 1485 }, { "epoch": 0.18903447398549802, "ewc_loss": 0.146484375, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.652571678161621, "learning_rate": 3.1475201356506995e-05, "loss": 0.5426, "mean_token_accuracy": 0.8666164875030518, "num_tokens": 56461974.0, "step": 1486 }, { "epoch": 0.18916168426408853, "ewc_loss": 0.162109375, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 0.000152587890625, "grad_norm": 5.099569797515869, "learning_rate": 3.1496396778295886e-05, "loss": 0.6055, "mean_token_accuracy": 0.8536527156829834, "num_tokens": 56507205.0, "step": 1487 }, { "epoch": 0.18928889454267905, "ewc_loss": 0.142578125, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 0.00013256072998046875, "grad_norm": 4.536675453186035, "learning_rate": 3.1517592200084784e-05, "loss": 0.5946, "mean_token_accuracy": 0.8472459316253662, "num_tokens": 56542076.0, "step": 1488 }, { "epoch": 0.18941610482126955, "ewc_loss": 0.1669921875, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 0.00015735626220703125, "grad_norm": 5.251255989074707, "learning_rate": 3.1538787621873675e-05, "loss": 0.6141, "mean_token_accuracy": 0.8552531599998474, "num_tokens": 56575980.0, "step": 1489 }, { "epoch": 0.18954331509986008, "ewc_loss": 0.1337890625, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 0.0001239776611328125, "grad_norm": 4.378526210784912, "learning_rate": 3.155998304366257e-05, "loss": 0.6154, "mean_token_accuracy": 0.8467827439308167, "num_tokens": 56613410.0, "step": 1490 }, { "epoch": 0.18967052537845058, "ewc_loss": 0.181640625, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 0.000171661376953125, "grad_norm": 5.486497402191162, "learning_rate": 3.1581178465451464e-05, "loss": 0.6513, "mean_token_accuracy": 0.8437877297401428, "num_tokens": 56649105.0, "step": 1491 }, { "epoch": 0.18979773565704108, "ewc_loss": 0.1337890625, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 0.0001239776611328125, "grad_norm": 4.245108127593994, "learning_rate": 3.1602373887240355e-05, "loss": 0.6081, "mean_token_accuracy": 0.8411329388618469, "num_tokens": 56686049.0, "step": 1492 }, { "epoch": 0.1899249459356316, "ewc_loss": 0.203125, "ewc_loss_diag": 0.00010204315185546875, "ewc_loss_parallel": 0.00019359588623046875, "grad_norm": 5.778017520904541, "learning_rate": 3.162356930902925e-05, "loss": 0.6432, "mean_token_accuracy": 0.8523234724998474, "num_tokens": 56725659.0, "step": 1493 }, { "epoch": 0.1900521562142221, "ewc_loss": 0.1435546875, "ewc_loss_diag": 0.00010204315185546875, "ewc_loss_parallel": 0.000133514404296875, "grad_norm": 4.255666255950928, "learning_rate": 3.1644764730818144e-05, "loss": 0.6672, "mean_token_accuracy": 0.8331609964370728, "num_tokens": 56760612.0, "step": 1494 }, { "epoch": 0.1901793664928126, "ewc_loss": 0.2265625, "ewc_loss_diag": 0.00010204315185546875, "ewc_loss_parallel": 0.00021648406982421875, "grad_norm": 5.91934871673584, "learning_rate": 3.166596015260704e-05, "loss": 0.63, "mean_token_accuracy": 0.8622980117797852, "num_tokens": 56796081.0, "step": 1495 }, { "epoch": 0.19030657677140314, "ewc_loss": 0.16796875, "ewc_loss_diag": 0.00010204315185546875, "ewc_loss_parallel": 0.0001583099365234375, "grad_norm": 4.577234268188477, "learning_rate": 3.168715557439593e-05, "loss": 0.6218, "mean_token_accuracy": 0.8504738807678223, "num_tokens": 56833477.0, "step": 1496 }, { "epoch": 0.19043378704999364, "ewc_loss": 0.2119140625, "ewc_loss_diag": 0.00010204315185546875, "ewc_loss_parallel": 0.000202178955078125, "grad_norm": 5.658257007598877, "learning_rate": 3.1708350996184823e-05, "loss": 0.6506, "mean_token_accuracy": 0.8565695285797119, "num_tokens": 56870898.0, "step": 1497 }, { "epoch": 0.19056099732858414, "ewc_loss": 0.2001953125, "ewc_loss_diag": 0.00010204315185546875, "ewc_loss_parallel": 0.00019073486328125, "grad_norm": 5.033886909484863, "learning_rate": 3.172954641797372e-05, "loss": 0.608, "mean_token_accuracy": 0.8617748618125916, "num_tokens": 56916980.0, "step": 1498 }, { "epoch": 0.19068820760717467, "ewc_loss": 0.166015625, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 0.000156402587890625, "grad_norm": 4.9482879638671875, "learning_rate": 3.175074183976261e-05, "loss": 0.6255, "mean_token_accuracy": 0.842460572719574, "num_tokens": 56952564.0, "step": 1499 }, { "epoch": 0.19081541788576517, "ewc_loss": 0.208984375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 0.0001983642578125, "grad_norm": 5.546536445617676, "learning_rate": 3.17719372615515e-05, "loss": 0.6303, "mean_token_accuracy": 0.8558375835418701, "num_tokens": 56987381.0, "step": 1500 }, { "epoch": 0.19094262816435567, "ewc_loss": 0.140625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 0.00012969970703125, "grad_norm": 4.223855018615723, "learning_rate": 3.17931326833404e-05, "loss": 0.5647, "mean_token_accuracy": 0.8590877652168274, "num_tokens": 57023402.0, "step": 1501 }, { "epoch": 0.1910698384429462, "ewc_loss": 0.2431640625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 6.254749298095703, "learning_rate": 3.181432810512929e-05, "loss": 0.663, "mean_token_accuracy": 0.8595504760742188, "num_tokens": 57066464.0, "step": 1502 }, { "epoch": 0.1911970487215367, "ewc_loss": 0.1474609375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 0.00013637542724609375, "grad_norm": 4.100294589996338, "learning_rate": 3.183552352691819e-05, "loss": 0.5822, "mean_token_accuracy": 0.8549877405166626, "num_tokens": 57104495.0, "step": 1503 }, { "epoch": 0.1913242590001272, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.907550811767578, "learning_rate": 3.185671894870708e-05, "loss": 0.7457, "mean_token_accuracy": 0.8498137593269348, "num_tokens": 57146246.0, "step": 1504 }, { "epoch": 0.19145146927871773, "ewc_loss": 0.2451171875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 5.07260799407959, "learning_rate": 3.187791437049598e-05, "loss": 0.6873, "mean_token_accuracy": 0.8520628213882446, "num_tokens": 57184861.0, "step": 1505 }, { "epoch": 0.19157867955730823, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.665746212005615, "learning_rate": 3.189910979228487e-05, "loss": 0.7743, "mean_token_accuracy": 0.8335605263710022, "num_tokens": 57219395.0, "step": 1506 }, { "epoch": 0.19170588983589873, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 5.725413799285889, "learning_rate": 3.192030521407376e-05, "loss": 0.7588, "mean_token_accuracy": 0.8467361330986023, "num_tokens": 57255816.0, "step": 1507 }, { "epoch": 0.19183310011448926, "ewc_loss": 0.2216796875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.00021076202392578125, "grad_norm": 5.880666732788086, "learning_rate": 3.194150063586265e-05, "loss": 0.6813, "mean_token_accuracy": 0.8483511209487915, "num_tokens": 57296347.0, "step": 1508 }, { "epoch": 0.19196031039307976, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 5.731057643890381, "learning_rate": 3.196269605765155e-05, "loss": 0.7219, "mean_token_accuracy": 0.8442385792732239, "num_tokens": 57341789.0, "step": 1509 }, { "epoch": 0.19208752067167026, "ewc_loss": 0.1982421875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.00018787384033203125, "grad_norm": 5.281753063201904, "learning_rate": 3.198389147944044e-05, "loss": 0.6227, "mean_token_accuracy": 0.8594031929969788, "num_tokens": 57379399.0, "step": 1510 }, { "epoch": 0.1922147309502608, "ewc_loss": 0.220703125, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.000209808349609375, "grad_norm": 5.642792701721191, "learning_rate": 3.200508690122934e-05, "loss": 0.6853, "mean_token_accuracy": 0.8458696007728577, "num_tokens": 57418936.0, "step": 1511 }, { "epoch": 0.1923419412288513, "ewc_loss": 0.1904296875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.000179290771484375, "grad_norm": 5.042748928070068, "learning_rate": 3.202628232301823e-05, "loss": 0.6393, "mean_token_accuracy": 0.8485082387924194, "num_tokens": 57461772.0, "step": 1512 }, { "epoch": 0.1924691515074418, "ewc_loss": 0.1904296875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.000179290771484375, "grad_norm": 5.386663913726807, "learning_rate": 3.204747774480713e-05, "loss": 0.6189, "mean_token_accuracy": 0.8581494092941284, "num_tokens": 57498388.0, "step": 1513 }, { "epoch": 0.19259636178603232, "ewc_loss": 0.1806640625, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.0001697540283203125, "grad_norm": 5.063668727874756, "learning_rate": 3.206867316659602e-05, "loss": 0.6179, "mean_token_accuracy": 0.8532060384750366, "num_tokens": 57534123.0, "step": 1514 }, { "epoch": 0.19272357206462282, "ewc_loss": 0.1669921875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.000156402587890625, "grad_norm": 5.046595096588135, "learning_rate": 3.208986858838491e-05, "loss": 0.5814, "mean_token_accuracy": 0.8625710010528564, "num_tokens": 57568351.0, "step": 1515 }, { "epoch": 0.19285078234321335, "ewc_loss": 0.1748046875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.000164031982421875, "grad_norm": 5.1515913009643555, "learning_rate": 3.21110640101738e-05, "loss": 0.6599, "mean_token_accuracy": 0.8382376432418823, "num_tokens": 57617842.0, "step": 1516 }, { "epoch": 0.19297799262180385, "ewc_loss": 0.154296875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.00014400482177734375, "grad_norm": 4.809430122375488, "learning_rate": 3.21322594319627e-05, "loss": 0.6556, "mean_token_accuracy": 0.838041365146637, "num_tokens": 57657658.0, "step": 1517 }, { "epoch": 0.19310520290039435, "ewc_loss": 0.17578125, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.00016498565673828125, "grad_norm": 5.259397029876709, "learning_rate": 3.215345485375159e-05, "loss": 0.6018, "mean_token_accuracy": 0.8565264344215393, "num_tokens": 57695964.0, "step": 1518 }, { "epoch": 0.19323241317898487, "ewc_loss": 0.150390625, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.0001392364501953125, "grad_norm": 4.643581390380859, "learning_rate": 3.217465027554049e-05, "loss": 0.6243, "mean_token_accuracy": 0.8426187038421631, "num_tokens": 57725722.0, "step": 1519 }, { "epoch": 0.19335962345757537, "ewc_loss": 0.1796875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.00016880035400390625, "grad_norm": 5.423376560211182, "learning_rate": 3.219584569732938e-05, "loss": 0.6673, "mean_token_accuracy": 0.8395867347717285, "num_tokens": 57761944.0, "step": 1520 }, { "epoch": 0.19348683373616588, "ewc_loss": 0.1435546875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.00013256072998046875, "grad_norm": 4.471909523010254, "learning_rate": 3.2217041119118276e-05, "loss": 0.5887, "mean_token_accuracy": 0.8504077196121216, "num_tokens": 57796853.0, "step": 1521 }, { "epoch": 0.1936140440147564, "ewc_loss": 0.1953125, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.0001850128173828125, "grad_norm": 5.642326354980469, "learning_rate": 3.223823654090716e-05, "loss": 0.6447, "mean_token_accuracy": 0.8510512709617615, "num_tokens": 57835295.0, "step": 1522 }, { "epoch": 0.1937412542933469, "ewc_loss": 0.1435546875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 0.00013256072998046875, "grad_norm": 4.349915027618408, "learning_rate": 3.225943196269606e-05, "loss": 0.569, "mean_token_accuracy": 0.8579360246658325, "num_tokens": 57876294.0, "step": 1523 }, { "epoch": 0.1938684645719374, "ewc_loss": 0.212890625, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 0.000202178955078125, "grad_norm": 5.928412437438965, "learning_rate": 3.228062738448495e-05, "loss": 0.6692, "mean_token_accuracy": 0.8505972623825073, "num_tokens": 57915332.0, "step": 1524 }, { "epoch": 0.19399567485052793, "ewc_loss": 0.15234375, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 0.000141143798828125, "grad_norm": 4.329560279846191, "learning_rate": 3.230182280627385e-05, "loss": 0.5642, "mean_token_accuracy": 0.8646947741508484, "num_tokens": 57955624.0, "step": 1525 }, { "epoch": 0.19412288512911843, "ewc_loss": 0.2421875, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 6.196897029876709, "learning_rate": 3.232301822806274e-05, "loss": 0.6504, "mean_token_accuracy": 0.8626394867897034, "num_tokens": 57993692.0, "step": 1526 }, { "epoch": 0.19425009540770893, "ewc_loss": 0.177734375, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 0.00016689300537109375, "grad_norm": 4.60594367980957, "learning_rate": 3.2344213649851636e-05, "loss": 0.6243, "mean_token_accuracy": 0.8564999103546143, "num_tokens": 58035813.0, "step": 1527 }, { "epoch": 0.19437730568629946, "ewc_loss": 0.25, "ewc_loss_diag": 0.0001068115234375, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.168057918548584, "learning_rate": 3.236540907164053e-05, "loss": 0.7361, "mean_token_accuracy": 0.8466167449951172, "num_tokens": 58077118.0, "step": 1528 }, { "epoch": 0.19450451596488996, "ewc_loss": 0.21875, "ewc_loss_diag": 0.0001068115234375, "ewc_loss_parallel": 0.0002079010009765625, "grad_norm": 5.049420356750488, "learning_rate": 3.2386604493429425e-05, "loss": 0.6791, "mean_token_accuracy": 0.8488754034042358, "num_tokens": 58120853.0, "step": 1529 }, { "epoch": 0.19463172624348046, "ewc_loss": 0.2177734375, "ewc_loss_diag": 0.0001068115234375, "ewc_loss_parallel": 0.00020694732666015625, "grad_norm": 5.688875198364258, "learning_rate": 3.240779991521831e-05, "loss": 0.69, "mean_token_accuracy": 0.8404510021209717, "num_tokens": 58160279.0, "step": 1530 }, { "epoch": 0.194758936522071, "ewc_loss": 0.2265625, "ewc_loss_diag": 0.0001068115234375, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 5.351737976074219, "learning_rate": 3.242899533700721e-05, "loss": 0.6497, "mean_token_accuracy": 0.8600703477859497, "num_tokens": 58200902.0, "step": 1531 }, { "epoch": 0.1948861468006615, "ewc_loss": 0.17578125, "ewc_loss_diag": 0.0001068115234375, "ewc_loss_parallel": 0.00016498565673828125, "grad_norm": 5.0521979331970215, "learning_rate": 3.24501907587961e-05, "loss": 0.5629, "mean_token_accuracy": 0.8713420033454895, "num_tokens": 58237179.0, "step": 1532 }, { "epoch": 0.195013357079252, "ewc_loss": 0.21875, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 0.0002079010009765625, "grad_norm": 5.5671515464782715, "learning_rate": 3.2471386180584996e-05, "loss": 0.6107, "mean_token_accuracy": 0.8670693635940552, "num_tokens": 58271072.0, "step": 1533 }, { "epoch": 0.19514056735784252, "ewc_loss": 0.1572265625, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 0.0001468658447265625, "grad_norm": 4.654187202453613, "learning_rate": 3.249258160237389e-05, "loss": 0.6944, "mean_token_accuracy": 0.8275816440582275, "num_tokens": 58317186.0, "step": 1534 }, { "epoch": 0.19526777763643302, "ewc_loss": 0.212890625, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 0.000202178955078125, "grad_norm": 5.712297439575195, "learning_rate": 3.2513777024162784e-05, "loss": 0.6522, "mean_token_accuracy": 0.8509997129440308, "num_tokens": 58353396.0, "step": 1535 }, { "epoch": 0.19539498791502352, "ewc_loss": 0.15625, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 0.00014591217041015625, "grad_norm": 4.587990760803223, "learning_rate": 3.2534972445951675e-05, "loss": 0.6715, "mean_token_accuracy": 0.8377804756164551, "num_tokens": 58392466.0, "step": 1536 }, { "epoch": 0.19552219819361405, "ewc_loss": 0.212890625, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 0.000202178955078125, "grad_norm": 5.710232734680176, "learning_rate": 3.255616786774057e-05, "loss": 0.6138, "mean_token_accuracy": 0.8648163676261902, "num_tokens": 58432745.0, "step": 1537 }, { "epoch": 0.19564940847220455, "ewc_loss": 0.1650390625, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 0.0001544952392578125, "grad_norm": 4.6952385902404785, "learning_rate": 3.2577363289529464e-05, "loss": 0.6146, "mean_token_accuracy": 0.8501218557357788, "num_tokens": 58464288.0, "step": 1538 }, { "epoch": 0.19577661875079505, "ewc_loss": 0.2080078125, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 0.00019741058349609375, "grad_norm": 5.625980377197266, "learning_rate": 3.2598558711318355e-05, "loss": 0.7127, "mean_token_accuracy": 0.8359763622283936, "num_tokens": 58501232.0, "step": 1539 }, { "epoch": 0.19590382902938558, "ewc_loss": 0.16796875, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 0.00015735626220703125, "grad_norm": 4.788624286651611, "learning_rate": 3.261975413310725e-05, "loss": 0.6696, "mean_token_accuracy": 0.8342393040657043, "num_tokens": 58539749.0, "step": 1540 }, { "epoch": 0.19603103930797608, "ewc_loss": 0.197265625, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.000186920166015625, "grad_norm": 5.504786014556885, "learning_rate": 3.2640949554896144e-05, "loss": 0.6286, "mean_token_accuracy": 0.8589062690734863, "num_tokens": 58576436.0, "step": 1541 }, { "epoch": 0.1961582495865666, "ewc_loss": 0.1689453125, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.0001583099365234375, "grad_norm": 4.836124420166016, "learning_rate": 3.2662144976685035e-05, "loss": 0.6114, "mean_token_accuracy": 0.853061318397522, "num_tokens": 58618677.0, "step": 1542 }, { "epoch": 0.1962854598651571, "ewc_loss": 0.193359375, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.00018215179443359375, "grad_norm": 5.415135860443115, "learning_rate": 3.268334039847393e-05, "loss": 0.6534, "mean_token_accuracy": 0.846261203289032, "num_tokens": 58654404.0, "step": 1543 }, { "epoch": 0.1964126701437476, "ewc_loss": 0.1650390625, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.0001544952392578125, "grad_norm": 4.827336311340332, "learning_rate": 3.2704535820262824e-05, "loss": 0.624, "mean_token_accuracy": 0.8473212718963623, "num_tokens": 58697312.0, "step": 1544 }, { "epoch": 0.19653988042233814, "ewc_loss": 0.1865234375, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.00017547607421875, "grad_norm": 5.376216888427734, "learning_rate": 3.2725731242051715e-05, "loss": 0.6531, "mean_token_accuracy": 0.8454279899597168, "num_tokens": 58738940.0, "step": 1545 }, { "epoch": 0.19666709070092864, "ewc_loss": 0.158203125, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.00014781951904296875, "grad_norm": 4.7560296058654785, "learning_rate": 3.274692666384061e-05, "loss": 0.6293, "mean_token_accuracy": 0.846611499786377, "num_tokens": 58779626.0, "step": 1546 }, { "epoch": 0.19679430097951914, "ewc_loss": 0.189453125, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.00017833709716796875, "grad_norm": 5.437037944793701, "learning_rate": 3.2768122085629504e-05, "loss": 0.6607, "mean_token_accuracy": 0.8437200784683228, "num_tokens": 58818938.0, "step": 1547 }, { "epoch": 0.19692151125810967, "ewc_loss": 0.1533203125, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.0001430511474609375, "grad_norm": 4.670226097106934, "learning_rate": 3.27893175074184e-05, "loss": 0.6385, "mean_token_accuracy": 0.8376206755638123, "num_tokens": 58855973.0, "step": 1548 }, { "epoch": 0.19704872153670017, "ewc_loss": 0.193359375, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.00018215179443359375, "grad_norm": 5.535055637359619, "learning_rate": 3.281051292920729e-05, "loss": 0.6222, "mean_token_accuracy": 0.8546551465988159, "num_tokens": 58894624.0, "step": 1549 }, { "epoch": 0.19717593181529067, "ewc_loss": 0.1533203125, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.00014209747314453125, "grad_norm": 4.591221332550049, "learning_rate": 3.283170835099619e-05, "loss": 0.5881, "mean_token_accuracy": 0.8564531803131104, "num_tokens": 58934440.0, "step": 1550 }, { "epoch": 0.1973031420938812, "ewc_loss": 0.201171875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.00019073486328125, "grad_norm": 5.649736404418945, "learning_rate": 3.285290377278508e-05, "loss": 0.641, "mean_token_accuracy": 0.8532464504241943, "num_tokens": 58972150.0, "step": 1551 }, { "epoch": 0.1974303523724717, "ewc_loss": 0.1533203125, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.0001430511474609375, "grad_norm": 4.583000659942627, "learning_rate": 3.287409919457398e-05, "loss": 0.5704, "mean_token_accuracy": 0.859078049659729, "num_tokens": 59006179.0, "step": 1552 }, { "epoch": 0.1975575626510622, "ewc_loss": 0.20703125, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.0001964569091796875, "grad_norm": 5.712006092071533, "learning_rate": 3.2895294616362864e-05, "loss": 0.6744, "mean_token_accuracy": 0.8447118401527405, "num_tokens": 59046583.0, "step": 1553 }, { "epoch": 0.19768477292965272, "ewc_loss": 0.1611328125, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.0001506805419921875, "grad_norm": 4.684953212738037, "learning_rate": 3.291649003815176e-05, "loss": 0.6214, "mean_token_accuracy": 0.8475483655929565, "num_tokens": 59078023.0, "step": 1554 }, { "epoch": 0.19781198320824323, "ewc_loss": 0.203125, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.0001926422119140625, "grad_norm": 5.592868328094482, "learning_rate": 3.293768545994065e-05, "loss": 0.6703, "mean_token_accuracy": 0.8448238968849182, "num_tokens": 59121160.0, "step": 1555 }, { "epoch": 0.19793919348683373, "ewc_loss": 0.1728515625, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.0001621246337890625, "grad_norm": 4.893041133880615, "learning_rate": 3.295888088172955e-05, "loss": 0.608, "mean_token_accuracy": 0.8521961569786072, "num_tokens": 59164564.0, "step": 1556 }, { "epoch": 0.19806640376542425, "ewc_loss": 0.1796875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.00016880035400390625, "grad_norm": 5.274825572967529, "learning_rate": 3.298007630351844e-05, "loss": 0.6163, "mean_token_accuracy": 0.8578479290008545, "num_tokens": 59211558.0, "step": 1557 }, { "epoch": 0.19819361404401475, "ewc_loss": 0.1796875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.00016880035400390625, "grad_norm": 5.155818462371826, "learning_rate": 3.300127172530734e-05, "loss": 0.6156, "mean_token_accuracy": 0.8556087017059326, "num_tokens": 59251260.0, "step": 1558 }, { "epoch": 0.19832082432260525, "ewc_loss": 0.15625, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.00014591217041015625, "grad_norm": 4.864866733551025, "learning_rate": 3.302246714709623e-05, "loss": 0.6504, "mean_token_accuracy": 0.8386152386665344, "num_tokens": 59293949.0, "step": 1559 }, { "epoch": 0.19844803460119578, "ewc_loss": 0.1884765625, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.0001773834228515625, "grad_norm": 5.436628818511963, "learning_rate": 3.304366256888512e-05, "loss": 0.6949, "mean_token_accuracy": 0.8430375456809998, "num_tokens": 59341549.0, "step": 1560 }, { "epoch": 0.19857524487978628, "ewc_loss": 0.1435546875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 0.00013256072998046875, "grad_norm": 4.54406213760376, "learning_rate": 3.306485799067401e-05, "loss": 0.5911, "mean_token_accuracy": 0.8499315977096558, "num_tokens": 59379355.0, "step": 1561 }, { "epoch": 0.19870245515837678, "ewc_loss": 0.1982421875, "ewc_loss_diag": 0.0001087188720703125, "ewc_loss_parallel": 0.00018787384033203125, "grad_norm": 5.72485876083374, "learning_rate": 3.308605341246291e-05, "loss": 0.6879, "mean_token_accuracy": 0.8374260663986206, "num_tokens": 59416340.0, "step": 1562 }, { "epoch": 0.1988296654369673, "ewc_loss": 0.140625, "ewc_loss_diag": 0.0001087188720703125, "ewc_loss_parallel": 0.00012969970703125, "grad_norm": 4.345405578613281, "learning_rate": 3.31072488342518e-05, "loss": 0.5606, "mean_token_accuracy": 0.8599839806556702, "num_tokens": 59453100.0, "step": 1563 }, { "epoch": 0.1989568757155578, "ewc_loss": 0.2255859375, "ewc_loss_diag": 0.0001087188720703125, "ewc_loss_parallel": 0.00021457672119140625, "grad_norm": 6.052728652954102, "learning_rate": 3.31284442560407e-05, "loss": 0.6866, "mean_token_accuracy": 0.8457680344581604, "num_tokens": 59491164.0, "step": 1564 }, { "epoch": 0.19908408599414834, "ewc_loss": 0.1484375, "ewc_loss_diag": 0.00010919570922851562, "ewc_loss_parallel": 0.0001373291015625, "grad_norm": 4.346738815307617, "learning_rate": 3.314963967782959e-05, "loss": 0.6176, "mean_token_accuracy": 0.8456416130065918, "num_tokens": 59525752.0, "step": 1565 }, { "epoch": 0.19921129627273884, "ewc_loss": 0.251953125, "ewc_loss_diag": 0.00010919570922851562, "ewc_loss_parallel": 0.00024127960205078125, "grad_norm": 6.2804789543151855, "learning_rate": 3.317083509961849e-05, "loss": 0.7575, "mean_token_accuracy": 0.8396817445755005, "num_tokens": 59565206.0, "step": 1566 }, { "epoch": 0.19933850655132934, "ewc_loss": 0.1865234375, "ewc_loss_diag": 0.00010919570922851562, "ewc_loss_parallel": 0.00017547607421875, "grad_norm": 4.710386276245117, "learning_rate": 3.319203052140738e-05, "loss": 0.6298, "mean_token_accuracy": 0.8523372411727905, "num_tokens": 59597578.0, "step": 1567 }, { "epoch": 0.19946571682991987, "ewc_loss": 0.248046875, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00023746490478515625, "grad_norm": 6.0769124031066895, "learning_rate": 3.321322594319627e-05, "loss": 0.6674, "mean_token_accuracy": 0.8587669134140015, "num_tokens": 59640073.0, "step": 1568 }, { "epoch": 0.19959292710851037, "ewc_loss": 0.2294921875, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00021839141845703125, "grad_norm": 5.2156596183776855, "learning_rate": 3.323442136498516e-05, "loss": 0.6989, "mean_token_accuracy": 0.842710018157959, "num_tokens": 59680259.0, "step": 1569 }, { "epoch": 0.19972013738710087, "ewc_loss": 0.2001953125, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00018978118896484375, "grad_norm": 5.398242950439453, "learning_rate": 3.325561678677406e-05, "loss": 0.6597, "mean_token_accuracy": 0.8463625311851501, "num_tokens": 59716372.0, "step": 1570 }, { "epoch": 0.1998473476656914, "ewc_loss": 0.234375, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00022411346435546875, "grad_norm": 5.617259979248047, "learning_rate": 3.327681220856295e-05, "loss": 0.7095, "mean_token_accuracy": 0.8416389226913452, "num_tokens": 59751294.0, "step": 1571 }, { "epoch": 0.1999745579442819, "ewc_loss": 0.16015625, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00014972686767578125, "grad_norm": 4.755255222320557, "learning_rate": 3.329800763035185e-05, "loss": 0.6689, "mean_token_accuracy": 0.83855801820755, "num_tokens": 59794395.0, "step": 1572 }, { "epoch": 0.2001017682228724, "ewc_loss": 0.2314453125, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00022029876708984375, "grad_norm": 5.946971893310547, "learning_rate": 3.331920305214074e-05, "loss": 0.7651, "mean_token_accuracy": 0.8317310810089111, "num_tokens": 59825503.0, "step": 1573 }, { "epoch": 0.20022897850146293, "ewc_loss": 0.158203125, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00014781951904296875, "grad_norm": 4.561115741729736, "learning_rate": 3.3340398473929636e-05, "loss": 0.5781, "mean_token_accuracy": 0.8609017133712769, "num_tokens": 59864634.0, "step": 1574 }, { "epoch": 0.20035618878005343, "ewc_loss": 0.2353515625, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.000225067138671875, "grad_norm": 6.025915622711182, "learning_rate": 3.336159389571852e-05, "loss": 0.6774, "mean_token_accuracy": 0.8507558703422546, "num_tokens": 59898693.0, "step": 1575 }, { "epoch": 0.20048339905864393, "ewc_loss": 0.1796875, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00016880035400390625, "grad_norm": 4.7974066734313965, "learning_rate": 3.338278931750742e-05, "loss": 0.6422, "mean_token_accuracy": 0.8476514220237732, "num_tokens": 59930234.0, "step": 1576 }, { "epoch": 0.20061060933723446, "ewc_loss": 0.2236328125, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00021266937255859375, "grad_norm": 5.742315769195557, "learning_rate": 3.340398473929631e-05, "loss": 0.6787, "mean_token_accuracy": 0.8491411209106445, "num_tokens": 59970932.0, "step": 1577 }, { "epoch": 0.20073781961582496, "ewc_loss": 0.1962890625, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00018596649169921875, "grad_norm": 5.088613986968994, "learning_rate": 3.342518016108521e-05, "loss": 0.6486, "mean_token_accuracy": 0.848996639251709, "num_tokens": 60007282.0, "step": 1578 }, { "epoch": 0.20086502989441546, "ewc_loss": 0.193359375, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00018215179443359375, "grad_norm": 5.326603889465332, "learning_rate": 3.34463755828741e-05, "loss": 0.6268, "mean_token_accuracy": 0.8581923246383667, "num_tokens": 60043336.0, "step": 1579 }, { "epoch": 0.200992240173006, "ewc_loss": 0.1962890625, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00018596649169921875, "grad_norm": 5.259820938110352, "learning_rate": 3.3467571004662996e-05, "loss": 0.6572, "mean_token_accuracy": 0.8506444692611694, "num_tokens": 60083284.0, "step": 1580 }, { "epoch": 0.2011194504515965, "ewc_loss": 0.171875, "ewc_loss_diag": 0.00010967254638671875, "ewc_loss_parallel": 0.00016117095947265625, "grad_norm": 5.081647872924805, "learning_rate": 3.348876642645189e-05, "loss": 0.6314, "mean_token_accuracy": 0.8462169170379639, "num_tokens": 60125115.0, "step": 1581 }, { "epoch": 0.201246660730187, "ewc_loss": 0.181640625, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 0.00017070770263671875, "grad_norm": 5.247406959533691, "learning_rate": 3.3509961848240785e-05, "loss": 0.5601, "mean_token_accuracy": 0.8719294667243958, "num_tokens": 60161497.0, "step": 1582 }, { "epoch": 0.20137387100877752, "ewc_loss": 0.166015625, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 0.00015544891357421875, "grad_norm": 5.060825824737549, "learning_rate": 3.3531157270029676e-05, "loss": 0.5744, "mean_token_accuracy": 0.8633401393890381, "num_tokens": 60202282.0, "step": 1583 }, { "epoch": 0.20150108128736802, "ewc_loss": 0.16796875, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 0.00015735626220703125, "grad_norm": 5.116189002990723, "learning_rate": 3.355235269181857e-05, "loss": 0.6072, "mean_token_accuracy": 0.8527855277061462, "num_tokens": 60241211.0, "step": 1584 }, { "epoch": 0.20162829156595852, "ewc_loss": 0.1748046875, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 0.000164031982421875, "grad_norm": 5.172457695007324, "learning_rate": 3.357354811360746e-05, "loss": 0.6579, "mean_token_accuracy": 0.8421551585197449, "num_tokens": 60273677.0, "step": 1585 }, { "epoch": 0.20175550184454905, "ewc_loss": 0.1591796875, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 0.000148773193359375, "grad_norm": 4.929821968078613, "learning_rate": 3.3594743535396356e-05, "loss": 0.5954, "mean_token_accuracy": 0.8550059795379639, "num_tokens": 60313904.0, "step": 1586 }, { "epoch": 0.20188271212313955, "ewc_loss": 0.1748046875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 0.000164031982421875, "grad_norm": 5.31424617767334, "learning_rate": 3.361593895718525e-05, "loss": 0.611, "mean_token_accuracy": 0.8547654151916504, "num_tokens": 60351534.0, "step": 1587 }, { "epoch": 0.20200992240173005, "ewc_loss": 0.150390625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 0.0001392364501953125, "grad_norm": 4.7520060539245605, "learning_rate": 3.3637134378974145e-05, "loss": 0.631, "mean_token_accuracy": 0.8442800641059875, "num_tokens": 60393362.0, "step": 1588 }, { "epoch": 0.20213713268032057, "ewc_loss": 0.1845703125, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 0.0001735687255859375, "grad_norm": 5.518918991088867, "learning_rate": 3.3658329800763036e-05, "loss": 0.6357, "mean_token_accuracy": 0.8468074798583984, "num_tokens": 60437735.0, "step": 1589 }, { "epoch": 0.20226434295891108, "ewc_loss": 0.1416015625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 0.00013065338134765625, "grad_norm": 4.571838855743408, "learning_rate": 3.3679525222551934e-05, "loss": 0.6516, "mean_token_accuracy": 0.8287991285324097, "num_tokens": 60476032.0, "step": 1590 }, { "epoch": 0.2023915532375016, "ewc_loss": 0.19921875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 0.0001888275146484375, "grad_norm": 5.785215377807617, "learning_rate": 3.3700720644340825e-05, "loss": 0.6638, "mean_token_accuracy": 0.8477280735969543, "num_tokens": 60513330.0, "step": 1591 }, { "epoch": 0.2025187635160921, "ewc_loss": 0.1376953125, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 0.00012683868408203125, "grad_norm": 4.346299648284912, "learning_rate": 3.3721916066129716e-05, "loss": 0.6289, "mean_token_accuracy": 0.8372423648834229, "num_tokens": 60552829.0, "step": 1592 }, { "epoch": 0.2026459737946826, "ewc_loss": 0.23046875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 0.0002193450927734375, "grad_norm": 6.223150253295898, "learning_rate": 3.3743111487918614e-05, "loss": 0.6554, "mean_token_accuracy": 0.85325026512146, "num_tokens": 60585333.0, "step": 1593 }, { "epoch": 0.20277318407327313, "ewc_loss": 0.1455078125, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 0.00013446807861328125, "grad_norm": 4.219889163970947, "learning_rate": 3.3764306909707505e-05, "loss": 0.5983, "mean_token_accuracy": 0.8468567132949829, "num_tokens": 60621091.0, "step": 1594 }, { "epoch": 0.20290039435186363, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00011157989501953125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.7504448890686035, "learning_rate": 3.37855023314964e-05, "loss": 0.7404, "mean_token_accuracy": 0.8506834506988525, "num_tokens": 60661602.0, "step": 1595 }, { "epoch": 0.20302760463045413, "ewc_loss": 0.193359375, "ewc_loss_diag": 0.00011157989501953125, "ewc_loss_parallel": 0.00018215179443359375, "grad_norm": 4.615419387817383, "learning_rate": 3.3806697753285293e-05, "loss": 0.6571, "mean_token_accuracy": 0.8435426950454712, "num_tokens": 60689085.0, "step": 1596 }, { "epoch": 0.20315481490904466, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00011157989501953125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.904338836669922, "learning_rate": 3.382789317507419e-05, "loss": 0.7628, "mean_token_accuracy": 0.85038161277771, "num_tokens": 60727259.0, "step": 1597 }, { "epoch": 0.20328202518763516, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00011157989501953125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 5.4481635093688965, "learning_rate": 3.3849088596863076e-05, "loss": 0.7016, "mean_token_accuracy": 0.8562861680984497, "num_tokens": 60761880.0, "step": 1598 }, { "epoch": 0.20340923546622566, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.00011157989501953125, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.255281925201416, "learning_rate": 3.387028401865197e-05, "loss": 0.7331, "mean_token_accuracy": 0.8447163701057434, "num_tokens": 60795416.0, "step": 1599 }, { "epoch": 0.2035364457448162, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00011157989501953125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 5.921863555908203, "learning_rate": 3.3891479440440864e-05, "loss": 0.7442, "mean_token_accuracy": 0.8517361283302307, "num_tokens": 60833373.0, "step": 1600 }, { "epoch": 0.2036636560234067, "ewc_loss": 0.1953125, "ewc_loss_diag": 0.00011157989501953125, "ewc_loss_parallel": 0.0001850128173828125, "grad_norm": 5.264523506164551, "learning_rate": 3.391267486222976e-05, "loss": 0.6606, "mean_token_accuracy": 0.8439954519271851, "num_tokens": 60870024.0, "step": 1601 }, { "epoch": 0.2037908663019972, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.347452640533447, "learning_rate": 3.393387028401865e-05, "loss": 0.739, "mean_token_accuracy": 0.8493367433547974, "num_tokens": 60909790.0, "step": 1602 }, { "epoch": 0.20391807658058772, "ewc_loss": 0.18359375, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 0.000171661376953125, "grad_norm": 4.735749244689941, "learning_rate": 3.395506570580755e-05, "loss": 0.6507, "mean_token_accuracy": 0.8448677062988281, "num_tokens": 60948793.0, "step": 1603 }, { "epoch": 0.20404528685917822, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.801462650299072, "learning_rate": 3.397626112759644e-05, "loss": 0.7876, "mean_token_accuracy": 0.836891770362854, "num_tokens": 60991370.0, "step": 1604 }, { "epoch": 0.20417249713776872, "ewc_loss": 0.234375, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 5.033227443695068, "learning_rate": 3.399745654938534e-05, "loss": 0.6939, "mean_token_accuracy": 0.8464521765708923, "num_tokens": 61028137.0, "step": 1605 }, { "epoch": 0.20429970741635925, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.783636569976807, "learning_rate": 3.4018651971174224e-05, "loss": 0.7135, "mean_token_accuracy": 0.8607580065727234, "num_tokens": 61060843.0, "step": 1606 }, { "epoch": 0.20442691769494975, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 5.572187423706055, "learning_rate": 3.403984739296312e-05, "loss": 0.7428, "mean_token_accuracy": 0.8499886989593506, "num_tokens": 61096272.0, "step": 1607 }, { "epoch": 0.20455412797354025, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.240300178527832, "learning_rate": 3.406104281475201e-05, "loss": 0.7357, "mean_token_accuracy": 0.8485503792762756, "num_tokens": 61139119.0, "step": 1608 }, { "epoch": 0.20468133825213078, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 5.698833465576172, "learning_rate": 3.408223823654091e-05, "loss": 0.7854, "mean_token_accuracy": 0.8316681981086731, "num_tokens": 61183327.0, "step": 1609 }, { "epoch": 0.20480854853072128, "ewc_loss": 0.220703125, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 0.00020885467529296875, "grad_norm": 5.684882164001465, "learning_rate": 3.41034336583298e-05, "loss": 0.6714, "mean_token_accuracy": 0.8506940603256226, "num_tokens": 61218787.0, "step": 1610 }, { "epoch": 0.20493575880931178, "ewc_loss": 0.23828125, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 0.0002269744873046875, "grad_norm": 5.625555038452148, "learning_rate": 3.41246290801187e-05, "loss": 0.694, "mean_token_accuracy": 0.8494521379470825, "num_tokens": 61256517.0, "step": 1611 }, { "epoch": 0.2050629690879023, "ewc_loss": 0.203125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 0.00019168853759765625, "grad_norm": 5.399820327758789, "learning_rate": 3.414582450190759e-05, "loss": 0.6496, "mean_token_accuracy": 0.8523356914520264, "num_tokens": 61301065.0, "step": 1612 }, { "epoch": 0.2051901793664928, "ewc_loss": 0.2109375, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 0.00019931793212890625, "grad_norm": 5.442714691162109, "learning_rate": 3.416701992369648e-05, "loss": 0.6574, "mean_token_accuracy": 0.8502814769744873, "num_tokens": 61336368.0, "step": 1613 }, { "epoch": 0.2053173896450833, "ewc_loss": 0.1953125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 0.00018310546875, "grad_norm": 5.334085941314697, "learning_rate": 3.418821534548537e-05, "loss": 0.7238, "mean_token_accuracy": 0.8278152942657471, "num_tokens": 61373742.0, "step": 1614 }, { "epoch": 0.20544459992367384, "ewc_loss": 0.1845703125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 0.00017261505126953125, "grad_norm": 5.2729926109313965, "learning_rate": 3.420941076727427e-05, "loss": 0.6195, "mean_token_accuracy": 0.854040801525116, "num_tokens": 61408454.0, "step": 1615 }, { "epoch": 0.20557181020226434, "ewc_loss": 0.185546875, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 0.0001735687255859375, "grad_norm": 5.293575286865234, "learning_rate": 3.423060618906316e-05, "loss": 0.588, "mean_token_accuracy": 0.8650997281074524, "num_tokens": 61445420.0, "step": 1616 }, { "epoch": 0.20569902048085487, "ewc_loss": 0.169921875, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 0.0001583099365234375, "grad_norm": 5.1667585372924805, "learning_rate": 3.425180161085206e-05, "loss": 0.6091, "mean_token_accuracy": 0.85433030128479, "num_tokens": 61487607.0, "step": 1617 }, { "epoch": 0.20582623075944537, "ewc_loss": 0.181640625, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 0.0001697540283203125, "grad_norm": 5.318074703216553, "learning_rate": 3.427299703264095e-05, "loss": 0.6248, "mean_token_accuracy": 0.8521291017532349, "num_tokens": 61521053.0, "step": 1618 }, { "epoch": 0.20595344103803587, "ewc_loss": 0.1640625, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 0.000152587890625, "grad_norm": 5.037691116333008, "learning_rate": 3.429419245442985e-05, "loss": 0.6667, "mean_token_accuracy": 0.8334150910377502, "num_tokens": 61562211.0, "step": 1619 }, { "epoch": 0.2060806513166264, "ewc_loss": 0.1826171875, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 0.00017070770263671875, "grad_norm": 5.428936958312988, "learning_rate": 3.431538787621874e-05, "loss": 0.6966, "mean_token_accuracy": 0.8362135291099548, "num_tokens": 61600089.0, "step": 1620 }, { "epoch": 0.2062078615952169, "ewc_loss": 0.158203125, "ewc_loss_diag": 0.00011348724365234375, "ewc_loss_parallel": 0.0001468658447265625, "grad_norm": 4.868630886077881, "learning_rate": 3.433658329800763e-05, "loss": 0.5906, "mean_token_accuracy": 0.8536951541900635, "num_tokens": 61636226.0, "step": 1621 }, { "epoch": 0.2063350718738074, "ewc_loss": 0.189453125, "ewc_loss_diag": 0.00011348724365234375, "ewc_loss_parallel": 0.0001773834228515625, "grad_norm": 5.599445343017578, "learning_rate": 3.435777871979652e-05, "loss": 0.6524, "mean_token_accuracy": 0.8486962914466858, "num_tokens": 61677224.0, "step": 1622 }, { "epoch": 0.20646228215239792, "ewc_loss": 0.150390625, "ewc_loss_diag": 0.00011348724365234375, "ewc_loss_parallel": 0.00013828277587890625, "grad_norm": 4.688699722290039, "learning_rate": 3.437897414158542e-05, "loss": 0.5906, "mean_token_accuracy": 0.855236828327179, "num_tokens": 61715174.0, "step": 1623 }, { "epoch": 0.20658949243098843, "ewc_loss": 0.20703125, "ewc_loss_diag": 0.00011348724365234375, "ewc_loss_parallel": 0.00019550323486328125, "grad_norm": 5.866634845733643, "learning_rate": 3.440016956337431e-05, "loss": 0.6337, "mean_token_accuracy": 0.8582558631896973, "num_tokens": 61760819.0, "step": 1624 }, { "epoch": 0.20671670270957893, "ewc_loss": 0.1455078125, "ewc_loss_diag": 0.00011348724365234375, "ewc_loss_parallel": 0.000133514404296875, "grad_norm": 4.476333141326904, "learning_rate": 3.442136498516321e-05, "loss": 0.6812, "mean_token_accuracy": 0.8264197707176208, "num_tokens": 61800502.0, "step": 1625 }, { "epoch": 0.20684391298816945, "ewc_loss": 0.236328125, "ewc_loss_diag": 0.00011348724365234375, "ewc_loss_parallel": 0.000225067138671875, "grad_norm": 6.292537212371826, "learning_rate": 3.44425604069521e-05, "loss": 0.6562, "mean_token_accuracy": 0.8607702255249023, "num_tokens": 61834922.0, "step": 1626 }, { "epoch": 0.20697112326675995, "ewc_loss": 0.150390625, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 0.00013828277587890625, "grad_norm": 4.343372344970703, "learning_rate": 3.4463755828741e-05, "loss": 0.628, "mean_token_accuracy": 0.8428825736045837, "num_tokens": 61874675.0, "step": 1627 }, { "epoch": 0.20709833354535045, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.798276901245117, "learning_rate": 3.448495125052989e-05, "loss": 0.7144, "mean_token_accuracy": 0.8592448234558105, "num_tokens": 61914310.0, "step": 1628 }, { "epoch": 0.20722554382394098, "ewc_loss": 0.197265625, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 0.00018596649169921875, "grad_norm": 4.707148551940918, "learning_rate": 3.450614667231878e-05, "loss": 0.624, "mean_token_accuracy": 0.8549426794052124, "num_tokens": 61949142.0, "step": 1629 }, { "epoch": 0.20735275410253148, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.924307346343994, "learning_rate": 3.452734209410767e-05, "loss": 0.7534, "mean_token_accuracy": 0.8558019399642944, "num_tokens": 61989939.0, "step": 1630 }, { "epoch": 0.20747996438112198, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 5.515754222869873, "learning_rate": 3.454853751589657e-05, "loss": 0.7265, "mean_token_accuracy": 0.8481553792953491, "num_tokens": 62029263.0, "step": 1631 }, { "epoch": 0.2076071746597125, "ewc_loss": 0.265625, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.239612579345703, "learning_rate": 3.456973293768546e-05, "loss": 0.7039, "mean_token_accuracy": 0.8536509871482849, "num_tokens": 62067963.0, "step": 1632 }, { "epoch": 0.207734384938303, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 5.986697196960449, "learning_rate": 3.4590928359474357e-05, "loss": 0.8122, "mean_token_accuracy": 0.8289916515350342, "num_tokens": 62105313.0, "step": 1633 }, { "epoch": 0.2078615952168935, "ewc_loss": 0.1943359375, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 0.00018215179443359375, "grad_norm": 5.284135341644287, "learning_rate": 3.461212378126325e-05, "loss": 0.6862, "mean_token_accuracy": 0.8499346971511841, "num_tokens": 62140836.0, "step": 1634 }, { "epoch": 0.20798880549548404, "ewc_loss": 0.28125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.382429599761963, "learning_rate": 3.4633319203052145e-05, "loss": 0.6903, "mean_token_accuracy": 0.8644243478775024, "num_tokens": 62178951.0, "step": 1635 }, { "epoch": 0.20811601577407454, "ewc_loss": 0.185546875, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 0.0001735687255859375, "grad_norm": 4.8342084884643555, "learning_rate": 3.4654514624841036e-05, "loss": 0.629, "mean_token_accuracy": 0.851825475692749, "num_tokens": 62216941.0, "step": 1636 }, { "epoch": 0.20824322605266504, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.704573631286621, "learning_rate": 3.467571004662993e-05, "loss": 0.7688, "mean_token_accuracy": 0.8427096605300903, "num_tokens": 62257676.0, "step": 1637 }, { "epoch": 0.20837043633125557, "ewc_loss": 0.23046875, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 0.00021839141845703125, "grad_norm": 5.10390043258667, "learning_rate": 3.4696905468418825e-05, "loss": 0.7225, "mean_token_accuracy": 0.838618814945221, "num_tokens": 62296878.0, "step": 1638 }, { "epoch": 0.20849764660984607, "ewc_loss": 0.28125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.563199996948242, "learning_rate": 3.4718100890207716e-05, "loss": 0.6961, "mean_token_accuracy": 0.8615165948867798, "num_tokens": 62335292.0, "step": 1639 }, { "epoch": 0.2086248568884366, "ewc_loss": 0.265625, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.525640964508057, "learning_rate": 3.4739296311996614e-05, "loss": 0.7171, "mean_token_accuracy": 0.8555598258972168, "num_tokens": 62371182.0, "step": 1640 }, { "epoch": 0.2087520671670271, "ewc_loss": 0.244140625, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 6.040395259857178, "learning_rate": 3.4760491733785505e-05, "loss": 0.6874, "mean_token_accuracy": 0.8536808490753174, "num_tokens": 62407803.0, "step": 1641 }, { "epoch": 0.2088792774456176, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.633883953094482, "learning_rate": 3.4781687155574396e-05, "loss": 0.6904, "mean_token_accuracy": 0.8579256534576416, "num_tokens": 62446320.0, "step": 1642 }, { "epoch": 0.20900648772420813, "ewc_loss": 0.21484375, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 0.00020313262939453125, "grad_norm": 5.6137847900390625, "learning_rate": 3.4802882577363294e-05, "loss": 0.6892, "mean_token_accuracy": 0.8449313640594482, "num_tokens": 62486919.0, "step": 1643 }, { "epoch": 0.20913369800279863, "ewc_loss": 0.2294921875, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 0.000217437744140625, "grad_norm": 5.548964023590088, "learning_rate": 3.4824077999152185e-05, "loss": 0.6834, "mean_token_accuracy": 0.8515610098838806, "num_tokens": 62521511.0, "step": 1644 }, { "epoch": 0.20926090828138913, "ewc_loss": 0.201171875, "ewc_loss_diag": 0.00011682510375976562, "ewc_loss_parallel": 0.00018978118896484375, "grad_norm": 5.445041656494141, "learning_rate": 3.4845273420941076e-05, "loss": 0.7117, "mean_token_accuracy": 0.8282646536827087, "num_tokens": 62564702.0, "step": 1645 }, { "epoch": 0.20938811855997966, "ewc_loss": 0.203125, "ewc_loss_diag": 0.00011682510375976562, "ewc_loss_parallel": 0.00019168853759765625, "grad_norm": 5.366342544555664, "learning_rate": 3.4866468842729974e-05, "loss": 0.628, "mean_token_accuracy": 0.8594836592674255, "num_tokens": 62605850.0, "step": 1646 }, { "epoch": 0.20951532883857016, "ewc_loss": 0.197265625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 0.00018596649169921875, "grad_norm": 5.443728446960449, "learning_rate": 3.4887664264518865e-05, "loss": 0.649, "mean_token_accuracy": 0.8482158780097961, "num_tokens": 62650497.0, "step": 1647 }, { "epoch": 0.20964253911716066, "ewc_loss": 0.1806640625, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 0.00016880035400390625, "grad_norm": 5.235715866088867, "learning_rate": 3.490885968630776e-05, "loss": 0.6825, "mean_token_accuracy": 0.8378357291221619, "num_tokens": 62685538.0, "step": 1648 }, { "epoch": 0.2097697493957512, "ewc_loss": 0.1884765625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 0.00017642974853515625, "grad_norm": 5.4247612953186035, "learning_rate": 3.4930055108096654e-05, "loss": 0.6493, "mean_token_accuracy": 0.848626971244812, "num_tokens": 62725076.0, "step": 1649 }, { "epoch": 0.2098969596743417, "ewc_loss": 0.171875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 0.00016021728515625, "grad_norm": 5.178469657897949, "learning_rate": 3.495125052988555e-05, "loss": 0.6495, "mean_token_accuracy": 0.8441874384880066, "num_tokens": 62762335.0, "step": 1650 }, { "epoch": 0.2100241699529322, "ewc_loss": 0.1845703125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 0.00017261505126953125, "grad_norm": 5.413545608520508, "learning_rate": 3.4972445951674436e-05, "loss": 0.6197, "mean_token_accuracy": 0.8553441166877747, "num_tokens": 62804507.0, "step": 1651 }, { "epoch": 0.21015138023152272, "ewc_loss": 0.16796875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 0.000156402587890625, "grad_norm": 5.110737323760986, "learning_rate": 3.4993641373463334e-05, "loss": 0.5779, "mean_token_accuracy": 0.8632218837738037, "num_tokens": 62844499.0, "step": 1652 }, { "epoch": 0.21027859051011322, "ewc_loss": 0.1826171875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 0.00017070770263671875, "grad_norm": 5.477734565734863, "learning_rate": 3.5014836795252225e-05, "loss": 0.6304, "mean_token_accuracy": 0.8582851886749268, "num_tokens": 62880061.0, "step": 1653 }, { "epoch": 0.21040580078870372, "ewc_loss": 0.1630859375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 0.00015163421630859375, "grad_norm": 4.995574474334717, "learning_rate": 3.503603221704112e-05, "loss": 0.5602, "mean_token_accuracy": 0.866803765296936, "num_tokens": 62916726.0, "step": 1654 }, { "epoch": 0.21053301106729425, "ewc_loss": 0.19140625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 0.000179290771484375, "grad_norm": 5.614362716674805, "learning_rate": 3.5057227638830014e-05, "loss": 0.6334, "mean_token_accuracy": 0.8519074320793152, "num_tokens": 62953341.0, "step": 1655 }, { "epoch": 0.21066022134588475, "ewc_loss": 0.1533203125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 0.000141143798828125, "grad_norm": 4.8463263511657715, "learning_rate": 3.507842306061891e-05, "loss": 0.5899, "mean_token_accuracy": 0.8533328771591187, "num_tokens": 62995047.0, "step": 1656 }, { "epoch": 0.21078743162447525, "ewc_loss": 0.2001953125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 0.0001888275146484375, "grad_norm": 5.837770938873291, "learning_rate": 3.50996184824078e-05, "loss": 0.6644, "mean_token_accuracy": 0.847053050994873, "num_tokens": 63039520.0, "step": 1657 }, { "epoch": 0.21091464190306577, "ewc_loss": 0.150390625, "ewc_loss_diag": 0.00011873245239257812, "ewc_loss_parallel": 0.00013828277587890625, "grad_norm": 4.638594627380371, "learning_rate": 3.51208139041967e-05, "loss": 0.633, "mean_token_accuracy": 0.8365373611450195, "num_tokens": 63071243.0, "step": 1658 }, { "epoch": 0.21104185218165628, "ewc_loss": 0.2275390625, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 6.205641746520996, "learning_rate": 3.5142009325985584e-05, "loss": 0.7065, "mean_token_accuracy": 0.8410149812698364, "num_tokens": 63114496.0, "step": 1659 }, { "epoch": 0.21116906246024678, "ewc_loss": 0.146484375, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 0.00013446807861328125, "grad_norm": 4.39941930770874, "learning_rate": 3.516320474777448e-05, "loss": 0.5378, "mean_token_accuracy": 0.86407470703125, "num_tokens": 63158111.0, "step": 1660 }, { "epoch": 0.2112962727388373, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.751516819000244, "learning_rate": 3.518440016956337e-05, "loss": 0.7626, "mean_token_accuracy": 0.8376714587211609, "num_tokens": 63198305.0, "step": 1661 }, { "epoch": 0.2114234830174278, "ewc_loss": 0.1640625, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.000152587890625, "grad_norm": 4.473147392272949, "learning_rate": 3.520559559135227e-05, "loss": 0.6003, "mean_token_accuracy": 0.8541818261146545, "num_tokens": 63235053.0, "step": 1662 }, { "epoch": 0.2115506932960183, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.165938854217529, "learning_rate": 3.522679101314116e-05, "loss": 0.7995, "mean_token_accuracy": 0.8436403870582581, "num_tokens": 63275495.0, "step": 1663 }, { "epoch": 0.21167790357460883, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 5.276245594024658, "learning_rate": 3.524798643493006e-05, "loss": 0.7434, "mean_token_accuracy": 0.8421599864959717, "num_tokens": 63314338.0, "step": 1664 }, { "epoch": 0.21180511385319933, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.7945075035095215, "learning_rate": 3.526918185671895e-05, "loss": 0.7313, "mean_token_accuracy": 0.85589998960495, "num_tokens": 63348582.0, "step": 1665 }, { "epoch": 0.21193232413178986, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 5.999121189117432, "learning_rate": 3.529037727850784e-05, "loss": 0.7839, "mean_token_accuracy": 0.8373258113861084, "num_tokens": 63388496.0, "step": 1666 }, { "epoch": 0.21205953441038036, "ewc_loss": 0.2216796875, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.000209808349609375, "grad_norm": 5.73935079574585, "learning_rate": 3.531157270029673e-05, "loss": 0.6487, "mean_token_accuracy": 0.860119104385376, "num_tokens": 63426216.0, "step": 1667 }, { "epoch": 0.21218674468897086, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.456759452819824, "learning_rate": 3.533276812208563e-05, "loss": 0.7302, "mean_token_accuracy": 0.8546556234359741, "num_tokens": 63460627.0, "step": 1668 }, { "epoch": 0.2123139549675614, "ewc_loss": 0.1865234375, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.00017452239990234375, "grad_norm": 4.9571614265441895, "learning_rate": 3.535396354387452e-05, "loss": 0.6469, "mean_token_accuracy": 0.8479021787643433, "num_tokens": 63501911.0, "step": 1669 }, { "epoch": 0.2124411652461519, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.003958702087402, "learning_rate": 3.537515896566342e-05, "loss": 0.7908, "mean_token_accuracy": 0.8432737588882446, "num_tokens": 63548764.0, "step": 1670 }, { "epoch": 0.2125683755247424, "ewc_loss": 0.2353515625, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 5.136275768280029, "learning_rate": 3.539635438745231e-05, "loss": 0.6712, "mean_token_accuracy": 0.8562153577804565, "num_tokens": 63583347.0, "step": 1671 }, { "epoch": 0.21269558580333292, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.076205730438232, "learning_rate": 3.541754980924121e-05, "loss": 0.8287, "mean_token_accuracy": 0.833401083946228, "num_tokens": 63624115.0, "step": 1672 }, { "epoch": 0.21282279608192342, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 5.7547502517700195, "learning_rate": 3.54387452310301e-05, "loss": 0.7286, "mean_token_accuracy": 0.8599035739898682, "num_tokens": 63660850.0, "step": 1673 }, { "epoch": 0.21295000636051392, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.513882637023926, "learning_rate": 3.545994065281899e-05, "loss": 0.7524, "mean_token_accuracy": 0.8415839672088623, "num_tokens": 63701530.0, "step": 1674 }, { "epoch": 0.21307721663910445, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 5.918456554412842, "learning_rate": 3.548113607460788e-05, "loss": 0.7633, "mean_token_accuracy": 0.8421875238418579, "num_tokens": 63740400.0, "step": 1675 }, { "epoch": 0.21320442691769495, "ewc_loss": 0.2373046875, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.00022602081298828125, "grad_norm": 5.9460906982421875, "learning_rate": 3.550233149639678e-05, "loss": 0.6978, "mean_token_accuracy": 0.8491265773773193, "num_tokens": 63775019.0, "step": 1676 }, { "epoch": 0.21333163719628545, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 5.797165393829346, "learning_rate": 3.552352691818567e-05, "loss": 0.7406, "mean_token_accuracy": 0.8460195660591125, "num_tokens": 63813154.0, "step": 1677 }, { "epoch": 0.21345884747487598, "ewc_loss": 0.2265625, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.00021457672119140625, "grad_norm": 5.671823978424072, "learning_rate": 3.554472233997457e-05, "loss": 0.6691, "mean_token_accuracy": 0.8549826145172119, "num_tokens": 63851982.0, "step": 1678 }, { "epoch": 0.21358605775346648, "ewc_loss": 0.22265625, "ewc_loss_diag": 0.00011968612670898438, "ewc_loss_parallel": 0.00021076202392578125, "grad_norm": 5.540438175201416, "learning_rate": 3.556591776176346e-05, "loss": 0.7327, "mean_token_accuracy": 0.830970823764801, "num_tokens": 63888983.0, "step": 1679 }, { "epoch": 0.21371326803205698, "ewc_loss": 0.2197265625, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 0.0002079010009765625, "grad_norm": 5.655033588409424, "learning_rate": 3.558711318355236e-05, "loss": 0.6196, "mean_token_accuracy": 0.8668075799942017, "num_tokens": 63927131.0, "step": 1680 }, { "epoch": 0.2138404783106475, "ewc_loss": 0.1923828125, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 0.00018024444580078125, "grad_norm": 5.286970615386963, "learning_rate": 3.560830860534125e-05, "loss": 0.725, "mean_token_accuracy": 0.8251502513885498, "num_tokens": 63962362.0, "step": 1681 }, { "epoch": 0.213967688589238, "ewc_loss": 0.2138671875, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 0.000202178955078125, "grad_norm": 5.699882984161377, "learning_rate": 3.562950402713014e-05, "loss": 0.6272, "mean_token_accuracy": 0.86312335729599, "num_tokens": 63997189.0, "step": 1682 }, { "epoch": 0.2140948988678285, "ewc_loss": 0.1748046875, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 0.00016307830810546875, "grad_norm": 5.12047004699707, "learning_rate": 3.565069944891904e-05, "loss": 0.6212, "mean_token_accuracy": 0.8520797491073608, "num_tokens": 64032763.0, "step": 1683 }, { "epoch": 0.21422210914641904, "ewc_loss": 0.2119140625, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 0.0002002716064453125, "grad_norm": 5.793238162994385, "learning_rate": 3.567189487070793e-05, "loss": 0.6563, "mean_token_accuracy": 0.8515427112579346, "num_tokens": 64070176.0, "step": 1684 }, { "epoch": 0.21434931942500954, "ewc_loss": 0.1650390625, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 0.00015354156494140625, "grad_norm": 4.957438945770264, "learning_rate": 3.569309029249682e-05, "loss": 0.6065, "mean_token_accuracy": 0.8558326959609985, "num_tokens": 64105632.0, "step": 1685 }, { "epoch": 0.21447652970360004, "ewc_loss": 0.2216796875, "ewc_loss_diag": 0.00012111663818359375, "ewc_loss_parallel": 0.000209808349609375, "grad_norm": 5.9930219650268555, "learning_rate": 3.571428571428572e-05, "loss": 0.6353, "mean_token_accuracy": 0.8632608652114868, "num_tokens": 64148729.0, "step": 1686 }, { "epoch": 0.21460373998219057, "ewc_loss": 0.1591796875, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 0.00014781951904296875, "grad_norm": 4.780116558074951, "learning_rate": 3.573548113607461e-05, "loss": 0.6006, "mean_token_accuracy": 0.8547235727310181, "num_tokens": 64189466.0, "step": 1687 }, { "epoch": 0.21473095026078107, "ewc_loss": 0.2431640625, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 6.2809343338012695, "learning_rate": 3.5756676557863506e-05, "loss": 0.6621, "mean_token_accuracy": 0.8593330383300781, "num_tokens": 64226325.0, "step": 1688 }, { "epoch": 0.21485816053937157, "ewc_loss": 0.162109375, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 0.0001506805419921875, "grad_norm": 4.687721252441406, "learning_rate": 3.57778719796524e-05, "loss": 0.623, "mean_token_accuracy": 0.8443411588668823, "num_tokens": 64262178.0, "step": 1689 }, { "epoch": 0.2149853708179621, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.541293621063232, "learning_rate": 3.579906740144129e-05, "loss": 0.7156, "mean_token_accuracy": 0.8506819605827332, "num_tokens": 64299933.0, "step": 1690 }, { "epoch": 0.2151125810965526, "ewc_loss": 0.1845703125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 0.00017261505126953125, "grad_norm": 4.822930335998535, "learning_rate": 3.5820262823230186e-05, "loss": 0.5876, "mean_token_accuracy": 0.8622699975967407, "num_tokens": 64335836.0, "step": 1691 }, { "epoch": 0.21523979137514312, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.624460220336914, "learning_rate": 3.584145824501908e-05, "loss": 0.7166, "mean_token_accuracy": 0.859618067741394, "num_tokens": 64377190.0, "step": 1692 }, { "epoch": 0.21536700165373363, "ewc_loss": 0.2353515625, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 5.208371162414551, "learning_rate": 3.5862653666807975e-05, "loss": 0.7115, "mean_token_accuracy": 0.8450611233711243, "num_tokens": 64414693.0, "step": 1693 }, { "epoch": 0.21549421193232413, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.295908451080322, "learning_rate": 3.5883849088596866e-05, "loss": 0.6647, "mean_token_accuracy": 0.8743969202041626, "num_tokens": 64454256.0, "step": 1694 }, { "epoch": 0.21562142221091465, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.619823932647705, "learning_rate": 3.590504451038576e-05, "loss": 0.7068, "mean_token_accuracy": 0.8549058437347412, "num_tokens": 64491513.0, "step": 1695 }, { "epoch": 0.21574863248950515, "ewc_loss": 0.22265625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.000209808349609375, "grad_norm": 5.634167194366455, "learning_rate": 3.5926239932174654e-05, "loss": 0.6906, "mean_token_accuracy": 0.8495000004768372, "num_tokens": 64531349.0, "step": 1696 }, { "epoch": 0.21587584276809565, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00024318695068359375, "grad_norm": 5.921483993530273, "learning_rate": 3.5947435353963545e-05, "loss": 0.7054, "mean_token_accuracy": 0.8537507057189941, "num_tokens": 64568007.0, "step": 1697 }, { "epoch": 0.21600305304668618, "ewc_loss": 0.181640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00016880035400390625, "grad_norm": 5.122953414916992, "learning_rate": 3.5968630775752436e-05, "loss": 0.63, "mean_token_accuracy": 0.8514114618301392, "num_tokens": 64610287.0, "step": 1698 }, { "epoch": 0.21613026332527668, "ewc_loss": 0.251953125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.139101982116699, "learning_rate": 3.5989826197541334e-05, "loss": 0.7035, "mean_token_accuracy": 0.8519254922866821, "num_tokens": 64648757.0, "step": 1699 }, { "epoch": 0.21625747360386718, "ewc_loss": 0.171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00016021728515625, "grad_norm": 4.9385905265808105, "learning_rate": 3.6011021619330225e-05, "loss": 0.6212, "mean_token_accuracy": 0.8547414541244507, "num_tokens": 64690137.0, "step": 1700 }, { "epoch": 0.2163846838824577, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.239386081695557, "learning_rate": 3.603221704111912e-05, "loss": 0.7025, "mean_token_accuracy": 0.8506622314453125, "num_tokens": 64730218.0, "step": 1701 }, { "epoch": 0.2165118941610482, "ewc_loss": 0.18359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00017070770263671875, "grad_norm": 5.000397682189941, "learning_rate": 3.6053412462908014e-05, "loss": 0.6104, "mean_token_accuracy": 0.8560091257095337, "num_tokens": 64762213.0, "step": 1702 }, { "epoch": 0.2166391044396387, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.227336406707764, "learning_rate": 3.607460788469691e-05, "loss": 0.7278, "mean_token_accuracy": 0.842904806137085, "num_tokens": 64802471.0, "step": 1703 }, { "epoch": 0.21676631471822924, "ewc_loss": 0.205078125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00019359588623046875, "grad_norm": 5.110073089599609, "learning_rate": 3.6095803306485796e-05, "loss": 0.6205, "mean_token_accuracy": 0.8594251275062561, "num_tokens": 64838139.0, "step": 1704 }, { "epoch": 0.21689352499681974, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.172302722930908, "learning_rate": 3.6116998728274694e-05, "loss": 0.7116, "mean_token_accuracy": 0.8491058349609375, "num_tokens": 64877721.0, "step": 1705 }, { "epoch": 0.21702073527541024, "ewc_loss": 0.2109375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00019931793212890625, "grad_norm": 5.196938991546631, "learning_rate": 3.6138194150063585e-05, "loss": 0.6826, "mean_token_accuracy": 0.8422163128852844, "num_tokens": 64916315.0, "step": 1706 }, { "epoch": 0.21714794555400077, "ewc_loss": 0.248046875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00023651123046875, "grad_norm": 6.08041524887085, "learning_rate": 3.615938957185248e-05, "loss": 0.6737, "mean_token_accuracy": 0.8600507974624634, "num_tokens": 64951303.0, "step": 1707 }, { "epoch": 0.21727515583259127, "ewc_loss": 0.212890625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0002002716064453125, "grad_norm": 5.25167179107666, "learning_rate": 3.6180584993641374e-05, "loss": 0.6959, "mean_token_accuracy": 0.8392678499221802, "num_tokens": 64984677.0, "step": 1708 }, { "epoch": 0.21740236611118177, "ewc_loss": 0.236328125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.000225067138671875, "grad_norm": 5.97632360458374, "learning_rate": 3.620178041543027e-05, "loss": 0.6107, "mean_token_accuracy": 0.8731198310852051, "num_tokens": 65021960.0, "step": 1709 }, { "epoch": 0.2175295763897723, "ewc_loss": 0.208984375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0001964569091796875, "grad_norm": 5.231728553771973, "learning_rate": 3.622297583721916e-05, "loss": 0.6845, "mean_token_accuracy": 0.8429005146026611, "num_tokens": 65065576.0, "step": 1710 }, { "epoch": 0.2176567866683628, "ewc_loss": 0.23828125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0002269744873046875, "grad_norm": 5.969231128692627, "learning_rate": 3.624417125900806e-05, "loss": 0.6836, "mean_token_accuracy": 0.8553839921951294, "num_tokens": 65107188.0, "step": 1711 }, { "epoch": 0.2177839969469533, "ewc_loss": 0.1953125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00018310546875, "grad_norm": 5.166690826416016, "learning_rate": 3.6265366680796945e-05, "loss": 0.611, "mean_token_accuracy": 0.8600019216537476, "num_tokens": 65145740.0, "step": 1712 }, { "epoch": 0.21791120722554383, "ewc_loss": 0.2421875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00022983551025390625, "grad_norm": 6.007043838500977, "learning_rate": 3.628656210258584e-05, "loss": 0.6449, "mean_token_accuracy": 0.8643732070922852, "num_tokens": 65176572.0, "step": 1713 }, { "epoch": 0.21803841750413433, "ewc_loss": 0.1875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00017452239990234375, "grad_norm": 5.129668712615967, "learning_rate": 3.6307757524374734e-05, "loss": 0.6165, "mean_token_accuracy": 0.8555215001106262, "num_tokens": 65214839.0, "step": 1714 }, { "epoch": 0.21816562778272486, "ewc_loss": 0.23828125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0002269744873046875, "grad_norm": 6.016958713531494, "learning_rate": 3.632895294616363e-05, "loss": 0.6637, "mean_token_accuracy": 0.8585523366928101, "num_tokens": 65251438.0, "step": 1715 }, { "epoch": 0.21829283806131536, "ewc_loss": 0.1875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00017452239990234375, "grad_norm": 5.134998321533203, "learning_rate": 3.635014836795252e-05, "loss": 0.6508, "mean_token_accuracy": 0.8485569953918457, "num_tokens": 65294256.0, "step": 1716 }, { "epoch": 0.21842004833990586, "ewc_loss": 0.236328125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00022411346435546875, "grad_norm": 5.99534797668457, "learning_rate": 3.637134378974142e-05, "loss": 0.746, "mean_token_accuracy": 0.8348580598831177, "num_tokens": 65332917.0, "step": 1717 }, { "epoch": 0.2185472586184964, "ewc_loss": 0.193359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0001811981201171875, "grad_norm": 5.196318626403809, "learning_rate": 3.639253921153031e-05, "loss": 0.6146, "mean_token_accuracy": 0.8574967980384827, "num_tokens": 65366678.0, "step": 1718 }, { "epoch": 0.2186744688970869, "ewc_loss": 0.228515625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 5.893619060516357, "learning_rate": 3.641373463331921e-05, "loss": 0.7396, "mean_token_accuracy": 0.8315978050231934, "num_tokens": 65404797.0, "step": 1719 }, { "epoch": 0.2188016791756774, "ewc_loss": 0.1953125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00018405914306640625, "grad_norm": 5.2929792404174805, "learning_rate": 3.6434930055108093e-05, "loss": 0.6164, "mean_token_accuracy": 0.8601794838905334, "num_tokens": 65441855.0, "step": 1720 }, { "epoch": 0.21892888945426792, "ewc_loss": 0.21484375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.000202178955078125, "grad_norm": 5.704007625579834, "learning_rate": 3.645612547689699e-05, "loss": 0.6791, "mean_token_accuracy": 0.8507479429244995, "num_tokens": 65486895.0, "step": 1721 }, { "epoch": 0.21905609973285842, "ewc_loss": 0.201171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0001888275146484375, "grad_norm": 5.440519332885742, "learning_rate": 3.647732089868588e-05, "loss": 0.648, "mean_token_accuracy": 0.854766309261322, "num_tokens": 65524862.0, "step": 1722 }, { "epoch": 0.21918331001144892, "ewc_loss": 0.197265625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0001850128173828125, "grad_norm": 5.49425745010376, "learning_rate": 3.649851632047478e-05, "loss": 0.6412, "mean_token_accuracy": 0.8510504961013794, "num_tokens": 65561538.0, "step": 1723 }, { "epoch": 0.21931052029003945, "ewc_loss": 0.203125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00019073486328125, "grad_norm": 5.545565128326416, "learning_rate": 3.651971174226367e-05, "loss": 0.6672, "mean_token_accuracy": 0.8457598686218262, "num_tokens": 65602376.0, "step": 1724 }, { "epoch": 0.21943773056862995, "ewc_loss": 0.18359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00017070770263671875, "grad_norm": 5.328698635101318, "learning_rate": 3.654090716405257e-05, "loss": 0.5966, "mean_token_accuracy": 0.8588266372680664, "num_tokens": 65640817.0, "step": 1725 }, { "epoch": 0.21956494084722045, "ewc_loss": 0.19921875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.000186920166015625, "grad_norm": 5.655272006988525, "learning_rate": 3.656210258584146e-05, "loss": 0.6387, "mean_token_accuracy": 0.8520885705947876, "num_tokens": 65679238.0, "step": 1726 }, { "epoch": 0.21969215112581097, "ewc_loss": 0.173828125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0001621246337890625, "grad_norm": 5.2026777267456055, "learning_rate": 3.658329800763035e-05, "loss": 0.6393, "mean_token_accuracy": 0.8420683145523071, "num_tokens": 65720141.0, "step": 1727 }, { "epoch": 0.21981936140440148, "ewc_loss": 0.203125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00019168853759765625, "grad_norm": 5.761203765869141, "learning_rate": 3.660449342941924e-05, "loss": 0.6375, "mean_token_accuracy": 0.8549354672431946, "num_tokens": 65758511.0, "step": 1728 }, { "epoch": 0.21994657168299198, "ewc_loss": 0.166015625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00015354156494140625, "grad_norm": 5.077871322631836, "learning_rate": 3.662568885120814e-05, "loss": 0.7219, "mean_token_accuracy": 0.8218173980712891, "num_tokens": 65798485.0, "step": 1729 }, { "epoch": 0.2200737819615825, "ewc_loss": 0.2109375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.0001983642578125, "grad_norm": 5.902945041656494, "learning_rate": 3.664688427299703e-05, "loss": 0.6746, "mean_token_accuracy": 0.8476177453994751, "num_tokens": 65840414.0, "step": 1730 }, { "epoch": 0.220200992240173, "ewc_loss": 0.16015625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.000148773193359375, "grad_norm": 4.955323219299316, "learning_rate": 3.666807969478593e-05, "loss": 0.6004, "mean_token_accuracy": 0.8533827066421509, "num_tokens": 65878425.0, "step": 1731 }, { "epoch": 0.2203282025187635, "ewc_loss": 0.21875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00020694732666015625, "grad_norm": 6.081787109375, "learning_rate": 3.668927511657482e-05, "loss": 0.6638, "mean_token_accuracy": 0.8506044149398804, "num_tokens": 65916958.0, "step": 1732 }, { "epoch": 0.22045541279735403, "ewc_loss": 0.15625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 0.00014400482177734375, "grad_norm": 4.818219184875488, "learning_rate": 3.671047053836372e-05, "loss": 0.6549, "mean_token_accuracy": 0.8349159955978394, "num_tokens": 65958511.0, "step": 1733 }, { "epoch": 0.22058262307594453, "ewc_loss": 0.2412109375, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.0002288818359375, "grad_norm": 6.351569175720215, "learning_rate": 3.673166596015261e-05, "loss": 0.6868, "mean_token_accuracy": 0.853918194770813, "num_tokens": 65994735.0, "step": 1734 }, { "epoch": 0.22070983335453503, "ewc_loss": 0.158203125, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.00014591217041015625, "grad_norm": 4.709537506103516, "learning_rate": 3.67528613819415e-05, "loss": 0.6608, "mean_token_accuracy": 0.8344370126724243, "num_tokens": 66036017.0, "step": 1735 }, { "epoch": 0.22083704363312556, "ewc_loss": 0.265625, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.614739894866943, "learning_rate": 3.67740568037304e-05, "loss": 0.7178, "mean_token_accuracy": 0.8479475975036621, "num_tokens": 66073203.0, "step": 1736 }, { "epoch": 0.22096425391171606, "ewc_loss": 0.1767578125, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.000164031982421875, "grad_norm": 4.817018032073975, "learning_rate": 3.679525222551929e-05, "loss": 0.648, "mean_token_accuracy": 0.8424910306930542, "num_tokens": 66109623.0, "step": 1737 }, { "epoch": 0.22109146419030656, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.641030311584473, "learning_rate": 3.6816447647308186e-05, "loss": 0.7259, "mean_token_accuracy": 0.8532364964485168, "num_tokens": 66148426.0, "step": 1738 }, { "epoch": 0.2212186744688971, "ewc_loss": 0.2255859375, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.00021266937255859375, "grad_norm": 5.226675987243652, "learning_rate": 3.683764306909708e-05, "loss": 0.6285, "mean_token_accuracy": 0.8662261962890625, "num_tokens": 66189164.0, "step": 1739 }, { "epoch": 0.2213458847474876, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.17813777923584, "learning_rate": 3.6858838490885975e-05, "loss": 0.6824, "mean_token_accuracy": 0.8571253418922424, "num_tokens": 66223523.0, "step": 1740 }, { "epoch": 0.22147309502607812, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.781332015991211, "learning_rate": 3.6880033912674866e-05, "loss": 0.6687, "mean_token_accuracy": 0.8643972873687744, "num_tokens": 66256819.0, "step": 1741 }, { "epoch": 0.22160030530466862, "ewc_loss": 0.201171875, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.0001888275146484375, "grad_norm": 5.3429365158081055, "learning_rate": 3.690122933446376e-05, "loss": 0.6401, "mean_token_accuracy": 0.8537201881408691, "num_tokens": 66291619.0, "step": 1742 }, { "epoch": 0.22172751558325912, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.368652820587158, "learning_rate": 3.692242475625265e-05, "loss": 0.7171, "mean_token_accuracy": 0.8533546924591064, "num_tokens": 66334034.0, "step": 1743 }, { "epoch": 0.22185472586184965, "ewc_loss": 0.1630859375, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.0001506805419921875, "grad_norm": 4.720298767089844, "learning_rate": 3.6943620178041546e-05, "loss": 0.5948, "mean_token_accuracy": 0.8592671155929565, "num_tokens": 66368786.0, "step": 1744 }, { "epoch": 0.22198193614044015, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.950750350952148, "learning_rate": 3.696481559983044e-05, "loss": 0.7172, "mean_token_accuracy": 0.8575652241706848, "num_tokens": 66400034.0, "step": 1745 }, { "epoch": 0.22210914641903065, "ewc_loss": 0.1826171875, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.0001697540283203125, "grad_norm": 4.707690715789795, "learning_rate": 3.6986011021619335e-05, "loss": 0.6411, "mean_token_accuracy": 0.8500651121139526, "num_tokens": 66436954.0, "step": 1746 }, { "epoch": 0.22223635669762118, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.215824127197266, "learning_rate": 3.7007206443408226e-05, "loss": 0.7486, "mean_token_accuracy": 0.8590590357780457, "num_tokens": 66476360.0, "step": 1747 }, { "epoch": 0.22236356697621168, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.417546272277832, "learning_rate": 3.7028401865197124e-05, "loss": 0.6761, "mean_token_accuracy": 0.8648801445960999, "num_tokens": 66516205.0, "step": 1748 }, { "epoch": 0.22249077725480218, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.7537007331848145, "learning_rate": 3.7049597286986015e-05, "loss": 0.7699, "mean_token_accuracy": 0.8428406119346619, "num_tokens": 66548732.0, "step": 1749 }, { "epoch": 0.2226179875333927, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000125885009765625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.0331645011901855, "learning_rate": 3.7070792708774906e-05, "loss": 0.8107, "mean_token_accuracy": 0.8316478133201599, "num_tokens": 66587888.0, "step": 1750 }, { "epoch": 0.2227451978119832, "ewc_loss": 0.2470703125, "ewc_loss_diag": 0.000125885009765625, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 5.869201183319092, "learning_rate": 3.70919881305638e-05, "loss": 0.6873, "mean_token_accuracy": 0.8608915209770203, "num_tokens": 66617846.0, "step": 1751 }, { "epoch": 0.2228724080905737, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.000125885009765625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.317078113555908, "learning_rate": 3.7113183552352695e-05, "loss": 0.7411, "mean_token_accuracy": 0.8496350049972534, "num_tokens": 66659703.0, "step": 1752 }, { "epoch": 0.22299961836916424, "ewc_loss": 0.2236328125, "ewc_loss_diag": 0.000125885009765625, "ewc_loss_parallel": 0.00021076202392578125, "grad_norm": 5.37923002243042, "learning_rate": 3.7134378974141586e-05, "loss": 0.6891, "mean_token_accuracy": 0.850039541721344, "num_tokens": 66697604.0, "step": 1753 }, { "epoch": 0.22312682864775474, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.000125885009765625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.282622814178467, "learning_rate": 3.7155574395930484e-05, "loss": 0.7676, "mean_token_accuracy": 0.8379439115524292, "num_tokens": 66736839.0, "step": 1754 }, { "epoch": 0.22325403892634524, "ewc_loss": 0.232421875, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 0.0002193450927734375, "grad_norm": 5.4737420082092285, "learning_rate": 3.7176769817719375e-05, "loss": 0.6926, "mean_token_accuracy": 0.847866415977478, "num_tokens": 66777477.0, "step": 1755 }, { "epoch": 0.22338124920493577, "ewc_loss": 0.2373046875, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 0.000225067138671875, "grad_norm": 5.83794641494751, "learning_rate": 3.719796523950827e-05, "loss": 0.6434, "mean_token_accuracy": 0.8626813888549805, "num_tokens": 66816294.0, "step": 1756 }, { "epoch": 0.22350845948352627, "ewc_loss": 0.244140625, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 5.765073299407959, "learning_rate": 3.721916066129716e-05, "loss": 0.7142, "mean_token_accuracy": 0.8441245555877686, "num_tokens": 66852740.0, "step": 1757 }, { "epoch": 0.22363566976211677, "ewc_loss": 0.203125, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 0.00019073486328125, "grad_norm": 5.395163059234619, "learning_rate": 3.7240356083086054e-05, "loss": 0.6471, "mean_token_accuracy": 0.8490983843803406, "num_tokens": 66890511.0, "step": 1758 }, { "epoch": 0.2237628800407073, "ewc_loss": 0.244140625, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 6.038090229034424, "learning_rate": 3.7261551504874945e-05, "loss": 0.6536, "mean_token_accuracy": 0.8614236116409302, "num_tokens": 66924724.0, "step": 1759 }, { "epoch": 0.2238900903192978, "ewc_loss": 0.1796875, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 0.00016689300537109375, "grad_norm": 5.105269908905029, "learning_rate": 3.728274692666384e-05, "loss": 0.6715, "mean_token_accuracy": 0.8384681344032288, "num_tokens": 66964102.0, "step": 1760 }, { "epoch": 0.2240173005978883, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.273462295532227, "learning_rate": 3.7303942348452734e-05, "loss": 0.6303, "mean_token_accuracy": 0.8720853328704834, "num_tokens": 67003234.0, "step": 1761 }, { "epoch": 0.22414451087647883, "ewc_loss": 0.1767578125, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 0.000164031982421875, "grad_norm": 4.968194007873535, "learning_rate": 3.732513777024163e-05, "loss": 0.7018, "mean_token_accuracy": 0.8286631107330322, "num_tokens": 67045952.0, "step": 1762 }, { "epoch": 0.22427172115506933, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.5205912590026855, "learning_rate": 3.734633319203052e-05, "loss": 0.7863, "mean_token_accuracy": 0.8338155746459961, "num_tokens": 67090169.0, "step": 1763 }, { "epoch": 0.22439893143365983, "ewc_loss": 0.1865234375, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 0.0001735687255859375, "grad_norm": 4.907416343688965, "learning_rate": 3.736752861381942e-05, "loss": 0.6171, "mean_token_accuracy": 0.8572797775268555, "num_tokens": 67125075.0, "step": 1764 }, { "epoch": 0.22452614171225035, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.754100322723389, "learning_rate": 3.7388724035608305e-05, "loss": 0.7731, "mean_token_accuracy": 0.8434700965881348, "num_tokens": 67162892.0, "step": 1765 }, { "epoch": 0.22465335199084085, "ewc_loss": 0.2158203125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.00020313262939453125, "grad_norm": 5.025537014007568, "learning_rate": 3.74099194573972e-05, "loss": 0.6757, "mean_token_accuracy": 0.8491319417953491, "num_tokens": 67198324.0, "step": 1766 }, { "epoch": 0.22478056226943138, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.824469566345215, "learning_rate": 3.7431114879186094e-05, "loss": 0.7424, "mean_token_accuracy": 0.8506259918212891, "num_tokens": 67239548.0, "step": 1767 }, { "epoch": 0.22490777254802188, "ewc_loss": 0.2451171875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 5.311011791229248, "learning_rate": 3.745231030097499e-05, "loss": 0.65, "mean_token_accuracy": 0.8608188033103943, "num_tokens": 67273373.0, "step": 1768 }, { "epoch": 0.22503498282661238, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.670563697814941, "learning_rate": 3.747350572276388e-05, "loss": 0.7888, "mean_token_accuracy": 0.8382174968719482, "num_tokens": 67307074.0, "step": 1769 }, { "epoch": 0.2251621931052029, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.563972473144531, "learning_rate": 3.749470114455278e-05, "loss": 0.7314, "mean_token_accuracy": 0.8444919586181641, "num_tokens": 67347905.0, "step": 1770 }, { "epoch": 0.2252894033837934, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.356155872344971, "learning_rate": 3.751589656634167e-05, "loss": 0.7233, "mean_token_accuracy": 0.8525791764259338, "num_tokens": 67382594.0, "step": 1771 }, { "epoch": 0.2254166136623839, "ewc_loss": 0.265625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 5.690937519073486, "learning_rate": 3.753709198813057e-05, "loss": 0.7433, "mean_token_accuracy": 0.8411551713943481, "num_tokens": 67422867.0, "step": 1772 }, { "epoch": 0.22554382394097444, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.059665203094482, "learning_rate": 3.7558287409919454e-05, "loss": 0.7345, "mean_token_accuracy": 0.841840922832489, "num_tokens": 67463524.0, "step": 1773 }, { "epoch": 0.22567103421956494, "ewc_loss": 0.24609375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.00023365020751953125, "grad_norm": 5.6439738273620605, "learning_rate": 3.757948283170835e-05, "loss": 0.6708, "mean_token_accuracy": 0.851035475730896, "num_tokens": 67499224.0, "step": 1774 }, { "epoch": 0.22579824449815544, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.003894329071045, "learning_rate": 3.760067825349724e-05, "loss": 0.7149, "mean_token_accuracy": 0.8470386266708374, "num_tokens": 67536745.0, "step": 1775 }, { "epoch": 0.22592545477674597, "ewc_loss": 0.2275390625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.00021457672119140625, "grad_norm": 5.471695423126221, "learning_rate": 3.762187367528614e-05, "loss": 0.7026, "mean_token_accuracy": 0.8474364876747131, "num_tokens": 67572725.0, "step": 1776 }, { "epoch": 0.22605266505533647, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.079862117767334, "learning_rate": 3.764306909707503e-05, "loss": 0.6955, "mean_token_accuracy": 0.8588192462921143, "num_tokens": 67607588.0, "step": 1777 }, { "epoch": 0.22617987533392697, "ewc_loss": 0.20703125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.000194549560546875, "grad_norm": 5.333339214324951, "learning_rate": 3.766426451886393e-05, "loss": 0.6718, "mean_token_accuracy": 0.8458362817764282, "num_tokens": 67643064.0, "step": 1778 }, { "epoch": 0.2263070856125175, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.167405605316162, "learning_rate": 3.768545994065282e-05, "loss": 0.7271, "mean_token_accuracy": 0.8432518243789673, "num_tokens": 67677985.0, "step": 1779 }, { "epoch": 0.226434295891108, "ewc_loss": 0.203125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.00019073486328125, "grad_norm": 5.271762371063232, "learning_rate": 3.770665536244171e-05, "loss": 0.6589, "mean_token_accuracy": 0.8487972021102905, "num_tokens": 67713814.0, "step": 1780 }, { "epoch": 0.2265615061696985, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.188928127288818, "learning_rate": 3.772785078423061e-05, "loss": 0.7145, "mean_token_accuracy": 0.8477806448936462, "num_tokens": 67752806.0, "step": 1781 }, { "epoch": 0.22668871644828903, "ewc_loss": 0.2080078125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.00019550323486328125, "grad_norm": 5.3165059089660645, "learning_rate": 3.77490462060195e-05, "loss": 0.6487, "mean_token_accuracy": 0.8553758263587952, "num_tokens": 67792804.0, "step": 1782 }, { "epoch": 0.22681592672687953, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.1065168380737305, "learning_rate": 3.77702416278084e-05, "loss": 0.6575, "mean_token_accuracy": 0.8647897839546204, "num_tokens": 67831361.0, "step": 1783 }, { "epoch": 0.22694313700547003, "ewc_loss": 0.212890625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.0002002716064453125, "grad_norm": 5.431339263916016, "learning_rate": 3.779143704959729e-05, "loss": 0.704, "mean_token_accuracy": 0.8412407636642456, "num_tokens": 67868981.0, "step": 1784 }, { "epoch": 0.22707034728406056, "ewc_loss": 0.240234375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.00022792816162109375, "grad_norm": 5.942626953125, "learning_rate": 3.781263247138618e-05, "loss": 0.7147, "mean_token_accuracy": 0.8409214019775391, "num_tokens": 67905501.0, "step": 1785 }, { "epoch": 0.22719755756265106, "ewc_loss": 0.2158203125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.00020313262939453125, "grad_norm": 5.533565044403076, "learning_rate": 3.783382789317508e-05, "loss": 0.6546, "mean_token_accuracy": 0.8558750152587891, "num_tokens": 67946001.0, "step": 1786 }, { "epoch": 0.22732476784124156, "ewc_loss": 0.2255859375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.00021266937255859375, "grad_norm": 5.77353572845459, "learning_rate": 3.785502331496397e-05, "loss": 0.6901, "mean_token_accuracy": 0.8487801551818848, "num_tokens": 67983868.0, "step": 1787 }, { "epoch": 0.2274519781198321, "ewc_loss": 0.2138671875, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.00020122528076171875, "grad_norm": 5.6201605796813965, "learning_rate": 3.787621873675286e-05, "loss": 0.7324, "mean_token_accuracy": 0.832145094871521, "num_tokens": 68023664.0, "step": 1788 }, { "epoch": 0.2275791883984226, "ewc_loss": 0.208984375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.0001964569091796875, "grad_norm": 5.619208335876465, "learning_rate": 3.789741415854176e-05, "loss": 0.6646, "mean_token_accuracy": 0.8512833714485168, "num_tokens": 68063950.0, "step": 1789 }, { "epoch": 0.2277063986770131, "ewc_loss": 0.20703125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.000194549560546875, "grad_norm": 5.702240467071533, "learning_rate": 3.791860958033065e-05, "loss": 0.6633, "mean_token_accuracy": 0.8509601354598999, "num_tokens": 68097812.0, "step": 1790 }, { "epoch": 0.22783360895560362, "ewc_loss": 0.1962890625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.00018310546875, "grad_norm": 5.503189563751221, "learning_rate": 3.793980500211955e-05, "loss": 0.6515, "mean_token_accuracy": 0.8486226797103882, "num_tokens": 68135219.0, "step": 1791 }, { "epoch": 0.22796081923419412, "ewc_loss": 0.20703125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.000194549560546875, "grad_norm": 5.7808990478515625, "learning_rate": 3.796100042390844e-05, "loss": 0.7135, "mean_token_accuracy": 0.8376179337501526, "num_tokens": 68169757.0, "step": 1792 }, { "epoch": 0.22808802951278465, "ewc_loss": 0.1875, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 0.00017452239990234375, "grad_norm": 5.382812976837158, "learning_rate": 3.7982195845697336e-05, "loss": 0.6403, "mean_token_accuracy": 0.8491748571395874, "num_tokens": 68204807.0, "step": 1793 }, { "epoch": 0.22821523979137515, "ewc_loss": 0.2109375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.0001983642578125, "grad_norm": 5.881872177124023, "learning_rate": 3.8003391267486227e-05, "loss": 0.6566, "mean_token_accuracy": 0.8553104996681213, "num_tokens": 68245037.0, "step": 1794 }, { "epoch": 0.22834245006996565, "ewc_loss": 0.1767578125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.000164031982421875, "grad_norm": 5.267087936401367, "learning_rate": 3.802458668927512e-05, "loss": 0.5932, "mean_token_accuracy": 0.8603906631469727, "num_tokens": 68281938.0, "step": 1795 }, { "epoch": 0.22846966034855618, "ewc_loss": 0.2177734375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.00020503997802734375, "grad_norm": 6.039638042449951, "learning_rate": 3.804578211106401e-05, "loss": 0.6199, "mean_token_accuracy": 0.8648264408111572, "num_tokens": 68319363.0, "step": 1796 }, { "epoch": 0.22859687062714668, "ewc_loss": 0.1669921875, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.0001544952392578125, "grad_norm": 5.102083206176758, "learning_rate": 3.8066977532852906e-05, "loss": 0.5944, "mean_token_accuracy": 0.8559860587120056, "num_tokens": 68357428.0, "step": 1797 }, { "epoch": 0.22872408090573718, "ewc_loss": 0.23046875, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.000217437744140625, "grad_norm": 6.298673152923584, "learning_rate": 3.80881729546418e-05, "loss": 0.6619, "mean_token_accuracy": 0.8574457168579102, "num_tokens": 68390558.0, "step": 1798 }, { "epoch": 0.2288512911843277, "ewc_loss": 0.16015625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.00014781951904296875, "grad_norm": 4.868936538696289, "learning_rate": 3.8109368376430695e-05, "loss": 0.627, "mean_token_accuracy": 0.8460235595703125, "num_tokens": 68432250.0, "step": 1799 }, { "epoch": 0.2289785014629182, "ewc_loss": 0.265625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.699697494506836, "learning_rate": 3.8130563798219586e-05, "loss": 0.7346, "mean_token_accuracy": 0.8450561761856079, "num_tokens": 68469427.0, "step": 1800 }, { "epoch": 0.2291057117415087, "ewc_loss": 0.1611328125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.000148773193359375, "grad_norm": 4.67586612701416, "learning_rate": 3.8151759220008484e-05, "loss": 0.6069, "mean_token_accuracy": 0.8527182936668396, "num_tokens": 68503442.0, "step": 1801 }, { "epoch": 0.22923292202009923, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.135162830352783, "learning_rate": 3.8172954641797375e-05, "loss": 0.737, "mean_token_accuracy": 0.8599427938461304, "num_tokens": 68540836.0, "step": 1802 }, { "epoch": 0.22936013229868973, "ewc_loss": 0.1962890625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.00018405914306640625, "grad_norm": 4.8716020584106445, "learning_rate": 3.8194150063586266e-05, "loss": 0.6618, "mean_token_accuracy": 0.8449053764343262, "num_tokens": 68581118.0, "step": 1803 }, { "epoch": 0.22948734257728023, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.227725505828857, "learning_rate": 3.821534548537516e-05, "loss": 0.7696, "mean_token_accuracy": 0.8540911674499512, "num_tokens": 68620988.0, "step": 1804 }, { "epoch": 0.22961455285587076, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 5.62760591506958, "learning_rate": 3.8236540907164055e-05, "loss": 0.6878, "mean_token_accuracy": 0.8617649078369141, "num_tokens": 68662141.0, "step": 1805 }, { "epoch": 0.22974176313446126, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.47918701171875, "learning_rate": 3.8257736328952946e-05, "loss": 0.7788, "mean_token_accuracy": 0.8401302695274353, "num_tokens": 68700922.0, "step": 1806 }, { "epoch": 0.22986897341305176, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.395870208740234, "learning_rate": 3.8278931750741844e-05, "loss": 0.7808, "mean_token_accuracy": 0.8474545478820801, "num_tokens": 68741137.0, "step": 1807 }, { "epoch": 0.2299961836916423, "ewc_loss": 0.203125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.00019073486328125, "grad_norm": 5.234397888183594, "learning_rate": 3.8300127172530735e-05, "loss": 0.6706, "mean_token_accuracy": 0.8458565473556519, "num_tokens": 68785679.0, "step": 1808 }, { "epoch": 0.2301233939702328, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.386580467224121, "learning_rate": 3.832132259431963e-05, "loss": 0.7953, "mean_token_accuracy": 0.8513090014457703, "num_tokens": 68823482.0, "step": 1809 }, { "epoch": 0.2302506042488233, "ewc_loss": 0.1962890625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.00018310546875, "grad_norm": 4.736588954925537, "learning_rate": 3.8342518016108524e-05, "loss": 0.7047, "mean_token_accuracy": 0.8287777304649353, "num_tokens": 68863854.0, "step": 1810 }, { "epoch": 0.23037781452741382, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.370121002197266, "learning_rate": 3.8363713437897415e-05, "loss": 0.8463, "mean_token_accuracy": 0.8622596263885498, "num_tokens": 68901341.0, "step": 1811 }, { "epoch": 0.23050502480600432, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.055786609649658, "learning_rate": 3.8384908859686306e-05, "loss": 0.7303, "mean_token_accuracy": 0.8668752908706665, "num_tokens": 68931774.0, "step": 1812 }, { "epoch": 0.23063223508459482, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.100796699523926, "learning_rate": 3.8406104281475204e-05, "loss": 0.8219, "mean_token_accuracy": 0.8589040040969849, "num_tokens": 68972114.0, "step": 1813 }, { "epoch": 0.23075944536318535, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 6.857379913330078, "learning_rate": 3.8427299703264095e-05, "loss": 0.8832, "mean_token_accuracy": 0.8511105179786682, "num_tokens": 69007272.0, "step": 1814 }, { "epoch": 0.23088665564177585, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.276812553405762, "learning_rate": 3.844849512505299e-05, "loss": 0.7694, "mean_token_accuracy": 0.8534489870071411, "num_tokens": 69050011.0, "step": 1815 }, { "epoch": 0.23101386592036638, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.648660659790039, "learning_rate": 3.8469690546841884e-05, "loss": 0.8117, "mean_token_accuracy": 0.8464173078536987, "num_tokens": 69087914.0, "step": 1816 }, { "epoch": 0.23114107619895688, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.710649490356445, "learning_rate": 3.849088596863078e-05, "loss": 0.7695, "mean_token_accuracy": 0.8482904434204102, "num_tokens": 69125301.0, "step": 1817 }, { "epoch": 0.23126828647754738, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.237997055053711, "learning_rate": 3.8512081390419666e-05, "loss": 0.7539, "mean_token_accuracy": 0.8471570014953613, "num_tokens": 69159268.0, "step": 1818 }, { "epoch": 0.2313954967561379, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.555273056030273, "learning_rate": 3.8533276812208563e-05, "loss": 0.7522, "mean_token_accuracy": 0.8526839017868042, "num_tokens": 69202207.0, "step": 1819 }, { "epoch": 0.2315227070347284, "ewc_loss": 0.244140625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 5.7137932777404785, "learning_rate": 3.8554472233997454e-05, "loss": 0.6655, "mean_token_accuracy": 0.8580589890480042, "num_tokens": 69234757.0, "step": 1820 }, { "epoch": 0.2316499173133189, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.784395694732666, "learning_rate": 3.857566765578635e-05, "loss": 0.735, "mean_token_accuracy": 0.8613276481628418, "num_tokens": 69274995.0, "step": 1821 }, { "epoch": 0.23177712759190944, "ewc_loss": 0.2265625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.00021266937255859375, "grad_norm": 5.385595798492432, "learning_rate": 3.859686307757524e-05, "loss": 0.7153, "mean_token_accuracy": 0.8386558294296265, "num_tokens": 69318706.0, "step": 1822 }, { "epoch": 0.23190433787049994, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.115307331085205, "learning_rate": 3.861805849936414e-05, "loss": 0.7668, "mean_token_accuracy": 0.8520598411560059, "num_tokens": 69361895.0, "step": 1823 }, { "epoch": 0.23203154814909044, "ewc_loss": 0.248046875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 5.477883815765381, "learning_rate": 3.863925392115303e-05, "loss": 0.7023, "mean_token_accuracy": 0.8462969660758972, "num_tokens": 69400877.0, "step": 1824 }, { "epoch": 0.23215875842768097, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.247723579406738, "learning_rate": 3.866044934294193e-05, "loss": 0.8254, "mean_token_accuracy": 0.8450791835784912, "num_tokens": 69437875.0, "step": 1825 }, { "epoch": 0.23228596870627147, "ewc_loss": 0.28125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 5.745939254760742, "learning_rate": 3.868164476473082e-05, "loss": 0.6763, "mean_token_accuracy": 0.8679942488670349, "num_tokens": 69481770.0, "step": 1826 }, { "epoch": 0.23241317898486197, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.139252185821533, "learning_rate": 3.870284018651971e-05, "loss": 0.7592, "mean_token_accuracy": 0.8576080799102783, "num_tokens": 69514643.0, "step": 1827 }, { "epoch": 0.2325403892634525, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.0465545654296875, "learning_rate": 3.87240356083086e-05, "loss": 0.6862, "mean_token_accuracy": 0.871918261051178, "num_tokens": 69551906.0, "step": 1828 }, { "epoch": 0.232667599542043, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.677324295043945, "learning_rate": 3.87452310300975e-05, "loss": 0.7735, "mean_token_accuracy": 0.8521633148193359, "num_tokens": 69593621.0, "step": 1829 }, { "epoch": 0.2327948098206335, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.164332389831543, "learning_rate": 3.876642645188639e-05, "loss": 0.6907, "mean_token_accuracy": 0.8660719394683838, "num_tokens": 69629586.0, "step": 1830 }, { "epoch": 0.23292202009922403, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.201244354248047, "learning_rate": 3.878762187367529e-05, "loss": 0.783, "mean_token_accuracy": 0.8357638120651245, "num_tokens": 69667409.0, "step": 1831 }, { "epoch": 0.23304923037781453, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.036994457244873, "learning_rate": 3.880881729546418e-05, "loss": 0.6731, "mean_token_accuracy": 0.8640720248222351, "num_tokens": 69704645.0, "step": 1832 }, { "epoch": 0.23317644065640503, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.049795150756836, "learning_rate": 3.883001271725307e-05, "loss": 0.7005, "mean_token_accuracy": 0.855774998664856, "num_tokens": 69742216.0, "step": 1833 }, { "epoch": 0.23330365093499555, "ewc_loss": 0.2373046875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.00022411346435546875, "grad_norm": 5.765433311462402, "learning_rate": 3.885120813904197e-05, "loss": 0.6321, "mean_token_accuracy": 0.8671256303787231, "num_tokens": 69779119.0, "step": 1834 }, { "epoch": 0.23343086121358606, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.175334930419922, "learning_rate": 3.887240356083086e-05, "loss": 0.74, "mean_token_accuracy": 0.845299243927002, "num_tokens": 69819980.0, "step": 1835 }, { "epoch": 0.23355807149217656, "ewc_loss": 0.21484375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.00020122528076171875, "grad_norm": 5.476433753967285, "learning_rate": 3.889359898261976e-05, "loss": 0.6659, "mean_token_accuracy": 0.8484464287757874, "num_tokens": 69859840.0, "step": 1836 }, { "epoch": 0.23368528177076708, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.453292369842529, "learning_rate": 3.891479440440865e-05, "loss": 0.7221, "mean_token_accuracy": 0.8509459495544434, "num_tokens": 69893446.0, "step": 1837 }, { "epoch": 0.23381249204935758, "ewc_loss": 0.197265625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.00018405914306640625, "grad_norm": 5.370159149169922, "learning_rate": 3.893598982619755e-05, "loss": 0.602, "mean_token_accuracy": 0.8611520528793335, "num_tokens": 69930258.0, "step": 1838 }, { "epoch": 0.23393970232794808, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.558010101318359, "learning_rate": 3.895718524798644e-05, "loss": 0.7912, "mean_token_accuracy": 0.8335659503936768, "num_tokens": 69960691.0, "step": 1839 }, { "epoch": 0.2340669126065386, "ewc_loss": 0.208984375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.00019550323486328125, "grad_norm": 5.2995285987854, "learning_rate": 3.8978380669775336e-05, "loss": 0.6647, "mean_token_accuracy": 0.8473352789878845, "num_tokens": 69998503.0, "step": 1840 }, { "epoch": 0.2341941228851291, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.701127052307129, "learning_rate": 3.899957609156422e-05, "loss": 0.7534, "mean_token_accuracy": 0.8476781845092773, "num_tokens": 70032381.0, "step": 1841 }, { "epoch": 0.23432133316371964, "ewc_loss": 0.2275390625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.000213623046875, "grad_norm": 5.418070316314697, "learning_rate": 3.902077151335312e-05, "loss": 0.6448, "mean_token_accuracy": 0.8596328496932983, "num_tokens": 70070616.0, "step": 1842 }, { "epoch": 0.23444854344231014, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.618498802185059, "learning_rate": 3.904196693514201e-05, "loss": 0.7358, "mean_token_accuracy": 0.8523759841918945, "num_tokens": 70111709.0, "step": 1843 }, { "epoch": 0.23457575372090064, "ewc_loss": 0.251953125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.0002384185791015625, "grad_norm": 5.702011585235596, "learning_rate": 3.906316235693091e-05, "loss": 0.7362, "mean_token_accuracy": 0.8376317024230957, "num_tokens": 70148516.0, "step": 1844 }, { "epoch": 0.23470296399949117, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.214147090911865, "learning_rate": 3.90843577787198e-05, "loss": 0.7234, "mean_token_accuracy": 0.8494566679000854, "num_tokens": 70185688.0, "step": 1845 }, { "epoch": 0.23483017427808167, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 5.988809108734131, "learning_rate": 3.9105553200508696e-05, "loss": 0.7286, "mean_token_accuracy": 0.8434057235717773, "num_tokens": 70228901.0, "step": 1846 }, { "epoch": 0.23495738455667217, "ewc_loss": 0.2353515625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 0.00022125244140625, "grad_norm": 5.713308811187744, "learning_rate": 3.912674862229759e-05, "loss": 0.6816, "mean_token_accuracy": 0.8531845808029175, "num_tokens": 70268341.0, "step": 1847 }, { "epoch": 0.2350845948352627, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.237487316131592, "learning_rate": 3.914794404408648e-05, "loss": 0.7069, "mean_token_accuracy": 0.8510821461677551, "num_tokens": 70309285.0, "step": 1848 }, { "epoch": 0.2352118051138532, "ewc_loss": 0.203125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 0.00018978118896484375, "grad_norm": 5.387398719787598, "learning_rate": 3.916913946587537e-05, "loss": 0.682, "mean_token_accuracy": 0.8374178409576416, "num_tokens": 70339694.0, "step": 1849 }, { "epoch": 0.2353390153924437, "ewc_loss": 0.265625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.3872904777526855, "learning_rate": 3.919033488766427e-05, "loss": 0.7035, "mean_token_accuracy": 0.8539937734603882, "num_tokens": 70377684.0, "step": 1850 }, { "epoch": 0.23546622567103423, "ewc_loss": 0.1962890625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 0.00018215179443359375, "grad_norm": 5.257950305938721, "learning_rate": 3.921153030945316e-05, "loss": 0.6717, "mean_token_accuracy": 0.843896746635437, "num_tokens": 70417690.0, "step": 1851 }, { "epoch": 0.23559343594962473, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.506834506988525, "learning_rate": 3.9232725731242056e-05, "loss": 0.7364, "mean_token_accuracy": 0.8437464833259583, "num_tokens": 70454485.0, "step": 1852 }, { "epoch": 0.23572064622821523, "ewc_loss": 0.189453125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00017547607421875, "grad_norm": 5.219038963317871, "learning_rate": 3.925392115303095e-05, "loss": 0.6772, "mean_token_accuracy": 0.8422325253486633, "num_tokens": 70490337.0, "step": 1853 }, { "epoch": 0.23584785650680576, "ewc_loss": 0.28125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.586365699768066, "learning_rate": 3.9275116574819844e-05, "loss": 0.7477, "mean_token_accuracy": 0.8483240604400635, "num_tokens": 70535194.0, "step": 1854 }, { "epoch": 0.23597506678539626, "ewc_loss": 0.2001953125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.000186920166015625, "grad_norm": 5.288898944854736, "learning_rate": 3.9296311996608736e-05, "loss": 0.6727, "mean_token_accuracy": 0.8428910374641418, "num_tokens": 70577046.0, "step": 1855 }, { "epoch": 0.23610227706398676, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.547066688537598, "learning_rate": 3.9317507418397627e-05, "loss": 0.7122, "mean_token_accuracy": 0.8559960126876831, "num_tokens": 70616483.0, "step": 1856 }, { "epoch": 0.2362294873425773, "ewc_loss": 0.2236328125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.000209808349609375, "grad_norm": 5.467652320861816, "learning_rate": 3.933870284018652e-05, "loss": 0.6506, "mean_token_accuracy": 0.8538780212402344, "num_tokens": 70654073.0, "step": 1857 }, { "epoch": 0.2363566976211678, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.32733678817749, "learning_rate": 3.9359898261975415e-05, "loss": 0.7275, "mean_token_accuracy": 0.8461762070655823, "num_tokens": 70687530.0, "step": 1858 }, { "epoch": 0.2364839078997583, "ewc_loss": 0.2373046875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00022411346435546875, "grad_norm": 5.692169189453125, "learning_rate": 3.9381093683764306e-05, "loss": 0.6453, "mean_token_accuracy": 0.8663413524627686, "num_tokens": 70729156.0, "step": 1859 }, { "epoch": 0.23661111817834882, "ewc_loss": 0.248046875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 6.045773983001709, "learning_rate": 3.9402289105553204e-05, "loss": 0.7283, "mean_token_accuracy": 0.8412741422653198, "num_tokens": 70769781.0, "step": 1860 }, { "epoch": 0.23673832845693932, "ewc_loss": 0.236328125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00022220611572265625, "grad_norm": 5.840644836425781, "learning_rate": 3.9423484527342095e-05, "loss": 0.6508, "mean_token_accuracy": 0.8628590106964111, "num_tokens": 70811822.0, "step": 1861 }, { "epoch": 0.23686553873552982, "ewc_loss": 0.2294921875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 5.866498947143555, "learning_rate": 3.944467994913099e-05, "loss": 0.6926, "mean_token_accuracy": 0.8441835641860962, "num_tokens": 70850470.0, "step": 1862 }, { "epoch": 0.23699274901412035, "ewc_loss": 0.22265625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00020885467529296875, "grad_norm": 5.858617305755615, "learning_rate": 3.9465875370919884e-05, "loss": 0.6754, "mean_token_accuracy": 0.8486534357070923, "num_tokens": 70889270.0, "step": 1863 }, { "epoch": 0.23711995929271085, "ewc_loss": 0.2197265625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00020599365234375, "grad_norm": 5.853806495666504, "learning_rate": 3.9487070792708775e-05, "loss": 0.6625, "mean_token_accuracy": 0.8516378402709961, "num_tokens": 70928787.0, "step": 1864 }, { "epoch": 0.23724716957130135, "ewc_loss": 0.2099609375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0001964569091796875, "grad_norm": 5.741548538208008, "learning_rate": 3.9508266214497666e-05, "loss": 0.6854, "mean_token_accuracy": 0.8462512493133545, "num_tokens": 70965726.0, "step": 1865 }, { "epoch": 0.23737437984989188, "ewc_loss": 0.2275390625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.000213623046875, "grad_norm": 6.010528087615967, "learning_rate": 3.9529461636286564e-05, "loss": 0.6713, "mean_token_accuracy": 0.8488229513168335, "num_tokens": 71002603.0, "step": 1866 }, { "epoch": 0.23750159012848238, "ewc_loss": 0.1943359375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00018024444580078125, "grad_norm": 5.503372669219971, "learning_rate": 3.9550657058075455e-05, "loss": 0.6307, "mean_token_accuracy": 0.8580799698829651, "num_tokens": 71039930.0, "step": 1867 }, { "epoch": 0.2376288004070729, "ewc_loss": 0.2392578125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00022602081298828125, "grad_norm": 6.300936222076416, "learning_rate": 3.957185247986435e-05, "loss": 0.6821, "mean_token_accuracy": 0.8524031043052673, "num_tokens": 71077789.0, "step": 1868 }, { "epoch": 0.2377560106856634, "ewc_loss": 0.1796875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0001659393310546875, "grad_norm": 5.192333698272705, "learning_rate": 3.9593047901653244e-05, "loss": 0.6652, "mean_token_accuracy": 0.8396304845809937, "num_tokens": 71121325.0, "step": 1869 }, { "epoch": 0.2378832209642539, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.715717792510986, "learning_rate": 3.961424332344214e-05, "loss": 0.7526, "mean_token_accuracy": 0.8438361883163452, "num_tokens": 71159764.0, "step": 1870 }, { "epoch": 0.23801043124284443, "ewc_loss": 0.169921875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.000156402587890625, "grad_norm": 4.878870010375977, "learning_rate": 3.963543874523103e-05, "loss": 0.56, "mean_token_accuracy": 0.8714207410812378, "num_tokens": 71196559.0, "step": 1871 }, { "epoch": 0.23813764152143493, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.248307228088379, "learning_rate": 3.9656634167019924e-05, "loss": 0.7391, "mean_token_accuracy": 0.8536204099655151, "num_tokens": 71229925.0, "step": 1872 }, { "epoch": 0.23826485180002543, "ewc_loss": 0.1845703125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00017070770263671875, "grad_norm": 4.840546607971191, "learning_rate": 3.9677829588808815e-05, "loss": 0.6196, "mean_token_accuracy": 0.8553577661514282, "num_tokens": 71273764.0, "step": 1873 }, { "epoch": 0.23839206207861596, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.59213924407959, "learning_rate": 3.969902501059771e-05, "loss": 0.8196, "mean_token_accuracy": 0.8460460901260376, "num_tokens": 71309090.0, "step": 1874 }, { "epoch": 0.23851927235720646, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.49994421005249, "learning_rate": 3.9720220432386604e-05, "loss": 0.7303, "mean_token_accuracy": 0.8483177423477173, "num_tokens": 71348867.0, "step": 1875 }, { "epoch": 0.23864648263579696, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.126723289489746, "learning_rate": 3.97414158541755e-05, "loss": 0.737, "mean_token_accuracy": 0.8620845079421997, "num_tokens": 71385707.0, "step": 1876 }, { "epoch": 0.2387736929143875, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.409384727478027, "learning_rate": 3.976261127596439e-05, "loss": 0.7127, "mean_token_accuracy": 0.8712900876998901, "num_tokens": 71428516.0, "step": 1877 }, { "epoch": 0.238900903192978, "ewc_loss": 0.25, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00023746490478515625, "grad_norm": 5.7945637702941895, "learning_rate": 3.978380669775329e-05, "loss": 0.6518, "mean_token_accuracy": 0.8645909428596497, "num_tokens": 71466913.0, "step": 1878 }, { "epoch": 0.2390281134715685, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.376438140869141, "learning_rate": 3.980500211954218e-05, "loss": 0.8128, "mean_token_accuracy": 0.8534572124481201, "num_tokens": 71506362.0, "step": 1879 }, { "epoch": 0.23915532375015902, "ewc_loss": 0.197265625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.00018310546875, "grad_norm": 4.873411178588867, "learning_rate": 3.982619754133107e-05, "loss": 0.6214, "mean_token_accuracy": 0.8577910661697388, "num_tokens": 71544303.0, "step": 1880 }, { "epoch": 0.23928253402874952, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.538748741149902, "learning_rate": 3.984739296311997e-05, "loss": 0.8683, "mean_token_accuracy": 0.8622965812683105, "num_tokens": 71582295.0, "step": 1881 }, { "epoch": 0.23940974430734002, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 5.876430034637451, "learning_rate": 3.986858838490886e-05, "loss": 0.7539, "mean_token_accuracy": 0.857134222984314, "num_tokens": 71621880.0, "step": 1882 }, { "epoch": 0.23953695458593055, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.645208358764648, "learning_rate": 3.988978380669776e-05, "loss": 0.8585, "mean_token_accuracy": 0.8634289503097534, "num_tokens": 71657997.0, "step": 1883 }, { "epoch": 0.23966416486452105, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.181725978851318, "learning_rate": 3.991097922848665e-05, "loss": 1.0242, "mean_token_accuracy": 0.8220566511154175, "num_tokens": 71697622.0, "step": 1884 }, { "epoch": 0.23979137514311155, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.662613391876221, "learning_rate": 3.993217465027554e-05, "loss": 0.7628, "mean_token_accuracy": 0.86280357837677, "num_tokens": 71733354.0, "step": 1885 }, { "epoch": 0.23991858542170208, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.056562900543213, "learning_rate": 3.995337007206443e-05, "loss": 0.8646, "mean_token_accuracy": 0.8478495478630066, "num_tokens": 71777172.0, "step": 1886 }, { "epoch": 0.24004579570029258, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.9756340980529785, "learning_rate": 3.997456549385333e-05, "loss": 0.784, "mean_token_accuracy": 0.8487905263900757, "num_tokens": 71816165.0, "step": 1887 }, { "epoch": 0.24017300597888308, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.655295372009277, "learning_rate": 3.999576091564222e-05, "loss": 0.7869, "mean_token_accuracy": 0.8439423441886902, "num_tokens": 71852313.0, "step": 1888 }, { "epoch": 0.2403002162574736, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.710516929626465, "learning_rate": 4.001695633743112e-05, "loss": 0.7466, "mean_token_accuracy": 0.8586346507072449, "num_tokens": 71892462.0, "step": 1889 }, { "epoch": 0.2404274265360641, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.092724323272705, "learning_rate": 4.003815175922001e-05, "loss": 0.7021, "mean_token_accuracy": 0.855563759803772, "num_tokens": 71924079.0, "step": 1890 }, { "epoch": 0.24055463681465464, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.86643648147583, "learning_rate": 4.005934718100891e-05, "loss": 0.7356, "mean_token_accuracy": 0.8577524423599243, "num_tokens": 71959546.0, "step": 1891 }, { "epoch": 0.24068184709324514, "ewc_loss": 0.2373046875, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 5.611138820648193, "learning_rate": 4.00805426027978e-05, "loss": 0.6893, "mean_token_accuracy": 0.8481364846229553, "num_tokens": 72002565.0, "step": 1892 }, { "epoch": 0.24080905737183564, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.2333574295043945, "learning_rate": 4.0101738024586697e-05, "loss": 0.7471, "mean_token_accuracy": 0.860283374786377, "num_tokens": 72037178.0, "step": 1893 }, { "epoch": 0.24093626765042617, "ewc_loss": 0.240234375, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.0002269744873046875, "grad_norm": 5.484077453613281, "learning_rate": 4.012293344637558e-05, "loss": 0.6753, "mean_token_accuracy": 0.8537773489952087, "num_tokens": 72076562.0, "step": 1894 }, { "epoch": 0.24106347792901667, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.542385578155518, "learning_rate": 4.014412886816448e-05, "loss": 0.8083, "mean_token_accuracy": 0.8564987182617188, "num_tokens": 72113379.0, "step": 1895 }, { "epoch": 0.24119068820760717, "ewc_loss": 0.28125, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 5.772640705108643, "learning_rate": 4.016532428995337e-05, "loss": 0.658, "mean_token_accuracy": 0.8715952634811401, "num_tokens": 72149225.0, "step": 1896 }, { "epoch": 0.2413178984861977, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.531059741973877, "learning_rate": 4.018651971174227e-05, "loss": 0.8032, "mean_token_accuracy": 0.8569719791412354, "num_tokens": 72185774.0, "step": 1897 }, { "epoch": 0.2414451087647882, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.214497089385986, "learning_rate": 4.020771513353116e-05, "loss": 0.7313, "mean_token_accuracy": 0.8599638938903809, "num_tokens": 72224270.0, "step": 1898 }, { "epoch": 0.2415723190433787, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.047105312347412, "learning_rate": 4.0228910555320056e-05, "loss": 0.7649, "mean_token_accuracy": 0.8549759387969971, "num_tokens": 72260629.0, "step": 1899 }, { "epoch": 0.24169952932196923, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.406036376953125, "learning_rate": 4.025010597710895e-05, "loss": 0.7917, "mean_token_accuracy": 0.8416387438774109, "num_tokens": 72300040.0, "step": 1900 }, { "epoch": 0.24182673960055973, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00013637542724609375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.4773712158203125, "learning_rate": 4.0271301398897845e-05, "loss": 0.7347, "mean_token_accuracy": 0.8538521528244019, "num_tokens": 72335396.0, "step": 1901 }, { "epoch": 0.24195394987915023, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.317740440368652, "learning_rate": 4.029249682068673e-05, "loss": 0.7665, "mean_token_accuracy": 0.8443719148635864, "num_tokens": 72375215.0, "step": 1902 }, { "epoch": 0.24208116015774075, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.23802375793457, "learning_rate": 4.031369224247563e-05, "loss": 0.6841, "mean_token_accuracy": 0.866040825843811, "num_tokens": 72416407.0, "step": 1903 }, { "epoch": 0.24220837043633126, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.047727584838867, "learning_rate": 4.033488766426452e-05, "loss": 0.6507, "mean_token_accuracy": 0.8705275058746338, "num_tokens": 72455031.0, "step": 1904 }, { "epoch": 0.24233558071492176, "ewc_loss": 0.28125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.320465564727783, "learning_rate": 4.0356083086053416e-05, "loss": 0.7356, "mean_token_accuracy": 0.850961446762085, "num_tokens": 72487186.0, "step": 1905 }, { "epoch": 0.24246279099351228, "ewc_loss": 0.23046875, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 0.00021648406982421875, "grad_norm": 5.7217912673950195, "learning_rate": 4.037727850784231e-05, "loss": 0.6532, "mean_token_accuracy": 0.857734203338623, "num_tokens": 72525659.0, "step": 1906 }, { "epoch": 0.24259000127210278, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.607905864715576, "learning_rate": 4.0398473929631205e-05, "loss": 0.7073, "mean_token_accuracy": 0.8616088628768921, "num_tokens": 72567912.0, "step": 1907 }, { "epoch": 0.24271721155069328, "ewc_loss": 0.2099609375, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 0.0001964569091796875, "grad_norm": 5.442991733551025, "learning_rate": 4.0419669351420096e-05, "loss": 0.5871, "mean_token_accuracy": 0.8722858428955078, "num_tokens": 72599525.0, "step": 1908 }, { "epoch": 0.2428444218292838, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.973028659820557, "learning_rate": 4.044086477320899e-05, "loss": 0.7397, "mean_token_accuracy": 0.8539592623710632, "num_tokens": 72635391.0, "step": 1909 }, { "epoch": 0.2429716321078743, "ewc_loss": 0.21484375, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 0.00020122528076171875, "grad_norm": 5.358060359954834, "learning_rate": 4.046206019499788e-05, "loss": 0.6857, "mean_token_accuracy": 0.8423663973808289, "num_tokens": 72666058.0, "step": 1910 }, { "epoch": 0.24309884238646481, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.181156635284424, "learning_rate": 4.0483255616786776e-05, "loss": 0.8035, "mean_token_accuracy": 0.8386036157608032, "num_tokens": 72699869.0, "step": 1911 }, { "epoch": 0.24322605266505534, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 0.000244140625, "grad_norm": 5.650754451751709, "learning_rate": 4.050445103857567e-05, "loss": 0.7368, "mean_token_accuracy": 0.8432103395462036, "num_tokens": 72735660.0, "step": 1912 }, { "epoch": 0.24335326294364584, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.938732624053955, "learning_rate": 4.0525646460364565e-05, "loss": 0.7629, "mean_token_accuracy": 0.8528144955635071, "num_tokens": 72774334.0, "step": 1913 }, { "epoch": 0.24348047322223634, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.139068126678467, "learning_rate": 4.0546841882153456e-05, "loss": 0.7586, "mean_token_accuracy": 0.8445344567298889, "num_tokens": 72808387.0, "step": 1914 }, { "epoch": 0.24360768350082687, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.229714870452881, "learning_rate": 4.0568037303942353e-05, "loss": 0.6885, "mean_token_accuracy": 0.8610544800758362, "num_tokens": 72846144.0, "step": 1915 }, { "epoch": 0.24373489377941737, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.507037162780762, "learning_rate": 4.0589232725731245e-05, "loss": 0.6781, "mean_token_accuracy": 0.868100643157959, "num_tokens": 72887381.0, "step": 1916 }, { "epoch": 0.2438621040580079, "ewc_loss": 0.2314453125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000217437744140625, "grad_norm": 5.609182357788086, "learning_rate": 4.0610428147520136e-05, "loss": 0.7005, "mean_token_accuracy": 0.8414677977561951, "num_tokens": 72927126.0, "step": 1917 }, { "epoch": 0.2439893143365984, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.795751094818115, "learning_rate": 4.0631623569309027e-05, "loss": 0.7433, "mean_token_accuracy": 0.8513727188110352, "num_tokens": 72968770.0, "step": 1918 }, { "epoch": 0.2441165246151889, "ewc_loss": 0.2060546875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0001926422119140625, "grad_norm": 5.333927631378174, "learning_rate": 4.0652818991097924e-05, "loss": 0.6313, "mean_token_accuracy": 0.8608114719390869, "num_tokens": 73008831.0, "step": 1919 }, { "epoch": 0.24424373489377943, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.033118724822998, "learning_rate": 4.0674014412886815e-05, "loss": 0.7661, "mean_token_accuracy": 0.8513106107711792, "num_tokens": 73048469.0, "step": 1920 }, { "epoch": 0.24437094517236993, "ewc_loss": 0.21875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.00020503997802734375, "grad_norm": 5.326361179351807, "learning_rate": 4.069520983467571e-05, "loss": 0.6598, "mean_token_accuracy": 0.8534361124038696, "num_tokens": 73096067.0, "step": 1921 }, { "epoch": 0.24449815545096043, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.118109703063965, "learning_rate": 4.0716405256464604e-05, "loss": 0.8193, "mean_token_accuracy": 0.840559184551239, "num_tokens": 73132680.0, "step": 1922 }, { "epoch": 0.24462536572955096, "ewc_loss": 0.248046875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 5.572113513946533, "learning_rate": 4.07376006782535e-05, "loss": 0.7029, "mean_token_accuracy": 0.8481537103652954, "num_tokens": 73168436.0, "step": 1923 }, { "epoch": 0.24475257600814146, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.925266742706299, "learning_rate": 4.075879610004239e-05, "loss": 0.7475, "mean_token_accuracy": 0.8533539772033691, "num_tokens": 73210961.0, "step": 1924 }, { "epoch": 0.24487978628673196, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.929389476776123, "learning_rate": 4.0779991521831284e-05, "loss": 0.7081, "mean_token_accuracy": 0.8588886260986328, "num_tokens": 73244580.0, "step": 1925 }, { "epoch": 0.2450069965653225, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.476056098937988, "learning_rate": 4.080118694362018e-05, "loss": 0.7695, "mean_token_accuracy": 0.8421013951301575, "num_tokens": 73278176.0, "step": 1926 }, { "epoch": 0.245134206843913, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.142721176147461, "learning_rate": 4.082238236540907e-05, "loss": 0.7241, "mean_token_accuracy": 0.8504091501235962, "num_tokens": 73315678.0, "step": 1927 }, { "epoch": 0.2452614171225035, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.164407730102539, "learning_rate": 4.084357778719797e-05, "loss": 0.6723, "mean_token_accuracy": 0.8658024072647095, "num_tokens": 73356336.0, "step": 1928 }, { "epoch": 0.24538862740109402, "ewc_loss": 0.265625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.114293575286865, "learning_rate": 4.086477320898686e-05, "loss": 0.6839, "mean_token_accuracy": 0.8579060435295105, "num_tokens": 73396605.0, "step": 1929 }, { "epoch": 0.24551583767968452, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.185758590698242, "learning_rate": 4.088596863077575e-05, "loss": 0.7876, "mean_token_accuracy": 0.8286916613578796, "num_tokens": 73444278.0, "step": 1930 }, { "epoch": 0.24564304795827502, "ewc_loss": 0.2314453125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000217437744140625, "grad_norm": 5.860301971435547, "learning_rate": 4.090716405256465e-05, "loss": 0.7164, "mean_token_accuracy": 0.8413264751434326, "num_tokens": 73483755.0, "step": 1931 }, { "epoch": 0.24577025823686555, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.4463276863098145, "learning_rate": 4.092835947435354e-05, "loss": 0.7665, "mean_token_accuracy": 0.8368735909461975, "num_tokens": 73522819.0, "step": 1932 }, { "epoch": 0.24589746851545605, "ewc_loss": 0.2060546875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0001926422119140625, "grad_norm": 5.52638053894043, "learning_rate": 4.094955489614243e-05, "loss": 0.6088, "mean_token_accuracy": 0.8619458675384521, "num_tokens": 73554973.0, "step": 1933 }, { "epoch": 0.24602467879404655, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.822849273681641, "learning_rate": 4.097075031793133e-05, "loss": 0.7809, "mean_token_accuracy": 0.8390415906906128, "num_tokens": 73592541.0, "step": 1934 }, { "epoch": 0.24615188907263708, "ewc_loss": 0.1943359375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.00018024444580078125, "grad_norm": 5.251867294311523, "learning_rate": 4.099194573972022e-05, "loss": 0.6691, "mean_token_accuracy": 0.8405320048332214, "num_tokens": 73629328.0, "step": 1935 }, { "epoch": 0.24627909935122758, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.174074172973633, "learning_rate": 4.101314116150912e-05, "loss": 0.7641, "mean_token_accuracy": 0.8519948720932007, "num_tokens": 73664944.0, "step": 1936 }, { "epoch": 0.24640630962981808, "ewc_loss": 0.2119140625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0001983642578125, "grad_norm": 5.256043910980225, "learning_rate": 4.103433658329801e-05, "loss": 0.6956, "mean_token_accuracy": 0.8386555910110474, "num_tokens": 73712403.0, "step": 1937 }, { "epoch": 0.2465335199084086, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.275559425354004, "learning_rate": 4.105553200508691e-05, "loss": 0.6914, "mean_token_accuracy": 0.8735408186912537, "num_tokens": 73752716.0, "step": 1938 }, { "epoch": 0.2466607301869991, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.646552562713623, "learning_rate": 4.10767274268758e-05, "loss": 0.7507, "mean_token_accuracy": 0.8379923105239868, "num_tokens": 73796900.0, "step": 1939 }, { "epoch": 0.2467879404655896, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.891220569610596, "learning_rate": 4.109792284866469e-05, "loss": 0.774, "mean_token_accuracy": 0.843507707118988, "num_tokens": 73836426.0, "step": 1940 }, { "epoch": 0.24691515074418013, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.192695617675781, "learning_rate": 4.111911827045358e-05, "loss": 0.6922, "mean_token_accuracy": 0.8632107973098755, "num_tokens": 73872147.0, "step": 1941 }, { "epoch": 0.24704236102277063, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.139137268066406, "learning_rate": 4.114031369224248e-05, "loss": 0.7077, "mean_token_accuracy": 0.8525372743606567, "num_tokens": 73912878.0, "step": 1942 }, { "epoch": 0.24716957130136116, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.6430535316467285, "learning_rate": 4.116150911403137e-05, "loss": 0.7599, "mean_token_accuracy": 0.8481416702270508, "num_tokens": 73952152.0, "step": 1943 }, { "epoch": 0.24729678157995166, "ewc_loss": 0.2275390625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000213623046875, "grad_norm": 5.557989597320557, "learning_rate": 4.118270453582027e-05, "loss": 0.6759, "mean_token_accuracy": 0.8528053760528564, "num_tokens": 73988760.0, "step": 1944 }, { "epoch": 0.24742399185854216, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.966071605682373, "learning_rate": 4.120389995760916e-05, "loss": 0.7253, "mean_token_accuracy": 0.8568629622459412, "num_tokens": 74028107.0, "step": 1945 }, { "epoch": 0.2475512021371327, "ewc_loss": 0.212890625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.00019931793212890625, "grad_norm": 5.3436994552612305, "learning_rate": 4.122509537939806e-05, "loss": 0.6784, "mean_token_accuracy": 0.8452647924423218, "num_tokens": 74071236.0, "step": 1946 }, { "epoch": 0.2476784124157232, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.1686835289001465, "learning_rate": 4.124629080118694e-05, "loss": 0.782, "mean_token_accuracy": 0.8521232604980469, "num_tokens": 74105734.0, "step": 1947 }, { "epoch": 0.2478056226943137, "ewc_loss": 0.234375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.00022029876708984375, "grad_norm": 5.396020889282227, "learning_rate": 4.126748622297584e-05, "loss": 0.6983, "mean_token_accuracy": 0.8457022905349731, "num_tokens": 74142569.0, "step": 1948 }, { "epoch": 0.24793283297290422, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.221235275268555, "learning_rate": 4.128868164476473e-05, "loss": 0.8152, "mean_token_accuracy": 0.8378961086273193, "num_tokens": 74185395.0, "step": 1949 }, { "epoch": 0.24806004325149472, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 5.657128810882568, "learning_rate": 4.130987706655363e-05, "loss": 0.7093, "mean_token_accuracy": 0.852049708366394, "num_tokens": 74223416.0, "step": 1950 }, { "epoch": 0.24818725353008522, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.029721260070801, "learning_rate": 4.133107248834252e-05, "loss": 0.7366, "mean_token_accuracy": 0.8640007376670837, "num_tokens": 74265356.0, "step": 1951 }, { "epoch": 0.24831446380867575, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 5.9847869873046875, "learning_rate": 4.135226791013142e-05, "loss": 0.7605, "mean_token_accuracy": 0.8425804376602173, "num_tokens": 74303807.0, "step": 1952 }, { "epoch": 0.24844167408726625, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.646849632263184, "learning_rate": 4.137346333192031e-05, "loss": 0.7278, "mean_token_accuracy": 0.8580813407897949, "num_tokens": 74337299.0, "step": 1953 }, { "epoch": 0.24856888436585675, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.156889915466309, "learning_rate": 4.1394658753709205e-05, "loss": 0.7314, "mean_token_accuracy": 0.8486003875732422, "num_tokens": 74370817.0, "step": 1954 }, { "epoch": 0.24869609464444728, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.38855504989624, "learning_rate": 4.141585417549809e-05, "loss": 0.7203, "mean_token_accuracy": 0.8592066168785095, "num_tokens": 74408009.0, "step": 1955 }, { "epoch": 0.24882330492303778, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.080251693725586, "learning_rate": 4.143704959728699e-05, "loss": 0.6848, "mean_token_accuracy": 0.860326886177063, "num_tokens": 74446669.0, "step": 1956 }, { "epoch": 0.24895051520162828, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.431817531585693, "learning_rate": 4.145824501907588e-05, "loss": 0.74, "mean_token_accuracy": 0.8493622541427612, "num_tokens": 74484264.0, "step": 1957 }, { "epoch": 0.2490777254802188, "ewc_loss": 0.2470703125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.00023365020751953125, "grad_norm": 5.833654403686523, "learning_rate": 4.1479440440864776e-05, "loss": 0.7244, "mean_token_accuracy": 0.845927357673645, "num_tokens": 74523956.0, "step": 1958 }, { "epoch": 0.2492049357588093, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.674300670623779, "learning_rate": 4.150063586265367e-05, "loss": 0.6646, "mean_token_accuracy": 0.8712863922119141, "num_tokens": 74560748.0, "step": 1959 }, { "epoch": 0.2493321460373998, "ewc_loss": 0.2275390625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000213623046875, "grad_norm": 5.574828624725342, "learning_rate": 4.1521831284442565e-05, "loss": 0.6527, "mean_token_accuracy": 0.857277512550354, "num_tokens": 74596465.0, "step": 1960 }, { "epoch": 0.24945935631599034, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.911744117736816, "learning_rate": 4.1543026706231456e-05, "loss": 0.7302, "mean_token_accuracy": 0.855077862739563, "num_tokens": 74632788.0, "step": 1961 }, { "epoch": 0.24958656659458084, "ewc_loss": 0.2236328125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000209808349609375, "grad_norm": 5.477980136871338, "learning_rate": 4.156422212802035e-05, "loss": 0.6568, "mean_token_accuracy": 0.8605647683143616, "num_tokens": 74667723.0, "step": 1962 }, { "epoch": 0.24971377687317134, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.038384914398193, "learning_rate": 4.158541754980924e-05, "loss": 0.7693, "mean_token_accuracy": 0.8496191501617432, "num_tokens": 74709205.0, "step": 1963 }, { "epoch": 0.24984098715176187, "ewc_loss": 0.2451171875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 5.58220100402832, "learning_rate": 4.1606612971598136e-05, "loss": 0.6514, "mean_token_accuracy": 0.8611999750137329, "num_tokens": 74749437.0, "step": 1964 }, { "epoch": 0.24996819743035237, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.950033664703369, "learning_rate": 4.162780839338703e-05, "loss": 0.7898, "mean_token_accuracy": 0.8401029109954834, "num_tokens": 74790602.0, "step": 1965 }, { "epoch": 0.2500954077089429, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.88727331161499, "learning_rate": 4.1649003815175925e-05, "loss": 0.7489, "mean_token_accuracy": 0.8414272665977478, "num_tokens": 74830845.0, "step": 1966 }, { "epoch": 0.25022261798753337, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.544372081756592, "learning_rate": 4.1670199236964816e-05, "loss": 0.7201, "mean_token_accuracy": 0.853415846824646, "num_tokens": 74864208.0, "step": 1967 }, { "epoch": 0.2503498282661239, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.246232509613037, "learning_rate": 4.1691394658753714e-05, "loss": 0.6848, "mean_token_accuracy": 0.8663380742073059, "num_tokens": 74897217.0, "step": 1968 }, { "epoch": 0.2504770385447144, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.00024318695068359375, "grad_norm": 6.022793292999268, "learning_rate": 4.1712590080542605e-05, "loss": 0.6843, "mean_token_accuracy": 0.8545360565185547, "num_tokens": 74932105.0, "step": 1969 }, { "epoch": 0.2506042488233049, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.559690952301025, "learning_rate": 4.1733785502331496e-05, "loss": 0.7127, "mean_token_accuracy": 0.8587974309921265, "num_tokens": 74965229.0, "step": 1970 }, { "epoch": 0.2507314591018954, "ewc_loss": 0.224609375, "ewc_loss_diag": 0.0001392364501953125, "ewc_loss_parallel": 0.00021076202392578125, "grad_norm": 5.662161350250244, "learning_rate": 4.1754980924120394e-05, "loss": 0.6988, "mean_token_accuracy": 0.842410147190094, "num_tokens": 75002290.0, "step": 1971 }, { "epoch": 0.25085866938048595, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0001392364501953125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.7894439697265625, "learning_rate": 4.1776176345909285e-05, "loss": 0.7411, "mean_token_accuracy": 0.8473273515701294, "num_tokens": 75038113.0, "step": 1972 }, { "epoch": 0.2509858796590764, "ewc_loss": 0.2099609375, "ewc_loss_diag": 0.0001392364501953125, "ewc_loss_parallel": 0.0001964569091796875, "grad_norm": 5.466750144958496, "learning_rate": 4.1797371767698176e-05, "loss": 0.6292, "mean_token_accuracy": 0.8590748310089111, "num_tokens": 75078145.0, "step": 1973 }, { "epoch": 0.25111308993766696, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0001392364501953125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.989033222198486, "learning_rate": 4.1818567189487074e-05, "loss": 0.8019, "mean_token_accuracy": 0.8376729488372803, "num_tokens": 75126473.0, "step": 1974 }, { "epoch": 0.2512403002162575, "ewc_loss": 0.2099609375, "ewc_loss_diag": 0.0001392364501953125, "ewc_loss_parallel": 0.0001964569091796875, "grad_norm": 5.397793769836426, "learning_rate": 4.1839762611275965e-05, "loss": 0.6527, "mean_token_accuracy": 0.8520441055297852, "num_tokens": 75166785.0, "step": 1975 }, { "epoch": 0.25136751049484796, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0001392364501953125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.13400411605835, "learning_rate": 4.186095803306486e-05, "loss": 0.7556, "mean_token_accuracy": 0.8506988286972046, "num_tokens": 75204599.0, "step": 1976 }, { "epoch": 0.2514947207734385, "ewc_loss": 0.234375, "ewc_loss_diag": 0.0001392364501953125, "ewc_loss_parallel": 0.00022029876708984375, "grad_norm": 5.48604679107666, "learning_rate": 4.1882153454853753e-05, "loss": 0.6348, "mean_token_accuracy": 0.8620855808258057, "num_tokens": 75235295.0, "step": 1977 }, { "epoch": 0.251621931052029, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0001392364501953125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.08600378036499, "learning_rate": 4.1903348876642645e-05, "loss": 0.7189, "mean_token_accuracy": 0.8662269115447998, "num_tokens": 75275549.0, "step": 1978 }, { "epoch": 0.25174914133061954, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.763141632080078, "learning_rate": 4.192454429843154e-05, "loss": 0.6658, "mean_token_accuracy": 0.8657845258712769, "num_tokens": 75316920.0, "step": 1979 }, { "epoch": 0.25187635160921, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.757240295410156, "learning_rate": 4.194573972022043e-05, "loss": 0.7452, "mean_token_accuracy": 0.8529664874076843, "num_tokens": 75360091.0, "step": 1980 }, { "epoch": 0.25200356188780054, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.10940408706665, "learning_rate": 4.196693514200933e-05, "loss": 0.746, "mean_token_accuracy": 0.8445348143577576, "num_tokens": 75397123.0, "step": 1981 }, { "epoch": 0.25213077216639107, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.324727535247803, "learning_rate": 4.198813056379822e-05, "loss": 0.7661, "mean_token_accuracy": 0.836822509765625, "num_tokens": 75429983.0, "step": 1982 }, { "epoch": 0.25225798244498154, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.345532417297363, "learning_rate": 4.200932598558712e-05, "loss": 0.7001, "mean_token_accuracy": 0.8594199419021606, "num_tokens": 75460322.0, "step": 1983 }, { "epoch": 0.25238519272357207, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 0.00024127960205078125, "grad_norm": 6.041557312011719, "learning_rate": 4.203052140737601e-05, "loss": 0.6654, "mean_token_accuracy": 0.8597770929336548, "num_tokens": 75499733.0, "step": 1984 }, { "epoch": 0.2525124030021626, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.38631010055542, "learning_rate": 4.20517168291649e-05, "loss": 0.689, "mean_token_accuracy": 0.8576635122299194, "num_tokens": 75541182.0, "step": 1985 }, { "epoch": 0.2526396132807531, "ewc_loss": 0.2421875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002288818359375, "grad_norm": 6.028238773345947, "learning_rate": 4.207291225095379e-05, "loss": 0.6736, "mean_token_accuracy": 0.8543819189071655, "num_tokens": 75585052.0, "step": 1986 }, { "epoch": 0.2527668235593436, "ewc_loss": 0.2470703125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.00023365020751953125, "grad_norm": 6.245053768157959, "learning_rate": 4.209410767274269e-05, "loss": 0.7658, "mean_token_accuracy": 0.8324542045593262, "num_tokens": 75625208.0, "step": 1987 }, { "epoch": 0.25289403383793413, "ewc_loss": 0.24609375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 6.176143646240234, "learning_rate": 4.211530309453158e-05, "loss": 0.6883, "mean_token_accuracy": 0.8521174192428589, "num_tokens": 75665205.0, "step": 1988 }, { "epoch": 0.2530212441165246, "ewc_loss": 0.2294921875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 6.025415420532227, "learning_rate": 4.213649851632048e-05, "loss": 0.6638, "mean_token_accuracy": 0.8551101684570312, "num_tokens": 75708278.0, "step": 1989 }, { "epoch": 0.25314845439511513, "ewc_loss": 0.25390625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.00023937225341796875, "grad_norm": 6.379258155822754, "learning_rate": 4.215769393810937e-05, "loss": 0.711, "mean_token_accuracy": 0.8490166068077087, "num_tokens": 75750565.0, "step": 1990 }, { "epoch": 0.25327566467370566, "ewc_loss": 0.212890625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.00019931793212890625, "grad_norm": 5.793074607849121, "learning_rate": 4.217888935989827e-05, "loss": 0.634, "mean_token_accuracy": 0.8591524362564087, "num_tokens": 75790519.0, "step": 1991 }, { "epoch": 0.25340287495229613, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.616981506347656, "learning_rate": 4.220008478168716e-05, "loss": 0.7136, "mean_token_accuracy": 0.8532983064651489, "num_tokens": 75828256.0, "step": 1992 }, { "epoch": 0.25353008523088666, "ewc_loss": 0.1982421875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0001850128173828125, "grad_norm": 5.550656795501709, "learning_rate": 4.222128020347605e-05, "loss": 0.6621, "mean_token_accuracy": 0.8492082357406616, "num_tokens": 75869238.0, "step": 1993 }, { "epoch": 0.2536572955094772, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.941531658172607, "learning_rate": 4.224247562526494e-05, "loss": 0.7285, "mean_token_accuracy": 0.8505886793136597, "num_tokens": 75906776.0, "step": 1994 }, { "epoch": 0.25378450578806766, "ewc_loss": 0.1884765625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.00017452239990234375, "grad_norm": 5.263795852661133, "learning_rate": 4.226367104705384e-05, "loss": 0.572, "mean_token_accuracy": 0.8723487854003906, "num_tokens": 75940271.0, "step": 1995 }, { "epoch": 0.2539117160666582, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.420875072479248, "learning_rate": 4.228486646884273e-05, "loss": 0.7408, "mean_token_accuracy": 0.8593990206718445, "num_tokens": 75974536.0, "step": 1996 }, { "epoch": 0.2540389263452487, "ewc_loss": 0.1904296875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.00017642974853515625, "grad_norm": 5.023309707641602, "learning_rate": 4.230606189063163e-05, "loss": 0.6585, "mean_token_accuracy": 0.8448503017425537, "num_tokens": 76013809.0, "step": 1997 }, { "epoch": 0.2541661366238392, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.972316741943359, "learning_rate": 4.232725731242052e-05, "loss": 0.887, "mean_token_accuracy": 0.8387176990509033, "num_tokens": 76049559.0, "step": 1998 }, { "epoch": 0.2542933469024297, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 5.293262004852295, "learning_rate": 4.234845273420942e-05, "loss": 0.6933, "mean_token_accuracy": 0.8535811901092529, "num_tokens": 76087091.0, "step": 1999 }, { "epoch": 0.25442055718102025, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.08829402923584, "learning_rate": 4.23696481559983e-05, "loss": 0.8616, "mean_token_accuracy": 0.8594726324081421, "num_tokens": 76121466.0, "step": 2000 }, { "epoch": 0.2545477674596107, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.308533191680908, "learning_rate": 4.23908435777872e-05, "loss": 0.7977, "mean_token_accuracy": 0.8575440645217896, "num_tokens": 76158291.0, "step": 2001 }, { "epoch": 0.25467497773820125, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.073559284210205, "learning_rate": 4.241203899957609e-05, "loss": 0.7821, "mean_token_accuracy": 0.8501105308532715, "num_tokens": 76197350.0, "step": 2002 }, { "epoch": 0.2548021880167918, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.234945774078369, "learning_rate": 4.243323442136499e-05, "loss": 0.8629, "mean_token_accuracy": 0.8496795892715454, "num_tokens": 76235367.0, "step": 2003 }, { "epoch": 0.25492939829538225, "ewc_loss": 0.244140625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002307891845703125, "grad_norm": 5.593039035797119, "learning_rate": 4.245442984315388e-05, "loss": 0.7249, "mean_token_accuracy": 0.839918851852417, "num_tokens": 76270462.0, "step": 2004 }, { "epoch": 0.2550566085739728, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.39687442779541, "learning_rate": 4.247562526494278e-05, "loss": 0.9272, "mean_token_accuracy": 0.849376916885376, "num_tokens": 76315032.0, "step": 2005 }, { "epoch": 0.2551838188525633, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.3499956130981445, "learning_rate": 4.249682068673167e-05, "loss": 0.7116, "mean_token_accuracy": 0.8530658483505249, "num_tokens": 76356337.0, "step": 2006 }, { "epoch": 0.2553110291311538, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.251365661621094, "learning_rate": 4.2518016108520566e-05, "loss": 1.0261, "mean_token_accuracy": 0.8463360071182251, "num_tokens": 76391329.0, "step": 2007 }, { "epoch": 0.2554382394097443, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.023151397705078, "learning_rate": 4.253921153030945e-05, "loss": 0.9248, "mean_token_accuracy": 0.8551381230354309, "num_tokens": 76429015.0, "step": 2008 }, { "epoch": 0.25556544968833483, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.510970115661621, "learning_rate": 4.256040695209835e-05, "loss": 0.909, "mean_token_accuracy": 0.8545728921890259, "num_tokens": 76470343.0, "step": 2009 }, { "epoch": 0.2556926599669253, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.5175981521606445, "learning_rate": 4.258160237388724e-05, "loss": 0.973, "mean_token_accuracy": 0.8515013456344604, "num_tokens": 76505140.0, "step": 2010 }, { "epoch": 0.25581987024551583, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.537214279174805, "learning_rate": 4.260279779567614e-05, "loss": 0.7909, "mean_token_accuracy": 0.866570234298706, "num_tokens": 76541701.0, "step": 2011 }, { "epoch": 0.25594708052410636, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.312607288360596, "learning_rate": 4.262399321746503e-05, "loss": 0.8923, "mean_token_accuracy": 0.848954439163208, "num_tokens": 76580208.0, "step": 2012 }, { "epoch": 0.25607429080269684, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 6.962599754333496, "learning_rate": 4.2645188639253926e-05, "loss": 0.7847, "mean_token_accuracy": 0.8606154322624207, "num_tokens": 76617750.0, "step": 2013 }, { "epoch": 0.25620150108128736, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.842844009399414, "learning_rate": 4.266638406104282e-05, "loss": 0.7084, "mean_token_accuracy": 0.8689043521881104, "num_tokens": 76652678.0, "step": 2014 }, { "epoch": 0.2563287113598779, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.926811218261719, "learning_rate": 4.268757948283171e-05, "loss": 0.7628, "mean_token_accuracy": 0.8586385250091553, "num_tokens": 76695375.0, "step": 2015 }, { "epoch": 0.25645592163846836, "ewc_loss": 0.28125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.211130142211914, "learning_rate": 4.27087749046206e-05, "loss": 0.7156, "mean_token_accuracy": 0.8551825284957886, "num_tokens": 76728695.0, "step": 2016 }, { "epoch": 0.2565831319170589, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.187130451202393, "learning_rate": 4.2729970326409497e-05, "loss": 0.7494, "mean_token_accuracy": 0.863858163356781, "num_tokens": 76761081.0, "step": 2017 }, { "epoch": 0.2567103421956494, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.784834861755371, "learning_rate": 4.275116574819839e-05, "loss": 0.7632, "mean_token_accuracy": 0.832356333732605, "num_tokens": 76800335.0, "step": 2018 }, { "epoch": 0.2568375524742399, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.572723388671875, "learning_rate": 4.2772361169987285e-05, "loss": 0.84, "mean_token_accuracy": 0.8443397283554077, "num_tokens": 76839653.0, "step": 2019 }, { "epoch": 0.2569647627528304, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.663623809814453, "learning_rate": 4.2793556591776176e-05, "loss": 0.7695, "mean_token_accuracy": 0.8335468769073486, "num_tokens": 76876557.0, "step": 2020 }, { "epoch": 0.25709197303142095, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.8940110206604, "learning_rate": 4.2814752013565074e-05, "loss": 0.7964, "mean_token_accuracy": 0.8710734844207764, "num_tokens": 76910000.0, "step": 2021 }, { "epoch": 0.2572191833100114, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 5.945772171020508, "learning_rate": 4.2835947435353965e-05, "loss": 0.7877, "mean_token_accuracy": 0.845064103603363, "num_tokens": 76946861.0, "step": 2022 }, { "epoch": 0.25734639358860195, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.866593837738037, "learning_rate": 4.2857142857142856e-05, "loss": 0.904, "mean_token_accuracy": 0.8399056792259216, "num_tokens": 76990831.0, "step": 2023 }, { "epoch": 0.2574736038671925, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.414398193359375, "learning_rate": 4.2878338278931754e-05, "loss": 0.7516, "mean_token_accuracy": 0.8678069710731506, "num_tokens": 77033145.0, "step": 2024 }, { "epoch": 0.25760081414578295, "ewc_loss": 0.375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.355353832244873, "learning_rate": 4.2899533700720645e-05, "loss": 0.811, "mean_token_accuracy": 0.8562188148498535, "num_tokens": 77076621.0, "step": 2025 }, { "epoch": 0.2577280244243735, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.656548500061035, "learning_rate": 4.292072912250954e-05, "loss": 0.8206, "mean_token_accuracy": 0.8425263166427612, "num_tokens": 77115951.0, "step": 2026 }, { "epoch": 0.257855234702964, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.764493942260742, "learning_rate": 4.2941924544298434e-05, "loss": 0.7985, "mean_token_accuracy": 0.8498789668083191, "num_tokens": 77154674.0, "step": 2027 }, { "epoch": 0.25798244498155454, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.61234712600708, "learning_rate": 4.296311996608733e-05, "loss": 0.764, "mean_token_accuracy": 0.8518391847610474, "num_tokens": 77192970.0, "step": 2028 }, { "epoch": 0.258109655260145, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.54587459564209, "learning_rate": 4.298431538787622e-05, "loss": 0.7861, "mean_token_accuracy": 0.845916211605072, "num_tokens": 77231258.0, "step": 2029 }, { "epoch": 0.25823686553873554, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.378155708312988, "learning_rate": 4.3005510809665114e-05, "loss": 0.7453, "mean_token_accuracy": 0.8509554862976074, "num_tokens": 77269829.0, "step": 2030 }, { "epoch": 0.25836407581732607, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.5884246826171875, "learning_rate": 4.3026706231454005e-05, "loss": 0.7299, "mean_token_accuracy": 0.8555859327316284, "num_tokens": 77309611.0, "step": 2031 }, { "epoch": 0.25849128609591654, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.140389442443848, "learning_rate": 4.30479016532429e-05, "loss": 0.6951, "mean_token_accuracy": 0.8583890199661255, "num_tokens": 77341937.0, "step": 2032 }, { "epoch": 0.25861849637450707, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.775893211364746, "learning_rate": 4.3069097075031794e-05, "loss": 0.7933, "mean_token_accuracy": 0.8414671421051025, "num_tokens": 77380484.0, "step": 2033 }, { "epoch": 0.2587457066530976, "ewc_loss": 0.2421875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 0.0002288818359375, "grad_norm": 5.8654046058654785, "learning_rate": 4.309029249682069e-05, "loss": 0.7321, "mean_token_accuracy": 0.8394951820373535, "num_tokens": 77421220.0, "step": 2034 }, { "epoch": 0.25887291693168807, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.092634677886963, "learning_rate": 4.311148791860958e-05, "loss": 0.7529, "mean_token_accuracy": 0.8580359816551208, "num_tokens": 77465136.0, "step": 2035 }, { "epoch": 0.2590001272102786, "ewc_loss": 0.2353515625, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 0.00022029876708984375, "grad_norm": 5.632843017578125, "learning_rate": 4.313268334039848e-05, "loss": 0.6806, "mean_token_accuracy": 0.8517025113105774, "num_tokens": 77500120.0, "step": 2036 }, { "epoch": 0.2591273374888691, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.437082767486572, "learning_rate": 4.315387876218737e-05, "loss": 0.7743, "mean_token_accuracy": 0.8532572984695435, "num_tokens": 77532860.0, "step": 2037 }, { "epoch": 0.2592545477674596, "ewc_loss": 0.24609375, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 5.631192207336426, "learning_rate": 4.317507418397626e-05, "loss": 0.7148, "mean_token_accuracy": 0.8462505340576172, "num_tokens": 77569154.0, "step": 2038 }, { "epoch": 0.2593817580460501, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.542109966278076, "learning_rate": 4.3196269605765154e-05, "loss": 0.7711, "mean_token_accuracy": 0.863166093826294, "num_tokens": 77602918.0, "step": 2039 }, { "epoch": 0.25950896832464065, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 5.993585586547852, "learning_rate": 4.321746502755405e-05, "loss": 0.7099, "mean_token_accuracy": 0.8592077493667603, "num_tokens": 77637233.0, "step": 2040 }, { "epoch": 0.2596361786032311, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.176778316497803, "learning_rate": 4.323866044934294e-05, "loss": 0.8361, "mean_token_accuracy": 0.835409939289093, "num_tokens": 77678344.0, "step": 2041 }, { "epoch": 0.25976338888182166, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.476765155792236, "learning_rate": 4.325985587113184e-05, "loss": 0.7296, "mean_token_accuracy": 0.8602110147476196, "num_tokens": 77712706.0, "step": 2042 }, { "epoch": 0.2598905991604122, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.463479042053223, "learning_rate": 4.328105129292073e-05, "loss": 0.7758, "mean_token_accuracy": 0.8397307395935059, "num_tokens": 77749841.0, "step": 2043 }, { "epoch": 0.26001780943900266, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.871337890625, "learning_rate": 4.330224671470963e-05, "loss": 0.7888, "mean_token_accuracy": 0.8422061204910278, "num_tokens": 77788464.0, "step": 2044 }, { "epoch": 0.2601450197175932, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 0.00024127960205078125, "grad_norm": 5.883208751678467, "learning_rate": 4.332344213649852e-05, "loss": 0.7058, "mean_token_accuracy": 0.8479740619659424, "num_tokens": 77826493.0, "step": 2045 }, { "epoch": 0.2602722299961837, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.192863464355469, "learning_rate": 4.334463755828741e-05, "loss": 0.7299, "mean_token_accuracy": 0.8605341911315918, "num_tokens": 77857814.0, "step": 2046 }, { "epoch": 0.2603994402747742, "ewc_loss": 0.23828125, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 5.61816930770874, "learning_rate": 4.33658329800763e-05, "loss": 0.6958, "mean_token_accuracy": 0.8481028079986572, "num_tokens": 77891263.0, "step": 2047 }, { "epoch": 0.2605266505533647, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.443016052246094, "learning_rate": 4.33870284018652e-05, "loss": 0.7645, "mean_token_accuracy": 0.860826849937439, "num_tokens": 77931683.0, "step": 2048 }, { "epoch": 0.26065386083195524, "ewc_loss": 0.24609375, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 5.619599342346191, "learning_rate": 4.340822382365409e-05, "loss": 0.6958, "mean_token_accuracy": 0.8495939373970032, "num_tokens": 77965204.0, "step": 2049 }, { "epoch": 0.2607810711105457, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00014400482177734375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.518572807312012, "learning_rate": 4.342941924544299e-05, "loss": 0.74, "mean_token_accuracy": 0.870805025100708, "num_tokens": 78003427.0, "step": 2050 }, { "epoch": 0.26090828138913624, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.00014400482177734375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 5.866643905639648, "learning_rate": 4.345061466723188e-05, "loss": 0.7249, "mean_token_accuracy": 0.8530047535896301, "num_tokens": 78045524.0, "step": 2051 }, { "epoch": 0.26103549166772677, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00014400482177734375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.3209381103515625, "learning_rate": 4.347181008902078e-05, "loss": 0.7652, "mean_token_accuracy": 0.8568862080574036, "num_tokens": 78085485.0, "step": 2052 }, { "epoch": 0.26116270194631724, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00014400482177734375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.248323917388916, "learning_rate": 4.349300551080966e-05, "loss": 0.756, "mean_token_accuracy": 0.8471707105636597, "num_tokens": 78128212.0, "step": 2053 }, { "epoch": 0.26128991222490777, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.849809169769287, "learning_rate": 4.351420093259856e-05, "loss": 0.7265, "mean_token_accuracy": 0.8631443977355957, "num_tokens": 78170959.0, "step": 2054 }, { "epoch": 0.2614171225034983, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.489613056182861, "learning_rate": 4.353539635438745e-05, "loss": 0.7113, "mean_token_accuracy": 0.8631239533424377, "num_tokens": 78211181.0, "step": 2055 }, { "epoch": 0.2615443327820888, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.541004180908203, "learning_rate": 4.355659177617635e-05, "loss": 0.7901, "mean_token_accuracy": 0.8331944942474365, "num_tokens": 78246281.0, "step": 2056 }, { "epoch": 0.2616715430606793, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.414559841156006, "learning_rate": 4.357778719796524e-05, "loss": 0.7026, "mean_token_accuracy": 0.8597681522369385, "num_tokens": 78279604.0, "step": 2057 }, { "epoch": 0.26179875333926983, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.62578821182251, "learning_rate": 4.359898261975414e-05, "loss": 0.7868, "mean_token_accuracy": 0.8338832259178162, "num_tokens": 78317534.0, "step": 2058 }, { "epoch": 0.2619259636178603, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.078258514404297, "learning_rate": 4.362017804154303e-05, "loss": 0.6385, "mean_token_accuracy": 0.8720923662185669, "num_tokens": 78349744.0, "step": 2059 }, { "epoch": 0.26205317389645083, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.970366477966309, "learning_rate": 4.3641373463331926e-05, "loss": 0.7496, "mean_token_accuracy": 0.8522824048995972, "num_tokens": 78390841.0, "step": 2060 }, { "epoch": 0.26218038417504136, "ewc_loss": 0.236328125, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 0.00022125244140625, "grad_norm": 5.712494850158691, "learning_rate": 4.366256888512081e-05, "loss": 0.6789, "mean_token_accuracy": 0.8510465025901794, "num_tokens": 78425604.0, "step": 2061 }, { "epoch": 0.26230759445363183, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.347804546356201, "learning_rate": 4.368376430690971e-05, "loss": 0.8163, "mean_token_accuracy": 0.8370726108551025, "num_tokens": 78469444.0, "step": 2062 }, { "epoch": 0.26243480473222236, "ewc_loss": 0.236328125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.00022125244140625, "grad_norm": 5.543458938598633, "learning_rate": 4.37049597286986e-05, "loss": 0.6775, "mean_token_accuracy": 0.8533697128295898, "num_tokens": 78507931.0, "step": 2063 }, { "epoch": 0.2625620150108129, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.595863342285156, "learning_rate": 4.37261551504875e-05, "loss": 0.7977, "mean_token_accuracy": 0.8544437289237976, "num_tokens": 78550572.0, "step": 2064 }, { "epoch": 0.26268922528940336, "ewc_loss": 0.265625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 5.730433464050293, "learning_rate": 4.374735057227639e-05, "loss": 0.6802, "mean_token_accuracy": 0.8595488667488098, "num_tokens": 78586975.0, "step": 2065 }, { "epoch": 0.2628164355679939, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.445793628692627, "learning_rate": 4.3768545994065286e-05, "loss": 0.8688, "mean_token_accuracy": 0.833916425704956, "num_tokens": 78630535.0, "step": 2066 }, { "epoch": 0.2629436458465844, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.240220546722412, "learning_rate": 4.378974141585418e-05, "loss": 0.7161, "mean_token_accuracy": 0.8643602132797241, "num_tokens": 78672486.0, "step": 2067 }, { "epoch": 0.2630708561251749, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.777674198150635, "learning_rate": 4.381093683764307e-05, "loss": 0.7211, "mean_token_accuracy": 0.862358808517456, "num_tokens": 78710208.0, "step": 2068 }, { "epoch": 0.2631980664037654, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.788849830627441, "learning_rate": 4.3832132259431966e-05, "loss": 0.777, "mean_token_accuracy": 0.8462717533111572, "num_tokens": 78754233.0, "step": 2069 }, { "epoch": 0.26332527668235595, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.028831481933594, "learning_rate": 4.385332768122086e-05, "loss": 0.7571, "mean_token_accuracy": 0.8415359258651733, "num_tokens": 78794530.0, "step": 2070 }, { "epoch": 0.2634524869609464, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.231140613555908, "learning_rate": 4.3874523103009755e-05, "loss": 0.7345, "mean_token_accuracy": 0.8596284985542297, "num_tokens": 78828620.0, "step": 2071 }, { "epoch": 0.26357969723953695, "ewc_loss": 0.236328125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.00022125244140625, "grad_norm": 5.6270833015441895, "learning_rate": 4.3895718524798646e-05, "loss": 0.6442, "mean_token_accuracy": 0.8636859655380249, "num_tokens": 78867658.0, "step": 2072 }, { "epoch": 0.2637069075181275, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.5069684982299805, "learning_rate": 4.391691394658754e-05, "loss": 0.808, "mean_token_accuracy": 0.848665714263916, "num_tokens": 78903744.0, "step": 2073 }, { "epoch": 0.26383411779671795, "ewc_loss": 0.244140625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.00022983551025390625, "grad_norm": 5.610409259796143, "learning_rate": 4.3938109368376435e-05, "loss": 0.7218, "mean_token_accuracy": 0.8392381072044373, "num_tokens": 78934098.0, "step": 2074 }, { "epoch": 0.2639613280753085, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.5837531089782715, "learning_rate": 4.3959304790165326e-05, "loss": 0.7773, "mean_token_accuracy": 0.8595594167709351, "num_tokens": 78969473.0, "step": 2075 }, { "epoch": 0.264088538353899, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.827417850494385, "learning_rate": 4.398050021195422e-05, "loss": 0.6874, "mean_token_accuracy": 0.862093448638916, "num_tokens": 79008799.0, "step": 2076 }, { "epoch": 0.2642157486324895, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.431145668029785, "learning_rate": 4.4001695633743114e-05, "loss": 0.7818, "mean_token_accuracy": 0.8599154949188232, "num_tokens": 79042502.0, "step": 2077 }, { "epoch": 0.26434295891108, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.189108371734619, "learning_rate": 4.4022891055532006e-05, "loss": 0.7852, "mean_token_accuracy": 0.8363839387893677, "num_tokens": 79079641.0, "step": 2078 }, { "epoch": 0.26447016918967053, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.030313014984131, "learning_rate": 4.40440864773209e-05, "loss": 0.7703, "mean_token_accuracy": 0.8541828393936157, "num_tokens": 79116120.0, "step": 2079 }, { "epoch": 0.26459737946826106, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.457822799682617, "learning_rate": 4.4065281899109794e-05, "loss": 0.7244, "mean_token_accuracy": 0.8623097538948059, "num_tokens": 79150322.0, "step": 2080 }, { "epoch": 0.26472458974685154, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.68051815032959, "learning_rate": 4.408647732089869e-05, "loss": 0.7445, "mean_token_accuracy": 0.851995587348938, "num_tokens": 79184944.0, "step": 2081 }, { "epoch": 0.26485180002544206, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.476587295532227, "learning_rate": 4.410767274268758e-05, "loss": 0.8063, "mean_token_accuracy": 0.828421413898468, "num_tokens": 79226137.0, "step": 2082 }, { "epoch": 0.2649790103040326, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.651429176330566, "learning_rate": 4.412886816447648e-05, "loss": 0.7202, "mean_token_accuracy": 0.8593830466270447, "num_tokens": 79267721.0, "step": 2083 }, { "epoch": 0.26510622058262306, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.241634845733643, "learning_rate": 4.4150063586265365e-05, "loss": 0.6985, "mean_token_accuracy": 0.8589264154434204, "num_tokens": 79305355.0, "step": 2084 }, { "epoch": 0.2652334308612136, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.875977993011475, "learning_rate": 4.417125900805426e-05, "loss": 0.7588, "mean_token_accuracy": 0.8470683097839355, "num_tokens": 79349714.0, "step": 2085 }, { "epoch": 0.2653606411398041, "ewc_loss": 0.25390625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 0.0002384185791015625, "grad_norm": 5.9766645431518555, "learning_rate": 4.4192454429843154e-05, "loss": 0.6605, "mean_token_accuracy": 0.8648630380630493, "num_tokens": 79387208.0, "step": 2086 }, { "epoch": 0.2654878514183946, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.108743667602539, "learning_rate": 4.421364985163205e-05, "loss": 0.716, "mean_token_accuracy": 0.869857907295227, "num_tokens": 79424396.0, "step": 2087 }, { "epoch": 0.2656150616969851, "ewc_loss": 0.2412109375, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 0.0002269744873046875, "grad_norm": 5.797616958618164, "learning_rate": 4.423484527342094e-05, "loss": 0.7628, "mean_token_accuracy": 0.8338584899902344, "num_tokens": 79465105.0, "step": 2088 }, { "epoch": 0.26574227197557565, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.296853542327881, "learning_rate": 4.425604069520984e-05, "loss": 0.7431, "mean_token_accuracy": 0.8613845705986023, "num_tokens": 79504000.0, "step": 2089 }, { "epoch": 0.2658694822541661, "ewc_loss": 0.24609375, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 5.76545524597168, "learning_rate": 4.427723611699873e-05, "loss": 0.7119, "mean_token_accuracy": 0.8443475365638733, "num_tokens": 79538954.0, "step": 2090 }, { "epoch": 0.26599669253275665, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.367477893829346, "learning_rate": 4.429843153878762e-05, "loss": 0.8289, "mean_token_accuracy": 0.8388316631317139, "num_tokens": 79577264.0, "step": 2091 }, { "epoch": 0.2661239028113472, "ewc_loss": 0.265625, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 5.893038749694824, "learning_rate": 4.4319626960576514e-05, "loss": 0.701, "mean_token_accuracy": 0.8533452749252319, "num_tokens": 79612803.0, "step": 2092 }, { "epoch": 0.26625111308993765, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.230015277862549, "learning_rate": 4.434082238236541e-05, "loss": 0.721, "mean_token_accuracy": 0.8693751692771912, "num_tokens": 79651595.0, "step": 2093 }, { "epoch": 0.2663783233685282, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.172093391418457, "learning_rate": 4.43620178041543e-05, "loss": 0.7149, "mean_token_accuracy": 0.8584332466125488, "num_tokens": 79691587.0, "step": 2094 }, { "epoch": 0.2665055336471187, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.865037441253662, "learning_rate": 4.43832132259432e-05, "loss": 0.7878, "mean_token_accuracy": 0.8406003713607788, "num_tokens": 79729803.0, "step": 2095 }, { "epoch": 0.2666327439257092, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.506720542907715, "learning_rate": 4.440440864773209e-05, "loss": 0.6926, "mean_token_accuracy": 0.8683494329452515, "num_tokens": 79765866.0, "step": 2096 }, { "epoch": 0.2667599542042997, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.398200988769531, "learning_rate": 4.442560406952099e-05, "loss": 0.6872, "mean_token_accuracy": 0.8627288937568665, "num_tokens": 79808973.0, "step": 2097 }, { "epoch": 0.26688716448289024, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.819834232330322, "learning_rate": 4.444679949130988e-05, "loss": 0.8039, "mean_token_accuracy": 0.834691047668457, "num_tokens": 79843677.0, "step": 2098 }, { "epoch": 0.2670143747614807, "ewc_loss": 0.248046875, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.00023365020751953125, "grad_norm": 6.010499954223633, "learning_rate": 4.446799491309877e-05, "loss": 0.6689, "mean_token_accuracy": 0.8597451448440552, "num_tokens": 79882680.0, "step": 2099 }, { "epoch": 0.26714158504007124, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.097228527069092, "learning_rate": 4.448919033488766e-05, "loss": 0.7469, "mean_token_accuracy": 0.8550996780395508, "num_tokens": 79929691.0, "step": 2100 }, { "epoch": 0.26726879531866177, "ewc_loss": 0.2265625, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.0002117156982421875, "grad_norm": 5.769188404083252, "learning_rate": 4.451038575667656e-05, "loss": 0.6595, "mean_token_accuracy": 0.8546066284179688, "num_tokens": 79970515.0, "step": 2101 }, { "epoch": 0.26739600559725224, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.356279373168945, "learning_rate": 4.453158117846545e-05, "loss": 0.848, "mean_token_accuracy": 0.8228164911270142, "num_tokens": 80007550.0, "step": 2102 }, { "epoch": 0.26752321587584277, "ewc_loss": 0.2255859375, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.00021076202392578125, "grad_norm": 5.602734088897705, "learning_rate": 4.455277660025435e-05, "loss": 0.667, "mean_token_accuracy": 0.8509399890899658, "num_tokens": 80044106.0, "step": 2103 }, { "epoch": 0.2676504261544333, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.640655040740967, "learning_rate": 4.457397202204324e-05, "loss": 0.802, "mean_token_accuracy": 0.8520655632019043, "num_tokens": 80084429.0, "step": 2104 }, { "epoch": 0.26777763643302377, "ewc_loss": 0.251953125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.00023651123046875, "grad_norm": 5.582864284515381, "learning_rate": 4.459516744383214e-05, "loss": 0.6281, "mean_token_accuracy": 0.8705073595046997, "num_tokens": 80123317.0, "step": 2105 }, { "epoch": 0.2679048467116143, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.7570672035217285, "learning_rate": 4.461636286562102e-05, "loss": 0.8572, "mean_token_accuracy": 0.8481703996658325, "num_tokens": 80157658.0, "step": 2106 }, { "epoch": 0.2680320569902048, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 5.916499137878418, "learning_rate": 4.463755828740992e-05, "loss": 0.7579, "mean_token_accuracy": 0.8410503268241882, "num_tokens": 80200482.0, "step": 2107 }, { "epoch": 0.2681592672687953, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.416988372802734, "learning_rate": 4.465875370919881e-05, "loss": 0.7631, "mean_token_accuracy": 0.8657686710357666, "num_tokens": 80231844.0, "step": 2108 }, { "epoch": 0.2682864775473858, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.488138675689697, "learning_rate": 4.467994913098771e-05, "loss": 0.7403, "mean_token_accuracy": 0.8596864938735962, "num_tokens": 80269734.0, "step": 2109 }, { "epoch": 0.26841368782597635, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.674907207489014, "learning_rate": 4.47011445527766e-05, "loss": 0.7078, "mean_token_accuracy": 0.8708048462867737, "num_tokens": 80306309.0, "step": 2110 }, { "epoch": 0.2685408981045668, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.974602699279785, "learning_rate": 4.47223399745655e-05, "loss": 0.7829, "mean_token_accuracy": 0.8483180403709412, "num_tokens": 80345390.0, "step": 2111 }, { "epoch": 0.26866810838315736, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.106998443603516, "learning_rate": 4.474353539635439e-05, "loss": 0.6434, "mean_token_accuracy": 0.8735932111740112, "num_tokens": 80384118.0, "step": 2112 }, { "epoch": 0.2687953186617479, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.149709224700928, "learning_rate": 4.4764730818143287e-05, "loss": 0.7534, "mean_token_accuracy": 0.8576066493988037, "num_tokens": 80416703.0, "step": 2113 }, { "epoch": 0.26892252894033836, "ewc_loss": 0.265625, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.044038772583008, "learning_rate": 4.478592623993218e-05, "loss": 0.644, "mean_token_accuracy": 0.8723138570785522, "num_tokens": 80453732.0, "step": 2114 }, { "epoch": 0.2690497392189289, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.978209972381592, "learning_rate": 4.480712166172107e-05, "loss": 0.7445, "mean_token_accuracy": 0.8561766147613525, "num_tokens": 80495265.0, "step": 2115 }, { "epoch": 0.2691769494975194, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.28794002532959, "learning_rate": 4.482831708350996e-05, "loss": 0.701, "mean_token_accuracy": 0.8560115694999695, "num_tokens": 80533789.0, "step": 2116 }, { "epoch": 0.2693041597761099, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.630970001220703, "learning_rate": 4.484951250529886e-05, "loss": 0.6887, "mean_token_accuracy": 0.8665494918823242, "num_tokens": 80570802.0, "step": 2117 }, { "epoch": 0.2694313700547004, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.540945529937744, "learning_rate": 4.487070792708775e-05, "loss": 0.6589, "mean_token_accuracy": 0.872932493686676, "num_tokens": 80608031.0, "step": 2118 }, { "epoch": 0.26955858033329094, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.380563259124756, "learning_rate": 4.4891903348876646e-05, "loss": 0.7242, "mean_token_accuracy": 0.848498523235321, "num_tokens": 80646271.0, "step": 2119 }, { "epoch": 0.2696857906118814, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.696832180023193, "learning_rate": 4.491309877066554e-05, "loss": 0.7438, "mean_token_accuracy": 0.8482460379600525, "num_tokens": 80683325.0, "step": 2120 }, { "epoch": 0.26981300089047194, "ewc_loss": 0.25, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.00023555755615234375, "grad_norm": 6.244969844818115, "learning_rate": 4.4934294192454435e-05, "loss": 0.7155, "mean_token_accuracy": 0.8476544618606567, "num_tokens": 80726403.0, "step": 2121 }, { "epoch": 0.26994021116906247, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.789411544799805, "learning_rate": 4.4955489614243326e-05, "loss": 0.7055, "mean_token_accuracy": 0.859383225440979, "num_tokens": 80767293.0, "step": 2122 }, { "epoch": 0.27006742144765294, "ewc_loss": 0.2392578125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.000225067138671875, "grad_norm": 6.142949104309082, "learning_rate": 4.497668503603222e-05, "loss": 0.6931, "mean_token_accuracy": 0.848781406879425, "num_tokens": 80806218.0, "step": 2123 }, { "epoch": 0.2701946317262435, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.889517307281494, "learning_rate": 4.4997880457821115e-05, "loss": 0.701, "mean_token_accuracy": 0.8621565103530884, "num_tokens": 80841694.0, "step": 2124 }, { "epoch": 0.270321842004834, "ewc_loss": 0.23046875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 6.029645919799805, "learning_rate": 4.5019075879610006e-05, "loss": 0.6556, "mean_token_accuracy": 0.8553956747055054, "num_tokens": 80874879.0, "step": 2125 }, { "epoch": 0.2704490522834245, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.07781457901001, "learning_rate": 4.5040271301398904e-05, "loss": 0.7352, "mean_token_accuracy": 0.854440450668335, "num_tokens": 80916317.0, "step": 2126 }, { "epoch": 0.270576262562015, "ewc_loss": 0.2177734375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.00020313262939453125, "grad_norm": 5.804709434509277, "learning_rate": 4.5061466723187795e-05, "loss": 0.6804, "mean_token_accuracy": 0.8485997915267944, "num_tokens": 80952304.0, "step": 2127 }, { "epoch": 0.27070347284060553, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.422530174255371, "learning_rate": 4.508266214497669e-05, "loss": 0.7267, "mean_token_accuracy": 0.862931489944458, "num_tokens": 80989886.0, "step": 2128 }, { "epoch": 0.27083068311919606, "ewc_loss": 0.20703125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.0001926422119140625, "grad_norm": 5.488976001739502, "learning_rate": 4.510385756676558e-05, "loss": 0.6662, "mean_token_accuracy": 0.8479951024055481, "num_tokens": 81031041.0, "step": 2129 }, { "epoch": 0.27095789339778653, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.9696455001831055, "learning_rate": 4.5125052988554475e-05, "loss": 0.8168, "mean_token_accuracy": 0.8519459962844849, "num_tokens": 81068915.0, "step": 2130 }, { "epoch": 0.27108510367637706, "ewc_loss": 0.2197265625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.00020503997802734375, "grad_norm": 5.2372822761535645, "learning_rate": 4.5146248410343366e-05, "loss": 0.6368, "mean_token_accuracy": 0.8631454706192017, "num_tokens": 81106044.0, "step": 2131 }, { "epoch": 0.2712123139549676, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.579557418823242, "learning_rate": 4.5167443832132264e-05, "loss": 0.8881, "mean_token_accuracy": 0.8580734133720398, "num_tokens": 81142589.0, "step": 2132 }, { "epoch": 0.27133952423355806, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 5.66265344619751, "learning_rate": 4.5188639253921155e-05, "loss": 0.6957, "mean_token_accuracy": 0.8645737171173096, "num_tokens": 81178415.0, "step": 2133 }, { "epoch": 0.2714667345121486, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.539871215820312, "learning_rate": 4.520983467571005e-05, "loss": 0.9598, "mean_token_accuracy": 0.8513598442077637, "num_tokens": 81216952.0, "step": 2134 }, { "epoch": 0.2715939447907391, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 6.878806114196777, "learning_rate": 4.5231030097498944e-05, "loss": 0.8444, "mean_token_accuracy": 0.8684306144714355, "num_tokens": 81257879.0, "step": 2135 }, { "epoch": 0.2717211550693296, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.219946384429932, "learning_rate": 4.525222551928784e-05, "loss": 0.8323, "mean_token_accuracy": 0.8482744693756104, "num_tokens": 81298045.0, "step": 2136 }, { "epoch": 0.2718483653479201, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.904190540313721, "learning_rate": 4.5273420941076726e-05, "loss": 0.8887, "mean_token_accuracy": 0.8721896409988403, "num_tokens": 81332342.0, "step": 2137 }, { "epoch": 0.27197557562651065, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.675359725952148, "learning_rate": 4.5294616362865623e-05, "loss": 0.7246, "mean_token_accuracy": 0.8459569811820984, "num_tokens": 81367270.0, "step": 2138 }, { "epoch": 0.2721027859051011, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.21542739868164, "learning_rate": 4.5315811784654514e-05, "loss": 0.9675, "mean_token_accuracy": 0.8636084198951721, "num_tokens": 81406644.0, "step": 2139 }, { "epoch": 0.27222999618369165, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 5.911933422088623, "learning_rate": 4.533700720644341e-05, "loss": 0.7772, "mean_token_accuracy": 0.8555756211280823, "num_tokens": 81443225.0, "step": 2140 }, { "epoch": 0.2723572064622822, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.852588653564453, "learning_rate": 4.53582026282323e-05, "loss": 1.0335, "mean_token_accuracy": 0.8577665090560913, "num_tokens": 81482168.0, "step": 2141 }, { "epoch": 0.27248441674087265, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 7.750400543212891, "learning_rate": 4.53793980500212e-05, "loss": 1.0669, "mean_token_accuracy": 0.8462274074554443, "num_tokens": 81518127.0, "step": 2142 }, { "epoch": 0.2726116270194632, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.831403732299805, "learning_rate": 4.540059347181009e-05, "loss": 0.9506, "mean_token_accuracy": 0.8454942107200623, "num_tokens": 81559857.0, "step": 2143 }, { "epoch": 0.2727388372980537, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.902960300445557, "learning_rate": 4.542178889359898e-05, "loss": 1.0203, "mean_token_accuracy": 0.8489407300949097, "num_tokens": 81596509.0, "step": 2144 }, { "epoch": 0.2728660475766442, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.940451622009277, "learning_rate": 4.5442984315387874e-05, "loss": 0.8684, "mean_token_accuracy": 0.8611493110656738, "num_tokens": 81632693.0, "step": 2145 }, { "epoch": 0.2729932578552347, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.526638984680176, "learning_rate": 4.546417973717677e-05, "loss": 0.8819, "mean_token_accuracy": 0.8529016375541687, "num_tokens": 81670067.0, "step": 2146 }, { "epoch": 0.27312046813382523, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.52785062789917, "learning_rate": 4.548537515896566e-05, "loss": 0.9171, "mean_token_accuracy": 0.841980516910553, "num_tokens": 81706902.0, "step": 2147 }, { "epoch": 0.2732476784124157, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.861575603485107, "learning_rate": 4.550657058075456e-05, "loss": 0.7299, "mean_token_accuracy": 0.8714002966880798, "num_tokens": 81743340.0, "step": 2148 }, { "epoch": 0.27337488869100623, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.666560649871826, "learning_rate": 4.552776600254345e-05, "loss": 0.8369, "mean_token_accuracy": 0.8621168732643127, "num_tokens": 81779573.0, "step": 2149 }, { "epoch": 0.27350209896959676, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.238164901733398, "learning_rate": 4.554896142433235e-05, "loss": 0.7382, "mean_token_accuracy": 0.856010377407074, "num_tokens": 81822004.0, "step": 2150 }, { "epoch": 0.27362930924818724, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.135628700256348, "learning_rate": 4.557015684612124e-05, "loss": 0.8922, "mean_token_accuracy": 0.8543137311935425, "num_tokens": 81860472.0, "step": 2151 }, { "epoch": 0.27375651952677776, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.082993507385254, "learning_rate": 4.559135226791013e-05, "loss": 0.8037, "mean_token_accuracy": 0.840248703956604, "num_tokens": 81899151.0, "step": 2152 }, { "epoch": 0.2738837298053683, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.457756042480469, "learning_rate": 4.561254768969902e-05, "loss": 0.9135, "mean_token_accuracy": 0.8538990616798401, "num_tokens": 81941213.0, "step": 2153 }, { "epoch": 0.27401094008395877, "ewc_loss": 0.375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.468696117401123, "learning_rate": 4.563374311148792e-05, "loss": 0.8439, "mean_token_accuracy": 0.8402049541473389, "num_tokens": 81978055.0, "step": 2154 }, { "epoch": 0.2741381503625493, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.347638130187988, "learning_rate": 4.565493853327681e-05, "loss": 0.9381, "mean_token_accuracy": 0.8473989367485046, "num_tokens": 82015599.0, "step": 2155 }, { "epoch": 0.2742653606411398, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 6.928998947143555, "learning_rate": 4.567613395506571e-05, "loss": 0.9091, "mean_token_accuracy": 0.8473343849182129, "num_tokens": 82055113.0, "step": 2156 }, { "epoch": 0.2743925709197303, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.805190563201904, "learning_rate": 4.56973293768546e-05, "loss": 0.8664, "mean_token_accuracy": 0.8533427119255066, "num_tokens": 82083052.0, "step": 2157 }, { "epoch": 0.2745197811983208, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.093975067138672, "learning_rate": 4.57185247986435e-05, "loss": 0.8469, "mean_token_accuracy": 0.8494096398353577, "num_tokens": 82118850.0, "step": 2158 }, { "epoch": 0.27464699147691135, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.19348669052124, "learning_rate": 4.573972022043239e-05, "loss": 0.8491, "mean_token_accuracy": 0.844653308391571, "num_tokens": 82151261.0, "step": 2159 }, { "epoch": 0.2747742017555018, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.99191427230835, "learning_rate": 4.576091564222128e-05, "loss": 0.8248, "mean_token_accuracy": 0.8506239652633667, "num_tokens": 82192018.0, "step": 2160 }, { "epoch": 0.27490141203409235, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.981133937835693, "learning_rate": 4.578211106401017e-05, "loss": 0.8381, "mean_token_accuracy": 0.8374968767166138, "num_tokens": 82231044.0, "step": 2161 }, { "epoch": 0.2750286223126829, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.669308185577393, "learning_rate": 4.580330648579907e-05, "loss": 0.7219, "mean_token_accuracy": 0.862558126449585, "num_tokens": 82265334.0, "step": 2162 }, { "epoch": 0.27515583259127335, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.132273197174072, "learning_rate": 4.582450190758796e-05, "loss": 0.7657, "mean_token_accuracy": 0.8555507063865662, "num_tokens": 82303995.0, "step": 2163 }, { "epoch": 0.2752830428698639, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.29345178604126, "learning_rate": 4.584569732937686e-05, "loss": 0.7228, "mean_token_accuracy": 0.8534353971481323, "num_tokens": 82345157.0, "step": 2164 }, { "epoch": 0.2754102531484544, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.497861385345459, "learning_rate": 4.586689275116575e-05, "loss": 0.8037, "mean_token_accuracy": 0.8539546132087708, "num_tokens": 82380723.0, "step": 2165 }, { "epoch": 0.2755374634270449, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.997981071472168, "learning_rate": 4.588808817295465e-05, "loss": 0.6705, "mean_token_accuracy": 0.8656484484672546, "num_tokens": 82418985.0, "step": 2166 }, { "epoch": 0.2756646737056354, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.9067158699035645, "learning_rate": 4.590928359474354e-05, "loss": 0.7905, "mean_token_accuracy": 0.8716283440589905, "num_tokens": 82455498.0, "step": 2167 }, { "epoch": 0.27579188398422594, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.009083271026611, "learning_rate": 4.593047901653243e-05, "loss": 0.7101, "mean_token_accuracy": 0.8588700294494629, "num_tokens": 82491117.0, "step": 2168 }, { "epoch": 0.2759190942628164, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.027974128723145, "learning_rate": 4.595167443832133e-05, "loss": 0.917, "mean_token_accuracy": 0.8468864560127258, "num_tokens": 82533640.0, "step": 2169 }, { "epoch": 0.27604630454140694, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.488353252410889, "learning_rate": 4.597286986011022e-05, "loss": 0.8583, "mean_token_accuracy": 0.8358255624771118, "num_tokens": 82568812.0, "step": 2170 }, { "epoch": 0.27617351481999747, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.476356029510498, "learning_rate": 4.5994065281899116e-05, "loss": 0.8994, "mean_token_accuracy": 0.8334530591964722, "num_tokens": 82611777.0, "step": 2171 }, { "epoch": 0.27630072509858794, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.022031307220459, "learning_rate": 4.601526070368801e-05, "loss": 0.8237, "mean_token_accuracy": 0.8509775400161743, "num_tokens": 82646938.0, "step": 2172 }, { "epoch": 0.27642793537717847, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.646615028381348, "learning_rate": 4.60364561254769e-05, "loss": 0.7192, "mean_token_accuracy": 0.8661659955978394, "num_tokens": 82684316.0, "step": 2173 }, { "epoch": 0.276555145655769, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.342503547668457, "learning_rate": 4.6057651547265796e-05, "loss": 0.7987, "mean_token_accuracy": 0.8499672412872314, "num_tokens": 82723204.0, "step": 2174 }, { "epoch": 0.27668235593435947, "ewc_loss": 0.28125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.168581485748291, "learning_rate": 4.6078846969054687e-05, "loss": 0.7174, "mean_token_accuracy": 0.855850100517273, "num_tokens": 82759634.0, "step": 2175 }, { "epoch": 0.27680956621295, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.523355007171631, "learning_rate": 4.610004239084358e-05, "loss": 0.8807, "mean_token_accuracy": 0.8272881507873535, "num_tokens": 82800446.0, "step": 2176 }, { "epoch": 0.2769367764915405, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.037348747253418, "learning_rate": 4.6121237812632475e-05, "loss": 0.7199, "mean_token_accuracy": 0.8508049845695496, "num_tokens": 82840754.0, "step": 2177 }, { "epoch": 0.277063986770131, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.607059955596924, "learning_rate": 4.6142433234421366e-05, "loss": 0.803, "mean_token_accuracy": 0.8543398976325989, "num_tokens": 82881859.0, "step": 2178 }, { "epoch": 0.2771911970487215, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.073861598968506, "learning_rate": 4.6163628656210264e-05, "loss": 0.7669, "mean_token_accuracy": 0.8400264978408813, "num_tokens": 82922146.0, "step": 2179 }, { "epoch": 0.27731840732731206, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.579229354858398, "learning_rate": 4.6184824077999155e-05, "loss": 0.8365, "mean_token_accuracy": 0.8435287475585938, "num_tokens": 82960017.0, "step": 2180 }, { "epoch": 0.2774456176059026, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.233888149261475, "learning_rate": 4.620601949978805e-05, "loss": 0.7044, "mean_token_accuracy": 0.8616873621940613, "num_tokens": 83001778.0, "step": 2181 }, { "epoch": 0.27757282788449306, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.382961273193359, "learning_rate": 4.622721492157694e-05, "loss": 0.7765, "mean_token_accuracy": 0.8562992811203003, "num_tokens": 83037023.0, "step": 2182 }, { "epoch": 0.2777000381630836, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.444187641143799, "learning_rate": 4.6248410343365835e-05, "loss": 0.7615, "mean_token_accuracy": 0.846550703048706, "num_tokens": 83075355.0, "step": 2183 }, { "epoch": 0.2778272484416741, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.120095729827881, "learning_rate": 4.6269605765154726e-05, "loss": 0.8284, "mean_token_accuracy": 0.8362913727760315, "num_tokens": 83116233.0, "step": 2184 }, { "epoch": 0.2779544587202646, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.565762042999268, "learning_rate": 4.6290801186943624e-05, "loss": 0.7156, "mean_token_accuracy": 0.8636079430580139, "num_tokens": 83152238.0, "step": 2185 }, { "epoch": 0.2780816689988551, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.001293182373047, "learning_rate": 4.6311996608732515e-05, "loss": 0.7325, "mean_token_accuracy": 0.8570243716239929, "num_tokens": 83186779.0, "step": 2186 }, { "epoch": 0.27820887927744564, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.517622470855713, "learning_rate": 4.633319203052141e-05, "loss": 0.772, "mean_token_accuracy": 0.8402311205863953, "num_tokens": 83226001.0, "step": 2187 }, { "epoch": 0.2783360895560361, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.069677829742432, "learning_rate": 4.6354387452310304e-05, "loss": 0.7571, "mean_token_accuracy": 0.8517760038375854, "num_tokens": 83257196.0, "step": 2188 }, { "epoch": 0.27846329983462664, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.316656112670898, "learning_rate": 4.63755828740992e-05, "loss": 0.722, "mean_token_accuracy": 0.8522374033927917, "num_tokens": 83293261.0, "step": 2189 }, { "epoch": 0.27859051011321717, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.30656623840332, "learning_rate": 4.6396778295888086e-05, "loss": 0.7869, "mean_token_accuracy": 0.8515260219573975, "num_tokens": 83327096.0, "step": 2190 }, { "epoch": 0.27871772039180764, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.078495979309082, "learning_rate": 4.6417973717676984e-05, "loss": 0.7357, "mean_token_accuracy": 0.8445995450019836, "num_tokens": 83360558.0, "step": 2191 }, { "epoch": 0.2788449306703982, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.588268280029297, "learning_rate": 4.6439169139465875e-05, "loss": 0.8006, "mean_token_accuracy": 0.8505266904830933, "num_tokens": 83400563.0, "step": 2192 }, { "epoch": 0.2789721409489887, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 5.923160076141357, "learning_rate": 4.646036456125477e-05, "loss": 0.7093, "mean_token_accuracy": 0.8463853001594543, "num_tokens": 83432961.0, "step": 2193 }, { "epoch": 0.2790993512275792, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.79957389831543, "learning_rate": 4.6481559983043664e-05, "loss": 0.7741, "mean_token_accuracy": 0.8657774925231934, "num_tokens": 83472183.0, "step": 2194 }, { "epoch": 0.2792265615061697, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.011669158935547, "learning_rate": 4.650275540483256e-05, "loss": 0.7862, "mean_token_accuracy": 0.8303137421607971, "num_tokens": 83512376.0, "step": 2195 }, { "epoch": 0.27935377178476023, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.681716442108154, "learning_rate": 4.652395082662145e-05, "loss": 0.8174, "mean_token_accuracy": 0.8540835976600647, "num_tokens": 83546141.0, "step": 2196 }, { "epoch": 0.2794809820633507, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.3957414627075195, "learning_rate": 4.6545146248410344e-05, "loss": 0.7341, "mean_token_accuracy": 0.8558751940727234, "num_tokens": 83584223.0, "step": 2197 }, { "epoch": 0.27960819234194123, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.186905384063721, "learning_rate": 4.6566341670199235e-05, "loss": 0.7951, "mean_token_accuracy": 0.8485442996025085, "num_tokens": 83621773.0, "step": 2198 }, { "epoch": 0.27973540262053176, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.871068477630615, "learning_rate": 4.658753709198813e-05, "loss": 0.7024, "mean_token_accuracy": 0.8729422092437744, "num_tokens": 83659129.0, "step": 2199 }, { "epoch": 0.27986261289912223, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.54766321182251, "learning_rate": 4.6608732513777023e-05, "loss": 0.7757, "mean_token_accuracy": 0.8435471653938293, "num_tokens": 83699146.0, "step": 2200 }, { "epoch": 0.27998982317771276, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.278806209564209, "learning_rate": 4.662992793556592e-05, "loss": 0.7871, "mean_token_accuracy": 0.8532949090003967, "num_tokens": 83733266.0, "step": 2201 }, { "epoch": 0.2801170334563033, "ewc_loss": 0.265625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.135993003845215, "learning_rate": 4.665112335735481e-05, "loss": 0.7175, "mean_token_accuracy": 0.8562135696411133, "num_tokens": 83766610.0, "step": 2202 }, { "epoch": 0.28024424373489376, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.540163040161133, "learning_rate": 4.667231877914371e-05, "loss": 0.814, "mean_token_accuracy": 0.8403670787811279, "num_tokens": 83797891.0, "step": 2203 }, { "epoch": 0.2803714540134843, "ewc_loss": 0.25, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.00023555755615234375, "grad_norm": 5.972582817077637, "learning_rate": 4.66935142009326e-05, "loss": 0.7207, "mean_token_accuracy": 0.8407230973243713, "num_tokens": 83837046.0, "step": 2204 }, { "epoch": 0.2804986642920748, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.694640636444092, "learning_rate": 4.671470962272149e-05, "loss": 0.8254, "mean_token_accuracy": 0.846205472946167, "num_tokens": 83876457.0, "step": 2205 }, { "epoch": 0.2806258745706653, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.000244140625, "grad_norm": 5.946807861328125, "learning_rate": 4.673590504451038e-05, "loss": 0.6601, "mean_token_accuracy": 0.8684744834899902, "num_tokens": 83912729.0, "step": 2206 }, { "epoch": 0.2807530848492558, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.7612104415893555, "learning_rate": 4.675710046629928e-05, "loss": 0.7806, "mean_token_accuracy": 0.8640920519828796, "num_tokens": 83945958.0, "step": 2207 }, { "epoch": 0.28088029512784635, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.08506965637207, "learning_rate": 4.677829588808817e-05, "loss": 0.6679, "mean_token_accuracy": 0.8648991584777832, "num_tokens": 83976767.0, "step": 2208 }, { "epoch": 0.2810075054064368, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.61417293548584, "learning_rate": 4.679949130987707e-05, "loss": 0.7662, "mean_token_accuracy": 0.868563711643219, "num_tokens": 84017996.0, "step": 2209 }, { "epoch": 0.28113471568502735, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.391607761383057, "learning_rate": 4.682068673166596e-05, "loss": 0.8147, "mean_token_accuracy": 0.8332338333129883, "num_tokens": 84057528.0, "step": 2210 }, { "epoch": 0.2812619259636179, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.239974021911621, "learning_rate": 4.684188215345486e-05, "loss": 0.7416, "mean_token_accuracy": 0.8627042174339294, "num_tokens": 84091009.0, "step": 2211 }, { "epoch": 0.28138913624220835, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.716668128967285, "learning_rate": 4.686307757524375e-05, "loss": 0.788, "mean_token_accuracy": 0.8452717065811157, "num_tokens": 84125042.0, "step": 2212 }, { "epoch": 0.2815163465207989, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.857512474060059, "learning_rate": 4.688427299703264e-05, "loss": 0.7855, "mean_token_accuracy": 0.8431674242019653, "num_tokens": 84169340.0, "step": 2213 }, { "epoch": 0.2816435567993894, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.890513896942139, "learning_rate": 4.690546841882154e-05, "loss": 0.7455, "mean_token_accuracy": 0.8516982793807983, "num_tokens": 84205347.0, "step": 2214 }, { "epoch": 0.2817707670779799, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.7078938484191895, "learning_rate": 4.692666384061043e-05, "loss": 0.7202, "mean_token_accuracy": 0.854494571685791, "num_tokens": 84245788.0, "step": 2215 }, { "epoch": 0.2818979773565704, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.822700500488281, "learning_rate": 4.694785926239933e-05, "loss": 0.7312, "mean_token_accuracy": 0.8593019843101501, "num_tokens": 84286744.0, "step": 2216 }, { "epoch": 0.28202518763516093, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.815819263458252, "learning_rate": 4.696905468418822e-05, "loss": 0.7351, "mean_token_accuracy": 0.8516010046005249, "num_tokens": 84326031.0, "step": 2217 }, { "epoch": 0.2821523979137514, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.5846638679504395, "learning_rate": 4.699025010597711e-05, "loss": 0.7463, "mean_token_accuracy": 0.8480364084243774, "num_tokens": 84367697.0, "step": 2218 }, { "epoch": 0.28227960819234194, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.05385684967041, "learning_rate": 4.701144552776601e-05, "loss": 0.7136, "mean_token_accuracy": 0.8614431619644165, "num_tokens": 84410647.0, "step": 2219 }, { "epoch": 0.28240681847093246, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.304530143737793, "learning_rate": 4.70326409495549e-05, "loss": 0.694, "mean_token_accuracy": 0.8553153872489929, "num_tokens": 84446839.0, "step": 2220 }, { "epoch": 0.28253402874952294, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.361893653869629, "learning_rate": 4.705383637134379e-05, "loss": 0.7339, "mean_token_accuracy": 0.8590616583824158, "num_tokens": 84485017.0, "step": 2221 }, { "epoch": 0.28266123902811346, "ewc_loss": 0.236328125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.00022029876708984375, "grad_norm": 5.997554779052734, "learning_rate": 4.707503179313269e-05, "loss": 0.7489, "mean_token_accuracy": 0.8310966491699219, "num_tokens": 84524518.0, "step": 2222 }, { "epoch": 0.282788449306704, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.771373271942139, "learning_rate": 4.709622721492158e-05, "loss": 0.8434, "mean_token_accuracy": 0.8373665809631348, "num_tokens": 84559850.0, "step": 2223 }, { "epoch": 0.28291565958529447, "ewc_loss": 0.224609375, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 0.00020885467529296875, "grad_norm": 5.690123081207275, "learning_rate": 4.7117422636710476e-05, "loss": 0.6572, "mean_token_accuracy": 0.8519905209541321, "num_tokens": 84594332.0, "step": 2224 }, { "epoch": 0.283042869863885, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.275247573852539, "learning_rate": 4.713861805849937e-05, "loss": 0.8497, "mean_token_accuracy": 0.8540048599243164, "num_tokens": 84628759.0, "step": 2225 }, { "epoch": 0.2831700801424755, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 0.00024127960205078125, "grad_norm": 5.587466716766357, "learning_rate": 4.7159813480288265e-05, "loss": 0.7185, "mean_token_accuracy": 0.8480854630470276, "num_tokens": 84672022.0, "step": 2226 }, { "epoch": 0.283297290421066, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.545473098754883, "learning_rate": 4.7181008902077156e-05, "loss": 0.8834, "mean_token_accuracy": 0.8667498826980591, "num_tokens": 84709266.0, "step": 2227 }, { "epoch": 0.2834245006996565, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.175778865814209, "learning_rate": 4.720220432386605e-05, "loss": 0.7746, "mean_token_accuracy": 0.8513526320457458, "num_tokens": 84745502.0, "step": 2228 }, { "epoch": 0.28355171097824705, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.99039363861084, "learning_rate": 4.722339974565494e-05, "loss": 0.9187, "mean_token_accuracy": 0.8476395606994629, "num_tokens": 84786061.0, "step": 2229 }, { "epoch": 0.2836789212568376, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.170169353485107, "learning_rate": 4.7244595167443836e-05, "loss": 0.8955, "mean_token_accuracy": 0.8472041487693787, "num_tokens": 84822550.0, "step": 2230 }, { "epoch": 0.28380613153542805, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.604179382324219, "learning_rate": 4.726579058923273e-05, "loss": 0.776, "mean_token_accuracy": 0.8469973802566528, "num_tokens": 84860608.0, "step": 2231 }, { "epoch": 0.2839333418140186, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.1834135055542, "learning_rate": 4.7286986011021625e-05, "loss": 0.9446, "mean_token_accuracy": 0.8414765000343323, "num_tokens": 84903065.0, "step": 2232 }, { "epoch": 0.2840605520926091, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 5.600037097930908, "learning_rate": 4.7308181432810516e-05, "loss": 0.6679, "mean_token_accuracy": 0.8646253943443298, "num_tokens": 84938583.0, "step": 2233 }, { "epoch": 0.2841877623711996, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.090167045593262, "learning_rate": 4.7329376854599414e-05, "loss": 1.0393, "mean_token_accuracy": 0.8393101692199707, "num_tokens": 84975939.0, "step": 2234 }, { "epoch": 0.2843149726497901, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 5.891420364379883, "learning_rate": 4.73505722763883e-05, "loss": 0.7757, "mean_token_accuracy": 0.8453090190887451, "num_tokens": 85011304.0, "step": 2235 }, { "epoch": 0.28444218292838064, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.26031494140625, "learning_rate": 4.7371767698177196e-05, "loss": 1.0107, "mean_token_accuracy": 0.8554946780204773, "num_tokens": 85053163.0, "step": 2236 }, { "epoch": 0.2845693932069711, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.341940879821777, "learning_rate": 4.739296311996609e-05, "loss": 0.9672, "mean_token_accuracy": 0.8534834980964661, "num_tokens": 85085919.0, "step": 2237 }, { "epoch": 0.28469660348556164, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.039024353027344, "learning_rate": 4.7414158541754984e-05, "loss": 0.8862, "mean_token_accuracy": 0.8554841876029968, "num_tokens": 85124323.0, "step": 2238 }, { "epoch": 0.28482381376415217, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.074142456054688, "learning_rate": 4.7435353963543875e-05, "loss": 1.0154, "mean_token_accuracy": 0.847292959690094, "num_tokens": 85163735.0, "step": 2239 }, { "epoch": 0.28495102404274264, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.734061241149902, "learning_rate": 4.745654938533277e-05, "loss": 0.8155, "mean_token_accuracy": 0.8458545207977295, "num_tokens": 85206533.0, "step": 2240 }, { "epoch": 0.28507823432133317, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.45039176940918, "learning_rate": 4.7477744807121664e-05, "loss": 0.9995, "mean_token_accuracy": 0.8401615619659424, "num_tokens": 85248435.0, "step": 2241 }, { "epoch": 0.2852054445999237, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 6.641883373260498, "learning_rate": 4.749894022891056e-05, "loss": 0.8279, "mean_token_accuracy": 0.8533773422241211, "num_tokens": 85285686.0, "step": 2242 }, { "epoch": 0.28533265487851417, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.998391151428223, "learning_rate": 4.7520135650699446e-05, "loss": 0.9075, "mean_token_accuracy": 0.8489446640014648, "num_tokens": 85327518.0, "step": 2243 }, { "epoch": 0.2854598651571047, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.128208160400391, "learning_rate": 4.7541331072488344e-05, "loss": 0.8441, "mean_token_accuracy": 0.8624197840690613, "num_tokens": 85364309.0, "step": 2244 }, { "epoch": 0.2855870754356952, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.099082946777344, "learning_rate": 4.7562526494277235e-05, "loss": 0.7945, "mean_token_accuracy": 0.8556989431381226, "num_tokens": 85407231.0, "step": 2245 }, { "epoch": 0.2857142857142857, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.503316879272461, "learning_rate": 4.758372191606613e-05, "loss": 0.8113, "mean_token_accuracy": 0.8692470192909241, "num_tokens": 85443753.0, "step": 2246 }, { "epoch": 0.2858414959928762, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.561367511749268, "learning_rate": 4.7604917337855024e-05, "loss": 0.776, "mean_token_accuracy": 0.8450639247894287, "num_tokens": 85482874.0, "step": 2247 }, { "epoch": 0.28596870627146675, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.7454609870910645, "learning_rate": 4.762611275964392e-05, "loss": 0.9116, "mean_token_accuracy": 0.842918336391449, "num_tokens": 85521648.0, "step": 2248 }, { "epoch": 0.2860959165500572, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.3567633628845215, "learning_rate": 4.764730818143281e-05, "loss": 0.7234, "mean_token_accuracy": 0.8609209060668945, "num_tokens": 85553078.0, "step": 2249 }, { "epoch": 0.28622312682864776, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.925410270690918, "learning_rate": 4.766850360322171e-05, "loss": 0.8066, "mean_token_accuracy": 0.8659560680389404, "num_tokens": 85589034.0, "step": 2250 }, { "epoch": 0.2863503371072383, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.327579021453857, "learning_rate": 4.7689699025010595e-05, "loss": 0.7698, "mean_token_accuracy": 0.8520069122314453, "num_tokens": 85631541.0, "step": 2251 }, { "epoch": 0.28647754738582876, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.9253082275390625, "learning_rate": 4.771089444679949e-05, "loss": 0.8462, "mean_token_accuracy": 0.8584530353546143, "num_tokens": 85668498.0, "step": 2252 }, { "epoch": 0.2866047576644193, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.422118663787842, "learning_rate": 4.7732089868588384e-05, "loss": 0.7831, "mean_token_accuracy": 0.8496799468994141, "num_tokens": 85708419.0, "step": 2253 }, { "epoch": 0.2867319679430098, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.7637553215026855, "learning_rate": 4.775328529037728e-05, "loss": 0.8735, "mean_token_accuracy": 0.8462231159210205, "num_tokens": 85751564.0, "step": 2254 }, { "epoch": 0.2868591782216003, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.508513927459717, "learning_rate": 4.777448071216617e-05, "loss": 0.794, "mean_token_accuracy": 0.840786337852478, "num_tokens": 85789776.0, "step": 2255 }, { "epoch": 0.2869863885001908, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.6530327796936035, "learning_rate": 4.779567613395507e-05, "loss": 0.8685, "mean_token_accuracy": 0.8373001217842102, "num_tokens": 85828427.0, "step": 2256 }, { "epoch": 0.28711359877878134, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.475935935974121, "learning_rate": 4.781687155574396e-05, "loss": 0.7937, "mean_token_accuracy": 0.840614914894104, "num_tokens": 85877598.0, "step": 2257 }, { "epoch": 0.2872408090573718, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.6617112159729, "learning_rate": 4.783806697753285e-05, "loss": 0.8024, "mean_token_accuracy": 0.858504056930542, "num_tokens": 85908395.0, "step": 2258 }, { "epoch": 0.28736801933596234, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.391561508178711, "learning_rate": 4.785926239932175e-05, "loss": 0.7145, "mean_token_accuracy": 0.8625702857971191, "num_tokens": 85945344.0, "step": 2259 }, { "epoch": 0.28749522961455287, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00016021728515625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.71364688873291, "learning_rate": 4.788045782111064e-05, "loss": 0.7617, "mean_token_accuracy": 0.8699907064437866, "num_tokens": 85982484.0, "step": 2260 }, { "epoch": 0.28762243989314334, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00016021728515625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.331408977508545, "learning_rate": 4.790165324289953e-05, "loss": 0.7647, "mean_token_accuracy": 0.8448498845100403, "num_tokens": 86019496.0, "step": 2261 }, { "epoch": 0.2877496501717339, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.797393798828125, "learning_rate": 4.792284866468843e-05, "loss": 0.8499, "mean_token_accuracy": 0.8493403196334839, "num_tokens": 86059344.0, "step": 2262 }, { "epoch": 0.2878768604503244, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.334882736206055, "learning_rate": 4.794404408647732e-05, "loss": 0.7458, "mean_token_accuracy": 0.8504133820533752, "num_tokens": 86093090.0, "step": 2263 }, { "epoch": 0.2880040707289149, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.786092281341553, "learning_rate": 4.796523950826622e-05, "loss": 0.8365, "mean_token_accuracy": 0.852311372756958, "num_tokens": 86127149.0, "step": 2264 }, { "epoch": 0.2881312810075054, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.4744439125061035, "learning_rate": 4.798643493005511e-05, "loss": 0.7104, "mean_token_accuracy": 0.867160439491272, "num_tokens": 86163541.0, "step": 2265 }, { "epoch": 0.28825849128609593, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.580838203430176, "learning_rate": 4.8007630351844e-05, "loss": 0.8593, "mean_token_accuracy": 0.8436073064804077, "num_tokens": 86200124.0, "step": 2266 }, { "epoch": 0.2883857015646864, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.722346305847168, "learning_rate": 4.80288257736329e-05, "loss": 0.7652, "mean_token_accuracy": 0.8543606400489807, "num_tokens": 86236464.0, "step": 2267 }, { "epoch": 0.28851291184327693, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.214977741241455, "learning_rate": 4.805002119542179e-05, "loss": 0.79, "mean_token_accuracy": 0.8500386476516724, "num_tokens": 86277504.0, "step": 2268 }, { "epoch": 0.28864012212186746, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.967399597167969, "learning_rate": 4.807121661721069e-05, "loss": 0.84, "mean_token_accuracy": 0.8301971554756165, "num_tokens": 86314404.0, "step": 2269 }, { "epoch": 0.28876733240045793, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.887901306152344, "learning_rate": 4.809241203899958e-05, "loss": 0.7303, "mean_token_accuracy": 0.8570932149887085, "num_tokens": 86350690.0, "step": 2270 }, { "epoch": 0.28889454267904846, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.1237616539001465, "learning_rate": 4.811360746078848e-05, "loss": 0.7131, "mean_token_accuracy": 0.8695864677429199, "num_tokens": 86389294.0, "step": 2271 }, { "epoch": 0.289021752957639, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.735923767089844, "learning_rate": 4.813480288257737e-05, "loss": 0.7395, "mean_token_accuracy": 0.8523101210594177, "num_tokens": 86420234.0, "step": 2272 }, { "epoch": 0.28914896323622946, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.136753559112549, "learning_rate": 4.815599830436626e-05, "loss": 0.7437, "mean_token_accuracy": 0.8568363785743713, "num_tokens": 86458435.0, "step": 2273 }, { "epoch": 0.28927617351482, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.712084770202637, "learning_rate": 4.817719372615515e-05, "loss": 0.7247, "mean_token_accuracy": 0.8571609854698181, "num_tokens": 86499082.0, "step": 2274 }, { "epoch": 0.2894033837934105, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.103700160980225, "learning_rate": 4.819838914794405e-05, "loss": 0.7671, "mean_token_accuracy": 0.8454431295394897, "num_tokens": 86536989.0, "step": 2275 }, { "epoch": 0.289530594072001, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.7237868309021, "learning_rate": 4.821958456973294e-05, "loss": 0.7018, "mean_token_accuracy": 0.8597792387008667, "num_tokens": 86572154.0, "step": 2276 }, { "epoch": 0.2896578043505915, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.05595588684082, "learning_rate": 4.8240779991521836e-05, "loss": 0.6944, "mean_token_accuracy": 0.8682447075843811, "num_tokens": 86610908.0, "step": 2277 }, { "epoch": 0.28978501462918205, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.758440017700195, "learning_rate": 4.826197541331073e-05, "loss": 0.7964, "mean_token_accuracy": 0.8295572996139526, "num_tokens": 86653585.0, "step": 2278 }, { "epoch": 0.2899122249077726, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.00986909866333, "learning_rate": 4.8283170835099625e-05, "loss": 0.7433, "mean_token_accuracy": 0.8496357202529907, "num_tokens": 86690794.0, "step": 2279 }, { "epoch": 0.29003943518636305, "ewc_loss": 0.28125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.783299446105957, "learning_rate": 4.8304366256888516e-05, "loss": 0.745, "mean_token_accuracy": 0.8453625440597534, "num_tokens": 86729742.0, "step": 2280 }, { "epoch": 0.2901666454649536, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.987752437591553, "learning_rate": 4.832556167867741e-05, "loss": 0.7675, "mean_token_accuracy": 0.8476244211196899, "num_tokens": 86765837.0, "step": 2281 }, { "epoch": 0.2902938557435441, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.772560119628906, "learning_rate": 4.83467571004663e-05, "loss": 0.6903, "mean_token_accuracy": 0.8617873787879944, "num_tokens": 86799697.0, "step": 2282 }, { "epoch": 0.2904210660221346, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 7.017280101776123, "learning_rate": 4.8367952522255196e-05, "loss": 0.716, "mean_token_accuracy": 0.8544418811798096, "num_tokens": 86832264.0, "step": 2283 }, { "epoch": 0.2905482763007251, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.710293292999268, "learning_rate": 4.838914794404409e-05, "loss": 0.6909, "mean_token_accuracy": 0.8536130785942078, "num_tokens": 86864492.0, "step": 2284 }, { "epoch": 0.29067548657931563, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 7.118167877197266, "learning_rate": 4.8410343365832985e-05, "loss": 0.782, "mean_token_accuracy": 0.84149169921875, "num_tokens": 86898186.0, "step": 2285 }, { "epoch": 0.2908026968579061, "ewc_loss": 0.25390625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 0.0002384185791015625, "grad_norm": 6.567381381988525, "learning_rate": 4.8431538787621876e-05, "loss": 0.6969, "mean_token_accuracy": 0.8522666692733765, "num_tokens": 86935850.0, "step": 2286 }, { "epoch": 0.29092990713649663, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.299294948577881, "learning_rate": 4.8452734209410774e-05, "loss": 0.6709, "mean_token_accuracy": 0.8758745193481445, "num_tokens": 86972732.0, "step": 2287 }, { "epoch": 0.29105711741508716, "ewc_loss": 0.240234375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 0.00022411346435546875, "grad_norm": 6.347805500030518, "learning_rate": 4.847392963119966e-05, "loss": 0.6905, "mean_token_accuracy": 0.8495504260063171, "num_tokens": 87012538.0, "step": 2288 }, { "epoch": 0.29118432769367764, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.618790626525879, "learning_rate": 4.8495125052988556e-05, "loss": 0.8095, "mean_token_accuracy": 0.8377705812454224, "num_tokens": 87046917.0, "step": 2289 }, { "epoch": 0.29131153797226816, "ewc_loss": 0.224609375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 0.0002079010009765625, "grad_norm": 5.962673664093018, "learning_rate": 4.851632047477745e-05, "loss": 0.6789, "mean_token_accuracy": 0.8478138446807861, "num_tokens": 87086002.0, "step": 2290 }, { "epoch": 0.2914387482508587, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.254284858703613, "learning_rate": 4.8537515896566345e-05, "loss": 0.8194, "mean_token_accuracy": 0.8562837839126587, "num_tokens": 87126138.0, "step": 2291 }, { "epoch": 0.29156595852944917, "ewc_loss": 0.2001953125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 0.00018405914306640625, "grad_norm": 5.359683036804199, "learning_rate": 4.8558711318355236e-05, "loss": 0.6298, "mean_token_accuracy": 0.8563898801803589, "num_tokens": 87163062.0, "step": 2292 }, { "epoch": 0.2916931688080397, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.329122543334961, "learning_rate": 4.8579906740144134e-05, "loss": 0.9291, "mean_token_accuracy": 0.8581790924072266, "num_tokens": 87203836.0, "step": 2293 }, { "epoch": 0.2918203790866302, "ewc_loss": 0.25, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 5.152534008026123, "learning_rate": 4.8601102161933025e-05, "loss": 0.6741, "mean_token_accuracy": 0.8581826090812683, "num_tokens": 87236854.0, "step": 2294 }, { "epoch": 0.2919475893652207, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.422184944152832, "learning_rate": 4.862229758372192e-05, "loss": 1.0903, "mean_token_accuracy": 0.8617955446243286, "num_tokens": 87276316.0, "step": 2295 }, { "epoch": 0.2920747996438112, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.097036838531494, "learning_rate": 4.864349300551081e-05, "loss": 0.9462, "mean_token_accuracy": 0.8645812273025513, "num_tokens": 87314504.0, "step": 2296 }, { "epoch": 0.29220200992240175, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.837414741516113, "learning_rate": 4.8664688427299705e-05, "loss": 1.0478, "mean_token_accuracy": 0.8520834445953369, "num_tokens": 87348403.0, "step": 2297 }, { "epoch": 0.2923292202009922, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 8.66723918914795, "learning_rate": 4.8685883849088596e-05, "loss": 1.22, "mean_token_accuracy": 0.8455567359924316, "num_tokens": 87393044.0, "step": 2298 }, { "epoch": 0.29245643047958275, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.832792282104492, "learning_rate": 4.8707079270877493e-05, "loss": 0.9056, "mean_token_accuracy": 0.8382951021194458, "num_tokens": 87430930.0, "step": 2299 }, { "epoch": 0.2925836407581733, "ewc_loss": 0.71875, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 9.81805419921875, "learning_rate": 4.8728274692666384e-05, "loss": 1.1829, "mean_token_accuracy": 0.8444019556045532, "num_tokens": 87467941.0, "step": 2300 }, { "epoch": 0.29271085103676375, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.052558422088623, "learning_rate": 4.874947011445528e-05, "loss": 0.9178, "mean_token_accuracy": 0.8423645496368408, "num_tokens": 87507453.0, "step": 2301 }, { "epoch": 0.2928380613153543, "ewc_loss": 0.765625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 11.1016206741333, "learning_rate": 4.877066553624417e-05, "loss": 1.1652, "mean_token_accuracy": 0.8638210296630859, "num_tokens": 87545983.0, "step": 2302 }, { "epoch": 0.2929652715939448, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.000743865966796875, "grad_norm": 8.895439147949219, "learning_rate": 4.879186095803307e-05, "loss": 1.2411, "mean_token_accuracy": 0.8375213146209717, "num_tokens": 87582352.0, "step": 2303 }, { "epoch": 0.2930924818725353, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 10.288945198059082, "learning_rate": 4.8813056379821955e-05, "loss": 0.9971, "mean_token_accuracy": 0.8655219078063965, "num_tokens": 87616668.0, "step": 2304 }, { "epoch": 0.2932196921511258, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 8.603724479675293, "learning_rate": 4.883425180161085e-05, "loss": 1.1395, "mean_token_accuracy": 0.8453191518783569, "num_tokens": 87653008.0, "step": 2305 }, { "epoch": 0.29334690242971634, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.98560905456543, "learning_rate": 4.8855447223399744e-05, "loss": 1.0237, "mean_token_accuracy": 0.8609979152679443, "num_tokens": 87694985.0, "step": 2306 }, { "epoch": 0.2934741127083068, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.6898088455200195, "learning_rate": 4.887664264518864e-05, "loss": 0.962, "mean_token_accuracy": 0.8464596271514893, "num_tokens": 87730841.0, "step": 2307 }, { "epoch": 0.29360132298689734, "ewc_loss": 0.625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.995226860046387, "learning_rate": 4.889783806697753e-05, "loss": 1.1158, "mean_token_accuracy": 0.8434770107269287, "num_tokens": 87762215.0, "step": 2308 }, { "epoch": 0.29372853326548787, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.202211380004883, "learning_rate": 4.891903348876643e-05, "loss": 0.879, "mean_token_accuracy": 0.8545509576797485, "num_tokens": 87799150.0, "step": 2309 }, { "epoch": 0.29385574354407834, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.684518814086914, "learning_rate": 4.894022891055532e-05, "loss": 1.0822, "mean_token_accuracy": 0.8647736310958862, "num_tokens": 87842463.0, "step": 2310 }, { "epoch": 0.29398295382266887, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 7.972377300262451, "learning_rate": 4.896142433234421e-05, "loss": 0.943, "mean_token_accuracy": 0.8654886484146118, "num_tokens": 87880093.0, "step": 2311 }, { "epoch": 0.2941101641012594, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.940296173095703, "learning_rate": 4.898261975413311e-05, "loss": 1.087, "mean_token_accuracy": 0.8640332221984863, "num_tokens": 87912713.0, "step": 2312 }, { "epoch": 0.29423737437984987, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 9.055008888244629, "learning_rate": 4.9003815175922e-05, "loss": 1.1188, "mean_token_accuracy": 0.8471604585647583, "num_tokens": 87946357.0, "step": 2313 }, { "epoch": 0.2943645846584404, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.14964771270752, "learning_rate": 4.90250105977109e-05, "loss": 1.1187, "mean_token_accuracy": 0.8350212574005127, "num_tokens": 87981102.0, "step": 2314 }, { "epoch": 0.2944917949370309, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.931845664978027, "learning_rate": 4.904620601949979e-05, "loss": 1.0291, "mean_token_accuracy": 0.8567156791687012, "num_tokens": 88022325.0, "step": 2315 }, { "epoch": 0.2946190052156214, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.430651664733887, "learning_rate": 4.906740144128869e-05, "loss": 1.04, "mean_token_accuracy": 0.8427432179450989, "num_tokens": 88057254.0, "step": 2316 }, { "epoch": 0.2947462154942119, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.480683326721191, "learning_rate": 4.908859686307758e-05, "loss": 0.9303, "mean_token_accuracy": 0.8612990379333496, "num_tokens": 88095667.0, "step": 2317 }, { "epoch": 0.29487342577280246, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.816455841064453, "learning_rate": 4.910979228486647e-05, "loss": 0.9029, "mean_token_accuracy": 0.8697513341903687, "num_tokens": 88129634.0, "step": 2318 }, { "epoch": 0.29500063605139293, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.8841352462768555, "learning_rate": 4.913098770665536e-05, "loss": 0.8252, "mean_token_accuracy": 0.8640803098678589, "num_tokens": 88164592.0, "step": 2319 }, { "epoch": 0.29512784632998346, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.697555541992188, "learning_rate": 4.915218312844426e-05, "loss": 0.9489, "mean_token_accuracy": 0.8493001461029053, "num_tokens": 88202390.0, "step": 2320 }, { "epoch": 0.295255056608574, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.207282066345215, "learning_rate": 4.917337855023315e-05, "loss": 0.7719, "mean_token_accuracy": 0.8614616394042969, "num_tokens": 88243104.0, "step": 2321 }, { "epoch": 0.29538226688716446, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.09371280670166, "learning_rate": 4.919457397202205e-05, "loss": 0.9201, "mean_token_accuracy": 0.8669850826263428, "num_tokens": 88284038.0, "step": 2322 }, { "epoch": 0.295509477165755, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.0110392570495605, "learning_rate": 4.921576939381094e-05, "loss": 0.8312, "mean_token_accuracy": 0.8456916213035583, "num_tokens": 88320705.0, "step": 2323 }, { "epoch": 0.2956366874443455, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.564064979553223, "learning_rate": 4.923696481559984e-05, "loss": 1.0018, "mean_token_accuracy": 0.8598760962486267, "num_tokens": 88359636.0, "step": 2324 }, { "epoch": 0.295763897722936, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.594686985015869, "learning_rate": 4.925816023738873e-05, "loss": 0.8597, "mean_token_accuracy": 0.8515585064888, "num_tokens": 88402029.0, "step": 2325 }, { "epoch": 0.2958911080015265, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.367542266845703, "learning_rate": 4.927935565917762e-05, "loss": 0.9415, "mean_token_accuracy": 0.8756272792816162, "num_tokens": 88438245.0, "step": 2326 }, { "epoch": 0.29601831828011704, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.199126243591309, "learning_rate": 4.930055108096651e-05, "loss": 0.8925, "mean_token_accuracy": 0.8635355234146118, "num_tokens": 88477987.0, "step": 2327 }, { "epoch": 0.2961455285587075, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.431078910827637, "learning_rate": 4.932174650275541e-05, "loss": 0.8747, "mean_token_accuracy": 0.867021918296814, "num_tokens": 88510759.0, "step": 2328 }, { "epoch": 0.29627273883729804, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.104129791259766, "learning_rate": 4.93429419245443e-05, "loss": 0.8551, "mean_token_accuracy": 0.8606024980545044, "num_tokens": 88544819.0, "step": 2329 }, { "epoch": 0.2963999491158886, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.649880409240723, "learning_rate": 4.93641373463332e-05, "loss": 0.8196, "mean_token_accuracy": 0.8589254021644592, "num_tokens": 88584560.0, "step": 2330 }, { "epoch": 0.2965271593944791, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.679590225219727, "learning_rate": 4.938533276812209e-05, "loss": 0.7854, "mean_token_accuracy": 0.8626551032066345, "num_tokens": 88620733.0, "step": 2331 }, { "epoch": 0.2966543696730696, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.482339859008789, "learning_rate": 4.9406528189910986e-05, "loss": 0.8051, "mean_token_accuracy": 0.8567199110984802, "num_tokens": 88661306.0, "step": 2332 }, { "epoch": 0.2967815799516601, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.201812744140625, "learning_rate": 4.942772361169988e-05, "loss": 0.7448, "mean_token_accuracy": 0.8621318340301514, "num_tokens": 88696886.0, "step": 2333 }, { "epoch": 0.29690879023025063, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.539418697357178, "learning_rate": 4.944891903348877e-05, "loss": 0.7529, "mean_token_accuracy": 0.8692221641540527, "num_tokens": 88733242.0, "step": 2334 }, { "epoch": 0.2970360005088411, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.924888610839844, "learning_rate": 4.947011445527766e-05, "loss": 0.6984, "mean_token_accuracy": 0.8708105683326721, "num_tokens": 88769599.0, "step": 2335 }, { "epoch": 0.29716321078743163, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.596390724182129, "learning_rate": 4.9491309877066557e-05, "loss": 0.7202, "mean_token_accuracy": 0.8764917850494385, "num_tokens": 88802885.0, "step": 2336 }, { "epoch": 0.29729042106602216, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.74644660949707, "learning_rate": 4.951250529885545e-05, "loss": 0.7155, "mean_token_accuracy": 0.8564072251319885, "num_tokens": 88837061.0, "step": 2337 }, { "epoch": 0.29741763134461263, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.768021106719971, "learning_rate": 4.9533700720644345e-05, "loss": 0.7976, "mean_token_accuracy": 0.8519663214683533, "num_tokens": 88874672.0, "step": 2338 }, { "epoch": 0.29754484162320316, "ewc_loss": 0.28125, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.499014377593994, "learning_rate": 4.9554896142433236e-05, "loss": 0.6669, "mean_token_accuracy": 0.8691992163658142, "num_tokens": 88910520.0, "step": 2339 }, { "epoch": 0.2976720519017937, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.094721794128418, "learning_rate": 4.9576091564222134e-05, "loss": 0.8575, "mean_token_accuracy": 0.8506554961204529, "num_tokens": 88949495.0, "step": 2340 }, { "epoch": 0.29779926218038416, "ewc_loss": 0.265625, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.276318550109863, "learning_rate": 4.9597286986011025e-05, "loss": 0.7194, "mean_token_accuracy": 0.848588228225708, "num_tokens": 88988776.0, "step": 2341 }, { "epoch": 0.2979264724589747, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.516138076782227, "learning_rate": 4.9618482407799916e-05, "loss": 0.8559, "mean_token_accuracy": 0.8637245893478394, "num_tokens": 89018983.0, "step": 2342 }, { "epoch": 0.2980536827375652, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.342875003814697, "learning_rate": 4.963967782958881e-05, "loss": 0.7433, "mean_token_accuracy": 0.846784234046936, "num_tokens": 89055834.0, "step": 2343 }, { "epoch": 0.2981808930161557, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00016689300537109375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.517838478088379, "learning_rate": 4.9660873251377705e-05, "loss": 0.913, "mean_token_accuracy": 0.8547152876853943, "num_tokens": 89098612.0, "step": 2344 }, { "epoch": 0.2983081032947462, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.947793960571289, "learning_rate": 4.9682068673166596e-05, "loss": 0.8406, "mean_token_accuracy": 0.8396641612052917, "num_tokens": 89130169.0, "step": 2345 }, { "epoch": 0.29843531357333675, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.801962852478027, "learning_rate": 4.9703264094955494e-05, "loss": 0.7949, "mean_token_accuracy": 0.868444561958313, "num_tokens": 89170346.0, "step": 2346 }, { "epoch": 0.2985625238519272, "ewc_loss": 0.375, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.558482646942139, "learning_rate": 4.9724459516744385e-05, "loss": 0.8356, "mean_token_accuracy": 0.845454216003418, "num_tokens": 89211713.0, "step": 2347 }, { "epoch": 0.29868973413051775, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.867508411407471, "learning_rate": 4.974565493853328e-05, "loss": 0.7327, "mean_token_accuracy": 0.8618791699409485, "num_tokens": 89252842.0, "step": 2348 }, { "epoch": 0.2988169444091083, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.070549011230469, "learning_rate": 4.976685036032217e-05, "loss": 0.7908, "mean_token_accuracy": 0.8646969795227051, "num_tokens": 89291914.0, "step": 2349 }, { "epoch": 0.29894415468769875, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.354562759399414, "learning_rate": 4.9788045782111065e-05, "loss": 0.7652, "mean_token_accuracy": 0.8403803110122681, "num_tokens": 89333912.0, "step": 2350 }, { "epoch": 0.2990713649662893, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.500372886657715, "learning_rate": 4.9809241203899956e-05, "loss": 0.8561, "mean_token_accuracy": 0.8584805727005005, "num_tokens": 89363603.0, "step": 2351 }, { "epoch": 0.2991985752448798, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.331043243408203, "learning_rate": 4.9830436625688854e-05, "loss": 0.6897, "mean_token_accuracy": 0.8655885457992554, "num_tokens": 89404392.0, "step": 2352 }, { "epoch": 0.2993257855234703, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.69738483428955, "learning_rate": 4.9851632047477745e-05, "loss": 0.9309, "mean_token_accuracy": 0.8411797285079956, "num_tokens": 89444180.0, "step": 2353 }, { "epoch": 0.2994529958020608, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.801054000854492, "learning_rate": 4.987282746926664e-05, "loss": 0.8046, "mean_token_accuracy": 0.8404080867767334, "num_tokens": 89483875.0, "step": 2354 }, { "epoch": 0.29958020608065133, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.14090347290039, "learning_rate": 4.9894022891055534e-05, "loss": 0.8998, "mean_token_accuracy": 0.8476468920707703, "num_tokens": 89519544.0, "step": 2355 }, { "epoch": 0.2997074163592418, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.439366817474365, "learning_rate": 4.991521831284443e-05, "loss": 0.8244, "mean_token_accuracy": 0.8513955473899841, "num_tokens": 89552635.0, "step": 2356 }, { "epoch": 0.29983462663783234, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.358441352844238, "learning_rate": 4.993641373463332e-05, "loss": 0.7869, "mean_token_accuracy": 0.8626613616943359, "num_tokens": 89588514.0, "step": 2357 }, { "epoch": 0.29996183691642286, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.549065589904785, "learning_rate": 4.9957609156422214e-05, "loss": 0.7815, "mean_token_accuracy": 0.8589341640472412, "num_tokens": 89629029.0, "step": 2358 }, { "epoch": 0.30008904719501334, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.343868255615234, "learning_rate": 4.997880457821111e-05, "loss": 0.7148, "mean_token_accuracy": 0.8747256994247437, "num_tokens": 89659214.0, "step": 2359 }, { "epoch": 0.30021625747360386, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0001697540283203125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.960115909576416, "learning_rate": 5e-05, "loss": 0.7022, "mean_token_accuracy": 0.8654965758323669, "num_tokens": 89695342.0, "step": 2360 }, { "epoch": 0.3003434677521944, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0001697540283203125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.953915119171143, "learning_rate": 5e-05, "loss": 0.8264, "mean_token_accuracy": 0.85451340675354, "num_tokens": 89731209.0, "step": 2361 }, { "epoch": 0.30047067803078487, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.0001697540283203125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.23341703414917, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.8402093052864075, "num_tokens": 89771516.0, "step": 2362 }, { "epoch": 0.3005978883093754, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.679547309875488, "learning_rate": 5e-05, "loss": 0.9316, "mean_token_accuracy": 0.8433812856674194, "num_tokens": 89817929.0, "step": 2363 }, { "epoch": 0.3007250985879659, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.962745666503906, "learning_rate": 5e-05, "loss": 0.8197, "mean_token_accuracy": 0.8277225494384766, "num_tokens": 89850714.0, "step": 2364 }, { "epoch": 0.3008523088665564, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.001168251037598, "learning_rate": 5e-05, "loss": 0.9681, "mean_token_accuracy": 0.8532021045684814, "num_tokens": 89885686.0, "step": 2365 }, { "epoch": 0.3009795191451469, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.53363037109375, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8606777787208557, "num_tokens": 89926549.0, "step": 2366 }, { "epoch": 0.30110672942373745, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.555793762207031, "learning_rate": 5e-05, "loss": 0.9868, "mean_token_accuracy": 0.8402245044708252, "num_tokens": 89969031.0, "step": 2367 }, { "epoch": 0.3012339397023279, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.396517753601074, "learning_rate": 5e-05, "loss": 0.8633, "mean_token_accuracy": 0.8487945199012756, "num_tokens": 90000348.0, "step": 2368 }, { "epoch": 0.30136114998091845, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.582455158233643, "learning_rate": 5e-05, "loss": 0.8508, "mean_token_accuracy": 0.854091465473175, "num_tokens": 90033945.0, "step": 2369 }, { "epoch": 0.301488360259509, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.785886764526367, "learning_rate": 5e-05, "loss": 0.8295, "mean_token_accuracy": 0.8635777235031128, "num_tokens": 90069537.0, "step": 2370 }, { "epoch": 0.30161557053809945, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.031581401824951, "learning_rate": 5e-05, "loss": 0.741, "mean_token_accuracy": 0.8658125400543213, "num_tokens": 90107923.0, "step": 2371 }, { "epoch": 0.30174278081669, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.682680606842041, "learning_rate": 5e-05, "loss": 0.8573, "mean_token_accuracy": 0.8442972898483276, "num_tokens": 90152211.0, "step": 2372 }, { "epoch": 0.3018699910952805, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.033773899078369, "learning_rate": 5e-05, "loss": 0.771, "mean_token_accuracy": 0.8505821228027344, "num_tokens": 90193687.0, "step": 2373 }, { "epoch": 0.301997201373871, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.483839988708496, "learning_rate": 5e-05, "loss": 0.7468, "mean_token_accuracy": 0.8696072101593018, "num_tokens": 90232639.0, "step": 2374 }, { "epoch": 0.3021244116524615, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.994504451751709, "learning_rate": 5e-05, "loss": 0.7689, "mean_token_accuracy": 0.8468626737594604, "num_tokens": 90273344.0, "step": 2375 }, { "epoch": 0.30225162193105204, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.610727787017822, "learning_rate": 5e-05, "loss": 0.8046, "mean_token_accuracy": 0.8519977927207947, "num_tokens": 90309999.0, "step": 2376 }, { "epoch": 0.3023788322096425, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.653997898101807, "learning_rate": 5e-05, "loss": 0.7141, "mean_token_accuracy": 0.8575998544692993, "num_tokens": 90349049.0, "step": 2377 }, { "epoch": 0.30250604248823304, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.016242027282715, "learning_rate": 5e-05, "loss": 0.8503, "mean_token_accuracy": 0.8490397334098816, "num_tokens": 90388795.0, "step": 2378 }, { "epoch": 0.30263325276682357, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.334307670593262, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8376361131668091, "num_tokens": 90423785.0, "step": 2379 }, { "epoch": 0.3027604630454141, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.297240257263184, "learning_rate": 5e-05, "loss": 0.8688, "mean_token_accuracy": 0.8526561260223389, "num_tokens": 90461625.0, "step": 2380 }, { "epoch": 0.30288767332400457, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00017547607421875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.302511692047119, "learning_rate": 5e-05, "loss": 0.704, "mean_token_accuracy": 0.8611774444580078, "num_tokens": 90500286.0, "step": 2381 }, { "epoch": 0.3030148836025951, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.27221965789795, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.8578395843505859, "num_tokens": 90533463.0, "step": 2382 }, { "epoch": 0.3031420938811856, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.6060099601745605, "learning_rate": 5e-05, "loss": 0.7228, "mean_token_accuracy": 0.8617247343063354, "num_tokens": 90570536.0, "step": 2383 }, { "epoch": 0.3032693041597761, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.933130264282227, "learning_rate": 5e-05, "loss": 0.8935, "mean_token_accuracy": 0.8401273488998413, "num_tokens": 90603030.0, "step": 2384 }, { "epoch": 0.3033965144383666, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.953818321228027, "learning_rate": 5e-05, "loss": 0.7969, "mean_token_accuracy": 0.8426328301429749, "num_tokens": 90638459.0, "step": 2385 }, { "epoch": 0.30352372471695716, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00017833709716796875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.519023895263672, "learning_rate": 5e-05, "loss": 0.7916, "mean_token_accuracy": 0.8551661372184753, "num_tokens": 90671894.0, "step": 2386 }, { "epoch": 0.30365093499554763, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00017833709716796875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.175876617431641, "learning_rate": 5e-05, "loss": 0.7622, "mean_token_accuracy": 0.8590880036354065, "num_tokens": 90708583.0, "step": 2387 }, { "epoch": 0.30377814527413816, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00017833709716796875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.304224967956543, "learning_rate": 5e-05, "loss": 0.805, "mean_token_accuracy": 0.8450735211372375, "num_tokens": 90749914.0, "step": 2388 }, { "epoch": 0.3039053555527287, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00017833709716796875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.167001247406006, "learning_rate": 5e-05, "loss": 0.7671, "mean_token_accuracy": 0.8530367612838745, "num_tokens": 90791955.0, "step": 2389 }, { "epoch": 0.30403256583131916, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00017833709716796875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.3216233253479, "learning_rate": 5e-05, "loss": 0.7109, "mean_token_accuracy": 0.8725686073303223, "num_tokens": 90827666.0, "step": 2390 }, { "epoch": 0.3041597761099097, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00017833709716796875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.008423328399658, "learning_rate": 5e-05, "loss": 0.7575, "mean_token_accuracy": 0.8477153182029724, "num_tokens": 90860647.0, "step": 2391 }, { "epoch": 0.3042869863885002, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00017833709716796875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.476953506469727, "learning_rate": 5e-05, "loss": 0.8148, "mean_token_accuracy": 0.8385920524597168, "num_tokens": 90899737.0, "step": 2392 }, { "epoch": 0.3044141966670907, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00017833709716796875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.775803565979004, "learning_rate": 5e-05, "loss": 0.7024, "mean_token_accuracy": 0.860923171043396, "num_tokens": 90939934.0, "step": 2393 }, { "epoch": 0.3045414069456812, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00017833709716796875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.698399543762207, "learning_rate": 5e-05, "loss": 0.8083, "mean_token_accuracy": 0.8454629778862, "num_tokens": 90971538.0, "step": 2394 }, { "epoch": 0.30466861722427174, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.00017833709716796875, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.519234657287598, "learning_rate": 5e-05, "loss": 0.6989, "mean_token_accuracy": 0.857207179069519, "num_tokens": 91012119.0, "step": 2395 }, { "epoch": 0.3047958275028622, "ewc_loss": 0.375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.995377540588379, "learning_rate": 5e-05, "loss": 0.8221, "mean_token_accuracy": 0.8507236242294312, "num_tokens": 91051342.0, "step": 2396 }, { "epoch": 0.30492303778145274, "ewc_loss": 0.265625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.270097732543945, "learning_rate": 5e-05, "loss": 0.688, "mean_token_accuracy": 0.8583076000213623, "num_tokens": 91087650.0, "step": 2397 }, { "epoch": 0.30505024806004327, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.334175109863281, "learning_rate": 5e-05, "loss": 0.8395, "mean_token_accuracy": 0.8555279970169067, "num_tokens": 91129786.0, "step": 2398 }, { "epoch": 0.30517745833863374, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.139876365661621, "learning_rate": 5e-05, "loss": 0.7003, "mean_token_accuracy": 0.8541059494018555, "num_tokens": 91165273.0, "step": 2399 }, { "epoch": 0.3053046686172243, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.50430679321289, "learning_rate": 5e-05, "loss": 0.853, "mean_token_accuracy": 0.8647987246513367, "num_tokens": 91198034.0, "step": 2400 }, { "epoch": 0.3054318788958148, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.366058826446533, "learning_rate": 5e-05, "loss": 0.7293, "mean_token_accuracy": 0.8575707077980042, "num_tokens": 91241415.0, "step": 2401 }, { "epoch": 0.3055590891744053, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.215617179870605, "learning_rate": 5e-05, "loss": 0.9027, "mean_token_accuracy": 0.8487668037414551, "num_tokens": 91284713.0, "step": 2402 }, { "epoch": 0.3056862994529958, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.987987995147705, "learning_rate": 5e-05, "loss": 0.788, "mean_token_accuracy": 0.8493145704269409, "num_tokens": 91322143.0, "step": 2403 }, { "epoch": 0.30581350973158633, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.3457255363464355, "learning_rate": 5e-05, "loss": 0.8222, "mean_token_accuracy": 0.8483925461769104, "num_tokens": 91366825.0, "step": 2404 }, { "epoch": 0.3059407200101768, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.732253551483154, "learning_rate": 5e-05, "loss": 0.8905, "mean_token_accuracy": 0.8458178639411926, "num_tokens": 91404469.0, "step": 2405 }, { "epoch": 0.30606793028876733, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.470022678375244, "learning_rate": 5e-05, "loss": 0.7565, "mean_token_accuracy": 0.8443711996078491, "num_tokens": 91444799.0, "step": 2406 }, { "epoch": 0.30619514056735786, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.362082481384277, "learning_rate": 5e-05, "loss": 0.8802, "mean_token_accuracy": 0.8518499135971069, "num_tokens": 91483843.0, "step": 2407 }, { "epoch": 0.30632235084594833, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 5.995370864868164, "learning_rate": 5e-05, "loss": 0.7101, "mean_token_accuracy": 0.8529826998710632, "num_tokens": 91523922.0, "step": 2408 }, { "epoch": 0.30644956112453886, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.842031478881836, "learning_rate": 5e-05, "loss": 0.9126, "mean_token_accuracy": 0.8578800559043884, "num_tokens": 91561155.0, "step": 2409 }, { "epoch": 0.3065767714031294, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.012680530548096, "learning_rate": 5e-05, "loss": 0.7195, "mean_token_accuracy": 0.8549271821975708, "num_tokens": 91596072.0, "step": 2410 }, { "epoch": 0.30670398168171986, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.964312553405762, "learning_rate": 5e-05, "loss": 0.9879, "mean_token_accuracy": 0.8515983819961548, "num_tokens": 91640369.0, "step": 2411 }, { "epoch": 0.3068311919603104, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 6.723545551300049, "learning_rate": 5e-05, "loss": 0.8129, "mean_token_accuracy": 0.8561168909072876, "num_tokens": 91674941.0, "step": 2412 }, { "epoch": 0.3069584022389009, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.309698104858398, "learning_rate": 5e-05, "loss": 0.9495, "mean_token_accuracy": 0.8551168441772461, "num_tokens": 91709023.0, "step": 2413 }, { "epoch": 0.3070856125174914, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.564542293548584, "learning_rate": 5e-05, "loss": 0.9109, "mean_token_accuracy": 0.8542027473449707, "num_tokens": 91750684.0, "step": 2414 }, { "epoch": 0.3072128227960819, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.238002777099609, "learning_rate": 5e-05, "loss": 0.847, "mean_token_accuracy": 0.8513867855072021, "num_tokens": 91789369.0, "step": 2415 }, { "epoch": 0.30734003307467245, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.950619220733643, "learning_rate": 5e-05, "loss": 0.9421, "mean_token_accuracy": 0.841080904006958, "num_tokens": 91828228.0, "step": 2416 }, { "epoch": 0.3074672433532629, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.894795894622803, "learning_rate": 5e-05, "loss": 0.75, "mean_token_accuracy": 0.8631619215011597, "num_tokens": 91866191.0, "step": 2417 }, { "epoch": 0.30759445363185345, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.575647830963135, "learning_rate": 5e-05, "loss": 0.8286, "mean_token_accuracy": 0.8545621037483215, "num_tokens": 91907945.0, "step": 2418 }, { "epoch": 0.307721663910444, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.349552631378174, "learning_rate": 5e-05, "loss": 0.7772, "mean_token_accuracy": 0.8597475290298462, "num_tokens": 91938651.0, "step": 2419 }, { "epoch": 0.30784887418903445, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.843587875366211, "learning_rate": 5e-05, "loss": 0.7749, "mean_token_accuracy": 0.8538150191307068, "num_tokens": 91977557.0, "step": 2420 }, { "epoch": 0.307976084467625, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.911683082580566, "learning_rate": 5e-05, "loss": 0.8622, "mean_token_accuracy": 0.84555983543396, "num_tokens": 92012316.0, "step": 2421 }, { "epoch": 0.3081032947462155, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.366189956665039, "learning_rate": 5e-05, "loss": 0.7233, "mean_token_accuracy": 0.8526416420936584, "num_tokens": 92049262.0, "step": 2422 }, { "epoch": 0.308230505024806, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.280389785766602, "learning_rate": 5e-05, "loss": 0.8978, "mean_token_accuracy": 0.8480808734893799, "num_tokens": 92086048.0, "step": 2423 }, { "epoch": 0.3083577153033965, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.191096782684326, "learning_rate": 5e-05, "loss": 0.6917, "mean_token_accuracy": 0.8639333248138428, "num_tokens": 92123540.0, "step": 2424 }, { "epoch": 0.30848492558198704, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.568137168884277, "learning_rate": 5e-05, "loss": 0.974, "mean_token_accuracy": 0.8331009745597839, "num_tokens": 92158275.0, "step": 2425 }, { "epoch": 0.3086121358605775, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.2439866065979, "learning_rate": 5e-05, "loss": 0.6881, "mean_token_accuracy": 0.8706459403038025, "num_tokens": 92196444.0, "step": 2426 }, { "epoch": 0.30873934613916804, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.652185440063477, "learning_rate": 5e-05, "loss": 0.9199, "mean_token_accuracy": 0.8596757054328918, "num_tokens": 92233988.0, "step": 2427 }, { "epoch": 0.30886655641775856, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.672426223754883, "learning_rate": 5e-05, "loss": 0.7725, "mean_token_accuracy": 0.8632062673568726, "num_tokens": 92271102.0, "step": 2428 }, { "epoch": 0.3089937666963491, "ewc_loss": 0.5, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.260337829589844, "learning_rate": 5e-05, "loss": 0.9389, "mean_token_accuracy": 0.8514341115951538, "num_tokens": 92306112.0, "step": 2429 }, { "epoch": 0.30912097697493957, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.185612678527832, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.858984112739563, "num_tokens": 92347796.0, "step": 2430 }, { "epoch": 0.3092481872535301, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.649119853973389, "learning_rate": 5e-05, "loss": 0.8976, "mean_token_accuracy": 0.844868540763855, "num_tokens": 92382018.0, "step": 2431 }, { "epoch": 0.3093753975321206, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.3944220542907715, "learning_rate": 5e-05, "loss": 0.8649, "mean_token_accuracy": 0.8498048186302185, "num_tokens": 92428352.0, "step": 2432 }, { "epoch": 0.3095026078107111, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.359438896179199, "learning_rate": 5e-05, "loss": 0.8285, "mean_token_accuracy": 0.850479006767273, "num_tokens": 92462826.0, "step": 2433 }, { "epoch": 0.3096298180893016, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.285680294036865, "learning_rate": 5e-05, "loss": 0.7788, "mean_token_accuracy": 0.8604730367660522, "num_tokens": 92498371.0, "step": 2434 }, { "epoch": 0.30975702836789215, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.339351654052734, "learning_rate": 5e-05, "loss": 0.7927, "mean_token_accuracy": 0.8539206981658936, "num_tokens": 92536921.0, "step": 2435 }, { "epoch": 0.3098842386464826, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.142192840576172, "learning_rate": 5e-05, "loss": 0.811, "mean_token_accuracy": 0.8440825343132019, "num_tokens": 92570627.0, "step": 2436 }, { "epoch": 0.31001144892507315, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.299699783325195, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8593867421150208, "num_tokens": 92608683.0, "step": 2437 }, { "epoch": 0.3101386592036637, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.168791770935059, "learning_rate": 5e-05, "loss": 0.7469, "mean_token_accuracy": 0.861816942691803, "num_tokens": 92644689.0, "step": 2438 }, { "epoch": 0.31026586948225415, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.092877388000488, "learning_rate": 5e-05, "loss": 0.7683, "mean_token_accuracy": 0.853980302810669, "num_tokens": 92684471.0, "step": 2439 }, { "epoch": 0.3103930797608447, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.3780059814453125, "learning_rate": 5e-05, "loss": 0.8189, "mean_token_accuracy": 0.8442789316177368, "num_tokens": 92724991.0, "step": 2440 }, { "epoch": 0.3105202900394352, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.803969383239746, "learning_rate": 5e-05, "loss": 0.7437, "mean_token_accuracy": 0.8469184637069702, "num_tokens": 92759302.0, "step": 2441 }, { "epoch": 0.3106475003180257, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.650618076324463, "learning_rate": 5e-05, "loss": 0.8164, "mean_token_accuracy": 0.8479585647583008, "num_tokens": 92794585.0, "step": 2442 }, { "epoch": 0.3107747105966162, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.541032314300537, "learning_rate": 5e-05, "loss": 0.7534, "mean_token_accuracy": 0.8485549688339233, "num_tokens": 92832405.0, "step": 2443 }, { "epoch": 0.31090192087520674, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.93714714050293, "learning_rate": 5e-05, "loss": 0.8459, "mean_token_accuracy": 0.8450915813446045, "num_tokens": 92877960.0, "step": 2444 }, { "epoch": 0.3110291311537972, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.271804332733154, "learning_rate": 5e-05, "loss": 0.6726, "mean_token_accuracy": 0.8648419380187988, "num_tokens": 92914421.0, "step": 2445 }, { "epoch": 0.31115634143238774, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.315115928649902, "learning_rate": 5e-05, "loss": 0.8571, "mean_token_accuracy": 0.8519995212554932, "num_tokens": 92954844.0, "step": 2446 }, { "epoch": 0.31128355171097827, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.035618782043457, "learning_rate": 5e-05, "loss": 0.692, "mean_token_accuracy": 0.8672449588775635, "num_tokens": 92987719.0, "step": 2447 }, { "epoch": 0.31141076198956874, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.7228422164917, "learning_rate": 5e-05, "loss": 0.9192, "mean_token_accuracy": 0.8515872359275818, "num_tokens": 93027335.0, "step": 2448 }, { "epoch": 0.31153797226815927, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.075153350830078, "learning_rate": 5e-05, "loss": 0.7399, "mean_token_accuracy": 0.8553433418273926, "num_tokens": 93063325.0, "step": 2449 }, { "epoch": 0.3116651825467498, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.804497718811035, "learning_rate": 5e-05, "loss": 0.9792, "mean_token_accuracy": 0.8456298112869263, "num_tokens": 93104057.0, "step": 2450 }, { "epoch": 0.31179239282534027, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.733126640319824, "learning_rate": 5e-05, "loss": 0.8444, "mean_token_accuracy": 0.8466094732284546, "num_tokens": 93138912.0, "step": 2451 }, { "epoch": 0.3119196031039308, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.055356979370117, "learning_rate": 5e-05, "loss": 0.968, "mean_token_accuracy": 0.8417273759841919, "num_tokens": 93178658.0, "step": 2452 }, { "epoch": 0.3120468133825213, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.601134300231934, "learning_rate": 5e-05, "loss": 0.905, "mean_token_accuracy": 0.8563588857650757, "num_tokens": 93217723.0, "step": 2453 }, { "epoch": 0.3121740236611118, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.884521484375, "learning_rate": 5e-05, "loss": 0.7571, "mean_token_accuracy": 0.8605067133903503, "num_tokens": 93253917.0, "step": 2454 }, { "epoch": 0.3123012339397023, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.23397159576416, "learning_rate": 5e-05, "loss": 0.8968, "mean_token_accuracy": 0.859655499458313, "num_tokens": 93289735.0, "step": 2455 }, { "epoch": 0.31242844421829286, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.314690113067627, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8521229028701782, "num_tokens": 93331549.0, "step": 2456 }, { "epoch": 0.31255565449688333, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.406418800354004, "learning_rate": 5e-05, "loss": 0.9468, "mean_token_accuracy": 0.839130163192749, "num_tokens": 93372977.0, "step": 2457 }, { "epoch": 0.31268286477547386, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.445376396179199, "learning_rate": 5e-05, "loss": 0.7182, "mean_token_accuracy": 0.8658480048179626, "num_tokens": 93406842.0, "step": 2458 }, { "epoch": 0.3128100750540644, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.118425369262695, "learning_rate": 5e-05, "loss": 0.8943, "mean_token_accuracy": 0.8552728295326233, "num_tokens": 93444113.0, "step": 2459 }, { "epoch": 0.31293728533265486, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.855783462524414, "learning_rate": 5e-05, "loss": 0.8356, "mean_token_accuracy": 0.8427250385284424, "num_tokens": 93490618.0, "step": 2460 }, { "epoch": 0.3130644956112454, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.639626979827881, "learning_rate": 5e-05, "loss": 0.865, "mean_token_accuracy": 0.8455526828765869, "num_tokens": 93524711.0, "step": 2461 }, { "epoch": 0.3131917058898359, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.158557891845703, "learning_rate": 5e-05, "loss": 0.8098, "mean_token_accuracy": 0.8536450862884521, "num_tokens": 93559169.0, "step": 2462 }, { "epoch": 0.3133189161684264, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.291635036468506, "learning_rate": 5e-05, "loss": 0.7977, "mean_token_accuracy": 0.8545030951499939, "num_tokens": 93598941.0, "step": 2463 }, { "epoch": 0.3134461264470169, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.256014823913574, "learning_rate": 5e-05, "loss": 0.7796, "mean_token_accuracy": 0.8585026264190674, "num_tokens": 93634979.0, "step": 2464 }, { "epoch": 0.31357333672560744, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.198077201843262, "learning_rate": 5e-05, "loss": 0.8333, "mean_token_accuracy": 0.831351637840271, "num_tokens": 93676736.0, "step": 2465 }, { "epoch": 0.3137005470041979, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.148311614990234, "learning_rate": 5e-05, "loss": 0.7515, "mean_token_accuracy": 0.8647819757461548, "num_tokens": 93720118.0, "step": 2466 }, { "epoch": 0.31382775728278844, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.272092819213867, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8606467843055725, "num_tokens": 93760814.0, "step": 2467 }, { "epoch": 0.313954967561379, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.9748358726501465, "learning_rate": 5e-05, "loss": 0.6949, "mean_token_accuracy": 0.869989812374115, "num_tokens": 93795326.0, "step": 2468 }, { "epoch": 0.31408217783996945, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.3917670249938965, "learning_rate": 5e-05, "loss": 0.7698, "mean_token_accuracy": 0.8557339906692505, "num_tokens": 93832993.0, "step": 2469 }, { "epoch": 0.31420938811856, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.817494869232178, "learning_rate": 5e-05, "loss": 0.7074, "mean_token_accuracy": 0.8669935464859009, "num_tokens": 93871282.0, "step": 2470 }, { "epoch": 0.3143365983971505, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.55452299118042, "learning_rate": 5e-05, "loss": 0.8039, "mean_token_accuracy": 0.8492050170898438, "num_tokens": 93912698.0, "step": 2471 }, { "epoch": 0.314463808675741, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.5995965003967285, "learning_rate": 5e-05, "loss": 0.707, "mean_token_accuracy": 0.8581012487411499, "num_tokens": 93952499.0, "step": 2472 }, { "epoch": 0.3145910189543315, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.837584972381592, "learning_rate": 5e-05, "loss": 0.8464, "mean_token_accuracy": 0.8406754732131958, "num_tokens": 93992277.0, "step": 2473 }, { "epoch": 0.31471822923292203, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.297183513641357, "learning_rate": 5e-05, "loss": 0.6933, "mean_token_accuracy": 0.8559583425521851, "num_tokens": 94032012.0, "step": 2474 }, { "epoch": 0.3148454395115125, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.247796058654785, "learning_rate": 5e-05, "loss": 0.8097, "mean_token_accuracy": 0.8613312244415283, "num_tokens": 94072210.0, "step": 2475 }, { "epoch": 0.31497264979010303, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 5.975426197052002, "learning_rate": 5e-05, "loss": 0.7147, "mean_token_accuracy": 0.851766049861908, "num_tokens": 94112269.0, "step": 2476 }, { "epoch": 0.31509986006869356, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.75621509552002, "learning_rate": 5e-05, "loss": 0.9789, "mean_token_accuracy": 0.8317251205444336, "num_tokens": 94153596.0, "step": 2477 }, { "epoch": 0.31522707034728403, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 5.843165874481201, "learning_rate": 5e-05, "loss": 0.7499, "mean_token_accuracy": 0.8465464115142822, "num_tokens": 94195146.0, "step": 2478 }, { "epoch": 0.31535428062587456, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.091191291809082, "learning_rate": 5e-05, "loss": 0.9489, "mean_token_accuracy": 0.8638627529144287, "num_tokens": 94233196.0, "step": 2479 }, { "epoch": 0.3154814909044651, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.411239147186279, "learning_rate": 5e-05, "loss": 0.819, "mean_token_accuracy": 0.8508409261703491, "num_tokens": 94275314.0, "step": 2480 }, { "epoch": 0.3156087011830556, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.632102012634277, "learning_rate": 5e-05, "loss": 0.9736, "mean_token_accuracy": 0.8567939400672913, "num_tokens": 94316673.0, "step": 2481 }, { "epoch": 0.3157359114616461, "ewc_loss": 0.5, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 7.470492362976074, "learning_rate": 5e-05, "loss": 0.9194, "mean_token_accuracy": 0.8583617210388184, "num_tokens": 94354118.0, "step": 2482 }, { "epoch": 0.3158631217402366, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.2417755126953125, "learning_rate": 5e-05, "loss": 0.8271, "mean_token_accuracy": 0.847690761089325, "num_tokens": 94386476.0, "step": 2483 }, { "epoch": 0.31599033201882715, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.398930549621582, "learning_rate": 5e-05, "loss": 0.9604, "mean_token_accuracy": 0.8545277118682861, "num_tokens": 94420146.0, "step": 2484 }, { "epoch": 0.3161175422974176, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.173041820526123, "learning_rate": 5e-05, "loss": 0.7509, "mean_token_accuracy": 0.8515501022338867, "num_tokens": 94455057.0, "step": 2485 }, { "epoch": 0.31624475257600815, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.998332977294922, "learning_rate": 5e-05, "loss": 0.9581, "mean_token_accuracy": 0.8617357015609741, "num_tokens": 94500666.0, "step": 2486 }, { "epoch": 0.3163719628545987, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.245965957641602, "learning_rate": 5e-05, "loss": 0.7226, "mean_token_accuracy": 0.8679930567741394, "num_tokens": 94537623.0, "step": 2487 }, { "epoch": 0.31649917313318915, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.886938095092773, "learning_rate": 5e-05, "loss": 0.9526, "mean_token_accuracy": 0.8655822277069092, "num_tokens": 94568563.0, "step": 2488 }, { "epoch": 0.3166263834117797, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.102281093597412, "learning_rate": 5e-05, "loss": 0.868, "mean_token_accuracy": 0.8677304983139038, "num_tokens": 94600931.0, "step": 2489 }, { "epoch": 0.3167535936903702, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.9904632568359375, "learning_rate": 5e-05, "loss": 0.9794, "mean_token_accuracy": 0.8358198404312134, "num_tokens": 94634345.0, "step": 2490 }, { "epoch": 0.3168808039689607, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.722716808319092, "learning_rate": 5e-05, "loss": 0.9204, "mean_token_accuracy": 0.8664820790290833, "num_tokens": 94673157.0, "step": 2491 }, { "epoch": 0.3170080142475512, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.161334037780762, "learning_rate": 5e-05, "loss": 0.8594, "mean_token_accuracy": 0.8398929834365845, "num_tokens": 94712727.0, "step": 2492 }, { "epoch": 0.31713522452614173, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00018787384033203125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.818105220794678, "learning_rate": 5e-05, "loss": 0.9731, "mean_token_accuracy": 0.8363763689994812, "num_tokens": 94752341.0, "step": 2493 }, { "epoch": 0.3172624348047322, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.042259216308594, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.8447303175926208, "num_tokens": 94791287.0, "step": 2494 }, { "epoch": 0.31738964508332274, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.446728229522705, "learning_rate": 5e-05, "loss": 0.844, "mean_token_accuracy": 0.8480290174484253, "num_tokens": 94833029.0, "step": 2495 }, { "epoch": 0.31751685536191326, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.295006275177002, "learning_rate": 5e-05, "loss": 0.8155, "mean_token_accuracy": 0.851473331451416, "num_tokens": 94873400.0, "step": 2496 }, { "epoch": 0.31764406564050374, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.037215232849121, "learning_rate": 5e-05, "loss": 0.8178, "mean_token_accuracy": 0.834807813167572, "num_tokens": 94914603.0, "step": 2497 }, { "epoch": 0.31777127591909426, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.518797874450684, "learning_rate": 5e-05, "loss": 0.7586, "mean_token_accuracy": 0.8732458353042603, "num_tokens": 94955925.0, "step": 2498 }, { "epoch": 0.3178984861976848, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.794971466064453, "learning_rate": 5e-05, "loss": 0.6926, "mean_token_accuracy": 0.8717514276504517, "num_tokens": 94989420.0, "step": 2499 }, { "epoch": 0.31802569647627527, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.6866135597229, "learning_rate": 5e-05, "loss": 0.8404, "mean_token_accuracy": 0.8479582071304321, "num_tokens": 95028835.0, "step": 2500 }, { "epoch": 0.3181529067548658, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.56006383895874, "learning_rate": 5e-05, "loss": 0.8055, "mean_token_accuracy": 0.8410196304321289, "num_tokens": 95069246.0, "step": 2501 }, { "epoch": 0.3182801170334563, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.996501445770264, "learning_rate": 5e-05, "loss": 0.9494, "mean_token_accuracy": 0.8222394585609436, "num_tokens": 95110184.0, "step": 2502 }, { "epoch": 0.3184073273120468, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.252754211425781, "learning_rate": 5e-05, "loss": 0.7254, "mean_token_accuracy": 0.8567811250686646, "num_tokens": 95149118.0, "step": 2503 }, { "epoch": 0.3185345375906373, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.419379234313965, "learning_rate": 5e-05, "loss": 0.8772, "mean_token_accuracy": 0.8550687432289124, "num_tokens": 95195605.0, "step": 2504 }, { "epoch": 0.31866174786922785, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.047575950622559, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.8475536108016968, "num_tokens": 95234602.0, "step": 2505 }, { "epoch": 0.3187889581478183, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.781291961669922, "learning_rate": 5e-05, "loss": 0.9218, "mean_token_accuracy": 0.863747775554657, "num_tokens": 95275646.0, "step": 2506 }, { "epoch": 0.31891616842640885, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.208086013793945, "learning_rate": 5e-05, "loss": 0.7624, "mean_token_accuracy": 0.8508240580558777, "num_tokens": 95309948.0, "step": 2507 }, { "epoch": 0.3190433787049994, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.792867660522461, "learning_rate": 5e-05, "loss": 0.9907, "mean_token_accuracy": 0.8534902334213257, "num_tokens": 95347506.0, "step": 2508 }, { "epoch": 0.31917058898358985, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 6.817042350769043, "learning_rate": 5e-05, "loss": 0.8493, "mean_token_accuracy": 0.8536983132362366, "num_tokens": 95386809.0, "step": 2509 }, { "epoch": 0.3192977992621804, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.318572998046875, "learning_rate": 5e-05, "loss": 1.0011, "mean_token_accuracy": 0.8446155786514282, "num_tokens": 95429287.0, "step": 2510 }, { "epoch": 0.3194250095407709, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.360910892486572, "learning_rate": 5e-05, "loss": 0.8325, "mean_token_accuracy": 0.875342845916748, "num_tokens": 95466443.0, "step": 2511 }, { "epoch": 0.3195522198193614, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.513476371765137, "learning_rate": 5e-05, "loss": 0.8314, "mean_token_accuracy": 0.8612338900566101, "num_tokens": 95511593.0, "step": 2512 }, { "epoch": 0.3196794300979519, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.6211676597595215, "learning_rate": 5e-05, "loss": 0.8733, "mean_token_accuracy": 0.8555136322975159, "num_tokens": 95546756.0, "step": 2513 }, { "epoch": 0.31980664037654244, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.116142749786377, "learning_rate": 5e-05, "loss": 0.798, "mean_token_accuracy": 0.8539695739746094, "num_tokens": 95586434.0, "step": 2514 }, { "epoch": 0.3199338506551329, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.431369304656982, "learning_rate": 5e-05, "loss": 0.7702, "mean_token_accuracy": 0.8743646144866943, "num_tokens": 95616980.0, "step": 2515 }, { "epoch": 0.32006106093372344, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.271115779876709, "learning_rate": 5e-05, "loss": 0.8055, "mean_token_accuracy": 0.8551137447357178, "num_tokens": 95656919.0, "step": 2516 }, { "epoch": 0.32018827121231397, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.030343532562256, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8562906980514526, "num_tokens": 95691785.0, "step": 2517 }, { "epoch": 0.32031548149090444, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.523767948150635, "learning_rate": 5e-05, "loss": 0.8157, "mean_token_accuracy": 0.8502946496009827, "num_tokens": 95731055.0, "step": 2518 }, { "epoch": 0.32044269176949497, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.786009788513184, "learning_rate": 5e-05, "loss": 0.6929, "mean_token_accuracy": 0.8721301555633545, "num_tokens": 95767609.0, "step": 2519 }, { "epoch": 0.3205699020480855, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.58656120300293, "learning_rate": 5e-05, "loss": 0.7584, "mean_token_accuracy": 0.8659927845001221, "num_tokens": 95805156.0, "step": 2520 }, { "epoch": 0.32069711232667597, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.78262996673584, "learning_rate": 5e-05, "loss": 0.7134, "mean_token_accuracy": 0.8640718460083008, "num_tokens": 95848809.0, "step": 2521 }, { "epoch": 0.3208243226052665, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.520757675170898, "learning_rate": 5e-05, "loss": 0.8027, "mean_token_accuracy": 0.8473922610282898, "num_tokens": 95882961.0, "step": 2522 }, { "epoch": 0.320951532883857, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.797131538391113, "learning_rate": 5e-05, "loss": 0.6927, "mean_token_accuracy": 0.867902934551239, "num_tokens": 95916360.0, "step": 2523 }, { "epoch": 0.3210787431624475, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.508592128753662, "learning_rate": 5e-05, "loss": 0.7738, "mean_token_accuracy": 0.8571294546127319, "num_tokens": 95952064.0, "step": 2524 }, { "epoch": 0.32120595344103803, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.752954006195068, "learning_rate": 5e-05, "loss": 0.7385, "mean_token_accuracy": 0.8541829586029053, "num_tokens": 95993073.0, "step": 2525 }, { "epoch": 0.32133316371962856, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.599770545959473, "learning_rate": 5e-05, "loss": 0.8215, "mean_token_accuracy": 0.8427581191062927, "num_tokens": 96028659.0, "step": 2526 }, { "epoch": 0.32146037399821903, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.621626853942871, "learning_rate": 5e-05, "loss": 0.7771, "mean_token_accuracy": 0.8488268256187439, "num_tokens": 96064802.0, "step": 2527 }, { "epoch": 0.32158758427680956, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.760921478271484, "learning_rate": 5e-05, "loss": 0.8333, "mean_token_accuracy": 0.8469661474227905, "num_tokens": 96112980.0, "step": 2528 }, { "epoch": 0.3217147945554001, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.4865264892578125, "learning_rate": 5e-05, "loss": 0.7481, "mean_token_accuracy": 0.8438775539398193, "num_tokens": 96144114.0, "step": 2529 }, { "epoch": 0.3218420048339906, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.9442830085754395, "learning_rate": 5e-05, "loss": 0.8186, "mean_token_accuracy": 0.8570027351379395, "num_tokens": 96177383.0, "step": 2530 }, { "epoch": 0.3219692151125811, "ewc_loss": 0.28125, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.3371477127075195, "learning_rate": 5e-05, "loss": 0.7133, "mean_token_accuracy": 0.8597071170806885, "num_tokens": 96216513.0, "step": 2531 }, { "epoch": 0.3220964253911716, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.139120101928711, "learning_rate": 5e-05, "loss": 0.8425, "mean_token_accuracy": 0.8613337278366089, "num_tokens": 96256588.0, "step": 2532 }, { "epoch": 0.32222363566976214, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.300887107849121, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.851953387260437, "num_tokens": 96295365.0, "step": 2533 }, { "epoch": 0.3223508459483526, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.211033821105957, "learning_rate": 5e-05, "loss": 0.9011, "mean_token_accuracy": 0.8506163358688354, "num_tokens": 96328048.0, "step": 2534 }, { "epoch": 0.32247805622694314, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.496256351470947, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8545418381690979, "num_tokens": 96362088.0, "step": 2535 }, { "epoch": 0.32260526650553367, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00019359588623046875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.918920516967773, "learning_rate": 5e-05, "loss": 0.881, "mean_token_accuracy": 0.8506125211715698, "num_tokens": 96396161.0, "step": 2536 }, { "epoch": 0.32273247678412414, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00019359588623046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.958598613739014, "learning_rate": 5e-05, "loss": 0.7754, "mean_token_accuracy": 0.8546581268310547, "num_tokens": 96433402.0, "step": 2537 }, { "epoch": 0.3228596870627147, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00019359588623046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.236721515655518, "learning_rate": 5e-05, "loss": 0.7934, "mean_token_accuracy": 0.8475414514541626, "num_tokens": 96473011.0, "step": 2538 }, { "epoch": 0.3229868973413052, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00019359588623046875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.558846950531006, "learning_rate": 5e-05, "loss": 0.796, "mean_token_accuracy": 0.862064003944397, "num_tokens": 96511595.0, "step": 2539 }, { "epoch": 0.3231141076198957, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00019359588623046875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.521731376647949, "learning_rate": 5e-05, "loss": 0.7396, "mean_token_accuracy": 0.8505781292915344, "num_tokens": 96546192.0, "step": 2540 }, { "epoch": 0.3232413178984862, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00019359588623046875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.14242172241211, "learning_rate": 5e-05, "loss": 0.7637, "mean_token_accuracy": 0.879145622253418, "num_tokens": 96582613.0, "step": 2541 }, { "epoch": 0.32336852817707673, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.014102458953857, "learning_rate": 5e-05, "loss": 0.716, "mean_token_accuracy": 0.8501841425895691, "num_tokens": 96620219.0, "step": 2542 }, { "epoch": 0.3234957384556672, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.717527389526367, "learning_rate": 5e-05, "loss": 0.9213, "mean_token_accuracy": 0.85133957862854, "num_tokens": 96661126.0, "step": 2543 }, { "epoch": 0.32362294873425773, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.812305450439453, "learning_rate": 5e-05, "loss": 0.7197, "mean_token_accuracy": 0.8477020263671875, "num_tokens": 96697995.0, "step": 2544 }, { "epoch": 0.32375015901284826, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.114886283874512, "learning_rate": 5e-05, "loss": 0.9786, "mean_token_accuracy": 0.8572105169296265, "num_tokens": 96732542.0, "step": 2545 }, { "epoch": 0.32387736929143873, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.357052326202393, "learning_rate": 5e-05, "loss": 0.8201, "mean_token_accuracy": 0.8432302474975586, "num_tokens": 96761787.0, "step": 2546 }, { "epoch": 0.32400457957002926, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.72485065460205, "learning_rate": 5e-05, "loss": 0.9933, "mean_token_accuracy": 0.8485815525054932, "num_tokens": 96804918.0, "step": 2547 }, { "epoch": 0.3241317898486198, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.438002586364746, "learning_rate": 5e-05, "loss": 0.9226, "mean_token_accuracy": 0.8532463312149048, "num_tokens": 96843646.0, "step": 2548 }, { "epoch": 0.32425900012721026, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.399777412414551, "learning_rate": 5e-05, "loss": 0.8959, "mean_token_accuracy": 0.8362008333206177, "num_tokens": 96874796.0, "step": 2549 }, { "epoch": 0.3243862104058008, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.253543853759766, "learning_rate": 5e-05, "loss": 0.9345, "mean_token_accuracy": 0.8599369525909424, "num_tokens": 96912777.0, "step": 2550 }, { "epoch": 0.3245134206843913, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.422626972198486, "learning_rate": 5e-05, "loss": 0.7763, "mean_token_accuracy": 0.8471019268035889, "num_tokens": 96952045.0, "step": 2551 }, { "epoch": 0.3246406309629818, "ewc_loss": 0.5, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.543031692504883, "learning_rate": 5e-05, "loss": 0.9129, "mean_token_accuracy": 0.8593534827232361, "num_tokens": 96984625.0, "step": 2552 }, { "epoch": 0.3247678412415723, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.593056678771973, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8607305288314819, "num_tokens": 97022409.0, "step": 2553 }, { "epoch": 0.32489505152016285, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.922832012176514, "learning_rate": 5e-05, "loss": 0.9188, "mean_token_accuracy": 0.8475091457366943, "num_tokens": 97062461.0, "step": 2554 }, { "epoch": 0.3250222617987533, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.298162937164307, "learning_rate": 5e-05, "loss": 0.9328, "mean_token_accuracy": 0.8302968740463257, "num_tokens": 97102010.0, "step": 2555 }, { "epoch": 0.32514947207734385, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.041278839111328, "learning_rate": 5e-05, "loss": 0.765, "mean_token_accuracy": 0.8576185703277588, "num_tokens": 97140304.0, "step": 2556 }, { "epoch": 0.3252766823559344, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.768442153930664, "learning_rate": 5e-05, "loss": 0.799, "mean_token_accuracy": 0.8750149011611938, "num_tokens": 97180460.0, "step": 2557 }, { "epoch": 0.32540389263452485, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.643223762512207, "learning_rate": 5e-05, "loss": 0.7802, "mean_token_accuracy": 0.8447975516319275, "num_tokens": 97223439.0, "step": 2558 }, { "epoch": 0.3255311029131154, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.918445110321045, "learning_rate": 5e-05, "loss": 0.9159, "mean_token_accuracy": 0.8404698371887207, "num_tokens": 97257050.0, "step": 2559 }, { "epoch": 0.3256583131917059, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.6015496253967285, "learning_rate": 5e-05, "loss": 0.7692, "mean_token_accuracy": 0.8485369086265564, "num_tokens": 97296790.0, "step": 2560 }, { "epoch": 0.3257855234702964, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.86895751953125, "learning_rate": 5e-05, "loss": 0.8419, "mean_token_accuracy": 0.8523197174072266, "num_tokens": 97336511.0, "step": 2561 }, { "epoch": 0.3259127337488869, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.6484527587890625, "learning_rate": 5e-05, "loss": 0.7438, "mean_token_accuracy": 0.854853630065918, "num_tokens": 97373589.0, "step": 2562 }, { "epoch": 0.32603994402747744, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.822503566741943, "learning_rate": 5e-05, "loss": 0.7923, "mean_token_accuracy": 0.8696780204772949, "num_tokens": 97413986.0, "step": 2563 }, { "epoch": 0.3261671543060679, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.62872314453125, "learning_rate": 5e-05, "loss": 0.7654, "mean_token_accuracy": 0.8491767644882202, "num_tokens": 97457878.0, "step": 2564 }, { "epoch": 0.32629436458465844, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.897855758666992, "learning_rate": 5e-05, "loss": 0.8268, "mean_token_accuracy": 0.859703540802002, "num_tokens": 97495463.0, "step": 2565 }, { "epoch": 0.32642157486324896, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.4971537590026855, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8476731777191162, "num_tokens": 97533748.0, "step": 2566 }, { "epoch": 0.32654878514183944, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.08843994140625, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.8564175963401794, "num_tokens": 97572476.0, "step": 2567 }, { "epoch": 0.32667599542042997, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.331496715545654, "learning_rate": 5e-05, "loss": 0.7044, "mean_token_accuracy": 0.8599387407302856, "num_tokens": 97606658.0, "step": 2568 }, { "epoch": 0.3268032056990205, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.338944435119629, "learning_rate": 5e-05, "loss": 0.8677, "mean_token_accuracy": 0.8596406579017639, "num_tokens": 97645083.0, "step": 2569 }, { "epoch": 0.32693041597761097, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.250626564025879, "learning_rate": 5e-05, "loss": 0.7445, "mean_token_accuracy": 0.8515273332595825, "num_tokens": 97679440.0, "step": 2570 }, { "epoch": 0.3270576262562015, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.496028900146484, "learning_rate": 5e-05, "loss": 1.0053, "mean_token_accuracy": 0.8338987231254578, "num_tokens": 97719295.0, "step": 2571 }, { "epoch": 0.327184836534792, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.453921318054199, "learning_rate": 5e-05, "loss": 0.7413, "mean_token_accuracy": 0.863189697265625, "num_tokens": 97759417.0, "step": 2572 }, { "epoch": 0.3273120468133825, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.349738121032715, "learning_rate": 5e-05, "loss": 0.9681, "mean_token_accuracy": 0.8432748317718506, "num_tokens": 97804004.0, "step": 2573 }, { "epoch": 0.327439257091973, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.900383949279785, "learning_rate": 5e-05, "loss": 0.8938, "mean_token_accuracy": 0.8394678831100464, "num_tokens": 97845496.0, "step": 2574 }, { "epoch": 0.32756646737056355, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.791338920593262, "learning_rate": 5e-05, "loss": 0.9156, "mean_token_accuracy": 0.8527721166610718, "num_tokens": 97887184.0, "step": 2575 }, { "epoch": 0.327693677649154, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.332039833068848, "learning_rate": 5e-05, "loss": 0.8963, "mean_token_accuracy": 0.8434288501739502, "num_tokens": 97924498.0, "step": 2576 }, { "epoch": 0.32782088792774455, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.190126419067383, "learning_rate": 5e-05, "loss": 0.7885, "mean_token_accuracy": 0.860306978225708, "num_tokens": 97960225.0, "step": 2577 }, { "epoch": 0.3279480982063351, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.498355865478516, "learning_rate": 5e-05, "loss": 0.8798, "mean_token_accuracy": 0.839225709438324, "num_tokens": 98001746.0, "step": 2578 }, { "epoch": 0.3280753084849256, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.989691734313965, "learning_rate": 5e-05, "loss": 0.7598, "mean_token_accuracy": 0.8578199148178101, "num_tokens": 98038812.0, "step": 2579 }, { "epoch": 0.3282025187635161, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.348004341125488, "learning_rate": 5e-05, "loss": 0.8204, "mean_token_accuracy": 0.8521304130554199, "num_tokens": 98073782.0, "step": 2580 }, { "epoch": 0.3283297290421066, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.130471229553223, "learning_rate": 5e-05, "loss": 0.7335, "mean_token_accuracy": 0.8661399483680725, "num_tokens": 98109379.0, "step": 2581 }, { "epoch": 0.32845693932069714, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.0895466804504395, "learning_rate": 5e-05, "loss": 0.73, "mean_token_accuracy": 0.8659650087356567, "num_tokens": 98141661.0, "step": 2582 }, { "epoch": 0.3285841495992876, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.3074235916137695, "learning_rate": 5e-05, "loss": 0.8109, "mean_token_accuracy": 0.8441407084465027, "num_tokens": 98182634.0, "step": 2583 }, { "epoch": 0.32871135987787814, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.901480674743652, "learning_rate": 5e-05, "loss": 0.7378, "mean_token_accuracy": 0.8567216992378235, "num_tokens": 98225356.0, "step": 2584 }, { "epoch": 0.32883857015646867, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.400350570678711, "learning_rate": 5e-05, "loss": 0.7745, "mean_token_accuracy": 0.8544180393218994, "num_tokens": 98263575.0, "step": 2585 }, { "epoch": 0.32896578043505914, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.787732124328613, "learning_rate": 5e-05, "loss": 0.7414, "mean_token_accuracy": 0.8516430258750916, "num_tokens": 98304993.0, "step": 2586 }, { "epoch": 0.32909299071364967, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.496853351593018, "learning_rate": 5e-05, "loss": 0.7552, "mean_token_accuracy": 0.8591389060020447, "num_tokens": 98339963.0, "step": 2587 }, { "epoch": 0.3292202009922402, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.663668155670166, "learning_rate": 5e-05, "loss": 0.7032, "mean_token_accuracy": 0.8595820665359497, "num_tokens": 98375507.0, "step": 2588 }, { "epoch": 0.32934741127083067, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.644880294799805, "learning_rate": 5e-05, "loss": 0.7781, "mean_token_accuracy": 0.8563935160636902, "num_tokens": 98413577.0, "step": 2589 }, { "epoch": 0.3294746215494212, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.49852991104126, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8453953266143799, "num_tokens": 98451615.0, "step": 2590 }, { "epoch": 0.3296018318280117, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.8983259201049805, "learning_rate": 5e-05, "loss": 0.8961, "mean_token_accuracy": 0.8300856351852417, "num_tokens": 98490728.0, "step": 2591 }, { "epoch": 0.3297290421066022, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.24470853805542, "learning_rate": 5e-05, "loss": 0.6617, "mean_token_accuracy": 0.8677400946617126, "num_tokens": 98526645.0, "step": 2592 }, { "epoch": 0.3298562523851927, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.242681503295898, "learning_rate": 5e-05, "loss": 0.8621, "mean_token_accuracy": 0.8466004133224487, "num_tokens": 98568447.0, "step": 2593 }, { "epoch": 0.32998346266378326, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.037761688232422, "learning_rate": 5e-05, "loss": 0.7058, "mean_token_accuracy": 0.8551815748214722, "num_tokens": 98608952.0, "step": 2594 }, { "epoch": 0.33011067294237373, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.583523750305176, "learning_rate": 5e-05, "loss": 0.833, "mean_token_accuracy": 0.8727138042449951, "num_tokens": 98649180.0, "step": 2595 }, { "epoch": 0.33023788322096426, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.120761871337891, "learning_rate": 5e-05, "loss": 0.7445, "mean_token_accuracy": 0.8532817363739014, "num_tokens": 98692506.0, "step": 2596 }, { "epoch": 0.3303650934995548, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.521925926208496, "learning_rate": 5e-05, "loss": 0.9248, "mean_token_accuracy": 0.8517165184020996, "num_tokens": 98727817.0, "step": 2597 }, { "epoch": 0.33049230377814526, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.746677398681641, "learning_rate": 5e-05, "loss": 0.792, "mean_token_accuracy": 0.8581767082214355, "num_tokens": 98769714.0, "step": 2598 }, { "epoch": 0.3306195140567358, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.688255786895752, "learning_rate": 5e-05, "loss": 0.8362, "mean_token_accuracy": 0.8625129461288452, "num_tokens": 98811894.0, "step": 2599 }, { "epoch": 0.3307467243353263, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00019931793212890625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.638808727264404, "learning_rate": 5e-05, "loss": 0.8972, "mean_token_accuracy": 0.845912754535675, "num_tokens": 98847117.0, "step": 2600 }, { "epoch": 0.3308739346139168, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00019931793212890625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.495831489562988, "learning_rate": 5e-05, "loss": 0.7328, "mean_token_accuracy": 0.8568881750106812, "num_tokens": 98882745.0, "step": 2601 }, { "epoch": 0.3310011448925073, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.51024341583252, "learning_rate": 5e-05, "loss": 0.881, "mean_token_accuracy": 0.8620210886001587, "num_tokens": 98918039.0, "step": 2602 }, { "epoch": 0.33112835517109784, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.808072090148926, "learning_rate": 5e-05, "loss": 0.7054, "mean_token_accuracy": 0.8547167181968689, "num_tokens": 98958571.0, "step": 2603 }, { "epoch": 0.3312555654496883, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.211531639099121, "learning_rate": 5e-05, "loss": 0.9793, "mean_token_accuracy": 0.8558921813964844, "num_tokens": 99004970.0, "step": 2604 }, { "epoch": 0.33138277572827884, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.0433502197265625, "learning_rate": 5e-05, "loss": 0.8076, "mean_token_accuracy": 0.8430610299110413, "num_tokens": 99041112.0, "step": 2605 }, { "epoch": 0.3315099860068694, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.27724552154541, "learning_rate": 5e-05, "loss": 1.0184, "mean_token_accuracy": 0.8566723465919495, "num_tokens": 99076771.0, "step": 2606 }, { "epoch": 0.33163719628545985, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.2581467628479, "learning_rate": 5e-05, "loss": 0.9603, "mean_token_accuracy": 0.8524384498596191, "num_tokens": 99112183.0, "step": 2607 }, { "epoch": 0.3317644065640504, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00019931793212890625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.145699501037598, "learning_rate": 5e-05, "loss": 0.9231, "mean_token_accuracy": 0.8550248146057129, "num_tokens": 99148789.0, "step": 2608 }, { "epoch": 0.3318916168426409, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00019931793212890625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.197152137756348, "learning_rate": 5e-05, "loss": 1.073, "mean_token_accuracy": 0.8316391706466675, "num_tokens": 99189697.0, "step": 2609 }, { "epoch": 0.3320188271212314, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.755507946014404, "learning_rate": 5e-05, "loss": 0.8311, "mean_token_accuracy": 0.8505231738090515, "num_tokens": 99228269.0, "step": 2610 }, { "epoch": 0.3321460373998219, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.704916954040527, "learning_rate": 5e-05, "loss": 0.9722, "mean_token_accuracy": 0.8532827496528625, "num_tokens": 99264275.0, "step": 2611 }, { "epoch": 0.33227324767841243, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.633820056915283, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.8578444719314575, "num_tokens": 99294598.0, "step": 2612 }, { "epoch": 0.3324004579570029, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00019931793212890625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.207632064819336, "learning_rate": 5e-05, "loss": 0.906, "mean_token_accuracy": 0.8598412275314331, "num_tokens": 99336307.0, "step": 2613 }, { "epoch": 0.33252766823559343, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00019931793212890625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.345362663269043, "learning_rate": 5e-05, "loss": 0.8558, "mean_token_accuracy": 0.8677052855491638, "num_tokens": 99373983.0, "step": 2614 }, { "epoch": 0.33265487851418396, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.1872687339782715, "learning_rate": 5e-05, "loss": 0.8232, "mean_token_accuracy": 0.8491533398628235, "num_tokens": 99415794.0, "step": 2615 }, { "epoch": 0.33278208879277443, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.886017799377441, "learning_rate": 5e-05, "loss": 0.9169, "mean_token_accuracy": 0.849054217338562, "num_tokens": 99457572.0, "step": 2616 }, { "epoch": 0.33290929907136496, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.6306376457214355, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8549018502235413, "num_tokens": 99496074.0, "step": 2617 }, { "epoch": 0.3330365093499555, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.099180221557617, "learning_rate": 5e-05, "loss": 0.9144, "mean_token_accuracy": 0.8467424511909485, "num_tokens": 99539655.0, "step": 2618 }, { "epoch": 0.33316371962854596, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.588193893432617, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.8454141616821289, "num_tokens": 99574696.0, "step": 2619 }, { "epoch": 0.3332909299071365, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.073938369750977, "learning_rate": 5e-05, "loss": 0.8754, "mean_token_accuracy": 0.858726978302002, "num_tokens": 99612960.0, "step": 2620 }, { "epoch": 0.333418140185727, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.67787504196167, "learning_rate": 5e-05, "loss": 0.8594, "mean_token_accuracy": 0.8301221132278442, "num_tokens": 99644107.0, "step": 2621 }, { "epoch": 0.3335453504643175, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.967489719390869, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.851584792137146, "num_tokens": 99682014.0, "step": 2622 }, { "epoch": 0.333672560742908, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.715019702911377, "learning_rate": 5e-05, "loss": 0.8053, "mean_token_accuracy": 0.8473938703536987, "num_tokens": 99718969.0, "step": 2623 }, { "epoch": 0.33379977102149855, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.919248104095459, "learning_rate": 5e-05, "loss": 0.9043, "mean_token_accuracy": 0.8391258716583252, "num_tokens": 99756955.0, "step": 2624 }, { "epoch": 0.333926981300089, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.643553733825684, "learning_rate": 5e-05, "loss": 0.733, "mean_token_accuracy": 0.8623921275138855, "num_tokens": 99792216.0, "step": 2625 }, { "epoch": 0.33405419157867955, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.009639739990234, "learning_rate": 5e-05, "loss": 0.8795, "mean_token_accuracy": 0.8447224497795105, "num_tokens": 99823973.0, "step": 2626 }, { "epoch": 0.3341814018572701, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.484224319458008, "learning_rate": 5e-05, "loss": 0.773, "mean_token_accuracy": 0.848364531993866, "num_tokens": 99857206.0, "step": 2627 }, { "epoch": 0.33430861213586055, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.22741413116455, "learning_rate": 5e-05, "loss": 0.9479, "mean_token_accuracy": 0.8361256718635559, "num_tokens": 99891146.0, "step": 2628 }, { "epoch": 0.3344358224144511, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.340475082397461, "learning_rate": 5e-05, "loss": 0.811, "mean_token_accuracy": 0.8369666337966919, "num_tokens": 99931456.0, "step": 2629 }, { "epoch": 0.3345630326930416, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.436003684997559, "learning_rate": 5e-05, "loss": 0.9342, "mean_token_accuracy": 0.8496468663215637, "num_tokens": 99965968.0, "step": 2630 }, { "epoch": 0.33469024297163213, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.4238457679748535, "learning_rate": 5e-05, "loss": 0.8032, "mean_token_accuracy": 0.8428114652633667, "num_tokens": 100010874.0, "step": 2631 }, { "epoch": 0.3348174532502226, "ewc_loss": 0.5, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.380836486816406, "learning_rate": 5e-05, "loss": 0.9263, "mean_token_accuracy": 0.8593363165855408, "num_tokens": 100052166.0, "step": 2632 }, { "epoch": 0.33494466352881314, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 6.812752723693848, "learning_rate": 5e-05, "loss": 0.8157, "mean_token_accuracy": 0.8583992123603821, "num_tokens": 100093927.0, "step": 2633 }, { "epoch": 0.33507187380740366, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.926937103271484, "learning_rate": 5e-05, "loss": 0.9158, "mean_token_accuracy": 0.8478695750236511, "num_tokens": 100128414.0, "step": 2634 }, { "epoch": 0.33519908408599414, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.240403175354004, "learning_rate": 5e-05, "loss": 0.8226, "mean_token_accuracy": 0.8606624603271484, "num_tokens": 100160334.0, "step": 2635 }, { "epoch": 0.33532629436458466, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.302516460418701, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.8541727066040039, "num_tokens": 100195188.0, "step": 2636 }, { "epoch": 0.3354535046431752, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.479888439178467, "learning_rate": 5e-05, "loss": 0.839, "mean_token_accuracy": 0.8584486246109009, "num_tokens": 100235338.0, "step": 2637 }, { "epoch": 0.33558071492176567, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.988522052764893, "learning_rate": 5e-05, "loss": 0.7794, "mean_token_accuracy": 0.8562911152839661, "num_tokens": 100276267.0, "step": 2638 }, { "epoch": 0.3357079252003562, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.444517612457275, "learning_rate": 5e-05, "loss": 0.867, "mean_token_accuracy": 0.8368015289306641, "num_tokens": 100316201.0, "step": 2639 }, { "epoch": 0.3358351354789467, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.012705326080322, "learning_rate": 5e-05, "loss": 0.8243, "mean_token_accuracy": 0.8396119475364685, "num_tokens": 100353278.0, "step": 2640 }, { "epoch": 0.3359623457575372, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.259114742279053, "learning_rate": 5e-05, "loss": 0.8027, "mean_token_accuracy": 0.8493987917900085, "num_tokens": 100384185.0, "step": 2641 }, { "epoch": 0.3360895560361277, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.135455131530762, "learning_rate": 5e-05, "loss": 0.8183, "mean_token_accuracy": 0.8346129655838013, "num_tokens": 100423232.0, "step": 2642 }, { "epoch": 0.33621676631471825, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.096048831939697, "learning_rate": 5e-05, "loss": 0.8243, "mean_token_accuracy": 0.8320345282554626, "num_tokens": 100457912.0, "step": 2643 }, { "epoch": 0.3363439765933087, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.196072101593018, "learning_rate": 5e-05, "loss": 0.7596, "mean_token_accuracy": 0.8569552898406982, "num_tokens": 100499894.0, "step": 2644 }, { "epoch": 0.33647118687189925, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.063785552978516, "learning_rate": 5e-05, "loss": 0.7211, "mean_token_accuracy": 0.8670406341552734, "num_tokens": 100540961.0, "step": 2645 }, { "epoch": 0.3365983971504898, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.133152484893799, "learning_rate": 5e-05, "loss": 0.7645, "mean_token_accuracy": 0.8481833934783936, "num_tokens": 100581302.0, "step": 2646 }, { "epoch": 0.33672560742908025, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.091831684112549, "learning_rate": 5e-05, "loss": 0.721, "mean_token_accuracy": 0.8580533266067505, "num_tokens": 100615289.0, "step": 2647 }, { "epoch": 0.3368528177076708, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.068535327911377, "learning_rate": 5e-05, "loss": 0.714, "mean_token_accuracy": 0.8651478290557861, "num_tokens": 100659911.0, "step": 2648 }, { "epoch": 0.3369800279862613, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.125278472900391, "learning_rate": 5e-05, "loss": 0.7384, "mean_token_accuracy": 0.8565049171447754, "num_tokens": 100700861.0, "step": 2649 }, { "epoch": 0.3371072382648518, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.040407180786133, "learning_rate": 5e-05, "loss": 0.7492, "mean_token_accuracy": 0.8520358204841614, "num_tokens": 100737329.0, "step": 2650 }, { "epoch": 0.3372344485434423, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.113378047943115, "learning_rate": 5e-05, "loss": 0.6979, "mean_token_accuracy": 0.8692331314086914, "num_tokens": 100774278.0, "step": 2651 }, { "epoch": 0.33736165882203284, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.022280693054199, "learning_rate": 5e-05, "loss": 0.7466, "mean_token_accuracy": 0.8492789268493652, "num_tokens": 100819083.0, "step": 2652 }, { "epoch": 0.3374888691006233, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.127791404724121, "learning_rate": 5e-05, "loss": 0.8064, "mean_token_accuracy": 0.838753342628479, "num_tokens": 100859711.0, "step": 2653 }, { "epoch": 0.33761607937921384, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.9900360107421875, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8304827213287354, "num_tokens": 100893915.0, "step": 2654 }, { "epoch": 0.33774328965780437, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.151096820831299, "learning_rate": 5e-05, "loss": 0.7146, "mean_token_accuracy": 0.8616076111793518, "num_tokens": 100932609.0, "step": 2655 }, { "epoch": 0.33787049993639484, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.938844203948975, "learning_rate": 5e-05, "loss": 0.718, "mean_token_accuracy": 0.854500412940979, "num_tokens": 100966539.0, "step": 2656 }, { "epoch": 0.33799771021498537, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.224867343902588, "learning_rate": 5e-05, "loss": 0.7942, "mean_token_accuracy": 0.8414685130119324, "num_tokens": 100999979.0, "step": 2657 }, { "epoch": 0.3381249204935759, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.818537712097168, "learning_rate": 5e-05, "loss": 0.7465, "mean_token_accuracy": 0.8460984826087952, "num_tokens": 101036546.0, "step": 2658 }, { "epoch": 0.33825213077216637, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.385856628417969, "learning_rate": 5e-05, "loss": 0.7326, "mean_token_accuracy": 0.8525429964065552, "num_tokens": 101067861.0, "step": 2659 }, { "epoch": 0.3383793410507569, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 0.00024127960205078125, "grad_norm": 6.5747294425964355, "learning_rate": 5e-05, "loss": 0.6344, "mean_token_accuracy": 0.8720223307609558, "num_tokens": 101104268.0, "step": 2660 }, { "epoch": 0.3385065513293474, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.72874116897583, "learning_rate": 5e-05, "loss": 0.7076, "mean_token_accuracy": 0.8707231879234314, "num_tokens": 101143364.0, "step": 2661 }, { "epoch": 0.3386337616079379, "ewc_loss": 0.23828125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 0.00021648406982421875, "grad_norm": 6.160647392272949, "learning_rate": 5e-05, "loss": 0.7157, "mean_token_accuracy": 0.8419122099876404, "num_tokens": 101178359.0, "step": 2662 }, { "epoch": 0.33876097188652843, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.38850212097168, "learning_rate": 5e-05, "loss": 0.7934, "mean_token_accuracy": 0.866205632686615, "num_tokens": 101215993.0, "step": 2663 }, { "epoch": 0.33888818216511896, "ewc_loss": 0.20703125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 0.00018596649169921875, "grad_norm": 5.439488410949707, "learning_rate": 5e-05, "loss": 0.6002, "mean_token_accuracy": 0.8668472766876221, "num_tokens": 101254924.0, "step": 2664 }, { "epoch": 0.33901539244370943, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.655198097229004, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.8636956810951233, "num_tokens": 101295632.0, "step": 2665 }, { "epoch": 0.33914260272229996, "ewc_loss": 0.236328125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 0.00021457672119140625, "grad_norm": 4.938054084777832, "learning_rate": 5e-05, "loss": 0.642, "mean_token_accuracy": 0.860528290271759, "num_tokens": 101333295.0, "step": 2666 }, { "epoch": 0.3392698130008905, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.000827789306640625, "grad_norm": 11.492024421691895, "learning_rate": 5e-05, "loss": 1.2693, "mean_token_accuracy": 0.8582378625869751, "num_tokens": 101372319.0, "step": 2667 }, { "epoch": 0.33939702327948096, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 7.448543548583984, "learning_rate": 5e-05, "loss": 1.0335, "mean_token_accuracy": 0.8550141453742981, "num_tokens": 101405300.0, "step": 2668 }, { "epoch": 0.3395242335580715, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 11.710649490356445, "learning_rate": 5e-05, "loss": 1.3114, "mean_token_accuracy": 0.8411002159118652, "num_tokens": 101445576.0, "step": 2669 }, { "epoch": 0.339651443836662, "ewc_loss": 1.015625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.0009918212890625, "grad_norm": 9.645291328430176, "learning_rate": 5e-05, "loss": 1.5213, "mean_token_accuracy": 0.8357994556427002, "num_tokens": 101481840.0, "step": 2670 }, { "epoch": 0.3397786541152525, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.49692153930664, "learning_rate": 5e-05, "loss": 1.1061, "mean_token_accuracy": 0.8441917896270752, "num_tokens": 101511968.0, "step": 2671 }, { "epoch": 0.339905864393843, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 8.992878913879395, "learning_rate": 5e-05, "loss": 1.198, "mean_token_accuracy": 0.8578011393547058, "num_tokens": 101549909.0, "step": 2672 }, { "epoch": 0.34003307467243354, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.613527297973633, "learning_rate": 5e-05, "loss": 1.0746, "mean_token_accuracy": 0.8464717864990234, "num_tokens": 101580892.0, "step": 2673 }, { "epoch": 0.340160284951024, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.47040843963623, "learning_rate": 5e-05, "loss": 1.0111, "mean_token_accuracy": 0.8564066886901855, "num_tokens": 101614589.0, "step": 2674 }, { "epoch": 0.34028749522961454, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.553606986999512, "learning_rate": 5e-05, "loss": 0.9793, "mean_token_accuracy": 0.8578605055809021, "num_tokens": 101658988.0, "step": 2675 }, { "epoch": 0.3404147055082051, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.187511444091797, "learning_rate": 5e-05, "loss": 1.0238, "mean_token_accuracy": 0.837405800819397, "num_tokens": 101697838.0, "step": 2676 }, { "epoch": 0.34054191578679555, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.858870983123779, "learning_rate": 5e-05, "loss": 0.8599, "mean_token_accuracy": 0.8639330863952637, "num_tokens": 101733831.0, "step": 2677 }, { "epoch": 0.3406691260653861, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.00814437866211, "learning_rate": 5e-05, "loss": 0.91, "mean_token_accuracy": 0.8559077978134155, "num_tokens": 101772697.0, "step": 2678 }, { "epoch": 0.3407963363439766, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.3862223625183105, "learning_rate": 5e-05, "loss": 0.9083, "mean_token_accuracy": 0.8378037214279175, "num_tokens": 101810107.0, "step": 2679 }, { "epoch": 0.34092354662256713, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.854025840759277, "learning_rate": 5e-05, "loss": 0.8609, "mean_token_accuracy": 0.857147216796875, "num_tokens": 101847239.0, "step": 2680 }, { "epoch": 0.3410507569011576, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.097229957580566, "learning_rate": 5e-05, "loss": 0.8246, "mean_token_accuracy": 0.8447348475456238, "num_tokens": 101884508.0, "step": 2681 }, { "epoch": 0.34117796717974813, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.818499565124512, "learning_rate": 5e-05, "loss": 0.7744, "mean_token_accuracy": 0.8778045773506165, "num_tokens": 101925351.0, "step": 2682 }, { "epoch": 0.34130517745833866, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.839513778686523, "learning_rate": 5e-05, "loss": 0.7574, "mean_token_accuracy": 0.8566951751708984, "num_tokens": 101966086.0, "step": 2683 }, { "epoch": 0.34143238773692913, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.9811201095581055, "learning_rate": 5e-05, "loss": 0.8432, "mean_token_accuracy": 0.8589442372322083, "num_tokens": 102004935.0, "step": 2684 }, { "epoch": 0.34155959801551966, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.541163921356201, "learning_rate": 5e-05, "loss": 0.7548, "mean_token_accuracy": 0.853979766368866, "num_tokens": 102039026.0, "step": 2685 }, { "epoch": 0.3416868082941102, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.331513404846191, "learning_rate": 5e-05, "loss": 0.8925, "mean_token_accuracy": 0.8538063764572144, "num_tokens": 102078990.0, "step": 2686 }, { "epoch": 0.34181401857270066, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.248952388763428, "learning_rate": 5e-05, "loss": 0.7418, "mean_token_accuracy": 0.8568651080131531, "num_tokens": 102114631.0, "step": 2687 }, { "epoch": 0.3419412288512912, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.84623908996582, "learning_rate": 5e-05, "loss": 0.9686, "mean_token_accuracy": 0.8557690978050232, "num_tokens": 102161583.0, "step": 2688 }, { "epoch": 0.3420684391298817, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.321595668792725, "learning_rate": 5e-05, "loss": 0.7545, "mean_token_accuracy": 0.8594329357147217, "num_tokens": 102202183.0, "step": 2689 }, { "epoch": 0.3421956494084722, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.100550651550293, "learning_rate": 5e-05, "loss": 0.9835, "mean_token_accuracy": 0.8637170791625977, "num_tokens": 102235288.0, "step": 2690 }, { "epoch": 0.3423228596870627, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.1018500328063965, "learning_rate": 5e-05, "loss": 0.9241, "mean_token_accuracy": 0.8479719161987305, "num_tokens": 102272814.0, "step": 2691 }, { "epoch": 0.34245006996565325, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.35806655883789, "learning_rate": 5e-05, "loss": 0.947, "mean_token_accuracy": 0.8523603677749634, "num_tokens": 102312763.0, "step": 2692 }, { "epoch": 0.3425772802442437, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 7.933940887451172, "learning_rate": 5e-05, "loss": 0.9749, "mean_token_accuracy": 0.8568614721298218, "num_tokens": 102351490.0, "step": 2693 }, { "epoch": 0.34270449052283425, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.912674903869629, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.8526490926742554, "num_tokens": 102394495.0, "step": 2694 }, { "epoch": 0.3428317008014248, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.786978721618652, "learning_rate": 5e-05, "loss": 1.0143, "mean_token_accuracy": 0.8487640619277954, "num_tokens": 102431689.0, "step": 2695 }, { "epoch": 0.34295891108001525, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 5.972688674926758, "learning_rate": 5e-05, "loss": 0.7205, "mean_token_accuracy": 0.8630478382110596, "num_tokens": 102468653.0, "step": 2696 }, { "epoch": 0.3430861213586058, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.837010383605957, "learning_rate": 5e-05, "loss": 1.0208, "mean_token_accuracy": 0.8679423332214355, "num_tokens": 102504442.0, "step": 2697 }, { "epoch": 0.3432133316371963, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 6.641838550567627, "learning_rate": 5e-05, "loss": 0.8556, "mean_token_accuracy": 0.8610107898712158, "num_tokens": 102539401.0, "step": 2698 }, { "epoch": 0.3433405419157868, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.099316596984863, "learning_rate": 5e-05, "loss": 1.1306, "mean_token_accuracy": 0.8595646619796753, "num_tokens": 102579874.0, "step": 2699 }, { "epoch": 0.3434677521943773, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 8.382513046264648, "learning_rate": 5e-05, "loss": 1.1358, "mean_token_accuracy": 0.8504654765129089, "num_tokens": 102617906.0, "step": 2700 }, { "epoch": 0.34359496247296784, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.667202949523926, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.857667088508606, "num_tokens": 102657744.0, "step": 2701 }, { "epoch": 0.3437221727515583, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 8.747589111328125, "learning_rate": 5e-05, "loss": 1.0951, "mean_token_accuracy": 0.8491145372390747, "num_tokens": 102696650.0, "step": 2702 }, { "epoch": 0.34384938303014884, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.418449878692627, "learning_rate": 5e-05, "loss": 0.8676, "mean_token_accuracy": 0.8529777526855469, "num_tokens": 102729114.0, "step": 2703 }, { "epoch": 0.34397659330873936, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.80501937866211, "learning_rate": 5e-05, "loss": 1.0114, "mean_token_accuracy": 0.8552244901657104, "num_tokens": 102763562.0, "step": 2704 }, { "epoch": 0.34410380358732984, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.276488780975342, "learning_rate": 5e-05, "loss": 0.9123, "mean_token_accuracy": 0.8469216823577881, "num_tokens": 102803680.0, "step": 2705 }, { "epoch": 0.34423101386592037, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.202668190002441, "learning_rate": 5e-05, "loss": 0.9672, "mean_token_accuracy": 0.8420116305351257, "num_tokens": 102847932.0, "step": 2706 }, { "epoch": 0.3443582241445109, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.610042572021484, "learning_rate": 5e-05, "loss": 0.8994, "mean_token_accuracy": 0.8597897291183472, "num_tokens": 102888371.0, "step": 2707 }, { "epoch": 0.34448543442310137, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.315627098083496, "learning_rate": 5e-05, "loss": 0.8711, "mean_token_accuracy": 0.836889922618866, "num_tokens": 102928943.0, "step": 2708 }, { "epoch": 0.3446126447016919, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.9061808586120605, "learning_rate": 5e-05, "loss": 0.917, "mean_token_accuracy": 0.8526977896690369, "num_tokens": 102969773.0, "step": 2709 }, { "epoch": 0.3447398549802824, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.74407434463501, "learning_rate": 5e-05, "loss": 0.8055, "mean_token_accuracy": 0.8427152037620544, "num_tokens": 103012496.0, "step": 2710 }, { "epoch": 0.3448670652588729, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.303885459899902, "learning_rate": 5e-05, "loss": 0.9411, "mean_token_accuracy": 0.8484747409820557, "num_tokens": 103047259.0, "step": 2711 }, { "epoch": 0.3449942755374634, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.416491508483887, "learning_rate": 5e-05, "loss": 0.7261, "mean_token_accuracy": 0.8616856932640076, "num_tokens": 103087534.0, "step": 2712 }, { "epoch": 0.34512148581605395, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.790921211242676, "learning_rate": 5e-05, "loss": 0.9398, "mean_token_accuracy": 0.8633868098258972, "num_tokens": 103128782.0, "step": 2713 }, { "epoch": 0.3452486960946444, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.502470970153809, "learning_rate": 5e-05, "loss": 0.7889, "mean_token_accuracy": 0.8564857244491577, "num_tokens": 103161252.0, "step": 2714 }, { "epoch": 0.34537590637323495, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.8944673538208, "learning_rate": 5e-05, "loss": 0.9613, "mean_token_accuracy": 0.8611375093460083, "num_tokens": 103201524.0, "step": 2715 }, { "epoch": 0.3455031166518255, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 6.998884677886963, "learning_rate": 5e-05, "loss": 0.9056, "mean_token_accuracy": 0.8442428112030029, "num_tokens": 103235444.0, "step": 2716 }, { "epoch": 0.34563032693041595, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.492795944213867, "learning_rate": 5e-05, "loss": 0.9796, "mean_token_accuracy": 0.8412686586380005, "num_tokens": 103270927.0, "step": 2717 }, { "epoch": 0.3457575372090065, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00021266937255859375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.277289390563965, "learning_rate": 5e-05, "loss": 0.8789, "mean_token_accuracy": 0.8619190454483032, "num_tokens": 103310518.0, "step": 2718 }, { "epoch": 0.345884747487597, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.081932067871094, "learning_rate": 5e-05, "loss": 0.8381, "mean_token_accuracy": 0.8735038042068481, "num_tokens": 103347940.0, "step": 2719 }, { "epoch": 0.3460119577661875, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.130270957946777, "learning_rate": 5e-05, "loss": 0.8327, "mean_token_accuracy": 0.8653578758239746, "num_tokens": 103383995.0, "step": 2720 }, { "epoch": 0.346139168044778, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.081658363342285, "learning_rate": 5e-05, "loss": 0.824, "mean_token_accuracy": 0.8758170008659363, "num_tokens": 103416411.0, "step": 2721 }, { "epoch": 0.34626637832336854, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.771186828613281, "learning_rate": 5e-05, "loss": 0.8194, "mean_token_accuracy": 0.8476543426513672, "num_tokens": 103454199.0, "step": 2722 }, { "epoch": 0.346393588601959, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.326218605041504, "learning_rate": 5e-05, "loss": 0.9452, "mean_token_accuracy": 0.8453189134597778, "num_tokens": 103490996.0, "step": 2723 }, { "epoch": 0.34652079888054954, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.601049423217773, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.8617401123046875, "num_tokens": 103525987.0, "step": 2724 }, { "epoch": 0.34664800915914007, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.513993263244629, "learning_rate": 5e-05, "loss": 0.9802, "mean_token_accuracy": 0.8432695865631104, "num_tokens": 103565702.0, "step": 2725 }, { "epoch": 0.34677521943773054, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.797791004180908, "learning_rate": 5e-05, "loss": 0.825, "mean_token_accuracy": 0.8545415997505188, "num_tokens": 103601764.0, "step": 2726 }, { "epoch": 0.34690242971632107, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.267359733581543, "learning_rate": 5e-05, "loss": 0.9354, "mean_token_accuracy": 0.8497301340103149, "num_tokens": 103640293.0, "step": 2727 }, { "epoch": 0.3470296399949116, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.240738391876221, "learning_rate": 5e-05, "loss": 0.8645, "mean_token_accuracy": 0.852348804473877, "num_tokens": 103680464.0, "step": 2728 }, { "epoch": 0.3471568502735021, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.61660099029541, "learning_rate": 5e-05, "loss": 0.8468, "mean_token_accuracy": 0.8580126762390137, "num_tokens": 103717434.0, "step": 2729 }, { "epoch": 0.3472840605520926, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.500150203704834, "learning_rate": 5e-05, "loss": 0.834, "mean_token_accuracy": 0.8571968674659729, "num_tokens": 103755186.0, "step": 2730 }, { "epoch": 0.3474112708306831, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.161997318267822, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8693728446960449, "num_tokens": 103789516.0, "step": 2731 }, { "epoch": 0.34753848110927366, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.479940414428711, "learning_rate": 5e-05, "loss": 0.8366, "mean_token_accuracy": 0.8481667041778564, "num_tokens": 103826103.0, "step": 2732 }, { "epoch": 0.34766569138786413, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.099605083465576, "learning_rate": 5e-05, "loss": 0.7836, "mean_token_accuracy": 0.8550262451171875, "num_tokens": 103863059.0, "step": 2733 }, { "epoch": 0.34779290166645466, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.317370414733887, "learning_rate": 5e-05, "loss": 0.7525, "mean_token_accuracy": 0.8617690801620483, "num_tokens": 103897976.0, "step": 2734 }, { "epoch": 0.3479201119450452, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.152950763702393, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.838785707950592, "num_tokens": 103935485.0, "step": 2735 }, { "epoch": 0.34804732222363566, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.192816257476807, "learning_rate": 5e-05, "loss": 0.7477, "mean_token_accuracy": 0.8587323427200317, "num_tokens": 103971917.0, "step": 2736 }, { "epoch": 0.3481745325022262, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.196112632751465, "learning_rate": 5e-05, "loss": 0.8053, "mean_token_accuracy": 0.8461083769798279, "num_tokens": 104008797.0, "step": 2737 }, { "epoch": 0.3483017427808167, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.1017351150512695, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8456540703773499, "num_tokens": 104049551.0, "step": 2738 }, { "epoch": 0.3484289530594072, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.248713493347168, "learning_rate": 5e-05, "loss": 0.8115, "mean_token_accuracy": 0.8385555744171143, "num_tokens": 104089341.0, "step": 2739 }, { "epoch": 0.3485561633379977, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.002602577209473, "learning_rate": 5e-05, "loss": 0.7564, "mean_token_accuracy": 0.8515551090240479, "num_tokens": 104133445.0, "step": 2740 }, { "epoch": 0.34868337361658824, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.3012213706970215, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.8563302755355835, "num_tokens": 104170435.0, "step": 2741 }, { "epoch": 0.3488105838951787, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.922794342041016, "learning_rate": 5e-05, "loss": 0.6979, "mean_token_accuracy": 0.8644713163375854, "num_tokens": 104205660.0, "step": 2742 }, { "epoch": 0.34893779417376924, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.359487056732178, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.8634140491485596, "num_tokens": 104243749.0, "step": 2743 }, { "epoch": 0.3490650044523598, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.8199920654296875, "learning_rate": 5e-05, "loss": 0.7255, "mean_token_accuracy": 0.8565948009490967, "num_tokens": 104283411.0, "step": 2744 }, { "epoch": 0.34919221473095025, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.505317687988281, "learning_rate": 5e-05, "loss": 0.7486, "mean_token_accuracy": 0.8583871126174927, "num_tokens": 104324570.0, "step": 2745 }, { "epoch": 0.3493194250095408, "ewc_loss": 0.28125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.649472713470459, "learning_rate": 5e-05, "loss": 0.7409, "mean_token_accuracy": 0.85162752866745, "num_tokens": 104363619.0, "step": 2746 }, { "epoch": 0.3494466352881313, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.725711345672607, "learning_rate": 5e-05, "loss": 0.7582, "mean_token_accuracy": 0.8601183295249939, "num_tokens": 104405572.0, "step": 2747 }, { "epoch": 0.3495738455667218, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.370573043823242, "learning_rate": 5e-05, "loss": 0.7061, "mean_token_accuracy": 0.8484818339347839, "num_tokens": 104446772.0, "step": 2748 }, { "epoch": 0.3497010558453123, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.124469757080078, "learning_rate": 5e-05, "loss": 0.7925, "mean_token_accuracy": 0.8594674468040466, "num_tokens": 104490912.0, "step": 2749 }, { "epoch": 0.34982826612390283, "ewc_loss": 0.251953125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.00022983551025390625, "grad_norm": 5.997826099395752, "learning_rate": 5e-05, "loss": 0.647, "mean_token_accuracy": 0.8614979386329651, "num_tokens": 104525307.0, "step": 2750 }, { "epoch": 0.3499554764024933, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.727858543395996, "learning_rate": 5e-05, "loss": 0.9053, "mean_token_accuracy": 0.8516129851341248, "num_tokens": 104562431.0, "step": 2751 }, { "epoch": 0.35008268668108383, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.00023555755615234375, "grad_norm": 5.7157440185546875, "learning_rate": 5e-05, "loss": 0.6405, "mean_token_accuracy": 0.873160183429718, "num_tokens": 104596440.0, "step": 2752 }, { "epoch": 0.35020989695967436, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.254443168640137, "learning_rate": 5e-05, "loss": 0.9278, "mean_token_accuracy": 0.8663163781166077, "num_tokens": 104630804.0, "step": 2753 }, { "epoch": 0.35033710723826483, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.221584796905518, "learning_rate": 5e-05, "loss": 0.7374, "mean_token_accuracy": 0.8642456531524658, "num_tokens": 104667997.0, "step": 2754 }, { "epoch": 0.35046431751685536, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.922377586364746, "learning_rate": 5e-05, "loss": 0.9699, "mean_token_accuracy": 0.860078752040863, "num_tokens": 104711605.0, "step": 2755 }, { "epoch": 0.3505915277954459, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.52087926864624, "learning_rate": 5e-05, "loss": 0.9446, "mean_token_accuracy": 0.8492459058761597, "num_tokens": 104748427.0, "step": 2756 }, { "epoch": 0.35071873807403636, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.261025905609131, "learning_rate": 5e-05, "loss": 0.7936, "mean_token_accuracy": 0.8587307929992676, "num_tokens": 104784536.0, "step": 2757 }, { "epoch": 0.3508459483526269, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.837706565856934, "learning_rate": 5e-05, "loss": 0.9775, "mean_token_accuracy": 0.8615800142288208, "num_tokens": 104820096.0, "step": 2758 }, { "epoch": 0.3509731586312174, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.757670879364014, "learning_rate": 5e-05, "loss": 0.6923, "mean_token_accuracy": 0.8594822883605957, "num_tokens": 104857016.0, "step": 2759 }, { "epoch": 0.3511003689098079, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.372407913208008, "learning_rate": 5e-05, "loss": 1.1038, "mean_token_accuracy": 0.8543708324432373, "num_tokens": 104903999.0, "step": 2760 }, { "epoch": 0.3512275791883984, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 6.596224784851074, "learning_rate": 5e-05, "loss": 0.9384, "mean_token_accuracy": 0.8351718187332153, "num_tokens": 104943954.0, "step": 2761 }, { "epoch": 0.35135478946698895, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.800283432006836, "learning_rate": 5e-05, "loss": 1.1704, "mean_token_accuracy": 0.8458390235900879, "num_tokens": 104980028.0, "step": 2762 }, { "epoch": 0.3514819997455794, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 0.000732421875, "grad_norm": 8.802162170410156, "learning_rate": 5e-05, "loss": 1.1787, "mean_token_accuracy": 0.8590543866157532, "num_tokens": 105020471.0, "step": 2763 }, { "epoch": 0.35160921002416995, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.27916431427002, "learning_rate": 5e-05, "loss": 0.9794, "mean_token_accuracy": 0.855972409248352, "num_tokens": 105050419.0, "step": 2764 }, { "epoch": 0.3517364203027605, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 8.948678970336914, "learning_rate": 5e-05, "loss": 1.061, "mean_token_accuracy": 0.8626326322555542, "num_tokens": 105079656.0, "step": 2765 }, { "epoch": 0.35186363058135095, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.097660064697266, "learning_rate": 5e-05, "loss": 0.9125, "mean_token_accuracy": 0.8583985567092896, "num_tokens": 105119715.0, "step": 2766 }, { "epoch": 0.3519908408599415, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.79708480834961, "learning_rate": 5e-05, "loss": 1.0076, "mean_token_accuracy": 0.8486583232879639, "num_tokens": 105158281.0, "step": 2767 }, { "epoch": 0.352118051138532, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.753663539886475, "learning_rate": 5e-05, "loss": 0.9512, "mean_token_accuracy": 0.8478540778160095, "num_tokens": 105201372.0, "step": 2768 }, { "epoch": 0.3522452614171225, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.03178596496582, "learning_rate": 5e-05, "loss": 0.8715, "mean_token_accuracy": 0.8511916399002075, "num_tokens": 105240557.0, "step": 2769 }, { "epoch": 0.352372471695713, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.976181983947754, "learning_rate": 5e-05, "loss": 0.9429, "mean_token_accuracy": 0.8553811311721802, "num_tokens": 105276701.0, "step": 2770 }, { "epoch": 0.35249968197430354, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.09729528427124, "learning_rate": 5e-05, "loss": 0.8451, "mean_token_accuracy": 0.8441710472106934, "num_tokens": 105319185.0, "step": 2771 }, { "epoch": 0.352626892252894, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.416241645812988, "learning_rate": 5e-05, "loss": 0.8626, "mean_token_accuracy": 0.8728885650634766, "num_tokens": 105362800.0, "step": 2772 }, { "epoch": 0.35275410253148454, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.674692630767822, "learning_rate": 5e-05, "loss": 0.7917, "mean_token_accuracy": 0.8522104620933533, "num_tokens": 105400731.0, "step": 2773 }, { "epoch": 0.35288131281007507, "ewc_loss": 0.5, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.643783569335938, "learning_rate": 5e-05, "loss": 0.9747, "mean_token_accuracy": 0.8495175242424011, "num_tokens": 105439712.0, "step": 2774 }, { "epoch": 0.35300852308866554, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.73246955871582, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8572596311569214, "num_tokens": 105475159.0, "step": 2775 }, { "epoch": 0.35313573336725607, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.52016544342041, "learning_rate": 5e-05, "loss": 0.9487, "mean_token_accuracy": 0.8448729515075684, "num_tokens": 105511323.0, "step": 2776 }, { "epoch": 0.3532629436458466, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 6.997511386871338, "learning_rate": 5e-05, "loss": 0.8279, "mean_token_accuracy": 0.8563085794448853, "num_tokens": 105544056.0, "step": 2777 }, { "epoch": 0.35339015392443707, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.06407642364502, "learning_rate": 5e-05, "loss": 0.8297, "mean_token_accuracy": 0.8701189756393433, "num_tokens": 105582335.0, "step": 2778 }, { "epoch": 0.3535173642030276, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00021839141845703125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.228580474853516, "learning_rate": 5e-05, "loss": 0.8274, "mean_token_accuracy": 0.8601551055908203, "num_tokens": 105614046.0, "step": 2779 }, { "epoch": 0.3536445744816181, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.609871864318848, "learning_rate": 5e-05, "loss": 0.8027, "mean_token_accuracy": 0.8630478382110596, "num_tokens": 105648294.0, "step": 2780 }, { "epoch": 0.35377178476020865, "ewc_loss": 0.375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.244380950927734, "learning_rate": 5e-05, "loss": 0.8204, "mean_token_accuracy": 0.8520076870918274, "num_tokens": 105685707.0, "step": 2781 }, { "epoch": 0.3538989950387991, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.507914066314697, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.855873703956604, "num_tokens": 105724662.0, "step": 2782 }, { "epoch": 0.35402620531738965, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.965644836425781, "learning_rate": 5e-05, "loss": 0.816, "mean_token_accuracy": 0.8420993089675903, "num_tokens": 105756781.0, "step": 2783 }, { "epoch": 0.3541534155959802, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.793123245239258, "learning_rate": 5e-05, "loss": 0.8047, "mean_token_accuracy": 0.8643072247505188, "num_tokens": 105798466.0, "step": 2784 }, { "epoch": 0.35428062587457065, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.522558689117432, "learning_rate": 5e-05, "loss": 0.7232, "mean_token_accuracy": 0.8589583039283752, "num_tokens": 105835085.0, "step": 2785 }, { "epoch": 0.3544078361531612, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.306525230407715, "learning_rate": 5e-05, "loss": 0.8434, "mean_token_accuracy": 0.8652912378311157, "num_tokens": 105872268.0, "step": 2786 }, { "epoch": 0.3545350464317517, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.139058589935303, "learning_rate": 5e-05, "loss": 0.7179, "mean_token_accuracy": 0.8532673716545105, "num_tokens": 105905721.0, "step": 2787 }, { "epoch": 0.3546622567103422, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.898444175720215, "learning_rate": 5e-05, "loss": 1.0047, "mean_token_accuracy": 0.8399455547332764, "num_tokens": 105945573.0, "step": 2788 }, { "epoch": 0.3547894669889327, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.151500701904297, "learning_rate": 5e-05, "loss": 0.7364, "mean_token_accuracy": 0.854896068572998, "num_tokens": 105978800.0, "step": 2789 }, { "epoch": 0.35491667726752324, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.177199363708496, "learning_rate": 5e-05, "loss": 1.021, "mean_token_accuracy": 0.8559079170227051, "num_tokens": 106014035.0, "step": 2790 }, { "epoch": 0.3550438875461137, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 6.971917152404785, "learning_rate": 5e-05, "loss": 0.9653, "mean_token_accuracy": 0.8291738033294678, "num_tokens": 106050721.0, "step": 2791 }, { "epoch": 0.35517109782470424, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.458650588989258, "learning_rate": 5e-05, "loss": 0.9299, "mean_token_accuracy": 0.8590948581695557, "num_tokens": 106086802.0, "step": 2792 }, { "epoch": 0.35529830810329477, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.848010063171387, "learning_rate": 5e-05, "loss": 0.9168, "mean_token_accuracy": 0.8655577898025513, "num_tokens": 106124174.0, "step": 2793 }, { "epoch": 0.35542551838188524, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.141668319702148, "learning_rate": 5e-05, "loss": 0.7985, "mean_token_accuracy": 0.8596606254577637, "num_tokens": 106166147.0, "step": 2794 }, { "epoch": 0.35555272866047577, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.4299898147583, "learning_rate": 5e-05, "loss": 0.9791, "mean_token_accuracy": 0.8463002443313599, "num_tokens": 106206758.0, "step": 2795 }, { "epoch": 0.3556799389390663, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.436549186706543, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.8517040610313416, "num_tokens": 106241823.0, "step": 2796 }, { "epoch": 0.35580714921765677, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.736272811889648, "learning_rate": 5e-05, "loss": 0.9342, "mean_token_accuracy": 0.8657257556915283, "num_tokens": 106275716.0, "step": 2797 }, { "epoch": 0.3559343594962473, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.520439147949219, "learning_rate": 5e-05, "loss": 0.7995, "mean_token_accuracy": 0.8470152616500854, "num_tokens": 106312595.0, "step": 2798 }, { "epoch": 0.3560615697748378, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.654953956604004, "learning_rate": 5e-05, "loss": 0.9942, "mean_token_accuracy": 0.8474941253662109, "num_tokens": 106346009.0, "step": 2799 }, { "epoch": 0.3561887800534283, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 6.9520039558410645, "learning_rate": 5e-05, "loss": 0.8054, "mean_token_accuracy": 0.8665604591369629, "num_tokens": 106384501.0, "step": 2800 }, { "epoch": 0.35631599033201883, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.259297370910645, "learning_rate": 5e-05, "loss": 0.9166, "mean_token_accuracy": 0.8575939536094666, "num_tokens": 106420099.0, "step": 2801 }, { "epoch": 0.35644320061060936, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.222006320953369, "learning_rate": 5e-05, "loss": 0.8259, "mean_token_accuracy": 0.869520366191864, "num_tokens": 106459697.0, "step": 2802 }, { "epoch": 0.35657041088919983, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.907047271728516, "learning_rate": 5e-05, "loss": 0.9771, "mean_token_accuracy": 0.8308866024017334, "num_tokens": 106499054.0, "step": 2803 }, { "epoch": 0.35669762116779036, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.1254472732543945, "learning_rate": 5e-05, "loss": 0.8234, "mean_token_accuracy": 0.8577570915222168, "num_tokens": 106538704.0, "step": 2804 }, { "epoch": 0.3568248314463809, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.884055137634277, "learning_rate": 5e-05, "loss": 0.8883, "mean_token_accuracy": 0.8468822836875916, "num_tokens": 106580012.0, "step": 2805 }, { "epoch": 0.35695204172497136, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.8514838218688965, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.8582525253295898, "num_tokens": 106618160.0, "step": 2806 }, { "epoch": 0.3570792520035619, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.044404029846191, "learning_rate": 5e-05, "loss": 0.9495, "mean_token_accuracy": 0.8307458162307739, "num_tokens": 106654387.0, "step": 2807 }, { "epoch": 0.3572064622821524, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.658057689666748, "learning_rate": 5e-05, "loss": 0.7265, "mean_token_accuracy": 0.8655716180801392, "num_tokens": 106693318.0, "step": 2808 }, { "epoch": 0.3573336725607429, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.120817184448242, "learning_rate": 5e-05, "loss": 0.8964, "mean_token_accuracy": 0.8485589623451233, "num_tokens": 106728779.0, "step": 2809 }, { "epoch": 0.3574608828393334, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.709972381591797, "learning_rate": 5e-05, "loss": 0.7672, "mean_token_accuracy": 0.8549467921257019, "num_tokens": 106771149.0, "step": 2810 }, { "epoch": 0.35758809311792394, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.007267951965332, "learning_rate": 5e-05, "loss": 0.9142, "mean_token_accuracy": 0.8416792750358582, "num_tokens": 106809132.0, "step": 2811 }, { "epoch": 0.3577153033965144, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.859174728393555, "learning_rate": 5e-05, "loss": 0.7117, "mean_token_accuracy": 0.8719643354415894, "num_tokens": 106842747.0, "step": 2812 }, { "epoch": 0.35784251367510495, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.795530319213867, "learning_rate": 5e-05, "loss": 0.8888, "mean_token_accuracy": 0.8416906595230103, "num_tokens": 106883519.0, "step": 2813 }, { "epoch": 0.3579697239536955, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.98577880859375, "learning_rate": 5e-05, "loss": 0.796, "mean_token_accuracy": 0.840585470199585, "num_tokens": 106914818.0, "step": 2814 }, { "epoch": 0.35809693423228595, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.6238579750061035, "learning_rate": 5e-05, "loss": 0.7939, "mean_token_accuracy": 0.8644985556602478, "num_tokens": 106954387.0, "step": 2815 }, { "epoch": 0.3582241445108765, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.042886734008789, "learning_rate": 5e-05, "loss": 0.7977, "mean_token_accuracy": 0.8476940393447876, "num_tokens": 106994264.0, "step": 2816 }, { "epoch": 0.358351354789467, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.470104217529297, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.8624112010002136, "num_tokens": 107035575.0, "step": 2817 }, { "epoch": 0.3584785650680575, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.084181785583496, "learning_rate": 5e-05, "loss": 0.7951, "mean_token_accuracy": 0.8447068929672241, "num_tokens": 107074644.0, "step": 2818 }, { "epoch": 0.358605775346648, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.365825176239014, "learning_rate": 5e-05, "loss": 0.7664, "mean_token_accuracy": 0.8577715158462524, "num_tokens": 107108758.0, "step": 2819 }, { "epoch": 0.35873298562523853, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.103623867034912, "learning_rate": 5e-05, "loss": 0.7034, "mean_token_accuracy": 0.8699479699134827, "num_tokens": 107147572.0, "step": 2820 }, { "epoch": 0.358860195903829, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.291823863983154, "learning_rate": 5e-05, "loss": 0.7952, "mean_token_accuracy": 0.8481073379516602, "num_tokens": 107184928.0, "step": 2821 }, { "epoch": 0.35898740618241953, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.100441932678223, "learning_rate": 5e-05, "loss": 0.7246, "mean_token_accuracy": 0.8669874668121338, "num_tokens": 107223462.0, "step": 2822 }, { "epoch": 0.35911461646101006, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.262819766998291, "learning_rate": 5e-05, "loss": 0.7476, "mean_token_accuracy": 0.8575586676597595, "num_tokens": 107263114.0, "step": 2823 }, { "epoch": 0.35924182673960053, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.08307409286499, "learning_rate": 5e-05, "loss": 0.6876, "mean_token_accuracy": 0.8745207786560059, "num_tokens": 107301234.0, "step": 2824 }, { "epoch": 0.35936903701819106, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.272552967071533, "learning_rate": 5e-05, "loss": 0.7413, "mean_token_accuracy": 0.8604278564453125, "num_tokens": 107334903.0, "step": 2825 }, { "epoch": 0.3594962472967816, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.0033135414123535, "learning_rate": 5e-05, "loss": 0.6689, "mean_token_accuracy": 0.8732944130897522, "num_tokens": 107374989.0, "step": 2826 }, { "epoch": 0.35962345757537206, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.346585750579834, "learning_rate": 5e-05, "loss": 0.7176, "mean_token_accuracy": 0.8640912175178528, "num_tokens": 107412545.0, "step": 2827 }, { "epoch": 0.3597506678539626, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.898039817810059, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8379999995231628, "num_tokens": 107447642.0, "step": 2828 }, { "epoch": 0.3598778781325531, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.449670314788818, "learning_rate": 5e-05, "loss": 0.7526, "mean_token_accuracy": 0.8580514788627625, "num_tokens": 107486514.0, "step": 2829 }, { "epoch": 0.36000508841114365, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.0002231597900390625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.769344329833984, "learning_rate": 5e-05, "loss": 0.7324, "mean_token_accuracy": 0.8479475975036621, "num_tokens": 107527351.0, "step": 2830 }, { "epoch": 0.3601322986897341, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.613247394561768, "learning_rate": 5e-05, "loss": 0.7558, "mean_token_accuracy": 0.8592983484268188, "num_tokens": 107558973.0, "step": 2831 }, { "epoch": 0.36025950896832465, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.5689167976379395, "learning_rate": 5e-05, "loss": 0.7173, "mean_token_accuracy": 0.8500202298164368, "num_tokens": 107597659.0, "step": 2832 }, { "epoch": 0.3603867192469152, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.874835968017578, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.8519025444984436, "num_tokens": 107635833.0, "step": 2833 }, { "epoch": 0.36051392952550565, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.00023937225341796875, "grad_norm": 6.2996134757995605, "learning_rate": 5e-05, "loss": 0.665, "mean_token_accuracy": 0.865619957447052, "num_tokens": 107673347.0, "step": 2834 }, { "epoch": 0.3606411398040962, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.255000114440918, "learning_rate": 5e-05, "loss": 0.8303, "mean_token_accuracy": 0.8535066246986389, "num_tokens": 107713820.0, "step": 2835 }, { "epoch": 0.3607683500826867, "ewc_loss": 0.25, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.0002288818359375, "grad_norm": 5.991525173187256, "learning_rate": 5e-05, "loss": 0.6643, "mean_token_accuracy": 0.8603832721710205, "num_tokens": 107745586.0, "step": 2836 }, { "epoch": 0.3608955603612772, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.717103958129883, "learning_rate": 5e-05, "loss": 0.8381, "mean_token_accuracy": 0.8678200244903564, "num_tokens": 107787649.0, "step": 2837 }, { "epoch": 0.3610227706398677, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.9224629402160645, "learning_rate": 5e-05, "loss": 0.7165, "mean_token_accuracy": 0.852033257484436, "num_tokens": 107831545.0, "step": 2838 }, { "epoch": 0.36114998091845824, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.86771011352539, "learning_rate": 5e-05, "loss": 0.992, "mean_token_accuracy": 0.8401460647583008, "num_tokens": 107875252.0, "step": 2839 }, { "epoch": 0.3612771911970487, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.538080215454102, "learning_rate": 5e-05, "loss": 0.8472, "mean_token_accuracy": 0.8341781497001648, "num_tokens": 107912060.0, "step": 2840 }, { "epoch": 0.36140440147563924, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.069012641906738, "learning_rate": 5e-05, "loss": 0.9065, "mean_token_accuracy": 0.8523175716400146, "num_tokens": 107949778.0, "step": 2841 }, { "epoch": 0.36153161175422976, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.62744665145874, "learning_rate": 5e-05, "loss": 0.8877, "mean_token_accuracy": 0.8563763499259949, "num_tokens": 107991160.0, "step": 2842 }, { "epoch": 0.36165882203282024, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.555243968963623, "learning_rate": 5e-05, "loss": 0.7786, "mean_token_accuracy": 0.8433512449264526, "num_tokens": 108027059.0, "step": 2843 }, { "epoch": 0.36178603231141077, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.897116661071777, "learning_rate": 5e-05, "loss": 0.9655, "mean_token_accuracy": 0.8525102138519287, "num_tokens": 108066738.0, "step": 2844 }, { "epoch": 0.3619132425900013, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.00023651123046875, "grad_norm": 5.540070056915283, "learning_rate": 5e-05, "loss": 0.7206, "mean_token_accuracy": 0.8443195819854736, "num_tokens": 108106498.0, "step": 2845 }, { "epoch": 0.36204045286859177, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.158266067504883, "learning_rate": 5e-05, "loss": 1.0614, "mean_token_accuracy": 0.8529796600341797, "num_tokens": 108139146.0, "step": 2846 }, { "epoch": 0.3621676631471823, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 6.346006870269775, "learning_rate": 5e-05, "loss": 0.7781, "mean_token_accuracy": 0.8683030605316162, "num_tokens": 108176070.0, "step": 2847 }, { "epoch": 0.3622948734257728, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 10.379304885864258, "learning_rate": 5e-05, "loss": 1.123, "mean_token_accuracy": 0.8638179302215576, "num_tokens": 108214859.0, "step": 2848 }, { "epoch": 0.3624220837043633, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 8.446494102478027, "learning_rate": 5e-05, "loss": 1.1277, "mean_token_accuracy": 0.8535715341567993, "num_tokens": 108250201.0, "step": 2849 }, { "epoch": 0.3625492939829538, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.68858814239502, "learning_rate": 5e-05, "loss": 0.9361, "mean_token_accuracy": 0.8601275682449341, "num_tokens": 108282111.0, "step": 2850 }, { "epoch": 0.36267650426154435, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.085078239440918, "learning_rate": 5e-05, "loss": 1.1278, "mean_token_accuracy": 0.8395618200302124, "num_tokens": 108317873.0, "step": 2851 }, { "epoch": 0.3628037145401348, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.263225555419922, "learning_rate": 5e-05, "loss": 0.8397, "mean_token_accuracy": 0.8667348623275757, "num_tokens": 108352874.0, "step": 2852 }, { "epoch": 0.36293092481872535, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.607561111450195, "learning_rate": 5e-05, "loss": 1.0682, "mean_token_accuracy": 0.8458167314529419, "num_tokens": 108387620.0, "step": 2853 }, { "epoch": 0.3630581350973159, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.417352676391602, "learning_rate": 5e-05, "loss": 0.9951, "mean_token_accuracy": 0.839725136756897, "num_tokens": 108428961.0, "step": 2854 }, { "epoch": 0.36318534537590635, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.0214262008667, "learning_rate": 5e-05, "loss": 1.0677, "mean_token_accuracy": 0.832223117351532, "num_tokens": 108471505.0, "step": 2855 }, { "epoch": 0.3633125556544969, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 7.881311893463135, "learning_rate": 5e-05, "loss": 0.9897, "mean_token_accuracy": 0.8533806204795837, "num_tokens": 108515874.0, "step": 2856 }, { "epoch": 0.3634397659330874, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.88566255569458, "learning_rate": 5e-05, "loss": 0.8738, "mean_token_accuracy": 0.8552409410476685, "num_tokens": 108558302.0, "step": 2857 }, { "epoch": 0.3635669762116779, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.138498306274414, "learning_rate": 5e-05, "loss": 0.9713, "mean_token_accuracy": 0.8478392362594604, "num_tokens": 108597668.0, "step": 2858 }, { "epoch": 0.3636941864902684, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.20890998840332, "learning_rate": 5e-05, "loss": 0.817, "mean_token_accuracy": 0.8540544509887695, "num_tokens": 108636623.0, "step": 2859 }, { "epoch": 0.36382139676885894, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.146981239318848, "learning_rate": 5e-05, "loss": 0.8198, "mean_token_accuracy": 0.8753755688667297, "num_tokens": 108673660.0, "step": 2860 }, { "epoch": 0.3639486070474494, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.077712059020996, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.8652701377868652, "num_tokens": 108710350.0, "step": 2861 }, { "epoch": 0.36407581732603994, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.840074062347412, "learning_rate": 5e-05, "loss": 0.8695, "mean_token_accuracy": 0.8473219275474548, "num_tokens": 108747028.0, "step": 2862 }, { "epoch": 0.36420302760463047, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.197956085205078, "learning_rate": 5e-05, "loss": 0.8625, "mean_token_accuracy": 0.843299388885498, "num_tokens": 108780315.0, "step": 2863 }, { "epoch": 0.36433023788322094, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.402654647827148, "learning_rate": 5e-05, "loss": 0.8079, "mean_token_accuracy": 0.8463922739028931, "num_tokens": 108811706.0, "step": 2864 }, { "epoch": 0.36445744816181147, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.436997890472412, "learning_rate": 5e-05, "loss": 0.8565, "mean_token_accuracy": 0.8460573554039001, "num_tokens": 108851545.0, "step": 2865 }, { "epoch": 0.364584658440402, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.955911636352539, "learning_rate": 5e-05, "loss": 0.7091, "mean_token_accuracy": 0.8667835593223572, "num_tokens": 108887013.0, "step": 2866 }, { "epoch": 0.36471186871899247, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.813145160675049, "learning_rate": 5e-05, "loss": 0.8553, "mean_token_accuracy": 0.8440415263175964, "num_tokens": 108934289.0, "step": 2867 }, { "epoch": 0.364839078997583, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.537715435028076, "learning_rate": 5e-05, "loss": 0.7088, "mean_token_accuracy": 0.8643670082092285, "num_tokens": 108973823.0, "step": 2868 }, { "epoch": 0.36496628927617353, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.296659469604492, "learning_rate": 5e-05, "loss": 0.8385, "mean_token_accuracy": 0.8622553944587708, "num_tokens": 109006947.0, "step": 2869 }, { "epoch": 0.365093499554764, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.181491374969482, "learning_rate": 5e-05, "loss": 0.7589, "mean_token_accuracy": 0.8439117670059204, "num_tokens": 109039839.0, "step": 2870 }, { "epoch": 0.36522070983335453, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.738324165344238, "learning_rate": 5e-05, "loss": 0.8879, "mean_token_accuracy": 0.864678680896759, "num_tokens": 109076348.0, "step": 2871 }, { "epoch": 0.36534792011194506, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.162712097167969, "learning_rate": 5e-05, "loss": 0.6974, "mean_token_accuracy": 0.8688618540763855, "num_tokens": 109117050.0, "step": 2872 }, { "epoch": 0.36547513039053553, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.892338752746582, "learning_rate": 5e-05, "loss": 0.9627, "mean_token_accuracy": 0.8526014089584351, "num_tokens": 109151131.0, "step": 2873 }, { "epoch": 0.36560234066912606, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.573888778686523, "learning_rate": 5e-05, "loss": 0.8194, "mean_token_accuracy": 0.8524087071418762, "num_tokens": 109194542.0, "step": 2874 }, { "epoch": 0.3657295509477166, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.569442749023438, "learning_rate": 5e-05, "loss": 0.9427, "mean_token_accuracy": 0.8540807962417603, "num_tokens": 109229636.0, "step": 2875 }, { "epoch": 0.36585676122630706, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.106186866760254, "learning_rate": 5e-05, "loss": 0.8974, "mean_token_accuracy": 0.8492070436477661, "num_tokens": 109264705.0, "step": 2876 }, { "epoch": 0.3659839715048976, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.948970317840576, "learning_rate": 5e-05, "loss": 0.8559, "mean_token_accuracy": 0.8624907732009888, "num_tokens": 109304220.0, "step": 2877 }, { "epoch": 0.3661111817834881, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.352072715759277, "learning_rate": 5e-05, "loss": 0.8874, "mean_token_accuracy": 0.8534949421882629, "num_tokens": 109347315.0, "step": 2878 }, { "epoch": 0.3662383920620786, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.558514595031738, "learning_rate": 5e-05, "loss": 0.8672, "mean_token_accuracy": 0.8430408239364624, "num_tokens": 109382679.0, "step": 2879 }, { "epoch": 0.3663656023406691, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.201160430908203, "learning_rate": 5e-05, "loss": 0.889, "mean_token_accuracy": 0.8386751413345337, "num_tokens": 109428491.0, "step": 2880 }, { "epoch": 0.36649281261925964, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.670351982116699, "learning_rate": 5e-05, "loss": 0.8549, "mean_token_accuracy": 0.8443461060523987, "num_tokens": 109469700.0, "step": 2881 }, { "epoch": 0.3666200228978502, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.796642780303955, "learning_rate": 5e-05, "loss": 0.7306, "mean_token_accuracy": 0.8632056713104248, "num_tokens": 109510245.0, "step": 2882 }, { "epoch": 0.36674723317644065, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.048299789428711, "learning_rate": 5e-05, "loss": 0.8762, "mean_token_accuracy": 0.8445971608161926, "num_tokens": 109544932.0, "step": 2883 }, { "epoch": 0.3668744434550312, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.467358589172363, "learning_rate": 5e-05, "loss": 0.7234, "mean_token_accuracy": 0.860909640789032, "num_tokens": 109579519.0, "step": 2884 }, { "epoch": 0.3670016537336217, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.317207336425781, "learning_rate": 5e-05, "loss": 0.8733, "mean_token_accuracy": 0.8626053333282471, "num_tokens": 109620313.0, "step": 2885 }, { "epoch": 0.3671288640122122, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.488131999969482, "learning_rate": 5e-05, "loss": 0.7998, "mean_token_accuracy": 0.8406570553779602, "num_tokens": 109663800.0, "step": 2886 }, { "epoch": 0.3672560742908027, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.207985877990723, "learning_rate": 5e-05, "loss": 0.8543, "mean_token_accuracy": 0.8642945885658264, "num_tokens": 109701118.0, "step": 2887 }, { "epoch": 0.36738328456939323, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.8515472412109375, "learning_rate": 5e-05, "loss": 0.7406, "mean_token_accuracy": 0.8711705803871155, "num_tokens": 109739486.0, "step": 2888 }, { "epoch": 0.3675104948479837, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.719764709472656, "learning_rate": 5e-05, "loss": 0.8075, "mean_token_accuracy": 0.8688997030258179, "num_tokens": 109781097.0, "step": 2889 }, { "epoch": 0.36763770512657423, "ewc_loss": 0.375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.287084102630615, "learning_rate": 5e-05, "loss": 0.783, "mean_token_accuracy": 0.8587981462478638, "num_tokens": 109813562.0, "step": 2890 }, { "epoch": 0.36776491540516476, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.136832237243652, "learning_rate": 5e-05, "loss": 0.8383, "mean_token_accuracy": 0.8363593816757202, "num_tokens": 109855326.0, "step": 2891 }, { "epoch": 0.36789212568375523, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.612155437469482, "learning_rate": 5e-05, "loss": 0.8333, "mean_token_accuracy": 0.8516156077384949, "num_tokens": 109890787.0, "step": 2892 }, { "epoch": 0.36801933596234576, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.805356025695801, "learning_rate": 5e-05, "loss": 0.7621, "mean_token_accuracy": 0.852386474609375, "num_tokens": 109932199.0, "step": 2893 }, { "epoch": 0.3681465462409363, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.721389293670654, "learning_rate": 5e-05, "loss": 0.7853, "mean_token_accuracy": 0.8594276309013367, "num_tokens": 109967388.0, "step": 2894 }, { "epoch": 0.36827375651952676, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.7280354499816895, "learning_rate": 5e-05, "loss": 0.6868, "mean_token_accuracy": 0.8662112951278687, "num_tokens": 110003437.0, "step": 2895 }, { "epoch": 0.3684009667981173, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.6914567947387695, "learning_rate": 5e-05, "loss": 0.8039, "mean_token_accuracy": 0.8501001596450806, "num_tokens": 110040822.0, "step": 2896 }, { "epoch": 0.3685281770767078, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.756533622741699, "learning_rate": 5e-05, "loss": 0.7158, "mean_token_accuracy": 0.8599668741226196, "num_tokens": 110079628.0, "step": 2897 }, { "epoch": 0.3686553873552983, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.615900993347168, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8656084537506104, "num_tokens": 110115794.0, "step": 2898 }, { "epoch": 0.3687825976338888, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.807686805725098, "learning_rate": 5e-05, "loss": 0.7589, "mean_token_accuracy": 0.8499672412872314, "num_tokens": 110153227.0, "step": 2899 }, { "epoch": 0.36890980791247935, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.538739204406738, "learning_rate": 5e-05, "loss": 0.8373, "mean_token_accuracy": 0.8366021513938904, "num_tokens": 110193186.0, "step": 2900 }, { "epoch": 0.3690370181910698, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.826371192932129, "learning_rate": 5e-05, "loss": 0.7782, "mean_token_accuracy": 0.8446766138076782, "num_tokens": 110232130.0, "step": 2901 }, { "epoch": 0.36916422846966035, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.560549736022949, "learning_rate": 5e-05, "loss": 0.7337, "mean_token_accuracy": 0.8672267198562622, "num_tokens": 110271376.0, "step": 2902 }, { "epoch": 0.3692914387482509, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.723034381866455, "learning_rate": 5e-05, "loss": 0.7525, "mean_token_accuracy": 0.8490272760391235, "num_tokens": 110314220.0, "step": 2903 }, { "epoch": 0.36941864902684135, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.711818695068359, "learning_rate": 5e-05, "loss": 0.8101, "mean_token_accuracy": 0.8494120240211487, "num_tokens": 110357292.0, "step": 2904 }, { "epoch": 0.3695458593054319, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.518881320953369, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8379939794540405, "num_tokens": 110397813.0, "step": 2905 }, { "epoch": 0.3696730695840224, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.985976696014404, "learning_rate": 5e-05, "loss": 0.7766, "mean_token_accuracy": 0.8664664030075073, "num_tokens": 110435773.0, "step": 2906 }, { "epoch": 0.3698002798626129, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.2482123374938965, "learning_rate": 5e-05, "loss": 0.6316, "mean_token_accuracy": 0.8774384260177612, "num_tokens": 110469518.0, "step": 2907 }, { "epoch": 0.3699274901412034, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.406079292297363, "learning_rate": 5e-05, "loss": 0.8644, "mean_token_accuracy": 0.8529051542282104, "num_tokens": 110503156.0, "step": 2908 }, { "epoch": 0.37005470041979394, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 5.996022701263428, "learning_rate": 5e-05, "loss": 0.7329, "mean_token_accuracy": 0.845553994178772, "num_tokens": 110543960.0, "step": 2909 }, { "epoch": 0.3701819106983844, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.780255317687988, "learning_rate": 5e-05, "loss": 0.9636, "mean_token_accuracy": 0.8436877727508545, "num_tokens": 110584811.0, "step": 2910 }, { "epoch": 0.37030912097697494, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.178794860839844, "learning_rate": 5e-05, "loss": 0.7754, "mean_token_accuracy": 0.8456933498382568, "num_tokens": 110622433.0, "step": 2911 }, { "epoch": 0.37043633125556547, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.622357368469238, "learning_rate": 5e-05, "loss": 0.9578, "mean_token_accuracy": 0.8512774109840393, "num_tokens": 110660994.0, "step": 2912 }, { "epoch": 0.37056354153415594, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 6.9470367431640625, "learning_rate": 5e-05, "loss": 0.9324, "mean_token_accuracy": 0.8327994346618652, "num_tokens": 110703022.0, "step": 2913 }, { "epoch": 0.37069075181274647, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.616589546203613, "learning_rate": 5e-05, "loss": 0.815, "mean_token_accuracy": 0.8653335571289062, "num_tokens": 110743864.0, "step": 2914 }, { "epoch": 0.370817962091337, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.883315086364746, "learning_rate": 5e-05, "loss": 0.9049, "mean_token_accuracy": 0.8525657653808594, "num_tokens": 110778591.0, "step": 2915 }, { "epoch": 0.37094517236992747, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.3944220542907715, "learning_rate": 5e-05, "loss": 0.7172, "mean_token_accuracy": 0.8623416423797607, "num_tokens": 110818402.0, "step": 2916 }, { "epoch": 0.371072382648518, "ewc_loss": 0.5, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.775876998901367, "learning_rate": 5e-05, "loss": 0.935, "mean_token_accuracy": 0.8564448952674866, "num_tokens": 110855129.0, "step": 2917 }, { "epoch": 0.3711995929271085, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.814914226531982, "learning_rate": 5e-05, "loss": 0.7011, "mean_token_accuracy": 0.8587362766265869, "num_tokens": 110896307.0, "step": 2918 }, { "epoch": 0.371326803205699, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.400053977966309, "learning_rate": 5e-05, "loss": 1.0037, "mean_token_accuracy": 0.8600080013275146, "num_tokens": 110940639.0, "step": 2919 }, { "epoch": 0.3714540134842895, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.340661525726318, "learning_rate": 5e-05, "loss": 0.8043, "mean_token_accuracy": 0.8562272787094116, "num_tokens": 110979134.0, "step": 2920 }, { "epoch": 0.37158122376288005, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.183380126953125, "learning_rate": 5e-05, "loss": 0.9927, "mean_token_accuracy": 0.8668690919876099, "num_tokens": 111013799.0, "step": 2921 }, { "epoch": 0.3717084340414705, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.561488151550293, "learning_rate": 5e-05, "loss": 0.9864, "mean_token_accuracy": 0.8570612072944641, "num_tokens": 111055486.0, "step": 2922 }, { "epoch": 0.37183564432006105, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.880358695983887, "learning_rate": 5e-05, "loss": 0.8877, "mean_token_accuracy": 0.858474850654602, "num_tokens": 111101332.0, "step": 2923 }, { "epoch": 0.3719628545986516, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.293188095092773, "learning_rate": 5e-05, "loss": 0.9858, "mean_token_accuracy": 0.858257532119751, "num_tokens": 111137285.0, "step": 2924 }, { "epoch": 0.37209006487724205, "ewc_loss": 0.375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.8454976081848145, "learning_rate": 5e-05, "loss": 0.8326, "mean_token_accuracy": 0.8469407558441162, "num_tokens": 111177254.0, "step": 2925 }, { "epoch": 0.3722172751558326, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.412571907043457, "learning_rate": 5e-05, "loss": 0.9436, "mean_token_accuracy": 0.8560795783996582, "num_tokens": 111217633.0, "step": 2926 }, { "epoch": 0.3723444854344231, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.003904342651367, "learning_rate": 5e-05, "loss": 0.9266, "mean_token_accuracy": 0.8247089385986328, "num_tokens": 111259273.0, "step": 2927 }, { "epoch": 0.3724716957130136, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.65248441696167, "learning_rate": 5e-05, "loss": 0.9005, "mean_token_accuracy": 0.8398782014846802, "num_tokens": 111295068.0, "step": 2928 }, { "epoch": 0.3725989059916041, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.595953464508057, "learning_rate": 5e-05, "loss": 0.9198, "mean_token_accuracy": 0.842011570930481, "num_tokens": 111336863.0, "step": 2929 }, { "epoch": 0.37272611627019464, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.853034019470215, "learning_rate": 5e-05, "loss": 0.8116, "mean_token_accuracy": 0.8410322070121765, "num_tokens": 111378432.0, "step": 2930 }, { "epoch": 0.37285332654878517, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.012444496154785, "learning_rate": 5e-05, "loss": 0.9014, "mean_token_accuracy": 0.847828209400177, "num_tokens": 111412481.0, "step": 2931 }, { "epoch": 0.37298053682737564, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.541004657745361, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8397579193115234, "num_tokens": 111450534.0, "step": 2932 }, { "epoch": 0.37310774710596617, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.247368812561035, "learning_rate": 5e-05, "loss": 0.8652, "mean_token_accuracy": 0.8630178570747375, "num_tokens": 111484278.0, "step": 2933 }, { "epoch": 0.3732349573845567, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.4192986488342285, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8505990505218506, "num_tokens": 111526797.0, "step": 2934 }, { "epoch": 0.37336216766314717, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.484477043151855, "learning_rate": 5e-05, "loss": 0.9552, "mean_token_accuracy": 0.8443161249160767, "num_tokens": 111568139.0, "step": 2935 }, { "epoch": 0.3734893779417377, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.448569297790527, "learning_rate": 5e-05, "loss": 0.8439, "mean_token_accuracy": 0.8310506343841553, "num_tokens": 111614519.0, "step": 2936 }, { "epoch": 0.3736165882203282, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.550520896911621, "learning_rate": 5e-05, "loss": 0.9244, "mean_token_accuracy": 0.8611080646514893, "num_tokens": 111647965.0, "step": 2937 }, { "epoch": 0.3737437984989187, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.652956485748291, "learning_rate": 5e-05, "loss": 0.8071, "mean_token_accuracy": 0.8549469709396362, "num_tokens": 111682492.0, "step": 2938 }, { "epoch": 0.37387100877750923, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.519807815551758, "learning_rate": 5e-05, "loss": 0.9833, "mean_token_accuracy": 0.8435789346694946, "num_tokens": 111717550.0, "step": 2939 }, { "epoch": 0.37399821905609976, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 6.7690110206604, "learning_rate": 5e-05, "loss": 0.8065, "mean_token_accuracy": 0.8632160425186157, "num_tokens": 111752173.0, "step": 2940 }, { "epoch": 0.37412542933469023, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.43369197845459, "learning_rate": 5e-05, "loss": 0.9531, "mean_token_accuracy": 0.8546138405799866, "num_tokens": 111790845.0, "step": 2941 }, { "epoch": 0.37425263961328076, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00023555755615234375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 6.848721981048584, "learning_rate": 5e-05, "loss": 0.8483, "mean_token_accuracy": 0.8508421182632446, "num_tokens": 111827837.0, "step": 2942 }, { "epoch": 0.3743798498918713, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00023555755615234375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.37848949432373, "learning_rate": 5e-05, "loss": 0.9901, "mean_token_accuracy": 0.8438692092895508, "num_tokens": 111868785.0, "step": 2943 }, { "epoch": 0.37450706017046176, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00023555755615234375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 6.918575763702393, "learning_rate": 5e-05, "loss": 0.8647, "mean_token_accuracy": 0.850824236869812, "num_tokens": 111907296.0, "step": 2944 }, { "epoch": 0.3746342704490523, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00023555755615234375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.259954452514648, "learning_rate": 5e-05, "loss": 0.9118, "mean_token_accuracy": 0.8691036701202393, "num_tokens": 111947619.0, "step": 2945 }, { "epoch": 0.3747614807276428, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00023555755615234375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.044394016265869, "learning_rate": 5e-05, "loss": 0.916, "mean_token_accuracy": 0.8377939462661743, "num_tokens": 111987255.0, "step": 2946 }, { "epoch": 0.3748886910062333, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.05163288116455, "learning_rate": 5e-05, "loss": 0.9415, "mean_token_accuracy": 0.8531973361968994, "num_tokens": 112023003.0, "step": 2947 }, { "epoch": 0.3750159012848238, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.149387359619141, "learning_rate": 5e-05, "loss": 0.8768, "mean_token_accuracy": 0.8454098701477051, "num_tokens": 112056292.0, "step": 2948 }, { "epoch": 0.37514311156341434, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.769892692565918, "learning_rate": 5e-05, "loss": 0.8943, "mean_token_accuracy": 0.8523820638656616, "num_tokens": 112089947.0, "step": 2949 }, { "epoch": 0.3752703218420048, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.206834316253662, "learning_rate": 5e-05, "loss": 0.9201, "mean_token_accuracy": 0.8303713798522949, "num_tokens": 112125415.0, "step": 2950 }, { "epoch": 0.37539753212059535, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.542877674102783, "learning_rate": 5e-05, "loss": 0.868, "mean_token_accuracy": 0.849244236946106, "num_tokens": 112164364.0, "step": 2951 }, { "epoch": 0.3755247423991859, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.172770023345947, "learning_rate": 5e-05, "loss": 0.787, "mean_token_accuracy": 0.8622939586639404, "num_tokens": 112208994.0, "step": 2952 }, { "epoch": 0.37565195267777635, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.476656913757324, "learning_rate": 5e-05, "loss": 0.8208, "mean_token_accuracy": 0.8555393218994141, "num_tokens": 112249902.0, "step": 2953 }, { "epoch": 0.3757791629563669, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.065556049346924, "learning_rate": 5e-05, "loss": 0.7666, "mean_token_accuracy": 0.8617765307426453, "num_tokens": 112290446.0, "step": 2954 }, { "epoch": 0.3759063732349574, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.495546340942383, "learning_rate": 5e-05, "loss": 0.82, "mean_token_accuracy": 0.8508758544921875, "num_tokens": 112323217.0, "step": 2955 }, { "epoch": 0.3760335835135479, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.996950149536133, "learning_rate": 5e-05, "loss": 0.7672, "mean_token_accuracy": 0.8569270372390747, "num_tokens": 112365863.0, "step": 2956 }, { "epoch": 0.3761607937921384, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.463799953460693, "learning_rate": 5e-05, "loss": 0.7963, "mean_token_accuracy": 0.8542328476905823, "num_tokens": 112400708.0, "step": 2957 }, { "epoch": 0.37628800407072893, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.974433898925781, "learning_rate": 5e-05, "loss": 0.7298, "mean_token_accuracy": 0.866200864315033, "num_tokens": 112440063.0, "step": 2958 }, { "epoch": 0.3764152143493194, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.442766189575195, "learning_rate": 5e-05, "loss": 0.7197, "mean_token_accuracy": 0.880416989326477, "num_tokens": 112478856.0, "step": 2959 }, { "epoch": 0.37654242462790993, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.968937397003174, "learning_rate": 5e-05, "loss": 0.764, "mean_token_accuracy": 0.8544262647628784, "num_tokens": 112515707.0, "step": 2960 }, { "epoch": 0.37666963490650046, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.416240215301514, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.8473012447357178, "num_tokens": 112551302.0, "step": 2961 }, { "epoch": 0.37679684518509093, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.935017108917236, "learning_rate": 5e-05, "loss": 0.7145, "mean_token_accuracy": 0.8645542860031128, "num_tokens": 112585821.0, "step": 2962 }, { "epoch": 0.37692405546368146, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.433379650115967, "learning_rate": 5e-05, "loss": 0.7716, "mean_token_accuracy": 0.8561855554580688, "num_tokens": 112624261.0, "step": 2963 }, { "epoch": 0.377051265742272, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.890079975128174, "learning_rate": 5e-05, "loss": 0.6842, "mean_token_accuracy": 0.8702282309532166, "num_tokens": 112656944.0, "step": 2964 }, { "epoch": 0.37717847602086246, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.484915733337402, "learning_rate": 5e-05, "loss": 0.755, "mean_token_accuracy": 0.8608416318893433, "num_tokens": 112694386.0, "step": 2965 }, { "epoch": 0.377305686299453, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.813548564910889, "learning_rate": 5e-05, "loss": 0.7467, "mean_token_accuracy": 0.8524420261383057, "num_tokens": 112731597.0, "step": 2966 }, { "epoch": 0.3774328965780435, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.5584187507629395, "learning_rate": 5e-05, "loss": 0.6977, "mean_token_accuracy": 0.8800426721572876, "num_tokens": 112768012.0, "step": 2967 }, { "epoch": 0.377560106856634, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.747287273406982, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8416877388954163, "num_tokens": 112798089.0, "step": 2968 }, { "epoch": 0.3776873171352245, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.613802433013916, "learning_rate": 5e-05, "loss": 0.7716, "mean_token_accuracy": 0.8581249713897705, "num_tokens": 112842332.0, "step": 2969 }, { "epoch": 0.37781452741381505, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.680851459503174, "learning_rate": 5e-05, "loss": 0.7424, "mean_token_accuracy": 0.8497737646102905, "num_tokens": 112882382.0, "step": 2970 }, { "epoch": 0.3779417376924055, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.690869331359863, "learning_rate": 5e-05, "loss": 0.8589, "mean_token_accuracy": 0.8325080871582031, "num_tokens": 112922627.0, "step": 2971 }, { "epoch": 0.37806894797099605, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.611906051635742, "learning_rate": 5e-05, "loss": 0.787, "mean_token_accuracy": 0.8307863473892212, "num_tokens": 112957571.0, "step": 2972 }, { "epoch": 0.3781961582495866, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.751529693603516, "learning_rate": 5e-05, "loss": 0.7924, "mean_token_accuracy": 0.8537116646766663, "num_tokens": 113001128.0, "step": 2973 }, { "epoch": 0.37832336852817705, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.592745304107666, "learning_rate": 5e-05, "loss": 0.7055, "mean_token_accuracy": 0.8612298965454102, "num_tokens": 113034515.0, "step": 2974 }, { "epoch": 0.3784505788067676, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.76600980758667, "learning_rate": 5e-05, "loss": 0.7427, "mean_token_accuracy": 0.8708548545837402, "num_tokens": 113071221.0, "step": 2975 }, { "epoch": 0.3785777890853581, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.634392261505127, "learning_rate": 5e-05, "loss": 0.7964, "mean_token_accuracy": 0.8362224102020264, "num_tokens": 113105804.0, "step": 2976 }, { "epoch": 0.3787049993639486, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.673799514770508, "learning_rate": 5e-05, "loss": 0.7653, "mean_token_accuracy": 0.8634389638900757, "num_tokens": 113146051.0, "step": 2977 }, { "epoch": 0.3788322096425391, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.771014213562012, "learning_rate": 5e-05, "loss": 0.6951, "mean_token_accuracy": 0.8670895099639893, "num_tokens": 113186696.0, "step": 2978 }, { "epoch": 0.37895941992112964, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.480402946472168, "learning_rate": 5e-05, "loss": 0.7965, "mean_token_accuracy": 0.8469222784042358, "num_tokens": 113228402.0, "step": 2979 }, { "epoch": 0.37908663019972016, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.98293399810791, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8435531258583069, "num_tokens": 113270407.0, "step": 2980 }, { "epoch": 0.37921384047831064, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.184176445007324, "learning_rate": 5e-05, "loss": 0.7696, "mean_token_accuracy": 0.8473717570304871, "num_tokens": 113304523.0, "step": 2981 }, { "epoch": 0.37934105075690117, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.279538631439209, "learning_rate": 5e-05, "loss": 0.7633, "mean_token_accuracy": 0.8564937114715576, "num_tokens": 113345224.0, "step": 2982 }, { "epoch": 0.3794682610354917, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.782714366912842, "learning_rate": 5e-05, "loss": 0.6926, "mean_token_accuracy": 0.8645856380462646, "num_tokens": 113382287.0, "step": 2983 }, { "epoch": 0.37959547131408217, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.710536956787109, "learning_rate": 5e-05, "loss": 0.8145, "mean_token_accuracy": 0.8447005748748779, "num_tokens": 113427308.0, "step": 2984 }, { "epoch": 0.3797226815926727, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 0.00023365020751953125, "grad_norm": 6.2852888107299805, "learning_rate": 5e-05, "loss": 0.6839, "mean_token_accuracy": 0.8550543785095215, "num_tokens": 113461272.0, "step": 2985 }, { "epoch": 0.3798498918712632, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.341949462890625, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8744429349899292, "num_tokens": 113499942.0, "step": 2986 }, { "epoch": 0.3799771021498537, "ewc_loss": 0.2265625, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.00020313262939453125, "grad_norm": 5.6712517738342285, "learning_rate": 5e-05, "loss": 0.6027, "mean_token_accuracy": 0.8713793754577637, "num_tokens": 113531801.0, "step": 2987 }, { "epoch": 0.3801043124284442, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.31453800201416, "learning_rate": 5e-05, "loss": 0.8996, "mean_token_accuracy": 0.8712752461433411, "num_tokens": 113567475.0, "step": 2988 }, { "epoch": 0.38023152270703475, "ewc_loss": 0.25390625, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.00022983551025390625, "grad_norm": 5.289417266845703, "learning_rate": 5e-05, "loss": 0.724, "mean_token_accuracy": 0.8427532315254211, "num_tokens": 113603767.0, "step": 2989 }, { "epoch": 0.3803587329856252, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00023746490478515625, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.358298301696777, "learning_rate": 5e-05, "loss": 1.106, "mean_token_accuracy": 0.8503477573394775, "num_tokens": 113641213.0, "step": 2990 }, { "epoch": 0.38048594326421575, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 6.668320655822754, "learning_rate": 5e-05, "loss": 0.9831, "mean_token_accuracy": 0.8306683301925659, "num_tokens": 113683364.0, "step": 2991 }, { "epoch": 0.3806131535428063, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.108736991882324, "learning_rate": 5e-05, "loss": 1.1057, "mean_token_accuracy": 0.8509950637817383, "num_tokens": 113718488.0, "step": 2992 }, { "epoch": 0.38074036382139675, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 8.585946083068848, "learning_rate": 5e-05, "loss": 1.1329, "mean_token_accuracy": 0.8667773008346558, "num_tokens": 113756664.0, "step": 2993 }, { "epoch": 0.3808675740999873, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.9651923179626465, "learning_rate": 5e-05, "loss": 0.8386, "mean_token_accuracy": 0.8602445125579834, "num_tokens": 113790711.0, "step": 2994 }, { "epoch": 0.3809947843785778, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 9.938523292541504, "learning_rate": 5e-05, "loss": 1.1824, "mean_token_accuracy": 0.8637909889221191, "num_tokens": 113827091.0, "step": 2995 }, { "epoch": 0.3811219946571683, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 6.78639030456543, "learning_rate": 5e-05, "loss": 0.879, "mean_token_accuracy": 0.8409925103187561, "num_tokens": 113862971.0, "step": 2996 }, { "epoch": 0.3812492049357588, "ewc_loss": 0.875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 11.737159729003906, "learning_rate": 5e-05, "loss": 1.3313, "mean_token_accuracy": 0.8481895923614502, "num_tokens": 113897430.0, "step": 2997 }, { "epoch": 0.38137641521434934, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00079345703125, "grad_norm": 9.129953384399414, "learning_rate": 5e-05, "loss": 1.2498, "mean_token_accuracy": 0.85451340675354, "num_tokens": 113926626.0, "step": 2998 }, { "epoch": 0.3815036254929398, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 11.046977043151855, "learning_rate": 5e-05, "loss": 1.0761, "mean_token_accuracy": 0.8643922805786133, "num_tokens": 113969656.0, "step": 2999 }, { "epoch": 0.38163083577153034, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 8.987475395202637, "learning_rate": 5e-05, "loss": 1.1385, "mean_token_accuracy": 0.8683704137802124, "num_tokens": 114006140.0, "step": 3000 }, { "epoch": 0.38175804605012087, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 11.009465217590332, "learning_rate": 5e-05, "loss": 1.1449, "mean_token_accuracy": 0.8603841066360474, "num_tokens": 114047300.0, "step": 3001 }, { "epoch": 0.38188525632871134, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.153790473937988, "learning_rate": 5e-05, "loss": 1.0236, "mean_token_accuracy": 0.848963737487793, "num_tokens": 114086129.0, "step": 3002 }, { "epoch": 0.38201246660730187, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000766754150390625, "grad_norm": 11.445043563842773, "learning_rate": 5e-05, "loss": 1.2165, "mean_token_accuracy": 0.856086254119873, "num_tokens": 114123345.0, "step": 3003 }, { "epoch": 0.3821396768858924, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.595423698425293, "learning_rate": 5e-05, "loss": 1.0378, "mean_token_accuracy": 0.8530105948448181, "num_tokens": 114158983.0, "step": 3004 }, { "epoch": 0.38226688716448287, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00077056884765625, "grad_norm": 11.748021125793457, "learning_rate": 5e-05, "loss": 1.292, "mean_token_accuracy": 0.8332856297492981, "num_tokens": 114199055.0, "step": 3005 }, { "epoch": 0.3823940974430734, "ewc_loss": 0.71875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 9.773627281188965, "learning_rate": 5e-05, "loss": 1.1369, "mean_token_accuracy": 0.8596944808959961, "num_tokens": 114236355.0, "step": 3006 }, { "epoch": 0.38252130772166393, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.96211051940918, "learning_rate": 5e-05, "loss": 1.1464, "mean_token_accuracy": 0.8501104712486267, "num_tokens": 114285111.0, "step": 3007 }, { "epoch": 0.3826485180002544, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 9.638128280639648, "learning_rate": 5e-05, "loss": 1.17, "mean_token_accuracy": 0.8306846618652344, "num_tokens": 114319306.0, "step": 3008 }, { "epoch": 0.38277572827884493, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.339186668395996, "learning_rate": 5e-05, "loss": 1.026, "mean_token_accuracy": 0.8667181134223938, "num_tokens": 114354117.0, "step": 3009 }, { "epoch": 0.38290293855743546, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.126214981079102, "learning_rate": 5e-05, "loss": 1.0529, "mean_token_accuracy": 0.8445033431053162, "num_tokens": 114395865.0, "step": 3010 }, { "epoch": 0.38303014883602593, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.868792533874512, "learning_rate": 5e-05, "loss": 1.0129, "mean_token_accuracy": 0.8611305952072144, "num_tokens": 114427862.0, "step": 3011 }, { "epoch": 0.38315735911461646, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.56769847869873, "learning_rate": 5e-05, "loss": 0.9213, "mean_token_accuracy": 0.8645703792572021, "num_tokens": 114462892.0, "step": 3012 }, { "epoch": 0.383284569393207, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.720537185668945, "learning_rate": 5e-05, "loss": 0.9756, "mean_token_accuracy": 0.8674407601356506, "num_tokens": 114500477.0, "step": 3013 }, { "epoch": 0.38341177967179746, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.069879531860352, "learning_rate": 5e-05, "loss": 0.8961, "mean_token_accuracy": 0.8528391122817993, "num_tokens": 114534695.0, "step": 3014 }, { "epoch": 0.383538989950388, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.937322616577148, "learning_rate": 5e-05, "loss": 1.0796, "mean_token_accuracy": 0.8386104106903076, "num_tokens": 114579969.0, "step": 3015 }, { "epoch": 0.3836662002289785, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.978382110595703, "learning_rate": 5e-05, "loss": 0.9428, "mean_token_accuracy": 0.8386017680168152, "num_tokens": 114614904.0, "step": 3016 }, { "epoch": 0.383793410507569, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.243232727050781, "learning_rate": 5e-05, "loss": 1.049, "mean_token_accuracy": 0.8661853075027466, "num_tokens": 114647126.0, "step": 3017 }, { "epoch": 0.3839206207861595, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.43825912475586, "learning_rate": 5e-05, "loss": 0.9259, "mean_token_accuracy": 0.855526328086853, "num_tokens": 114683529.0, "step": 3018 }, { "epoch": 0.38404783106475004, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.992558479309082, "learning_rate": 5e-05, "loss": 1.0344, "mean_token_accuracy": 0.8625581860542297, "num_tokens": 114722471.0, "step": 3019 }, { "epoch": 0.3841750413433405, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.7649507522583, "learning_rate": 5e-05, "loss": 0.9919, "mean_token_accuracy": 0.8490039110183716, "num_tokens": 114765243.0, "step": 3020 }, { "epoch": 0.38430225162193105, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.237894058227539, "learning_rate": 5e-05, "loss": 0.9903, "mean_token_accuracy": 0.8569509983062744, "num_tokens": 114806341.0, "step": 3021 }, { "epoch": 0.3844294619005216, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.470256805419922, "learning_rate": 5e-05, "loss": 1.0075, "mean_token_accuracy": 0.8284569978713989, "num_tokens": 114850014.0, "step": 3022 }, { "epoch": 0.38455667217911205, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.678664207458496, "learning_rate": 5e-05, "loss": 0.9593, "mean_token_accuracy": 0.8487737774848938, "num_tokens": 114884771.0, "step": 3023 }, { "epoch": 0.3846838824577026, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.872707843780518, "learning_rate": 5e-05, "loss": 0.8167, "mean_token_accuracy": 0.8618506193161011, "num_tokens": 114926319.0, "step": 3024 }, { "epoch": 0.3848110927362931, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.58537769317627, "learning_rate": 5e-05, "loss": 0.9426, "mean_token_accuracy": 0.8466163873672485, "num_tokens": 114971008.0, "step": 3025 }, { "epoch": 0.3849383030148836, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.315940856933594, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8521589636802673, "num_tokens": 115007294.0, "step": 3026 }, { "epoch": 0.3850655132934741, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.844690322875977, "learning_rate": 5e-05, "loss": 0.9186, "mean_token_accuracy": 0.8604447841644287, "num_tokens": 115043066.0, "step": 3027 }, { "epoch": 0.38519272357206463, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.095505714416504, "learning_rate": 5e-05, "loss": 0.7519, "mean_token_accuracy": 0.8605090379714966, "num_tokens": 115077404.0, "step": 3028 }, { "epoch": 0.3853199338506551, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.197410583496094, "learning_rate": 5e-05, "loss": 0.9492, "mean_token_accuracy": 0.8670909404754639, "num_tokens": 115114092.0, "step": 3029 }, { "epoch": 0.38544714412924563, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.431115627288818, "learning_rate": 5e-05, "loss": 0.7598, "mean_token_accuracy": 0.8692803382873535, "num_tokens": 115157095.0, "step": 3030 }, { "epoch": 0.38557435440783616, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.055638313293457, "learning_rate": 5e-05, "loss": 0.9405, "mean_token_accuracy": 0.8704417943954468, "num_tokens": 115189752.0, "step": 3031 }, { "epoch": 0.3857015646864267, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.942523956298828, "learning_rate": 5e-05, "loss": 0.9266, "mean_token_accuracy": 0.8340761065483093, "num_tokens": 115232650.0, "step": 3032 }, { "epoch": 0.38582877496501716, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.242886543273926, "learning_rate": 5e-05, "loss": 0.8603, "mean_token_accuracy": 0.8690418004989624, "num_tokens": 115267812.0, "step": 3033 }, { "epoch": 0.3859559852436077, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.957117080688477, "learning_rate": 5e-05, "loss": 0.8002, "mean_token_accuracy": 0.8621557950973511, "num_tokens": 115300905.0, "step": 3034 }, { "epoch": 0.3860831955221982, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.522693157196045, "learning_rate": 5e-05, "loss": 0.7893, "mean_token_accuracy": 0.8649558424949646, "num_tokens": 115341869.0, "step": 3035 }, { "epoch": 0.3862104058007887, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.722893238067627, "learning_rate": 5e-05, "loss": 0.8781, "mean_token_accuracy": 0.8341975212097168, "num_tokens": 115380499.0, "step": 3036 }, { "epoch": 0.3863376160793792, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.227327823638916, "learning_rate": 5e-05, "loss": 0.7394, "mean_token_accuracy": 0.864091694355011, "num_tokens": 115413573.0, "step": 3037 }, { "epoch": 0.38646482635796975, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.624849319458008, "learning_rate": 5e-05, "loss": 0.7601, "mean_token_accuracy": 0.8653578758239746, "num_tokens": 115452634.0, "step": 3038 }, { "epoch": 0.3865920366365602, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.013436794281006, "learning_rate": 5e-05, "loss": 0.8392, "mean_token_accuracy": 0.8299383521080017, "num_tokens": 115492509.0, "step": 3039 }, { "epoch": 0.38671924691515075, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.789981365203857, "learning_rate": 5e-05, "loss": 0.7497, "mean_token_accuracy": 0.8672087788581848, "num_tokens": 115531256.0, "step": 3040 }, { "epoch": 0.3868464571937413, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.738464832305908, "learning_rate": 5e-05, "loss": 0.7313, "mean_token_accuracy": 0.8473918437957764, "num_tokens": 115564173.0, "step": 3041 }, { "epoch": 0.38697366747233175, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.08768367767334, "learning_rate": 5e-05, "loss": 0.8114, "mean_token_accuracy": 0.8618887662887573, "num_tokens": 115598359.0, "step": 3042 }, { "epoch": 0.3871008777509223, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.556262016296387, "learning_rate": 5e-05, "loss": 0.7259, "mean_token_accuracy": 0.854412317276001, "num_tokens": 115637782.0, "step": 3043 }, { "epoch": 0.3872280880295128, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.321805000305176, "learning_rate": 5e-05, "loss": 0.9105, "mean_token_accuracy": 0.8410662412643433, "num_tokens": 115680809.0, "step": 3044 }, { "epoch": 0.3873552983081033, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.631221771240234, "learning_rate": 5e-05, "loss": 0.7292, "mean_token_accuracy": 0.8529781699180603, "num_tokens": 115712092.0, "step": 3045 }, { "epoch": 0.3874825085866938, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.266240119934082, "learning_rate": 5e-05, "loss": 0.8643, "mean_token_accuracy": 0.8555712699890137, "num_tokens": 115749400.0, "step": 3046 }, { "epoch": 0.38760971886528434, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.955582141876221, "learning_rate": 5e-05, "loss": 0.7578, "mean_token_accuracy": 0.8495098948478699, "num_tokens": 115783932.0, "step": 3047 }, { "epoch": 0.3877369291438748, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.903512477874756, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.8541945219039917, "num_tokens": 115819358.0, "step": 3048 }, { "epoch": 0.38786413942246534, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.225503921508789, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.8668771386146545, "num_tokens": 115855370.0, "step": 3049 }, { "epoch": 0.38799134970105587, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.538436412811279, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8389203548431396, "num_tokens": 115894722.0, "step": 3050 }, { "epoch": 0.38811855997964634, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.276557922363281, "learning_rate": 5e-05, "loss": 0.7246, "mean_token_accuracy": 0.86236572265625, "num_tokens": 115935276.0, "step": 3051 }, { "epoch": 0.38824577025823687, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.445851802825928, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8531818389892578, "num_tokens": 115971255.0, "step": 3052 }, { "epoch": 0.3883729805368274, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.066339492797852, "learning_rate": 5e-05, "loss": 0.6958, "mean_token_accuracy": 0.8665465712547302, "num_tokens": 116003818.0, "step": 3053 }, { "epoch": 0.38850019081541787, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.695230007171631, "learning_rate": 5e-05, "loss": 0.7664, "mean_token_accuracy": 0.8595257997512817, "num_tokens": 116040720.0, "step": 3054 }, { "epoch": 0.3886274010940084, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.689216613769531, "learning_rate": 5e-05, "loss": 0.6783, "mean_token_accuracy": 0.8676950931549072, "num_tokens": 116076485.0, "step": 3055 }, { "epoch": 0.3887546113725989, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.07154655456543, "learning_rate": 5e-05, "loss": 0.7716, "mean_token_accuracy": 0.869133710861206, "num_tokens": 116109802.0, "step": 3056 }, { "epoch": 0.3888818216511894, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.402860164642334, "learning_rate": 5e-05, "loss": 0.7073, "mean_token_accuracy": 0.8531414270401001, "num_tokens": 116149865.0, "step": 3057 }, { "epoch": 0.3890090319297799, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.383302688598633, "learning_rate": 5e-05, "loss": 0.8409, "mean_token_accuracy": 0.8573681116104126, "num_tokens": 116185887.0, "step": 3058 }, { "epoch": 0.38913624220837045, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.40586519241333, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8479586243629456, "num_tokens": 116233090.0, "step": 3059 }, { "epoch": 0.3892634524869609, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.316021919250488, "learning_rate": 5e-05, "loss": 0.8702, "mean_token_accuracy": 0.8538117408752441, "num_tokens": 116274433.0, "step": 3060 }, { "epoch": 0.38939066276555145, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.88028621673584, "learning_rate": 5e-05, "loss": 0.822, "mean_token_accuracy": 0.8371439576148987, "num_tokens": 116309104.0, "step": 3061 }, { "epoch": 0.389517873044142, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.620877265930176, "learning_rate": 5e-05, "loss": 0.8007, "mean_token_accuracy": 0.8608306646347046, "num_tokens": 116350248.0, "step": 3062 }, { "epoch": 0.38964508332273246, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.551278591156006, "learning_rate": 5e-05, "loss": 0.8172, "mean_token_accuracy": 0.8452932238578796, "num_tokens": 116391078.0, "step": 3063 }, { "epoch": 0.389772293601323, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.750682830810547, "learning_rate": 5e-05, "loss": 0.7316, "mean_token_accuracy": 0.8508824110031128, "num_tokens": 116430652.0, "step": 3064 }, { "epoch": 0.3898995038799135, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.157949447631836, "learning_rate": 5e-05, "loss": 0.7848, "mean_token_accuracy": 0.8678051233291626, "num_tokens": 116467024.0, "step": 3065 }, { "epoch": 0.390026714158504, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.2174272537231445, "learning_rate": 5e-05, "loss": 0.6664, "mean_token_accuracy": 0.8638277053833008, "num_tokens": 116504304.0, "step": 3066 }, { "epoch": 0.3901539244370945, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.62272834777832, "learning_rate": 5e-05, "loss": 0.8792, "mean_token_accuracy": 0.8495460748672485, "num_tokens": 116541079.0, "step": 3067 }, { "epoch": 0.39028113471568504, "ewc_loss": 0.265625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.058521270751953, "learning_rate": 5e-05, "loss": 0.7416, "mean_token_accuracy": 0.8461862802505493, "num_tokens": 116578690.0, "step": 3068 }, { "epoch": 0.3904083449942755, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.814188003540039, "learning_rate": 5e-05, "loss": 0.9174, "mean_token_accuracy": 0.8547975420951843, "num_tokens": 116614945.0, "step": 3069 }, { "epoch": 0.39053555527286604, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.48954439163208, "learning_rate": 5e-05, "loss": 0.6965, "mean_token_accuracy": 0.8684068918228149, "num_tokens": 116646795.0, "step": 3070 }, { "epoch": 0.39066276555145657, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.306838989257812, "learning_rate": 5e-05, "loss": 0.9248, "mean_token_accuracy": 0.8483181595802307, "num_tokens": 116685731.0, "step": 3071 }, { "epoch": 0.39078997583004704, "ewc_loss": 0.375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.278866291046143, "learning_rate": 5e-05, "loss": 0.7266, "mean_token_accuracy": 0.8770469427108765, "num_tokens": 116722559.0, "step": 3072 }, { "epoch": 0.39091718610863757, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.359309196472168, "learning_rate": 5e-05, "loss": 0.8618, "mean_token_accuracy": 0.8400754332542419, "num_tokens": 116762954.0, "step": 3073 }, { "epoch": 0.3910443963872281, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.827883243560791, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.8652410507202148, "num_tokens": 116801745.0, "step": 3074 }, { "epoch": 0.39117160666581857, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.839584827423096, "learning_rate": 5e-05, "loss": 0.7154, "mean_token_accuracy": 0.8683598637580872, "num_tokens": 116837650.0, "step": 3075 }, { "epoch": 0.3912988169444091, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.869245529174805, "learning_rate": 5e-05, "loss": 0.8537, "mean_token_accuracy": 0.8445578813552856, "num_tokens": 116872722.0, "step": 3076 }, { "epoch": 0.39142602722299963, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.957322120666504, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.8579400181770325, "num_tokens": 116911175.0, "step": 3077 }, { "epoch": 0.3915532375015901, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.4932098388671875, "learning_rate": 5e-05, "loss": 0.7784, "mean_token_accuracy": 0.8533176183700562, "num_tokens": 116946317.0, "step": 3078 }, { "epoch": 0.39168044778018063, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.279946804046631, "learning_rate": 5e-05, "loss": 0.8252, "mean_token_accuracy": 0.837968111038208, "num_tokens": 116988661.0, "step": 3079 }, { "epoch": 0.39180765805877116, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.111690521240234, "learning_rate": 5e-05, "loss": 0.7252, "mean_token_accuracy": 0.8638978004455566, "num_tokens": 117027309.0, "step": 3080 }, { "epoch": 0.3919348683373617, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.494465351104736, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8578274250030518, "num_tokens": 117065608.0, "step": 3081 }, { "epoch": 0.39206207861595216, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.916467189788818, "learning_rate": 5e-05, "loss": 0.7556, "mean_token_accuracy": 0.8512072563171387, "num_tokens": 117101824.0, "step": 3082 }, { "epoch": 0.3921892888945427, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.60609769821167, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8503905534744263, "num_tokens": 117141985.0, "step": 3083 }, { "epoch": 0.3923164991731332, "ewc_loss": 0.296875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.777097702026367, "learning_rate": 5e-05, "loss": 0.7059, "mean_token_accuracy": 0.8619729280471802, "num_tokens": 117177606.0, "step": 3084 }, { "epoch": 0.3924437094517237, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.764318943023682, "learning_rate": 5e-05, "loss": 0.7832, "mean_token_accuracy": 0.856451153755188, "num_tokens": 117220214.0, "step": 3085 }, { "epoch": 0.3925709197303142, "ewc_loss": 0.28125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.601673126220703, "learning_rate": 5e-05, "loss": 0.7346, "mean_token_accuracy": 0.8487646579742432, "num_tokens": 117259188.0, "step": 3086 }, { "epoch": 0.39269813000890474, "ewc_loss": 0.375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.970717430114746, "learning_rate": 5e-05, "loss": 0.8171, "mean_token_accuracy": 0.8519338965415955, "num_tokens": 117299146.0, "step": 3087 }, { "epoch": 0.3928253402874952, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.407732009887695, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8536639213562012, "num_tokens": 117337560.0, "step": 3088 }, { "epoch": 0.39295255056608575, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.269920349121094, "learning_rate": 5e-05, "loss": 0.8892, "mean_token_accuracy": 0.8407992720603943, "num_tokens": 117376935.0, "step": 3089 }, { "epoch": 0.3930797608446763, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00023937225341796875, "grad_norm": 6.161013126373291, "learning_rate": 5e-05, "loss": 0.7093, "mean_token_accuracy": 0.8572440147399902, "num_tokens": 117420273.0, "step": 3090 }, { "epoch": 0.39320697112326675, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.637812614440918, "learning_rate": 5e-05, "loss": 0.8357, "mean_token_accuracy": 0.8685631155967712, "num_tokens": 117457338.0, "step": 3091 }, { "epoch": 0.3933341814018573, "ewc_loss": 0.28125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.067482948303223, "learning_rate": 5e-05, "loss": 0.6749, "mean_token_accuracy": 0.8678473234176636, "num_tokens": 117491082.0, "step": 3092 }, { "epoch": 0.3934613916804478, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.876960754394531, "learning_rate": 5e-05, "loss": 0.8435, "mean_token_accuracy": 0.8847231864929199, "num_tokens": 117520425.0, "step": 3093 }, { "epoch": 0.3935886019590383, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.381371021270752, "learning_rate": 5e-05, "loss": 0.714, "mean_token_accuracy": 0.86729896068573, "num_tokens": 117559021.0, "step": 3094 }, { "epoch": 0.3937158122376288, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.60376262664795, "learning_rate": 5e-05, "loss": 0.9211, "mean_token_accuracy": 0.8644605875015259, "num_tokens": 117598705.0, "step": 3095 }, { "epoch": 0.39384302251621933, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.098198413848877, "learning_rate": 5e-05, "loss": 0.8181, "mean_token_accuracy": 0.8670019507408142, "num_tokens": 117636185.0, "step": 3096 }, { "epoch": 0.3939702327948098, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.7487263679504395, "learning_rate": 5e-05, "loss": 0.8283, "mean_token_accuracy": 0.8657025694847107, "num_tokens": 117674290.0, "step": 3097 }, { "epoch": 0.39409744307340033, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.7170023918151855, "learning_rate": 5e-05, "loss": 0.8589, "mean_token_accuracy": 0.8611123561859131, "num_tokens": 117708708.0, "step": 3098 }, { "epoch": 0.39422465335199086, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.971566677093506, "learning_rate": 5e-05, "loss": 0.785, "mean_token_accuracy": 0.854947566986084, "num_tokens": 117742528.0, "step": 3099 }, { "epoch": 0.39435186363058133, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.883238792419434, "learning_rate": 5e-05, "loss": 0.8826, "mean_token_accuracy": 0.8435839414596558, "num_tokens": 117779050.0, "step": 3100 }, { "epoch": 0.39447907390917186, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.885993003845215, "learning_rate": 5e-05, "loss": 0.7857, "mean_token_accuracy": 0.8453048467636108, "num_tokens": 117824004.0, "step": 3101 }, { "epoch": 0.3946062841877624, "ewc_loss": 0.375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.566976547241211, "learning_rate": 5e-05, "loss": 0.7589, "mean_token_accuracy": 0.8679065704345703, "num_tokens": 117864837.0, "step": 3102 }, { "epoch": 0.39473349446635286, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.207018852233887, "learning_rate": 5e-05, "loss": 0.7424, "mean_token_accuracy": 0.859581708908081, "num_tokens": 117903675.0, "step": 3103 }, { "epoch": 0.3948607047449434, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.161317825317383, "learning_rate": 5e-05, "loss": 0.7672, "mean_token_accuracy": 0.8527759313583374, "num_tokens": 117942561.0, "step": 3104 }, { "epoch": 0.3949879150235339, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.4550628662109375, "learning_rate": 5e-05, "loss": 0.8069, "mean_token_accuracy": 0.8503159284591675, "num_tokens": 117982439.0, "step": 3105 }, { "epoch": 0.3951151253021244, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.9551005363464355, "learning_rate": 5e-05, "loss": 0.7588, "mean_token_accuracy": 0.8523238301277161, "num_tokens": 118025153.0, "step": 3106 }, { "epoch": 0.3952423355807149, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.52029275894165, "learning_rate": 5e-05, "loss": 0.7915, "mean_token_accuracy": 0.8503157496452332, "num_tokens": 118064508.0, "step": 3107 }, { "epoch": 0.39536954585930545, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.91429328918457, "learning_rate": 5e-05, "loss": 0.7057, "mean_token_accuracy": 0.8620530366897583, "num_tokens": 118093368.0, "step": 3108 }, { "epoch": 0.3954967561378959, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.518714427947998, "learning_rate": 5e-05, "loss": 0.7212, "mean_token_accuracy": 0.8676049113273621, "num_tokens": 118129164.0, "step": 3109 }, { "epoch": 0.39562396641648645, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.872147560119629, "learning_rate": 5e-05, "loss": 0.7025, "mean_token_accuracy": 0.8640221357345581, "num_tokens": 118169092.0, "step": 3110 }, { "epoch": 0.395751176695077, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.578627586364746, "learning_rate": 5e-05, "loss": 0.7672, "mean_token_accuracy": 0.8569214940071106, "num_tokens": 118211616.0, "step": 3111 }, { "epoch": 0.39587838697366745, "ewc_loss": 0.296875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.775021553039551, "learning_rate": 5e-05, "loss": 0.7374, "mean_token_accuracy": 0.8505045771598816, "num_tokens": 118248483.0, "step": 3112 }, { "epoch": 0.396005597252258, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.695693492889404, "learning_rate": 5e-05, "loss": 0.7843, "mean_token_accuracy": 0.8567107915878296, "num_tokens": 118288540.0, "step": 3113 }, { "epoch": 0.3961328075308485, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.626924514770508, "learning_rate": 5e-05, "loss": 0.7193, "mean_token_accuracy": 0.8578789234161377, "num_tokens": 118326988.0, "step": 3114 }, { "epoch": 0.396260017809439, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.909262657165527, "learning_rate": 5e-05, "loss": 0.7973, "mean_token_accuracy": 0.8583539724349976, "num_tokens": 118366100.0, "step": 3115 }, { "epoch": 0.3963872280880295, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.385824680328369, "learning_rate": 5e-05, "loss": 0.7019, "mean_token_accuracy": 0.8540639877319336, "num_tokens": 118397243.0, "step": 3116 }, { "epoch": 0.39651443836662004, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.2680082321167, "learning_rate": 5e-05, "loss": 0.8866, "mean_token_accuracy": 0.8399550914764404, "num_tokens": 118438198.0, "step": 3117 }, { "epoch": 0.3966416486452105, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002384185791015625, "grad_norm": 6.058207988739014, "learning_rate": 5e-05, "loss": 0.7003, "mean_token_accuracy": 0.8512086868286133, "num_tokens": 118473496.0, "step": 3118 }, { "epoch": 0.39676885892380104, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.784466743469238, "learning_rate": 5e-05, "loss": 0.8766, "mean_token_accuracy": 0.8652414083480835, "num_tokens": 118511988.0, "step": 3119 }, { "epoch": 0.39689606920239157, "ewc_loss": 0.28125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.875050067901611, "learning_rate": 5e-05, "loss": 0.6698, "mean_token_accuracy": 0.8649472594261169, "num_tokens": 118548287.0, "step": 3120 }, { "epoch": 0.39702327948098204, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.178298950195312, "learning_rate": 5e-05, "loss": 0.956, "mean_token_accuracy": 0.8649375438690186, "num_tokens": 118580637.0, "step": 3121 }, { "epoch": 0.39715048975957257, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.372814655303955, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8474792242050171, "num_tokens": 118622316.0, "step": 3122 }, { "epoch": 0.3972777000381631, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.735516548156738, "learning_rate": 5e-05, "loss": 1.0013, "mean_token_accuracy": 0.8449214696884155, "num_tokens": 118660497.0, "step": 3123 }, { "epoch": 0.39740491031675357, "ewc_loss": 0.5, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.5480780601501465, "learning_rate": 5e-05, "loss": 0.9386, "mean_token_accuracy": 0.8523476719856262, "num_tokens": 118691733.0, "step": 3124 }, { "epoch": 0.3975321205953441, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.14136266708374, "learning_rate": 5e-05, "loss": 0.8216, "mean_token_accuracy": 0.8442268967628479, "num_tokens": 118728576.0, "step": 3125 }, { "epoch": 0.3976593308739346, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.7972993850708, "learning_rate": 5e-05, "loss": 1.0202, "mean_token_accuracy": 0.8445636034011841, "num_tokens": 118765394.0, "step": 3126 }, { "epoch": 0.3977865411525251, "ewc_loss": 0.28125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.806687831878662, "learning_rate": 5e-05, "loss": 0.7395, "mean_token_accuracy": 0.8438828587532043, "num_tokens": 118804856.0, "step": 3127 }, { "epoch": 0.3979137514311156, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.051407814025879, "learning_rate": 5e-05, "loss": 1.0645, "mean_token_accuracy": 0.8547248840332031, "num_tokens": 118839274.0, "step": 3128 }, { "epoch": 0.39804096170970615, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 6.296970844268799, "learning_rate": 5e-05, "loss": 0.8596, "mean_token_accuracy": 0.8459113240242004, "num_tokens": 118877866.0, "step": 3129 }, { "epoch": 0.3981681719882967, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.382272720336914, "learning_rate": 5e-05, "loss": 1.0979, "mean_token_accuracy": 0.8698213696479797, "num_tokens": 118908942.0, "step": 3130 }, { "epoch": 0.39829538226688715, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 8.32393741607666, "learning_rate": 5e-05, "loss": 1.1815, "mean_token_accuracy": 0.8421977162361145, "num_tokens": 118948957.0, "step": 3131 }, { "epoch": 0.3984225925454777, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.800137519836426, "learning_rate": 5e-05, "loss": 0.9553, "mean_token_accuracy": 0.862263560295105, "num_tokens": 118991246.0, "step": 3132 }, { "epoch": 0.3985498028240682, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 9.032320976257324, "learning_rate": 5e-05, "loss": 1.1482, "mean_token_accuracy": 0.8402862548828125, "num_tokens": 119025661.0, "step": 3133 }, { "epoch": 0.3986770131026587, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.324833393096924, "learning_rate": 5e-05, "loss": 0.8782, "mean_token_accuracy": 0.8503603339195251, "num_tokens": 119059524.0, "step": 3134 }, { "epoch": 0.3988042233812492, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.513230323791504, "learning_rate": 5e-05, "loss": 1.0767, "mean_token_accuracy": 0.8470340371131897, "num_tokens": 119101636.0, "step": 3135 }, { "epoch": 0.39893143365983974, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.424783229827881, "learning_rate": 5e-05, "loss": 0.9753, "mean_token_accuracy": 0.8477523922920227, "num_tokens": 119138448.0, "step": 3136 }, { "epoch": 0.3990586439384302, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.88658618927002, "learning_rate": 5e-05, "loss": 1.0028, "mean_token_accuracy": 0.8447785377502441, "num_tokens": 119183337.0, "step": 3137 }, { "epoch": 0.39918585421702074, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 7.988590717315674, "learning_rate": 5e-05, "loss": 0.9264, "mean_token_accuracy": 0.8808775544166565, "num_tokens": 119217503.0, "step": 3138 }, { "epoch": 0.39931306449561127, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.641962051391602, "learning_rate": 5e-05, "loss": 0.9041, "mean_token_accuracy": 0.8428026437759399, "num_tokens": 119255824.0, "step": 3139 }, { "epoch": 0.39944027477420174, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.453383445739746, "learning_rate": 5e-05, "loss": 0.9933, "mean_token_accuracy": 0.853039026260376, "num_tokens": 119292677.0, "step": 3140 }, { "epoch": 0.39956748505279227, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 6.9677629470825195, "learning_rate": 5e-05, "loss": 0.8038, "mean_token_accuracy": 0.8586037158966064, "num_tokens": 119330961.0, "step": 3141 }, { "epoch": 0.3996946953313828, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.66398811340332, "learning_rate": 5e-05, "loss": 1.0048, "mean_token_accuracy": 0.8431851267814636, "num_tokens": 119362122.0, "step": 3142 }, { "epoch": 0.39982190560997327, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.0035529136657715, "learning_rate": 5e-05, "loss": 0.8407, "mean_token_accuracy": 0.8576468229293823, "num_tokens": 119398018.0, "step": 3143 }, { "epoch": 0.3999491158885638, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.380692481994629, "learning_rate": 5e-05, "loss": 0.948, "mean_token_accuracy": 0.8469445109367371, "num_tokens": 119438552.0, "step": 3144 }, { "epoch": 0.40007632616715433, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.242035865783691, "learning_rate": 5e-05, "loss": 0.8884, "mean_token_accuracy": 0.8541274666786194, "num_tokens": 119470777.0, "step": 3145 }, { "epoch": 0.4002035364457448, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.829607963562012, "learning_rate": 5e-05, "loss": 0.8142, "mean_token_accuracy": 0.8691256046295166, "num_tokens": 119505062.0, "step": 3146 }, { "epoch": 0.40033074672433533, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.443253517150879, "learning_rate": 5e-05, "loss": 0.8777, "mean_token_accuracy": 0.8518631458282471, "num_tokens": 119540394.0, "step": 3147 }, { "epoch": 0.40045795700292586, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.295662879943848, "learning_rate": 5e-05, "loss": 0.7215, "mean_token_accuracy": 0.8777192831039429, "num_tokens": 119581657.0, "step": 3148 }, { "epoch": 0.40058516728151633, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.730798721313477, "learning_rate": 5e-05, "loss": 0.8186, "mean_token_accuracy": 0.8633930683135986, "num_tokens": 119617140.0, "step": 3149 }, { "epoch": 0.40071237756010686, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.885526180267334, "learning_rate": 5e-05, "loss": 0.8154, "mean_token_accuracy": 0.8401147127151489, "num_tokens": 119660155.0, "step": 3150 }, { "epoch": 0.4008395878386974, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.019604682922363, "learning_rate": 5e-05, "loss": 0.8985, "mean_token_accuracy": 0.8438697457313538, "num_tokens": 119706258.0, "step": 3151 }, { "epoch": 0.40096679811728786, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.634394645690918, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.8578802347183228, "num_tokens": 119742564.0, "step": 3152 }, { "epoch": 0.4010940083958784, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.226274490356445, "learning_rate": 5e-05, "loss": 0.868, "mean_token_accuracy": 0.8557881712913513, "num_tokens": 119781301.0, "step": 3153 }, { "epoch": 0.4012212186744689, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.53456449508667, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.850055456161499, "num_tokens": 119816513.0, "step": 3154 }, { "epoch": 0.4013484289530594, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.28072452545166, "learning_rate": 5e-05, "loss": 0.8697, "mean_token_accuracy": 0.8606052398681641, "num_tokens": 119858510.0, "step": 3155 }, { "epoch": 0.4014756392316499, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.614976406097412, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8591817617416382, "num_tokens": 119902243.0, "step": 3156 }, { "epoch": 0.40160284951024044, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.131000518798828, "learning_rate": 5e-05, "loss": 0.8439, "mean_token_accuracy": 0.864242434501648, "num_tokens": 119938956.0, "step": 3157 }, { "epoch": 0.4017300597888309, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.771874904632568, "learning_rate": 5e-05, "loss": 0.763, "mean_token_accuracy": 0.8651891350746155, "num_tokens": 119976070.0, "step": 3158 }, { "epoch": 0.40185727006742145, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.929597854614258, "learning_rate": 5e-05, "loss": 0.8179, "mean_token_accuracy": 0.8613066673278809, "num_tokens": 120011811.0, "step": 3159 }, { "epoch": 0.401984480346012, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.875367641448975, "learning_rate": 5e-05, "loss": 0.7869, "mean_token_accuracy": 0.849837064743042, "num_tokens": 120042278.0, "step": 3160 }, { "epoch": 0.40211169062460245, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.79972505569458, "learning_rate": 5e-05, "loss": 0.8219, "mean_token_accuracy": 0.8539656400680542, "num_tokens": 120078379.0, "step": 3161 }, { "epoch": 0.402238900903193, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.8646392822265625, "learning_rate": 5e-05, "loss": 0.7195, "mean_token_accuracy": 0.867304801940918, "num_tokens": 120118852.0, "step": 3162 }, { "epoch": 0.4023661111817835, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.804292678833008, "learning_rate": 5e-05, "loss": 0.7976, "mean_token_accuracy": 0.8667848110198975, "num_tokens": 120153825.0, "step": 3163 }, { "epoch": 0.402493321460374, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.763824939727783, "learning_rate": 5e-05, "loss": 0.7564, "mean_token_accuracy": 0.8525265455245972, "num_tokens": 120190945.0, "step": 3164 }, { "epoch": 0.4026205317389645, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.92238187789917, "learning_rate": 5e-05, "loss": 0.7972, "mean_token_accuracy": 0.8631002902984619, "num_tokens": 120221435.0, "step": 3165 }, { "epoch": 0.40274774201755503, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.6175150871276855, "learning_rate": 5e-05, "loss": 0.722, "mean_token_accuracy": 0.8631073236465454, "num_tokens": 120253553.0, "step": 3166 }, { "epoch": 0.4028749522961455, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.078086853027344, "learning_rate": 5e-05, "loss": 0.8524, "mean_token_accuracy": 0.8515958189964294, "num_tokens": 120291438.0, "step": 3167 }, { "epoch": 0.40300216257473603, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.54246187210083, "learning_rate": 5e-05, "loss": 0.7506, "mean_token_accuracy": 0.8515872955322266, "num_tokens": 120332005.0, "step": 3168 }, { "epoch": 0.40312937285332656, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.12890338897705, "learning_rate": 5e-05, "loss": 0.8915, "mean_token_accuracy": 0.840017557144165, "num_tokens": 120369280.0, "step": 3169 }, { "epoch": 0.40325658313191703, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.637552261352539, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.8577835559844971, "num_tokens": 120402350.0, "step": 3170 }, { "epoch": 0.40338379341050756, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.946475982666016, "learning_rate": 5e-05, "loss": 0.8366, "mean_token_accuracy": 0.8619186282157898, "num_tokens": 120441791.0, "step": 3171 }, { "epoch": 0.4035110036890981, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.91593599319458, "learning_rate": 5e-05, "loss": 0.7117, "mean_token_accuracy": 0.8700764179229736, "num_tokens": 120476895.0, "step": 3172 }, { "epoch": 0.40363821396768856, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.55574369430542, "learning_rate": 5e-05, "loss": 0.82, "mean_token_accuracy": 0.8490670919418335, "num_tokens": 120518016.0, "step": 3173 }, { "epoch": 0.4037654242462791, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.256564140319824, "learning_rate": 5e-05, "loss": 0.7393, "mean_token_accuracy": 0.869690477848053, "num_tokens": 120551500.0, "step": 3174 }, { "epoch": 0.4038926345248696, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.105841159820557, "learning_rate": 5e-05, "loss": 0.7171, "mean_token_accuracy": 0.8658878803253174, "num_tokens": 120588543.0, "step": 3175 }, { "epoch": 0.4040198448034601, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.598711967468262, "learning_rate": 5e-05, "loss": 0.8224, "mean_token_accuracy": 0.8544740676879883, "num_tokens": 120634394.0, "step": 3176 }, { "epoch": 0.4041470550820506, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.743600368499756, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8514331579208374, "num_tokens": 120672982.0, "step": 3177 }, { "epoch": 0.40427426536064115, "ewc_loss": 0.375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.880198001861572, "learning_rate": 5e-05, "loss": 0.8719, "mean_token_accuracy": 0.833378791809082, "num_tokens": 120712091.0, "step": 3178 }, { "epoch": 0.4044014756392316, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.513218402862549, "learning_rate": 5e-05, "loss": 0.6767, "mean_token_accuracy": 0.8674386739730835, "num_tokens": 120743530.0, "step": 3179 }, { "epoch": 0.40452868591782215, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.057687759399414, "learning_rate": 5e-05, "loss": 0.7788, "mean_token_accuracy": 0.8694692254066467, "num_tokens": 120778884.0, "step": 3180 }, { "epoch": 0.4046558961964127, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.438793659210205, "learning_rate": 5e-05, "loss": 0.7235, "mean_token_accuracy": 0.8522374033927917, "num_tokens": 120815951.0, "step": 3181 }, { "epoch": 0.4047831064750032, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.113550186157227, "learning_rate": 5e-05, "loss": 0.8287, "mean_token_accuracy": 0.8556973338127136, "num_tokens": 120853111.0, "step": 3182 }, { "epoch": 0.4049103167535937, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.480388164520264, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.850225567817688, "num_tokens": 120894041.0, "step": 3183 }, { "epoch": 0.4050375270321842, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.073598861694336, "learning_rate": 5e-05, "loss": 0.8152, "mean_token_accuracy": 0.8635604977607727, "num_tokens": 120930730.0, "step": 3184 }, { "epoch": 0.40516473731077474, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.614995002746582, "learning_rate": 5e-05, "loss": 0.7464, "mean_token_accuracy": 0.852935254573822, "num_tokens": 120965044.0, "step": 3185 }, { "epoch": 0.4052919475893652, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.904889106750488, "learning_rate": 5e-05, "loss": 0.8328, "mean_token_accuracy": 0.8531026840209961, "num_tokens": 120999777.0, "step": 3186 }, { "epoch": 0.40541915786795574, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.783576965332031, "learning_rate": 5e-05, "loss": 0.796, "mean_token_accuracy": 0.8394714593887329, "num_tokens": 121031910.0, "step": 3187 }, { "epoch": 0.40554636814654627, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.718731880187988, "learning_rate": 5e-05, "loss": 0.8055, "mean_token_accuracy": 0.8614122271537781, "num_tokens": 121069600.0, "step": 3188 }, { "epoch": 0.40567357842513674, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.909502029418945, "learning_rate": 5e-05, "loss": 0.7933, "mean_token_accuracy": 0.8421928882598877, "num_tokens": 121103790.0, "step": 3189 }, { "epoch": 0.40580078870372727, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.616337776184082, "learning_rate": 5e-05, "loss": 0.8069, "mean_token_accuracy": 0.8490513563156128, "num_tokens": 121135199.0, "step": 3190 }, { "epoch": 0.4059279989823178, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.892821311950684, "learning_rate": 5e-05, "loss": 0.7565, "mean_token_accuracy": 0.854636549949646, "num_tokens": 121176040.0, "step": 3191 }, { "epoch": 0.40605520926090827, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.662242412567139, "learning_rate": 5e-05, "loss": 0.8422, "mean_token_accuracy": 0.8436505794525146, "num_tokens": 121214675.0, "step": 3192 }, { "epoch": 0.4061824195394988, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.753571510314941, "learning_rate": 5e-05, "loss": 0.7862, "mean_token_accuracy": 0.8414832353591919, "num_tokens": 121252834.0, "step": 3193 }, { "epoch": 0.4063096298180893, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.825659275054932, "learning_rate": 5e-05, "loss": 0.8357, "mean_token_accuracy": 0.8529433012008667, "num_tokens": 121292837.0, "step": 3194 }, { "epoch": 0.4064368400966798, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.573239326477051, "learning_rate": 5e-05, "loss": 0.7148, "mean_token_accuracy": 0.8595676422119141, "num_tokens": 121332353.0, "step": 3195 }, { "epoch": 0.4065640503752703, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.011460304260254, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.8647514581680298, "num_tokens": 121371095.0, "step": 3196 }, { "epoch": 0.40669126065386085, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.473556995391846, "learning_rate": 5e-05, "loss": 0.7889, "mean_token_accuracy": 0.8384681344032288, "num_tokens": 121409653.0, "step": 3197 }, { "epoch": 0.4068184709324513, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.122772216796875, "learning_rate": 5e-05, "loss": 0.8073, "mean_token_accuracy": 0.8673278093338013, "num_tokens": 121449309.0, "step": 3198 }, { "epoch": 0.40694568121104185, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.5079450607299805, "learning_rate": 5e-05, "loss": 0.7171, "mean_token_accuracy": 0.8624542951583862, "num_tokens": 121495004.0, "step": 3199 }, { "epoch": 0.4070728914896324, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.041350364685059, "learning_rate": 5e-05, "loss": 0.8236, "mean_token_accuracy": 0.8635927438735962, "num_tokens": 121537679.0, "step": 3200 }, { "epoch": 0.40720010176822286, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.766236305236816, "learning_rate": 5e-05, "loss": 0.7555, "mean_token_accuracy": 0.8553140163421631, "num_tokens": 121579156.0, "step": 3201 }, { "epoch": 0.4073273120468134, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.659932613372803, "learning_rate": 5e-05, "loss": 0.8442, "mean_token_accuracy": 0.8511024117469788, "num_tokens": 121611056.0, "step": 3202 }, { "epoch": 0.4074545223254039, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.18238639831543, "learning_rate": 5e-05, "loss": 0.7505, "mean_token_accuracy": 0.8654756546020508, "num_tokens": 121642821.0, "step": 3203 }, { "epoch": 0.4075817326039944, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.101935863494873, "learning_rate": 5e-05, "loss": 0.7766, "mean_token_accuracy": 0.850888729095459, "num_tokens": 121674021.0, "step": 3204 }, { "epoch": 0.4077089428825849, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.643894672393799, "learning_rate": 5e-05, "loss": 0.7999, "mean_token_accuracy": 0.8585361242294312, "num_tokens": 121711434.0, "step": 3205 }, { "epoch": 0.40783615316117544, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.5772705078125, "learning_rate": 5e-05, "loss": 0.7647, "mean_token_accuracy": 0.8419010639190674, "num_tokens": 121755357.0, "step": 3206 }, { "epoch": 0.4079633634397659, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.094355583190918, "learning_rate": 5e-05, "loss": 0.8053, "mean_token_accuracy": 0.8634435534477234, "num_tokens": 121791703.0, "step": 3207 }, { "epoch": 0.40809057371835644, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.154201984405518, "learning_rate": 5e-05, "loss": 0.6652, "mean_token_accuracy": 0.8687467575073242, "num_tokens": 121824940.0, "step": 3208 }, { "epoch": 0.40821778399694697, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.598843574523926, "learning_rate": 5e-05, "loss": 0.8702, "mean_token_accuracy": 0.8593048453330994, "num_tokens": 121860434.0, "step": 3209 }, { "epoch": 0.40834499427553744, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 5.867735385894775, "learning_rate": 5e-05, "loss": 0.7358, "mean_token_accuracy": 0.843503475189209, "num_tokens": 121897536.0, "step": 3210 }, { "epoch": 0.40847220455412797, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.05853271484375, "learning_rate": 5e-05, "loss": 1.0022, "mean_token_accuracy": 0.8390159010887146, "num_tokens": 121936721.0, "step": 3211 }, { "epoch": 0.4085994148327185, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.127676486968994, "learning_rate": 5e-05, "loss": 0.7726, "mean_token_accuracy": 0.8511130213737488, "num_tokens": 121977407.0, "step": 3212 }, { "epoch": 0.40872662511130897, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.926566123962402, "learning_rate": 5e-05, "loss": 1.0479, "mean_token_accuracy": 0.8383979797363281, "num_tokens": 122015068.0, "step": 3213 }, { "epoch": 0.4088538353898995, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.1212968826293945, "learning_rate": 5e-05, "loss": 0.9091, "mean_token_accuracy": 0.8531767129898071, "num_tokens": 122058899.0, "step": 3214 }, { "epoch": 0.40898104566849003, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.735931396484375, "learning_rate": 5e-05, "loss": 0.8634, "mean_token_accuracy": 0.8595414161682129, "num_tokens": 122095407.0, "step": 3215 }, { "epoch": 0.4091082559470805, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.1648530960083, "learning_rate": 5e-05, "loss": 0.9344, "mean_token_accuracy": 0.8612092137336731, "num_tokens": 122134836.0, "step": 3216 }, { "epoch": 0.40923546622567103, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.430630683898926, "learning_rate": 5e-05, "loss": 0.7567, "mean_token_accuracy": 0.857993483543396, "num_tokens": 122165959.0, "step": 3217 }, { "epoch": 0.40936267650426156, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.920780181884766, "learning_rate": 5e-05, "loss": 0.9745, "mean_token_accuracy": 0.8503539562225342, "num_tokens": 122206223.0, "step": 3218 }, { "epoch": 0.40948988678285203, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.20037317276001, "learning_rate": 5e-05, "loss": 0.7325, "mean_token_accuracy": 0.8629801273345947, "num_tokens": 122245416.0, "step": 3219 }, { "epoch": 0.40961709706144256, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.872577667236328, "learning_rate": 5e-05, "loss": 1.0275, "mean_token_accuracy": 0.8401399850845337, "num_tokens": 122285264.0, "step": 3220 }, { "epoch": 0.4097443073400331, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 6.977945327758789, "learning_rate": 5e-05, "loss": 0.8842, "mean_token_accuracy": 0.8519461750984192, "num_tokens": 122326956.0, "step": 3221 }, { "epoch": 0.40987151761862356, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.920616149902344, "learning_rate": 5e-05, "loss": 0.8843, "mean_token_accuracy": 0.864111065864563, "num_tokens": 122362409.0, "step": 3222 }, { "epoch": 0.4099987278972141, "ewc_loss": 0.5, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.74757194519043, "learning_rate": 5e-05, "loss": 0.9914, "mean_token_accuracy": 0.8335961103439331, "num_tokens": 122405011.0, "step": 3223 }, { "epoch": 0.4101259381758046, "ewc_loss": 0.375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.019186019897461, "learning_rate": 5e-05, "loss": 0.7881, "mean_token_accuracy": 0.8610357046127319, "num_tokens": 122445482.0, "step": 3224 }, { "epoch": 0.4102531484543951, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.027199745178223, "learning_rate": 5e-05, "loss": 0.8846, "mean_token_accuracy": 0.8617538213729858, "num_tokens": 122485024.0, "step": 3225 }, { "epoch": 0.4103803587329856, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.870847702026367, "learning_rate": 5e-05, "loss": 0.7752, "mean_token_accuracy": 0.8577622771263123, "num_tokens": 122524766.0, "step": 3226 }, { "epoch": 0.41050756901157615, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.7709269523620605, "learning_rate": 5e-05, "loss": 0.8511, "mean_token_accuracy": 0.8573992848396301, "num_tokens": 122566664.0, "step": 3227 }, { "epoch": 0.4106347792901666, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.145397663116455, "learning_rate": 5e-05, "loss": 0.9083, "mean_token_accuracy": 0.8235048651695251, "num_tokens": 122606724.0, "step": 3228 }, { "epoch": 0.41076198956875715, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.3445611000061035, "learning_rate": 5e-05, "loss": 0.8146, "mean_token_accuracy": 0.8499305248260498, "num_tokens": 122641664.0, "step": 3229 }, { "epoch": 0.4108891998473477, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.360269546508789, "learning_rate": 5e-05, "loss": 0.7561, "mean_token_accuracy": 0.8744153380393982, "num_tokens": 122673370.0, "step": 3230 }, { "epoch": 0.4110164101259382, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.095583438873291, "learning_rate": 5e-05, "loss": 0.7649, "mean_token_accuracy": 0.8564857840538025, "num_tokens": 122711545.0, "step": 3231 }, { "epoch": 0.4111436204045287, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.443199634552002, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.8596082329750061, "num_tokens": 122743656.0, "step": 3232 }, { "epoch": 0.4112708306831192, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.974673271179199, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8508392572402954, "num_tokens": 122778233.0, "step": 3233 }, { "epoch": 0.41139804096170973, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.535121440887451, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8642308712005615, "num_tokens": 122821063.0, "step": 3234 }, { "epoch": 0.4115252512403002, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.82318639755249, "learning_rate": 5e-05, "loss": 0.7637, "mean_token_accuracy": 0.8524330854415894, "num_tokens": 122855871.0, "step": 3235 }, { "epoch": 0.41165246151889073, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.68540096282959, "learning_rate": 5e-05, "loss": 0.7921, "mean_token_accuracy": 0.8579322695732117, "num_tokens": 122893990.0, "step": 3236 }, { "epoch": 0.41177967179748126, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.632174015045166, "learning_rate": 5e-05, "loss": 0.7302, "mean_token_accuracy": 0.8595428466796875, "num_tokens": 122932017.0, "step": 3237 }, { "epoch": 0.41190688207607173, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.921759605407715, "learning_rate": 5e-05, "loss": 0.7991, "mean_token_accuracy": 0.8612960577011108, "num_tokens": 122975581.0, "step": 3238 }, { "epoch": 0.41203409235466226, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.381155967712402, "learning_rate": 5e-05, "loss": 0.7237, "mean_token_accuracy": 0.8529032468795776, "num_tokens": 123006051.0, "step": 3239 }, { "epoch": 0.4121613026332528, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.2855806350708, "learning_rate": 5e-05, "loss": 0.8562, "mean_token_accuracy": 0.8509483337402344, "num_tokens": 123049756.0, "step": 3240 }, { "epoch": 0.41228851291184326, "ewc_loss": 0.28125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.0508131980896, "learning_rate": 5e-05, "loss": 0.7347, "mean_token_accuracy": 0.8484257459640503, "num_tokens": 123091059.0, "step": 3241 }, { "epoch": 0.4124157231904338, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.818611145019531, "learning_rate": 5e-05, "loss": 0.8953, "mean_token_accuracy": 0.8620374202728271, "num_tokens": 123130369.0, "step": 3242 }, { "epoch": 0.4125429334690243, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 5.842141628265381, "learning_rate": 5e-05, "loss": 0.7201, "mean_token_accuracy": 0.8567206859588623, "num_tokens": 123173507.0, "step": 3243 }, { "epoch": 0.4126701437476148, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.31014633178711, "learning_rate": 5e-05, "loss": 0.9661, "mean_token_accuracy": 0.8668975830078125, "num_tokens": 123211428.0, "step": 3244 }, { "epoch": 0.4127973540262053, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.285638332366943, "learning_rate": 5e-05, "loss": 0.8259, "mean_token_accuracy": 0.8506869077682495, "num_tokens": 123256115.0, "step": 3245 }, { "epoch": 0.41292456430479585, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.194892883300781, "learning_rate": 5e-05, "loss": 1.014, "mean_token_accuracy": 0.8583639860153198, "num_tokens": 123297319.0, "step": 3246 }, { "epoch": 0.4130517745833863, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.391975402832031, "learning_rate": 5e-05, "loss": 1.0125, "mean_token_accuracy": 0.8403362035751343, "num_tokens": 123341241.0, "step": 3247 }, { "epoch": 0.41317898486197685, "ewc_loss": 0.5, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.014569282531738, "learning_rate": 5e-05, "loss": 0.9592, "mean_token_accuracy": 0.8450932502746582, "num_tokens": 123381832.0, "step": 3248 }, { "epoch": 0.4133061951405674, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.236674308776855, "learning_rate": 5e-05, "loss": 0.9834, "mean_token_accuracy": 0.85970139503479, "num_tokens": 123418393.0, "step": 3249 }, { "epoch": 0.41343340541915785, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.730903625488281, "learning_rate": 5e-05, "loss": 0.7383, "mean_token_accuracy": 0.8673884868621826, "num_tokens": 123449916.0, "step": 3250 }, { "epoch": 0.4135606156977484, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.70612907409668, "learning_rate": 5e-05, "loss": 0.9001, "mean_token_accuracy": 0.8766775727272034, "num_tokens": 123484025.0, "step": 3251 }, { "epoch": 0.4136878259763389, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.517652988433838, "learning_rate": 5e-05, "loss": 0.7799, "mean_token_accuracy": 0.858437180519104, "num_tokens": 123522345.0, "step": 3252 }, { "epoch": 0.4138150362549294, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.488125801086426, "learning_rate": 5e-05, "loss": 0.9717, "mean_token_accuracy": 0.8546889424324036, "num_tokens": 123562123.0, "step": 3253 }, { "epoch": 0.4139422465335199, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.023167133331299, "learning_rate": 5e-05, "loss": 0.8874, "mean_token_accuracy": 0.8490746021270752, "num_tokens": 123606014.0, "step": 3254 }, { "epoch": 0.41406945681211044, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.81560754776001, "learning_rate": 5e-05, "loss": 0.8544, "mean_token_accuracy": 0.8671661615371704, "num_tokens": 123643483.0, "step": 3255 }, { "epoch": 0.4141966670907009, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.413377285003662, "learning_rate": 5e-05, "loss": 0.8914, "mean_token_accuracy": 0.8586088418960571, "num_tokens": 123689303.0, "step": 3256 }, { "epoch": 0.41432387736929144, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.3139424324035645, "learning_rate": 5e-05, "loss": 0.7857, "mean_token_accuracy": 0.8630118370056152, "num_tokens": 123722388.0, "step": 3257 }, { "epoch": 0.41445108764788197, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.490610599517822, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.8557952642440796, "num_tokens": 123761761.0, "step": 3258 }, { "epoch": 0.41457829792647244, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.147068500518799, "learning_rate": 5e-05, "loss": 0.7154, "mean_token_accuracy": 0.8790686726570129, "num_tokens": 123799138.0, "step": 3259 }, { "epoch": 0.41470550820506297, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.456846714019775, "learning_rate": 5e-05, "loss": 0.7887, "mean_token_accuracy": 0.8649770021438599, "num_tokens": 123840803.0, "step": 3260 }, { "epoch": 0.4148327184836535, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.077769756317139, "learning_rate": 5e-05, "loss": 0.8006, "mean_token_accuracy": 0.8529025316238403, "num_tokens": 123875705.0, "step": 3261 }, { "epoch": 0.41495992876224397, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.4464945793151855, "learning_rate": 5e-05, "loss": 0.8421, "mean_token_accuracy": 0.8430404663085938, "num_tokens": 123915830.0, "step": 3262 }, { "epoch": 0.4150871390408345, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.0037384033203125, "learning_rate": 5e-05, "loss": 0.7922, "mean_token_accuracy": 0.8461324572563171, "num_tokens": 123953855.0, "step": 3263 }, { "epoch": 0.415214349319425, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.481564521789551, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.857349157333374, "num_tokens": 123985697.0, "step": 3264 }, { "epoch": 0.4153415595980155, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.925790309906006, "learning_rate": 5e-05, "loss": 0.7577, "mean_token_accuracy": 0.8528422117233276, "num_tokens": 124021682.0, "step": 3265 }, { "epoch": 0.415468769876606, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.5340375900268555, "learning_rate": 5e-05, "loss": 0.8266, "mean_token_accuracy": 0.841761589050293, "num_tokens": 124057472.0, "step": 3266 }, { "epoch": 0.41559598015519655, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.811702251434326, "learning_rate": 5e-05, "loss": 0.7336, "mean_token_accuracy": 0.8594451546669006, "num_tokens": 124100160.0, "step": 3267 }, { "epoch": 0.415723190433787, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.670871734619141, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8542996048927307, "num_tokens": 124138503.0, "step": 3268 }, { "epoch": 0.41585040071237755, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.636847972869873, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8450330495834351, "num_tokens": 124176082.0, "step": 3269 }, { "epoch": 0.4159776109909681, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.890928745269775, "learning_rate": 5e-05, "loss": 0.8396, "mean_token_accuracy": 0.8496757745742798, "num_tokens": 124213853.0, "step": 3270 }, { "epoch": 0.41610482126955856, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.368577480316162, "learning_rate": 5e-05, "loss": 0.6593, "mean_token_accuracy": 0.8728433847427368, "num_tokens": 124250562.0, "step": 3271 }, { "epoch": 0.4162320315481491, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.297155380249023, "learning_rate": 5e-05, "loss": 0.8864, "mean_token_accuracy": 0.8416441679000854, "num_tokens": 124282321.0, "step": 3272 }, { "epoch": 0.4163592418267396, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.010748386383057, "learning_rate": 5e-05, "loss": 0.702, "mean_token_accuracy": 0.8610672354698181, "num_tokens": 124323172.0, "step": 3273 }, { "epoch": 0.4164864521053301, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.864266395568848, "learning_rate": 5e-05, "loss": 0.927, "mean_token_accuracy": 0.8549270629882812, "num_tokens": 124361585.0, "step": 3274 }, { "epoch": 0.4166136623839206, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 5.856263637542725, "learning_rate": 5e-05, "loss": 0.8078, "mean_token_accuracy": 0.8320357799530029, "num_tokens": 124398926.0, "step": 3275 }, { "epoch": 0.41674087266251114, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.27706241607666, "learning_rate": 5e-05, "loss": 0.9538, "mean_token_accuracy": 0.8719237446784973, "num_tokens": 124437914.0, "step": 3276 }, { "epoch": 0.4168680829411016, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 6.461444854736328, "learning_rate": 5e-05, "loss": 0.8266, "mean_token_accuracy": 0.8583031892776489, "num_tokens": 124478850.0, "step": 3277 }, { "epoch": 0.41699529321969214, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.869470596313477, "learning_rate": 5e-05, "loss": 0.9476, "mean_token_accuracy": 0.8702725768089294, "num_tokens": 124519361.0, "step": 3278 }, { "epoch": 0.41712250349828267, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.59641170501709, "learning_rate": 5e-05, "loss": 0.9916, "mean_token_accuracy": 0.853684663772583, "num_tokens": 124555616.0, "step": 3279 }, { "epoch": 0.4172497137768732, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.4202189445495605, "learning_rate": 5e-05, "loss": 0.8897, "mean_token_accuracy": 0.838835597038269, "num_tokens": 124593045.0, "step": 3280 }, { "epoch": 0.41737692405546367, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.559940338134766, "learning_rate": 5e-05, "loss": 1.0458, "mean_token_accuracy": 0.833601176738739, "num_tokens": 124631842.0, "step": 3281 }, { "epoch": 0.4175041343340542, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.233154296875, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.8512265682220459, "num_tokens": 124674466.0, "step": 3282 }, { "epoch": 0.41763134461264473, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.320977210998535, "learning_rate": 5e-05, "loss": 0.9953, "mean_token_accuracy": 0.8598031401634216, "num_tokens": 124707893.0, "step": 3283 }, { "epoch": 0.4177585548912352, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.417480945587158, "learning_rate": 5e-05, "loss": 0.8191, "mean_token_accuracy": 0.8554502725601196, "num_tokens": 124746230.0, "step": 3284 }, { "epoch": 0.41788576516982573, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.226621627807617, "learning_rate": 5e-05, "loss": 1.0289, "mean_token_accuracy": 0.8559244871139526, "num_tokens": 124786988.0, "step": 3285 }, { "epoch": 0.41801297544841626, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 7.493001461029053, "learning_rate": 5e-05, "loss": 1.0172, "mean_token_accuracy": 0.8485805988311768, "num_tokens": 124831061.0, "step": 3286 }, { "epoch": 0.41814018572700673, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.105409622192383, "learning_rate": 5e-05, "loss": 0.9212, "mean_token_accuracy": 0.8527883291244507, "num_tokens": 124870486.0, "step": 3287 }, { "epoch": 0.41826739600559726, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.10369873046875, "learning_rate": 5e-05, "loss": 0.9555, "mean_token_accuracy": 0.8682390451431274, "num_tokens": 124909675.0, "step": 3288 }, { "epoch": 0.4183946062841878, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.180929660797119, "learning_rate": 5e-05, "loss": 0.8659, "mean_token_accuracy": 0.8457313179969788, "num_tokens": 124953854.0, "step": 3289 }, { "epoch": 0.41852181656277826, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.11179256439209, "learning_rate": 5e-05, "loss": 0.96, "mean_token_accuracy": 0.8546201586723328, "num_tokens": 124991117.0, "step": 3290 }, { "epoch": 0.4186490268413688, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.171537399291992, "learning_rate": 5e-05, "loss": 0.846, "mean_token_accuracy": 0.8534290790557861, "num_tokens": 125027160.0, "step": 3291 }, { "epoch": 0.4187762371199593, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.556731224060059, "learning_rate": 5e-05, "loss": 0.8217, "mean_token_accuracy": 0.8644028306007385, "num_tokens": 125061324.0, "step": 3292 }, { "epoch": 0.4189034473985498, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.53539514541626, "learning_rate": 5e-05, "loss": 0.8676, "mean_token_accuracy": 0.8548799753189087, "num_tokens": 125099616.0, "step": 3293 }, { "epoch": 0.4190306576771403, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.990725517272949, "learning_rate": 5e-05, "loss": 0.7858, "mean_token_accuracy": 0.8508151769638062, "num_tokens": 125132855.0, "step": 3294 }, { "epoch": 0.41915786795573085, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.807723522186279, "learning_rate": 5e-05, "loss": 0.9038, "mean_token_accuracy": 0.8485063314437866, "num_tokens": 125168040.0, "step": 3295 }, { "epoch": 0.4192850782343213, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.658317565917969, "learning_rate": 5e-05, "loss": 0.7048, "mean_token_accuracy": 0.864155113697052, "num_tokens": 125198788.0, "step": 3296 }, { "epoch": 0.41941228851291185, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.143301963806152, "learning_rate": 5e-05, "loss": 0.8571, "mean_token_accuracy": 0.863002598285675, "num_tokens": 125238387.0, "step": 3297 }, { "epoch": 0.4195394987915024, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.350896835327148, "learning_rate": 5e-05, "loss": 0.7062, "mean_token_accuracy": 0.8675180077552795, "num_tokens": 125273666.0, "step": 3298 }, { "epoch": 0.41966670907009285, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.626021385192871, "learning_rate": 5e-05, "loss": 0.9503, "mean_token_accuracy": 0.8473990559577942, "num_tokens": 125315961.0, "step": 3299 }, { "epoch": 0.4197939193486834, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.195605754852295, "learning_rate": 5e-05, "loss": 0.7796, "mean_token_accuracy": 0.8436705470085144, "num_tokens": 125358830.0, "step": 3300 }, { "epoch": 0.4199211296272739, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.951117515563965, "learning_rate": 5e-05, "loss": 0.9979, "mean_token_accuracy": 0.8496559262275696, "num_tokens": 125397497.0, "step": 3301 }, { "epoch": 0.4200483399058644, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.520605087280273, "learning_rate": 5e-05, "loss": 0.8806, "mean_token_accuracy": 0.8357535600662231, "num_tokens": 125437337.0, "step": 3302 }, { "epoch": 0.4201755501844549, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.825239181518555, "learning_rate": 5e-05, "loss": 0.948, "mean_token_accuracy": 0.8648728728294373, "num_tokens": 125470535.0, "step": 3303 }, { "epoch": 0.42030276046304543, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.066619396209717, "learning_rate": 5e-05, "loss": 0.9432, "mean_token_accuracy": 0.8424088954925537, "num_tokens": 125510692.0, "step": 3304 }, { "epoch": 0.4204299707416359, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.355847358703613, "learning_rate": 5e-05, "loss": 0.9968, "mean_token_accuracy": 0.839858889579773, "num_tokens": 125557328.0, "step": 3305 }, { "epoch": 0.42055718102022643, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.310094833374023, "learning_rate": 5e-05, "loss": 0.9288, "mean_token_accuracy": 0.8520868420600891, "num_tokens": 125594006.0, "step": 3306 }, { "epoch": 0.42068439129881696, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.021749496459961, "learning_rate": 5e-05, "loss": 0.9199, "mean_token_accuracy": 0.8511725068092346, "num_tokens": 125627692.0, "step": 3307 }, { "epoch": 0.42081160157740743, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.144634246826172, "learning_rate": 5e-05, "loss": 0.8318, "mean_token_accuracy": 0.8599940538406372, "num_tokens": 125660218.0, "step": 3308 }, { "epoch": 0.42093881185599796, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.092120170593262, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.8642807006835938, "num_tokens": 125695793.0, "step": 3309 }, { "epoch": 0.4210660221345885, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.837535381317139, "learning_rate": 5e-05, "loss": 0.8011, "mean_token_accuracy": 0.8596656322479248, "num_tokens": 125732438.0, "step": 3310 }, { "epoch": 0.42119323241317896, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.323811531066895, "learning_rate": 5e-05, "loss": 0.9084, "mean_token_accuracy": 0.8642998337745667, "num_tokens": 125764919.0, "step": 3311 }, { "epoch": 0.4213204426917695, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.770280361175537, "learning_rate": 5e-05, "loss": 0.8079, "mean_token_accuracy": 0.8535584211349487, "num_tokens": 125797172.0, "step": 3312 }, { "epoch": 0.42144765297036, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.2659273147583, "learning_rate": 5e-05, "loss": 0.9137, "mean_token_accuracy": 0.8604000806808472, "num_tokens": 125838144.0, "step": 3313 }, { "epoch": 0.4215748632489505, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.044813632965088, "learning_rate": 5e-05, "loss": 0.8024, "mean_token_accuracy": 0.866309642791748, "num_tokens": 125873897.0, "step": 3314 }, { "epoch": 0.421702073527541, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.822873115539551, "learning_rate": 5e-05, "loss": 0.9295, "mean_token_accuracy": 0.8439285159111023, "num_tokens": 125909773.0, "step": 3315 }, { "epoch": 0.42182928380613155, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.340205669403076, "learning_rate": 5e-05, "loss": 0.8433, "mean_token_accuracy": 0.8579386472702026, "num_tokens": 125947661.0, "step": 3316 }, { "epoch": 0.421956494084722, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.383472442626953, "learning_rate": 5e-05, "loss": 0.8141, "mean_token_accuracy": 0.8586737513542175, "num_tokens": 125979483.0, "step": 3317 }, { "epoch": 0.42208370436331255, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.389191150665283, "learning_rate": 5e-05, "loss": 0.7815, "mean_token_accuracy": 0.8671047687530518, "num_tokens": 126016907.0, "step": 3318 }, { "epoch": 0.4222109146419031, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.271936416625977, "learning_rate": 5e-05, "loss": 0.8057, "mean_token_accuracy": 0.851923406124115, "num_tokens": 126059475.0, "step": 3319 }, { "epoch": 0.42233812492049355, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.240900039672852, "learning_rate": 5e-05, "loss": 0.7327, "mean_token_accuracy": 0.8723596334457397, "num_tokens": 126097373.0, "step": 3320 }, { "epoch": 0.4224653351990841, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.348594665527344, "learning_rate": 5e-05, "loss": 0.7596, "mean_token_accuracy": 0.8679298758506775, "num_tokens": 126138884.0, "step": 3321 }, { "epoch": 0.4225925454776746, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.078033924102783, "learning_rate": 5e-05, "loss": 0.7563, "mean_token_accuracy": 0.8631031513214111, "num_tokens": 126178453.0, "step": 3322 }, { "epoch": 0.4227197557562651, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.443143844604492, "learning_rate": 5e-05, "loss": 0.804, "mean_token_accuracy": 0.8518012762069702, "num_tokens": 126208979.0, "step": 3323 }, { "epoch": 0.4228469660348556, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.957915782928467, "learning_rate": 5e-05, "loss": 0.7356, "mean_token_accuracy": 0.8598672747612, "num_tokens": 126252195.0, "step": 3324 }, { "epoch": 0.42297417631344614, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.494232654571533, "learning_rate": 5e-05, "loss": 0.761, "mean_token_accuracy": 0.861786961555481, "num_tokens": 126287526.0, "step": 3325 }, { "epoch": 0.4231013865920366, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.891081809997559, "learning_rate": 5e-05, "loss": 0.7072, "mean_token_accuracy": 0.8681331872940063, "num_tokens": 126325868.0, "step": 3326 }, { "epoch": 0.42322859687062714, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.537511825561523, "learning_rate": 5e-05, "loss": 0.7714, "mean_token_accuracy": 0.8603545427322388, "num_tokens": 126364613.0, "step": 3327 }, { "epoch": 0.42335580714921767, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.826725482940674, "learning_rate": 5e-05, "loss": 0.7027, "mean_token_accuracy": 0.8669956922531128, "num_tokens": 126404373.0, "step": 3328 }, { "epoch": 0.42348301742780814, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.623260021209717, "learning_rate": 5e-05, "loss": 0.8512, "mean_token_accuracy": 0.8377307653427124, "num_tokens": 126444557.0, "step": 3329 }, { "epoch": 0.42361022770639867, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.720698356628418, "learning_rate": 5e-05, "loss": 0.7334, "mean_token_accuracy": 0.8523575067520142, "num_tokens": 126479657.0, "step": 3330 }, { "epoch": 0.4237374379849892, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.7458882331848145, "learning_rate": 5e-05, "loss": 0.8273, "mean_token_accuracy": 0.8490058183670044, "num_tokens": 126518004.0, "step": 3331 }, { "epoch": 0.4238646482635797, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.594187259674072, "learning_rate": 5e-05, "loss": 0.761, "mean_token_accuracy": 0.849474310874939, "num_tokens": 126557854.0, "step": 3332 }, { "epoch": 0.4239918585421702, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.9172258377075195, "learning_rate": 5e-05, "loss": 0.8506, "mean_token_accuracy": 0.8527617454528809, "num_tokens": 126593657.0, "step": 3333 }, { "epoch": 0.4241190688207607, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.444789409637451, "learning_rate": 5e-05, "loss": 0.6567, "mean_token_accuracy": 0.8791726231575012, "num_tokens": 126635913.0, "step": 3334 }, { "epoch": 0.42424627909935125, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.130714416503906, "learning_rate": 5e-05, "loss": 0.876, "mean_token_accuracy": 0.8407767415046692, "num_tokens": 126668703.0, "step": 3335 }, { "epoch": 0.4243734893779417, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.364487171173096, "learning_rate": 5e-05, "loss": 0.6717, "mean_token_accuracy": 0.8700348138809204, "num_tokens": 126707952.0, "step": 3336 }, { "epoch": 0.42450069965653225, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.21574592590332, "learning_rate": 5e-05, "loss": 0.8542, "mean_token_accuracy": 0.8565526008605957, "num_tokens": 126748756.0, "step": 3337 }, { "epoch": 0.4246279099351228, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.505494117736816, "learning_rate": 5e-05, "loss": 0.7043, "mean_token_accuracy": 0.868147075176239, "num_tokens": 126784780.0, "step": 3338 }, { "epoch": 0.42475512021371326, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.987168788909912, "learning_rate": 5e-05, "loss": 0.8288, "mean_token_accuracy": 0.8604748249053955, "num_tokens": 126827371.0, "step": 3339 }, { "epoch": 0.4248823304923038, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.937496185302734, "learning_rate": 5e-05, "loss": 0.7832, "mean_token_accuracy": 0.8508113622665405, "num_tokens": 126862357.0, "step": 3340 }, { "epoch": 0.4250095407708943, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.366767883300781, "learning_rate": 5e-05, "loss": 0.7301, "mean_token_accuracy": 0.8762396574020386, "num_tokens": 126902433.0, "step": 3341 }, { "epoch": 0.4251367510494848, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.535669803619385, "learning_rate": 5e-05, "loss": 0.8369, "mean_token_accuracy": 0.8494781851768494, "num_tokens": 126946958.0, "step": 3342 }, { "epoch": 0.4252639613280753, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.604380130767822, "learning_rate": 5e-05, "loss": 0.6776, "mean_token_accuracy": 0.8683127164840698, "num_tokens": 126980267.0, "step": 3343 }, { "epoch": 0.42539117160666584, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.19855785369873, "learning_rate": 5e-05, "loss": 0.7887, "mean_token_accuracy": 0.8752750754356384, "num_tokens": 127018215.0, "step": 3344 }, { "epoch": 0.4255183818852563, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.00023651123046875, "grad_norm": 5.978799819946289, "learning_rate": 5e-05, "loss": 0.7214, "mean_token_accuracy": 0.8480774164199829, "num_tokens": 127055844.0, "step": 3345 }, { "epoch": 0.42564559216384684, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.949459075927734, "learning_rate": 5e-05, "loss": 0.8994, "mean_token_accuracy": 0.8623753786087036, "num_tokens": 127099373.0, "step": 3346 }, { "epoch": 0.42577280244243737, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 5.648760795593262, "learning_rate": 5e-05, "loss": 0.6879, "mean_token_accuracy": 0.8603485226631165, "num_tokens": 127143755.0, "step": 3347 }, { "epoch": 0.42590001272102784, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.611945152282715, "learning_rate": 5e-05, "loss": 1.002, "mean_token_accuracy": 0.8658457398414612, "num_tokens": 127180543.0, "step": 3348 }, { "epoch": 0.42602722299961837, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.397355079650879, "learning_rate": 5e-05, "loss": 0.8338, "mean_token_accuracy": 0.8493280410766602, "num_tokens": 127225203.0, "step": 3349 }, { "epoch": 0.4261544332782089, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.294973373413086, "learning_rate": 5e-05, "loss": 1.0953, "mean_token_accuracy": 0.8402979969978333, "num_tokens": 127260949.0, "step": 3350 }, { "epoch": 0.4262816435567994, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 7.880377769470215, "learning_rate": 5e-05, "loss": 0.9574, "mean_token_accuracy": 0.8717951774597168, "num_tokens": 127297487.0, "step": 3351 }, { "epoch": 0.4264088538353899, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.544856071472168, "learning_rate": 5e-05, "loss": 0.8995, "mean_token_accuracy": 0.842180609703064, "num_tokens": 127342612.0, "step": 3352 }, { "epoch": 0.42653606411398043, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.033527374267578, "learning_rate": 5e-05, "loss": 0.9963, "mean_token_accuracy": 0.8654251098632812, "num_tokens": 127383915.0, "step": 3353 }, { "epoch": 0.4266632743925709, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.191440105438232, "learning_rate": 5e-05, "loss": 0.7856, "mean_token_accuracy": 0.8513826131820679, "num_tokens": 127421564.0, "step": 3354 }, { "epoch": 0.42679048467116143, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.013998985290527, "learning_rate": 5e-05, "loss": 1.1304, "mean_token_accuracy": 0.8416741490364075, "num_tokens": 127463750.0, "step": 3355 }, { "epoch": 0.42691769494975196, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.014705181121826, "learning_rate": 5e-05, "loss": 0.9562, "mean_token_accuracy": 0.8494272232055664, "num_tokens": 127502793.0, "step": 3356 }, { "epoch": 0.42704490522834243, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.516029357910156, "learning_rate": 5e-05, "loss": 1.0515, "mean_token_accuracy": 0.8555237054824829, "num_tokens": 127542650.0, "step": 3357 }, { "epoch": 0.42717211550693296, "ewc_loss": 0.625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.1790132522583, "learning_rate": 5e-05, "loss": 1.0268, "mean_token_accuracy": 0.8672820329666138, "num_tokens": 127584753.0, "step": 3358 }, { "epoch": 0.4272993257855235, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.059017181396484, "learning_rate": 5e-05, "loss": 0.9157, "mean_token_accuracy": 0.8490033149719238, "num_tokens": 127624427.0, "step": 3359 }, { "epoch": 0.42742653606411396, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.552230834960938, "learning_rate": 5e-05, "loss": 0.985, "mean_token_accuracy": 0.8649208545684814, "num_tokens": 127665636.0, "step": 3360 }, { "epoch": 0.4275537463427045, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.333642482757568, "learning_rate": 5e-05, "loss": 0.8892, "mean_token_accuracy": 0.8506932258605957, "num_tokens": 127706066.0, "step": 3361 }, { "epoch": 0.427680956621295, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.36350154876709, "learning_rate": 5e-05, "loss": 0.907, "mean_token_accuracy": 0.8684098720550537, "num_tokens": 127741680.0, "step": 3362 }, { "epoch": 0.4278081668998855, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.458339214324951, "learning_rate": 5e-05, "loss": 0.9158, "mean_token_accuracy": 0.8518273830413818, "num_tokens": 127781147.0, "step": 3363 }, { "epoch": 0.427935377178476, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.566032886505127, "learning_rate": 5e-05, "loss": 0.8899, "mean_token_accuracy": 0.8378950357437134, "num_tokens": 127823787.0, "step": 3364 }, { "epoch": 0.42806258745706655, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.862081527709961, "learning_rate": 5e-05, "loss": 0.967, "mean_token_accuracy": 0.8343838453292847, "num_tokens": 127862239.0, "step": 3365 }, { "epoch": 0.428189797735657, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.860846519470215, "learning_rate": 5e-05, "loss": 0.8177, "mean_token_accuracy": 0.8443247079849243, "num_tokens": 127904376.0, "step": 3366 }, { "epoch": 0.42831700801424755, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.278847694396973, "learning_rate": 5e-05, "loss": 0.9473, "mean_token_accuracy": 0.844368577003479, "num_tokens": 127944132.0, "step": 3367 }, { "epoch": 0.4284442182928381, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.50808572769165, "learning_rate": 5e-05, "loss": 0.7827, "mean_token_accuracy": 0.8464586734771729, "num_tokens": 127983512.0, "step": 3368 }, { "epoch": 0.42857142857142855, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.628358840942383, "learning_rate": 5e-05, "loss": 0.8927, "mean_token_accuracy": 0.8653218746185303, "num_tokens": 128018864.0, "step": 3369 }, { "epoch": 0.4286986388500191, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.471556663513184, "learning_rate": 5e-05, "loss": 0.7193, "mean_token_accuracy": 0.8724861145019531, "num_tokens": 128056586.0, "step": 3370 }, { "epoch": 0.4288258491286096, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.756040573120117, "learning_rate": 5e-05, "loss": 0.9617, "mean_token_accuracy": 0.8514264225959778, "num_tokens": 128091131.0, "step": 3371 }, { "epoch": 0.4289530594072001, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.666175842285156, "learning_rate": 5e-05, "loss": 0.8381, "mean_token_accuracy": 0.8484614491462708, "num_tokens": 128125455.0, "step": 3372 }, { "epoch": 0.4290802696857906, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.581972122192383, "learning_rate": 5e-05, "loss": 0.9521, "mean_token_accuracy": 0.8532638549804688, "num_tokens": 128164357.0, "step": 3373 }, { "epoch": 0.42920747996438113, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 6.888472557067871, "learning_rate": 5e-05, "loss": 0.8676, "mean_token_accuracy": 0.8509615659713745, "num_tokens": 128203590.0, "step": 3374 }, { "epoch": 0.4293346902429716, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.364697456359863, "learning_rate": 5e-05, "loss": 0.935, "mean_token_accuracy": 0.8521741628646851, "num_tokens": 128238819.0, "step": 3375 }, { "epoch": 0.42946190052156213, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 6.896366119384766, "learning_rate": 5e-05, "loss": 0.7863, "mean_token_accuracy": 0.8723065853118896, "num_tokens": 128273879.0, "step": 3376 }, { "epoch": 0.42958911080015266, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.343417167663574, "learning_rate": 5e-05, "loss": 0.9249, "mean_token_accuracy": 0.8590875864028931, "num_tokens": 128309688.0, "step": 3377 }, { "epoch": 0.42971632107874314, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.781701564788818, "learning_rate": 5e-05, "loss": 0.8014, "mean_token_accuracy": 0.8587161302566528, "num_tokens": 128341515.0, "step": 3378 }, { "epoch": 0.42984353135733366, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.426166534423828, "learning_rate": 5e-05, "loss": 0.9303, "mean_token_accuracy": 0.8634415864944458, "num_tokens": 128382676.0, "step": 3379 }, { "epoch": 0.4299707416359242, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.8057451248168945, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.8746892809867859, "num_tokens": 128417512.0, "step": 3380 }, { "epoch": 0.4300979519145147, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.3603515625, "learning_rate": 5e-05, "loss": 0.9075, "mean_token_accuracy": 0.862504243850708, "num_tokens": 128457695.0, "step": 3381 }, { "epoch": 0.4302251621931052, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.007821083068848, "learning_rate": 5e-05, "loss": 0.8331, "mean_token_accuracy": 0.8573801517486572, "num_tokens": 128487459.0, "step": 3382 }, { "epoch": 0.4303523724716957, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.051734924316406, "learning_rate": 5e-05, "loss": 0.9655, "mean_token_accuracy": 0.8340378999710083, "num_tokens": 128526823.0, "step": 3383 }, { "epoch": 0.43047958275028625, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.232369422912598, "learning_rate": 5e-05, "loss": 0.8531, "mean_token_accuracy": 0.8546708822250366, "num_tokens": 128566841.0, "step": 3384 }, { "epoch": 0.4306067930288767, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.633156776428223, "learning_rate": 5e-05, "loss": 0.8786, "mean_token_accuracy": 0.8472722768783569, "num_tokens": 128608147.0, "step": 3385 }, { "epoch": 0.43073400330746725, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.362968921661377, "learning_rate": 5e-05, "loss": 0.8232, "mean_token_accuracy": 0.8597062230110168, "num_tokens": 128648277.0, "step": 3386 }, { "epoch": 0.4308612135860578, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.360780715942383, "learning_rate": 5e-05, "loss": 0.8719, "mean_token_accuracy": 0.8389937877655029, "num_tokens": 128682547.0, "step": 3387 }, { "epoch": 0.43098842386464825, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.329877853393555, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.8610913157463074, "num_tokens": 128715103.0, "step": 3388 }, { "epoch": 0.4311156341432388, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.326804161071777, "learning_rate": 5e-05, "loss": 0.7994, "mean_token_accuracy": 0.8523539304733276, "num_tokens": 128748924.0, "step": 3389 }, { "epoch": 0.4312428444218293, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.1902971267700195, "learning_rate": 5e-05, "loss": 0.7775, "mean_token_accuracy": 0.8516222238540649, "num_tokens": 128786879.0, "step": 3390 }, { "epoch": 0.4313700547004198, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.39999532699585, "learning_rate": 5e-05, "loss": 0.7861, "mean_token_accuracy": 0.8547818660736084, "num_tokens": 128822121.0, "step": 3391 }, { "epoch": 0.4314972649790103, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.064158916473389, "learning_rate": 5e-05, "loss": 0.8212, "mean_token_accuracy": 0.8366211652755737, "num_tokens": 128858594.0, "step": 3392 }, { "epoch": 0.43162447525760084, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.434816837310791, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.855438232421875, "num_tokens": 128896147.0, "step": 3393 }, { "epoch": 0.4317516855361913, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.978918552398682, "learning_rate": 5e-05, "loss": 0.7837, "mean_token_accuracy": 0.8438425064086914, "num_tokens": 128930055.0, "step": 3394 }, { "epoch": 0.43187889581478184, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.500126361846924, "learning_rate": 5e-05, "loss": 0.7847, "mean_token_accuracy": 0.8548708558082581, "num_tokens": 128968577.0, "step": 3395 }, { "epoch": 0.43200610609337237, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.879036903381348, "learning_rate": 5e-05, "loss": 0.7563, "mean_token_accuracy": 0.849402129650116, "num_tokens": 129007074.0, "step": 3396 }, { "epoch": 0.43213331637196284, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.615166664123535, "learning_rate": 5e-05, "loss": 0.8115, "mean_token_accuracy": 0.8484058976173401, "num_tokens": 129044722.0, "step": 3397 }, { "epoch": 0.43226052665055337, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.745434284210205, "learning_rate": 5e-05, "loss": 0.722, "mean_token_accuracy": 0.8582692742347717, "num_tokens": 129081567.0, "step": 3398 }, { "epoch": 0.4323877369291439, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.760879039764404, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.8520885705947876, "num_tokens": 129119354.0, "step": 3399 }, { "epoch": 0.43251494720773437, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.578592777252197, "learning_rate": 5e-05, "loss": 0.6926, "mean_token_accuracy": 0.8675022125244141, "num_tokens": 129157871.0, "step": 3400 }, { "epoch": 0.4326421574863249, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.020007133483887, "learning_rate": 5e-05, "loss": 0.7933, "mean_token_accuracy": 0.8638038635253906, "num_tokens": 129200164.0, "step": 3401 }, { "epoch": 0.4327693677649154, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.357840061187744, "learning_rate": 5e-05, "loss": 0.7286, "mean_token_accuracy": 0.8529905676841736, "num_tokens": 129239382.0, "step": 3402 }, { "epoch": 0.4328965780435059, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.306020736694336, "learning_rate": 5e-05, "loss": 0.8722, "mean_token_accuracy": 0.8498491644859314, "num_tokens": 129273201.0, "step": 3403 }, { "epoch": 0.4330237883220964, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.272995948791504, "learning_rate": 5e-05, "loss": 0.6982, "mean_token_accuracy": 0.8668699264526367, "num_tokens": 129309636.0, "step": 3404 }, { "epoch": 0.43315099860068695, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.388360023498535, "learning_rate": 5e-05, "loss": 0.8484, "mean_token_accuracy": 0.8644150495529175, "num_tokens": 129345803.0, "step": 3405 }, { "epoch": 0.4332782088792774, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.522232532501221, "learning_rate": 5e-05, "loss": 0.7225, "mean_token_accuracy": 0.8608255386352539, "num_tokens": 129384955.0, "step": 3406 }, { "epoch": 0.43340541915786795, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.040096282958984, "learning_rate": 5e-05, "loss": 0.8161, "mean_token_accuracy": 0.8725026845932007, "num_tokens": 129425129.0, "step": 3407 }, { "epoch": 0.4335326294364585, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.070738792419434, "learning_rate": 5e-05, "loss": 0.7924, "mean_token_accuracy": 0.8536010980606079, "num_tokens": 129465931.0, "step": 3408 }, { "epoch": 0.43365983971504896, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.274395942687988, "learning_rate": 5e-05, "loss": 0.7443, "mean_token_accuracy": 0.8644853830337524, "num_tokens": 129497296.0, "step": 3409 }, { "epoch": 0.4337870499936395, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.733860969543457, "learning_rate": 5e-05, "loss": 0.8153, "mean_token_accuracy": 0.8591822385787964, "num_tokens": 129539037.0, "step": 3410 }, { "epoch": 0.43391426027223, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.4748454093933105, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.850512683391571, "num_tokens": 129581036.0, "step": 3411 }, { "epoch": 0.4340414705508205, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.42491626739502, "learning_rate": 5e-05, "loss": 0.891, "mean_token_accuracy": 0.8504502177238464, "num_tokens": 129617515.0, "step": 3412 }, { "epoch": 0.434168680829411, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.00024127960205078125, "grad_norm": 5.912353038787842, "learning_rate": 5e-05, "loss": 0.6307, "mean_token_accuracy": 0.8768675327301025, "num_tokens": 129659348.0, "step": 3413 }, { "epoch": 0.43429589110800154, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.13372802734375, "learning_rate": 5e-05, "loss": 1.0187, "mean_token_accuracy": 0.8348110318183899, "num_tokens": 129695000.0, "step": 3414 }, { "epoch": 0.434423101386592, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 5.849857330322266, "learning_rate": 5e-05, "loss": 0.7367, "mean_token_accuracy": 0.8504788279533386, "num_tokens": 129734125.0, "step": 3415 }, { "epoch": 0.43455031166518254, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.473118782043457, "learning_rate": 5e-05, "loss": 1.0328, "mean_token_accuracy": 0.8556545972824097, "num_tokens": 129774799.0, "step": 3416 }, { "epoch": 0.43467752194377307, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 6.876956939697266, "learning_rate": 5e-05, "loss": 0.8524, "mean_token_accuracy": 0.8635830879211426, "num_tokens": 129814390.0, "step": 3417 }, { "epoch": 0.43480473222236354, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.669900894165039, "learning_rate": 5e-05, "loss": 0.9629, "mean_token_accuracy": 0.8643411993980408, "num_tokens": 129848172.0, "step": 3418 }, { "epoch": 0.43493194250095407, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.049135208129883, "learning_rate": 5e-05, "loss": 0.9396, "mean_token_accuracy": 0.8707231879234314, "num_tokens": 129878728.0, "step": 3419 }, { "epoch": 0.4350591527795446, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.0443902015686035, "learning_rate": 5e-05, "loss": 0.8143, "mean_token_accuracy": 0.8545234799385071, "num_tokens": 129916843.0, "step": 3420 }, { "epoch": 0.4351863630581351, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.877129554748535, "learning_rate": 5e-05, "loss": 1.022, "mean_token_accuracy": 0.8439155220985413, "num_tokens": 129954614.0, "step": 3421 }, { "epoch": 0.4353135733367256, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.4231276512146, "learning_rate": 5e-05, "loss": 0.8486, "mean_token_accuracy": 0.8396631479263306, "num_tokens": 129993084.0, "step": 3422 }, { "epoch": 0.43544078361531613, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.897770881652832, "learning_rate": 5e-05, "loss": 0.96, "mean_token_accuracy": 0.862838625907898, "num_tokens": 130028313.0, "step": 3423 }, { "epoch": 0.4355679938939066, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.075862884521484, "learning_rate": 5e-05, "loss": 0.8747, "mean_token_accuracy": 0.854830265045166, "num_tokens": 130066722.0, "step": 3424 }, { "epoch": 0.43569520417249713, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.875276565551758, "learning_rate": 5e-05, "loss": 0.8888, "mean_token_accuracy": 0.8528250455856323, "num_tokens": 130108628.0, "step": 3425 }, { "epoch": 0.43582241445108766, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.852547645568848, "learning_rate": 5e-05, "loss": 0.8784, "mean_token_accuracy": 0.8689772486686707, "num_tokens": 130145770.0, "step": 3426 }, { "epoch": 0.43594962472967813, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.923653602600098, "learning_rate": 5e-05, "loss": 0.7769, "mean_token_accuracy": 0.8554098010063171, "num_tokens": 130191539.0, "step": 3427 }, { "epoch": 0.43607683500826866, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.208915710449219, "learning_rate": 5e-05, "loss": 0.8845, "mean_token_accuracy": 0.8689638376235962, "num_tokens": 130229358.0, "step": 3428 }, { "epoch": 0.4362040452868592, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.7411603927612305, "learning_rate": 5e-05, "loss": 0.7723, "mean_token_accuracy": 0.8567516803741455, "num_tokens": 130272173.0, "step": 3429 }, { "epoch": 0.4363312555654497, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.112120628356934, "learning_rate": 5e-05, "loss": 0.8949, "mean_token_accuracy": 0.8534290194511414, "num_tokens": 130313135.0, "step": 3430 }, { "epoch": 0.4364584658440402, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.906225204467773, "learning_rate": 5e-05, "loss": 0.7574, "mean_token_accuracy": 0.8654851317405701, "num_tokens": 130350595.0, "step": 3431 }, { "epoch": 0.4365856761226307, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.825760364532471, "learning_rate": 5e-05, "loss": 0.827, "mean_token_accuracy": 0.8625085353851318, "num_tokens": 130390194.0, "step": 3432 }, { "epoch": 0.43671288640122125, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.017161846160889, "learning_rate": 5e-05, "loss": 0.7933, "mean_token_accuracy": 0.8572891354560852, "num_tokens": 130424768.0, "step": 3433 }, { "epoch": 0.4368400966798117, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.669688701629639, "learning_rate": 5e-05, "loss": 0.8731, "mean_token_accuracy": 0.8418775200843811, "num_tokens": 130469457.0, "step": 3434 }, { "epoch": 0.43696730695840225, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.93053674697876, "learning_rate": 5e-05, "loss": 0.7237, "mean_token_accuracy": 0.8719500303268433, "num_tokens": 130506250.0, "step": 3435 }, { "epoch": 0.4370945172369928, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.744868278503418, "learning_rate": 5e-05, "loss": 0.8186, "mean_token_accuracy": 0.8572295904159546, "num_tokens": 130538605.0, "step": 3436 }, { "epoch": 0.43722172751558325, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.704168796539307, "learning_rate": 5e-05, "loss": 0.6989, "mean_token_accuracy": 0.8732671141624451, "num_tokens": 130577172.0, "step": 3437 }, { "epoch": 0.4373489377941738, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.9695281982421875, "learning_rate": 5e-05, "loss": 0.8466, "mean_token_accuracy": 0.8516202569007874, "num_tokens": 130613067.0, "step": 3438 }, { "epoch": 0.4374761480727643, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.471449375152588, "learning_rate": 5e-05, "loss": 0.6753, "mean_token_accuracy": 0.8724100589752197, "num_tokens": 130649176.0, "step": 3439 }, { "epoch": 0.4376033583513548, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.227287292480469, "learning_rate": 5e-05, "loss": 0.8557, "mean_token_accuracy": 0.8592248558998108, "num_tokens": 130685904.0, "step": 3440 }, { "epoch": 0.4377305686299453, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.2895684242248535, "learning_rate": 5e-05, "loss": 0.6841, "mean_token_accuracy": 0.870272159576416, "num_tokens": 130727502.0, "step": 3441 }, { "epoch": 0.43785777890853583, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.492623329162598, "learning_rate": 5e-05, "loss": 0.9167, "mean_token_accuracy": 0.8489649891853333, "num_tokens": 130765038.0, "step": 3442 }, { "epoch": 0.4379849891871263, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.239773750305176, "learning_rate": 5e-05, "loss": 0.7069, "mean_token_accuracy": 0.8639562129974365, "num_tokens": 130810349.0, "step": 3443 }, { "epoch": 0.43811219946571683, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.662084579467773, "learning_rate": 5e-05, "loss": 0.9567, "mean_token_accuracy": 0.8501906394958496, "num_tokens": 130843383.0, "step": 3444 }, { "epoch": 0.43823940974430736, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.4816789627075195, "learning_rate": 5e-05, "loss": 0.7867, "mean_token_accuracy": 0.8536696434020996, "num_tokens": 130880399.0, "step": 3445 }, { "epoch": 0.43836662002289783, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.42853832244873, "learning_rate": 5e-05, "loss": 0.9916, "mean_token_accuracy": 0.8429874181747437, "num_tokens": 130915537.0, "step": 3446 }, { "epoch": 0.43849383030148836, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.004215717315674, "learning_rate": 5e-05, "loss": 0.8129, "mean_token_accuracy": 0.8668828010559082, "num_tokens": 130948560.0, "step": 3447 }, { "epoch": 0.4386210405800789, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.775855541229248, "learning_rate": 5e-05, "loss": 0.8423, "mean_token_accuracy": 0.8651690483093262, "num_tokens": 130982405.0, "step": 3448 }, { "epoch": 0.43874825085866936, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.477606296539307, "learning_rate": 5e-05, "loss": 0.8465, "mean_token_accuracy": 0.8618958592414856, "num_tokens": 131022673.0, "step": 3449 }, { "epoch": 0.4388754611372599, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.129321098327637, "learning_rate": 5e-05, "loss": 0.7867, "mean_token_accuracy": 0.8541295528411865, "num_tokens": 131054647.0, "step": 3450 }, { "epoch": 0.4390026714158504, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.745196342468262, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.8626744151115417, "num_tokens": 131091616.0, "step": 3451 }, { "epoch": 0.4391298816944409, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.816274642944336, "learning_rate": 5e-05, "loss": 0.8297, "mean_token_accuracy": 0.8331339955329895, "num_tokens": 131129663.0, "step": 3452 }, { "epoch": 0.4392570919730314, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.792332172393799, "learning_rate": 5e-05, "loss": 0.8187, "mean_token_accuracy": 0.8621039986610413, "num_tokens": 131170858.0, "step": 3453 }, { "epoch": 0.43938430225162195, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.7636399269104, "learning_rate": 5e-05, "loss": 0.7403, "mean_token_accuracy": 0.8545798659324646, "num_tokens": 131210892.0, "step": 3454 }, { "epoch": 0.4395115125302124, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.817505836486816, "learning_rate": 5e-05, "loss": 0.8417, "mean_token_accuracy": 0.8578086495399475, "num_tokens": 131246616.0, "step": 3455 }, { "epoch": 0.43963872280880295, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.700067520141602, "learning_rate": 5e-05, "loss": 0.7299, "mean_token_accuracy": 0.8588396310806274, "num_tokens": 131286515.0, "step": 3456 }, { "epoch": 0.4397659330873935, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.907496452331543, "learning_rate": 5e-05, "loss": 0.8381, "mean_token_accuracy": 0.8557456135749817, "num_tokens": 131321715.0, "step": 3457 }, { "epoch": 0.43989314336598395, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.574763774871826, "learning_rate": 5e-05, "loss": 0.7822, "mean_token_accuracy": 0.8471014499664307, "num_tokens": 131359235.0, "step": 3458 }, { "epoch": 0.4400203536445745, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.066675186157227, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8747838735580444, "num_tokens": 131390551.0, "step": 3459 }, { "epoch": 0.440147563923165, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.496570587158203, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8600068092346191, "num_tokens": 131424107.0, "step": 3460 }, { "epoch": 0.4402747742017555, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.156594276428223, "learning_rate": 5e-05, "loss": 0.8336, "mean_token_accuracy": 0.8648525476455688, "num_tokens": 131459797.0, "step": 3461 }, { "epoch": 0.440401984480346, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.532248020172119, "learning_rate": 5e-05, "loss": 0.7469, "mean_token_accuracy": 0.8558606505393982, "num_tokens": 131496877.0, "step": 3462 }, { "epoch": 0.44052919475893654, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.13512897491455, "learning_rate": 5e-05, "loss": 0.8438, "mean_token_accuracy": 0.8653415441513062, "num_tokens": 131528958.0, "step": 3463 }, { "epoch": 0.440656405037527, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.691628456115723, "learning_rate": 5e-05, "loss": 0.8257, "mean_token_accuracy": 0.8384063243865967, "num_tokens": 131572157.0, "step": 3464 }, { "epoch": 0.44078361531611754, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.921278476715088, "learning_rate": 5e-05, "loss": 0.8758, "mean_token_accuracy": 0.8516625165939331, "num_tokens": 131613855.0, "step": 3465 }, { "epoch": 0.44091082559470807, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.903542518615723, "learning_rate": 5e-05, "loss": 0.7823, "mean_token_accuracy": 0.855009138584137, "num_tokens": 131654056.0, "step": 3466 }, { "epoch": 0.44103803587329854, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.642724990844727, "learning_rate": 5e-05, "loss": 0.8878, "mean_token_accuracy": 0.8354178667068481, "num_tokens": 131694865.0, "step": 3467 }, { "epoch": 0.44116524615188907, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.089748859405518, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.8567154407501221, "num_tokens": 131732883.0, "step": 3468 }, { "epoch": 0.4412924564304796, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.504228591918945, "learning_rate": 5e-05, "loss": 0.8228, "mean_token_accuracy": 0.8507153987884521, "num_tokens": 131770384.0, "step": 3469 }, { "epoch": 0.44141966670907007, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.02843713760376, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8527287244796753, "num_tokens": 131801611.0, "step": 3470 }, { "epoch": 0.4415468769876606, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.555069923400879, "learning_rate": 5e-05, "loss": 0.9085, "mean_token_accuracy": 0.8302537202835083, "num_tokens": 131842522.0, "step": 3471 }, { "epoch": 0.4416740872662511, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.877920627593994, "learning_rate": 5e-05, "loss": 0.818, "mean_token_accuracy": 0.8408224582672119, "num_tokens": 131880041.0, "step": 3472 }, { "epoch": 0.4418012975448416, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.657304286956787, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8605809211730957, "num_tokens": 131917682.0, "step": 3473 }, { "epoch": 0.4419285078234321, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.75466775894165, "learning_rate": 5e-05, "loss": 0.7501, "mean_token_accuracy": 0.8529047966003418, "num_tokens": 131958149.0, "step": 3474 }, { "epoch": 0.44205571810202265, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.754539966583252, "learning_rate": 5e-05, "loss": 0.8469, "mean_token_accuracy": 0.8452823162078857, "num_tokens": 131996648.0, "step": 3475 }, { "epoch": 0.4421829283806131, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.673099994659424, "learning_rate": 5e-05, "loss": 0.7323, "mean_token_accuracy": 0.8596101999282837, "num_tokens": 132036524.0, "step": 3476 }, { "epoch": 0.44231013865920366, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.846344947814941, "learning_rate": 5e-05, "loss": 0.8151, "mean_token_accuracy": 0.8596391081809998, "num_tokens": 132077474.0, "step": 3477 }, { "epoch": 0.4424373489377942, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.60886287689209, "learning_rate": 5e-05, "loss": 0.7742, "mean_token_accuracy": 0.8422245979309082, "num_tokens": 132117153.0, "step": 3478 }, { "epoch": 0.44256455921638466, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.899768352508545, "learning_rate": 5e-05, "loss": 0.7781, "mean_token_accuracy": 0.870576024055481, "num_tokens": 132156832.0, "step": 3479 }, { "epoch": 0.4426917694949752, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.63870096206665, "learning_rate": 5e-05, "loss": 0.7427, "mean_token_accuracy": 0.8556958436965942, "num_tokens": 132189245.0, "step": 3480 }, { "epoch": 0.4428189797735657, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.816507339477539, "learning_rate": 5e-05, "loss": 0.8234, "mean_token_accuracy": 0.8549586534500122, "num_tokens": 132230572.0, "step": 3481 }, { "epoch": 0.44294619005215624, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.799018859863281, "learning_rate": 5e-05, "loss": 0.7376, "mean_token_accuracy": 0.8570986390113831, "num_tokens": 132271850.0, "step": 3482 }, { "epoch": 0.4430734003307467, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.5678582191467285, "learning_rate": 5e-05, "loss": 0.7895, "mean_token_accuracy": 0.8583868145942688, "num_tokens": 132312461.0, "step": 3483 }, { "epoch": 0.44320061060933724, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.062160968780518, "learning_rate": 5e-05, "loss": 0.7412, "mean_token_accuracy": 0.8622422814369202, "num_tokens": 132350628.0, "step": 3484 }, { "epoch": 0.44332782088792777, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.2216477394104, "learning_rate": 5e-05, "loss": 0.781, "mean_token_accuracy": 0.8531932830810547, "num_tokens": 132394040.0, "step": 3485 }, { "epoch": 0.44345503116651824, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.3727874755859375, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.8565860986709595, "num_tokens": 132435727.0, "step": 3486 }, { "epoch": 0.44358224144510877, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.818983554840088, "learning_rate": 5e-05, "loss": 0.7425, "mean_token_accuracy": 0.8527452349662781, "num_tokens": 132475769.0, "step": 3487 }, { "epoch": 0.4437094517236993, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.741859436035156, "learning_rate": 5e-05, "loss": 0.8067, "mean_token_accuracy": 0.8543184995651245, "num_tokens": 132520944.0, "step": 3488 }, { "epoch": 0.4438366620022898, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.408249378204346, "learning_rate": 5e-05, "loss": 0.7015, "mean_token_accuracy": 0.8550024032592773, "num_tokens": 132558138.0, "step": 3489 }, { "epoch": 0.4439638722808803, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.210926055908203, "learning_rate": 5e-05, "loss": 0.8961, "mean_token_accuracy": 0.8301119208335876, "num_tokens": 132594937.0, "step": 3490 }, { "epoch": 0.44409108255947083, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0002288818359375, "grad_norm": 5.986598968505859, "learning_rate": 5e-05, "loss": 0.7414, "mean_token_accuracy": 0.8386512398719788, "num_tokens": 132629168.0, "step": 3491 }, { "epoch": 0.4442182928380613, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.82433795928955, "learning_rate": 5e-05, "loss": 0.9895, "mean_token_accuracy": 0.8285265564918518, "num_tokens": 132673239.0, "step": 3492 }, { "epoch": 0.44434550311665183, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000244140625, "grad_norm": 5.693880081176758, "learning_rate": 5e-05, "loss": 0.6854, "mean_token_accuracy": 0.8637933731079102, "num_tokens": 132707429.0, "step": 3493 }, { "epoch": 0.44447271339524236, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.413261413574219, "learning_rate": 5e-05, "loss": 0.9966, "mean_token_accuracy": 0.8572055101394653, "num_tokens": 132745592.0, "step": 3494 }, { "epoch": 0.44459992367383283, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.201019763946533, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8644487857818604, "num_tokens": 132778260.0, "step": 3495 }, { "epoch": 0.44472713395242336, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.174556732177734, "learning_rate": 5e-05, "loss": 0.9287, "mean_token_accuracy": 0.8856204152107239, "num_tokens": 132813156.0, "step": 3496 }, { "epoch": 0.4448543442310139, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.583978652954102, "learning_rate": 5e-05, "loss": 0.9825, "mean_token_accuracy": 0.8547451496124268, "num_tokens": 132847654.0, "step": 3497 }, { "epoch": 0.44498155450960436, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.4890947341918945, "learning_rate": 5e-05, "loss": 0.818, "mean_token_accuracy": 0.8644393086433411, "num_tokens": 132889258.0, "step": 3498 }, { "epoch": 0.4451087647881949, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.921107292175293, "learning_rate": 5e-05, "loss": 1.0755, "mean_token_accuracy": 0.8365445733070374, "num_tokens": 132919962.0, "step": 3499 }, { "epoch": 0.4452359750667854, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 5.92133092880249, "learning_rate": 5e-05, "loss": 0.7724, "mean_token_accuracy": 0.8484753370285034, "num_tokens": 132957333.0, "step": 3500 }, { "epoch": 0.4453631853453759, "ewc_loss": 0.703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.394692420959473, "learning_rate": 5e-05, "loss": 1.1226, "mean_token_accuracy": 0.8557807207107544, "num_tokens": 132990645.0, "step": 3501 }, { "epoch": 0.4454903956239664, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 6.758787155151367, "learning_rate": 5e-05, "loss": 0.8687, "mean_token_accuracy": 0.8674764633178711, "num_tokens": 133026234.0, "step": 3502 }, { "epoch": 0.44561760590255695, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 10.615530014038086, "learning_rate": 5e-05, "loss": 1.1422, "mean_token_accuracy": 0.8581470251083374, "num_tokens": 133062833.0, "step": 3503 }, { "epoch": 0.4457448161811474, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 8.606169700622559, "learning_rate": 5e-05, "loss": 1.1586, "mean_token_accuracy": 0.8692783713340759, "num_tokens": 133097868.0, "step": 3504 }, { "epoch": 0.44587202645973795, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.23884105682373, "learning_rate": 5e-05, "loss": 0.9745, "mean_token_accuracy": 0.8631105422973633, "num_tokens": 133142062.0, "step": 3505 }, { "epoch": 0.4459992367383285, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 8.72315502166748, "learning_rate": 5e-05, "loss": 1.1058, "mean_token_accuracy": 0.8521683812141418, "num_tokens": 133181415.0, "step": 3506 }, { "epoch": 0.44612644701691895, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.156441688537598, "learning_rate": 5e-05, "loss": 0.9596, "mean_token_accuracy": 0.8577965497970581, "num_tokens": 133217849.0, "step": 3507 }, { "epoch": 0.4462536572955095, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.479633331298828, "learning_rate": 5e-05, "loss": 0.9844, "mean_token_accuracy": 0.8546600341796875, "num_tokens": 133257639.0, "step": 3508 }, { "epoch": 0.4463808675741, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 7.845980644226074, "learning_rate": 5e-05, "loss": 0.9675, "mean_token_accuracy": 0.8512576818466187, "num_tokens": 133299692.0, "step": 3509 }, { "epoch": 0.4465080778526905, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.71578311920166, "learning_rate": 5e-05, "loss": 0.9069, "mean_token_accuracy": 0.8403290510177612, "num_tokens": 133339315.0, "step": 3510 }, { "epoch": 0.446635288131281, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.083700180053711, "learning_rate": 5e-05, "loss": 0.9307, "mean_token_accuracy": 0.8608623743057251, "num_tokens": 133378477.0, "step": 3511 }, { "epoch": 0.44676249840987153, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.947960376739502, "learning_rate": 5e-05, "loss": 0.8744, "mean_token_accuracy": 0.831964373588562, "num_tokens": 133416041.0, "step": 3512 }, { "epoch": 0.446889708688462, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.500789642333984, "learning_rate": 5e-05, "loss": 0.9388, "mean_token_accuracy": 0.8614908456802368, "num_tokens": 133458201.0, "step": 3513 }, { "epoch": 0.44701691896705253, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.606758117675781, "learning_rate": 5e-05, "loss": 0.7968, "mean_token_accuracy": 0.8582507371902466, "num_tokens": 133495645.0, "step": 3514 }, { "epoch": 0.44714412924564306, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.791013717651367, "learning_rate": 5e-05, "loss": 0.9928, "mean_token_accuracy": 0.8514339923858643, "num_tokens": 133535385.0, "step": 3515 }, { "epoch": 0.44727133952423354, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 6.684277057647705, "learning_rate": 5e-05, "loss": 0.815, "mean_token_accuracy": 0.860720157623291, "num_tokens": 133575093.0, "step": 3516 }, { "epoch": 0.44739854980282406, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.809317588806152, "learning_rate": 5e-05, "loss": 0.984, "mean_token_accuracy": 0.8480812311172485, "num_tokens": 133610447.0, "step": 3517 }, { "epoch": 0.4475257600814146, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 6.97287654876709, "learning_rate": 5e-05, "loss": 0.8556, "mean_token_accuracy": 0.8644894957542419, "num_tokens": 133643506.0, "step": 3518 }, { "epoch": 0.44765297036000506, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.484661102294922, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.877424418926239, "num_tokens": 133683081.0, "step": 3519 }, { "epoch": 0.4477801806385956, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.222013473510742, "learning_rate": 5e-05, "loss": 0.9194, "mean_token_accuracy": 0.8469314575195312, "num_tokens": 133721836.0, "step": 3520 }, { "epoch": 0.4479073909171861, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.067200660705566, "learning_rate": 5e-05, "loss": 0.9071, "mean_token_accuracy": 0.8614046573638916, "num_tokens": 133756243.0, "step": 3521 }, { "epoch": 0.4480346011957766, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.184278964996338, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.8449289798736572, "num_tokens": 133807126.0, "step": 3522 }, { "epoch": 0.4481618114743671, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.979942798614502, "learning_rate": 5e-05, "loss": 0.9172, "mean_token_accuracy": 0.8536660671234131, "num_tokens": 133842036.0, "step": 3523 }, { "epoch": 0.44828902175295765, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.9062604904174805, "learning_rate": 5e-05, "loss": 0.8474, "mean_token_accuracy": 0.8414825201034546, "num_tokens": 133881948.0, "step": 3524 }, { "epoch": 0.4484162320315481, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.2194242477417, "learning_rate": 5e-05, "loss": 0.896, "mean_token_accuracy": 0.8639194965362549, "num_tokens": 133917765.0, "step": 3525 }, { "epoch": 0.44854344231013865, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.6553778648376465, "learning_rate": 5e-05, "loss": 0.7708, "mean_token_accuracy": 0.8533878326416016, "num_tokens": 133959199.0, "step": 3526 }, { "epoch": 0.4486706525887292, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.525842666625977, "learning_rate": 5e-05, "loss": 0.9937, "mean_token_accuracy": 0.8450343608856201, "num_tokens": 133991484.0, "step": 3527 }, { "epoch": 0.44879786286731965, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.6564860343933105, "learning_rate": 5e-05, "loss": 0.7572, "mean_token_accuracy": 0.8659825325012207, "num_tokens": 134025817.0, "step": 3528 }, { "epoch": 0.4489250731459102, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.6158447265625, "learning_rate": 5e-05, "loss": 0.9697, "mean_token_accuracy": 0.8568696975708008, "num_tokens": 134063956.0, "step": 3529 }, { "epoch": 0.4490522834245007, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.0408935546875, "learning_rate": 5e-05, "loss": 0.8603, "mean_token_accuracy": 0.8542661666870117, "num_tokens": 134101959.0, "step": 3530 }, { "epoch": 0.44917949370309124, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.177931785583496, "learning_rate": 5e-05, "loss": 0.9388, "mean_token_accuracy": 0.8522626161575317, "num_tokens": 134139961.0, "step": 3531 }, { "epoch": 0.4493067039816817, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.525426387786865, "learning_rate": 5e-05, "loss": 0.9233, "mean_token_accuracy": 0.8498882055282593, "num_tokens": 134180368.0, "step": 3532 }, { "epoch": 0.44943391426027224, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.350916385650635, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.8679995536804199, "num_tokens": 134216842.0, "step": 3533 }, { "epoch": 0.44956112453886277, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.825721740722656, "learning_rate": 5e-05, "loss": 0.8727, "mean_token_accuracy": 0.8608276844024658, "num_tokens": 134254089.0, "step": 3534 }, { "epoch": 0.44968833481745324, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.832843780517578, "learning_rate": 5e-05, "loss": 0.7352, "mean_token_accuracy": 0.8640361428260803, "num_tokens": 134289559.0, "step": 3535 }, { "epoch": 0.44981554509604377, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.882505893707275, "learning_rate": 5e-05, "loss": 0.8709, "mean_token_accuracy": 0.8557935953140259, "num_tokens": 134323029.0, "step": 3536 }, { "epoch": 0.4499427553746343, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.778762340545654, "learning_rate": 5e-05, "loss": 0.7669, "mean_token_accuracy": 0.8548575043678284, "num_tokens": 134362657.0, "step": 3537 }, { "epoch": 0.45006996565322477, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.7928385734558105, "learning_rate": 5e-05, "loss": 0.8141, "mean_token_accuracy": 0.8593826293945312, "num_tokens": 134403555.0, "step": 3538 }, { "epoch": 0.4501971759318153, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.814925193786621, "learning_rate": 5e-05, "loss": 0.7422, "mean_token_accuracy": 0.8610707521438599, "num_tokens": 134438780.0, "step": 3539 }, { "epoch": 0.4503243862104058, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.716811656951904, "learning_rate": 5e-05, "loss": 0.8322, "mean_token_accuracy": 0.85273277759552, "num_tokens": 134477331.0, "step": 3540 }, { "epoch": 0.4504515964889963, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.8546013832092285, "learning_rate": 5e-05, "loss": 0.6995, "mean_token_accuracy": 0.8707319498062134, "num_tokens": 134513898.0, "step": 3541 }, { "epoch": 0.4505788067675868, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.644149303436279, "learning_rate": 5e-05, "loss": 0.8647, "mean_token_accuracy": 0.8453214764595032, "num_tokens": 134548507.0, "step": 3542 }, { "epoch": 0.45070601704617735, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.874543190002441, "learning_rate": 5e-05, "loss": 0.7796, "mean_token_accuracy": 0.850861668586731, "num_tokens": 134590243.0, "step": 3543 }, { "epoch": 0.4508332273247678, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.645157337188721, "learning_rate": 5e-05, "loss": 0.7888, "mean_token_accuracy": 0.8614377975463867, "num_tokens": 134624693.0, "step": 3544 }, { "epoch": 0.45096043760335836, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.802087306976318, "learning_rate": 5e-05, "loss": 0.6861, "mean_token_accuracy": 0.8790807127952576, "num_tokens": 134661365.0, "step": 3545 }, { "epoch": 0.4510876478819489, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.7383270263671875, "learning_rate": 5e-05, "loss": 0.8097, "mean_token_accuracy": 0.8582501411437988, "num_tokens": 134698218.0, "step": 3546 }, { "epoch": 0.45121485816053936, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.677436828613281, "learning_rate": 5e-05, "loss": 0.7553, "mean_token_accuracy": 0.8503841161727905, "num_tokens": 134736011.0, "step": 3547 }, { "epoch": 0.4513420684391299, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.918482303619385, "learning_rate": 5e-05, "loss": 0.8075, "mean_token_accuracy": 0.8585703372955322, "num_tokens": 134774457.0, "step": 3548 }, { "epoch": 0.4514692787177204, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.493758201599121, "learning_rate": 5e-05, "loss": 0.7878, "mean_token_accuracy": 0.8359826803207397, "num_tokens": 134813788.0, "step": 3549 }, { "epoch": 0.4515964889963109, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.154336929321289, "learning_rate": 5e-05, "loss": 0.8758, "mean_token_accuracy": 0.8458186388015747, "num_tokens": 134851916.0, "step": 3550 }, { "epoch": 0.4517236992749014, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.34783935546875, "learning_rate": 5e-05, "loss": 0.6994, "mean_token_accuracy": 0.8633729219436646, "num_tokens": 134892655.0, "step": 3551 }, { "epoch": 0.45185090955349194, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.367478370666504, "learning_rate": 5e-05, "loss": 0.8864, "mean_token_accuracy": 0.8556362390518188, "num_tokens": 134931391.0, "step": 3552 }, { "epoch": 0.4519781198320824, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.3901896476745605, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8512142896652222, "num_tokens": 134966285.0, "step": 3553 }, { "epoch": 0.45210533011067294, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.319151878356934, "learning_rate": 5e-05, "loss": 0.9, "mean_token_accuracy": 0.8576023578643799, "num_tokens": 135003893.0, "step": 3554 }, { "epoch": 0.45223254038926347, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.737720012664795, "learning_rate": 5e-05, "loss": 0.7782, "mean_token_accuracy": 0.8585700392723083, "num_tokens": 135043245.0, "step": 3555 }, { "epoch": 0.45235975066785394, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.841371059417725, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.855115532875061, "num_tokens": 135082867.0, "step": 3556 }, { "epoch": 0.45248696094644447, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.267369270324707, "learning_rate": 5e-05, "loss": 0.7498, "mean_token_accuracy": 0.8794071674346924, "num_tokens": 135121153.0, "step": 3557 }, { "epoch": 0.452614171225035, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.136797904968262, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8524255752563477, "num_tokens": 135154221.0, "step": 3558 }, { "epoch": 0.4527413815036255, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.7395219802856445, "learning_rate": 5e-05, "loss": 0.8085, "mean_token_accuracy": 0.8604958653450012, "num_tokens": 135196065.0, "step": 3559 }, { "epoch": 0.452868591782216, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.621559143066406, "learning_rate": 5e-05, "loss": 0.71, "mean_token_accuracy": 0.8663445711135864, "num_tokens": 135233692.0, "step": 3560 }, { "epoch": 0.45299580206080653, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.053274154663086, "learning_rate": 5e-05, "loss": 0.7824, "mean_token_accuracy": 0.8742059469223022, "num_tokens": 135267538.0, "step": 3561 }, { "epoch": 0.453123012339397, "ewc_loss": 0.296875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.412044525146484, "learning_rate": 5e-05, "loss": 0.6896, "mean_token_accuracy": 0.8678083419799805, "num_tokens": 135303946.0, "step": 3562 }, { "epoch": 0.45325022261798753, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.173189163208008, "learning_rate": 5e-05, "loss": 0.8804, "mean_token_accuracy": 0.8479065299034119, "num_tokens": 135340940.0, "step": 3563 }, { "epoch": 0.45337743289657806, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.4232401847839355, "learning_rate": 5e-05, "loss": 0.7629, "mean_token_accuracy": 0.8493143320083618, "num_tokens": 135377472.0, "step": 3564 }, { "epoch": 0.45350464317516853, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.126689910888672, "learning_rate": 5e-05, "loss": 0.9143, "mean_token_accuracy": 0.8378888368606567, "num_tokens": 135411396.0, "step": 3565 }, { "epoch": 0.45363185345375906, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.602327823638916, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8463742136955261, "num_tokens": 135445323.0, "step": 3566 }, { "epoch": 0.4537590637323496, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.903264999389648, "learning_rate": 5e-05, "loss": 0.8302, "mean_token_accuracy": 0.861854076385498, "num_tokens": 135487364.0, "step": 3567 }, { "epoch": 0.45388627401094006, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.873409748077393, "learning_rate": 5e-05, "loss": 0.7707, "mean_token_accuracy": 0.8566235899925232, "num_tokens": 135529879.0, "step": 3568 }, { "epoch": 0.4540134842895306, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.566425800323486, "learning_rate": 5e-05, "loss": 0.7878, "mean_token_accuracy": 0.8620628118515015, "num_tokens": 135568812.0, "step": 3569 }, { "epoch": 0.4541406945681211, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.14301872253418, "learning_rate": 5e-05, "loss": 0.7973, "mean_token_accuracy": 0.8518998622894287, "num_tokens": 135608096.0, "step": 3570 }, { "epoch": 0.4542679048467116, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.273980140686035, "learning_rate": 5e-05, "loss": 0.8597, "mean_token_accuracy": 0.8321353197097778, "num_tokens": 135645796.0, "step": 3571 }, { "epoch": 0.4543951151253021, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.262478351593018, "learning_rate": 5e-05, "loss": 0.7979, "mean_token_accuracy": 0.8561724424362183, "num_tokens": 135686535.0, "step": 3572 }, { "epoch": 0.45452232540389265, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.180361747741699, "learning_rate": 5e-05, "loss": 0.7661, "mean_token_accuracy": 0.856626033782959, "num_tokens": 135719555.0, "step": 3573 }, { "epoch": 0.4546495356824831, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.186608791351318, "learning_rate": 5e-05, "loss": 0.7116, "mean_token_accuracy": 0.8686187267303467, "num_tokens": 135756498.0, "step": 3574 }, { "epoch": 0.45477674596107365, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.2930989265441895, "learning_rate": 5e-05, "loss": 0.8271, "mean_token_accuracy": 0.8389005661010742, "num_tokens": 135797438.0, "step": 3575 }, { "epoch": 0.4549039562396642, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.966463565826416, "learning_rate": 5e-05, "loss": 0.7583, "mean_token_accuracy": 0.8524215221405029, "num_tokens": 135836085.0, "step": 3576 }, { "epoch": 0.45503116651825465, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.526160717010498, "learning_rate": 5e-05, "loss": 0.7622, "mean_token_accuracy": 0.8611518144607544, "num_tokens": 135872272.0, "step": 3577 }, { "epoch": 0.4551583767968452, "ewc_loss": 0.296875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.676668167114258, "learning_rate": 5e-05, "loss": 0.7888, "mean_token_accuracy": 0.8373700380325317, "num_tokens": 135908667.0, "step": 3578 }, { "epoch": 0.4552855870754357, "ewc_loss": 0.375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.898110866546631, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8667397499084473, "num_tokens": 135939205.0, "step": 3579 }, { "epoch": 0.4554127973540262, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.243120193481445, "learning_rate": 5e-05, "loss": 0.6811, "mean_token_accuracy": 0.8596449494361877, "num_tokens": 135977084.0, "step": 3580 }, { "epoch": 0.4555400076326167, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.484725952148438, "learning_rate": 5e-05, "loss": 0.8753, "mean_token_accuracy": 0.8507468700408936, "num_tokens": 136012712.0, "step": 3581 }, { "epoch": 0.45566721791120723, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0002288818359375, "grad_norm": 5.785286903381348, "learning_rate": 5e-05, "loss": 0.675, "mean_token_accuracy": 0.8621155023574829, "num_tokens": 136050253.0, "step": 3582 }, { "epoch": 0.45579442818979776, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.25768756866455, "learning_rate": 5e-05, "loss": 0.9609, "mean_token_accuracy": 0.8560341596603394, "num_tokens": 136088072.0, "step": 3583 }, { "epoch": 0.45592163846838824, "ewc_loss": 0.296875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 5.773089408874512, "learning_rate": 5e-05, "loss": 0.7495, "mean_token_accuracy": 0.8529893159866333, "num_tokens": 136122434.0, "step": 3584 }, { "epoch": 0.45604884874697876, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.613003730773926, "learning_rate": 5e-05, "loss": 1.0218, "mean_token_accuracy": 0.8640621900558472, "num_tokens": 136164489.0, "step": 3585 }, { "epoch": 0.4561760590255693, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 6.975648880004883, "learning_rate": 5e-05, "loss": 0.9171, "mean_token_accuracy": 0.8547746539115906, "num_tokens": 136198534.0, "step": 3586 }, { "epoch": 0.45630326930415976, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.579538345336914, "learning_rate": 5e-05, "loss": 0.9688, "mean_token_accuracy": 0.8574174642562866, "num_tokens": 136242261.0, "step": 3587 }, { "epoch": 0.4564304795827503, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.37363338470459, "learning_rate": 5e-05, "loss": 1.0588, "mean_token_accuracy": 0.847916841506958, "num_tokens": 136277106.0, "step": 3588 }, { "epoch": 0.4565576898613408, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.493041515350342, "learning_rate": 5e-05, "loss": 0.7379, "mean_token_accuracy": 0.8609445095062256, "num_tokens": 136315872.0, "step": 3589 }, { "epoch": 0.4566849001399313, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.115514755249023, "learning_rate": 5e-05, "loss": 1.0806, "mean_token_accuracy": 0.8585647344589233, "num_tokens": 136359513.0, "step": 3590 }, { "epoch": 0.4568121104185218, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.090760231018066, "learning_rate": 5e-05, "loss": 0.8152, "mean_token_accuracy": 0.8501650094985962, "num_tokens": 136398575.0, "step": 3591 }, { "epoch": 0.45693932069711235, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000865936279296875, "grad_norm": 11.608267784118652, "learning_rate": 5e-05, "loss": 1.2836, "mean_token_accuracy": 0.8688328862190247, "num_tokens": 136439062.0, "step": 3592 }, { "epoch": 0.4570665309757028, "ewc_loss": 0.8515625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000823974609375, "grad_norm": 8.904336929321289, "learning_rate": 5e-05, "loss": 1.2857, "mean_token_accuracy": 0.854840874671936, "num_tokens": 136475044.0, "step": 3593 }, { "epoch": 0.45719374125429335, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.539493560791016, "learning_rate": 5e-05, "loss": 1.1186, "mean_token_accuracy": 0.8470693826675415, "num_tokens": 136513674.0, "step": 3594 }, { "epoch": 0.4573209515328839, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 9.087756156921387, "learning_rate": 5e-05, "loss": 1.2063, "mean_token_accuracy": 0.8613111972808838, "num_tokens": 136551218.0, "step": 3595 }, { "epoch": 0.45744816181147435, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.754844665527344, "learning_rate": 5e-05, "loss": 1.0205, "mean_token_accuracy": 0.870684027671814, "num_tokens": 136589088.0, "step": 3596 }, { "epoch": 0.4575753720900649, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.418458938598633, "learning_rate": 5e-05, "loss": 1.0019, "mean_token_accuracy": 0.8588495254516602, "num_tokens": 136625269.0, "step": 3597 }, { "epoch": 0.4577025823686554, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.099908828735352, "learning_rate": 5e-05, "loss": 1.031, "mean_token_accuracy": 0.8606411218643188, "num_tokens": 136660212.0, "step": 3598 }, { "epoch": 0.4578297926472459, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.794846534729004, "learning_rate": 5e-05, "loss": 0.9127, "mean_token_accuracy": 0.8492614030838013, "num_tokens": 136703372.0, "step": 3599 }, { "epoch": 0.4579570029258364, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.018064498901367, "learning_rate": 5e-05, "loss": 1.0422, "mean_token_accuracy": 0.8499277830123901, "num_tokens": 136741241.0, "step": 3600 }, { "epoch": 0.45808421320442694, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.295478820800781, "learning_rate": 5e-05, "loss": 0.8741, "mean_token_accuracy": 0.8555257320404053, "num_tokens": 136780329.0, "step": 3601 }, { "epoch": 0.4582114234830174, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.333368301391602, "learning_rate": 5e-05, "loss": 1.0744, "mean_token_accuracy": 0.8429766893386841, "num_tokens": 136812567.0, "step": 3602 }, { "epoch": 0.45833863376160794, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.276587963104248, "learning_rate": 5e-05, "loss": 0.9146, "mean_token_accuracy": 0.848725438117981, "num_tokens": 136854370.0, "step": 3603 }, { "epoch": 0.45846584404019847, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.505143165588379, "learning_rate": 5e-05, "loss": 1.0721, "mean_token_accuracy": 0.8539485931396484, "num_tokens": 136890536.0, "step": 3604 }, { "epoch": 0.45859305431878894, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.605857849121094, "learning_rate": 5e-05, "loss": 1.0323, "mean_token_accuracy": 0.8310929536819458, "num_tokens": 136929376.0, "step": 3605 }, { "epoch": 0.45872026459737947, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.349443435668945, "learning_rate": 5e-05, "loss": 1.0199, "mean_token_accuracy": 0.8637911677360535, "num_tokens": 136967857.0, "step": 3606 }, { "epoch": 0.45884747487597, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.862878322601318, "learning_rate": 5e-05, "loss": 0.966, "mean_token_accuracy": 0.8549922108650208, "num_tokens": 137007853.0, "step": 3607 }, { "epoch": 0.45897468515456047, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.905434608459473, "learning_rate": 5e-05, "loss": 1.0189, "mean_token_accuracy": 0.8597105741500854, "num_tokens": 137048172.0, "step": 3608 }, { "epoch": 0.459101895433151, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 7.887359619140625, "learning_rate": 5e-05, "loss": 0.9902, "mean_token_accuracy": 0.8397374153137207, "num_tokens": 137083598.0, "step": 3609 }, { "epoch": 0.4592291057117415, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.43114948272705, "learning_rate": 5e-05, "loss": 0.9942, "mean_token_accuracy": 0.8490321636199951, "num_tokens": 137117626.0, "step": 3610 }, { "epoch": 0.459356315990332, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.613224983215332, "learning_rate": 5e-05, "loss": 0.9172, "mean_token_accuracy": 0.8471775054931641, "num_tokens": 137161536.0, "step": 3611 }, { "epoch": 0.4594835262689225, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.378618240356445, "learning_rate": 5e-05, "loss": 0.9324, "mean_token_accuracy": 0.8652043342590332, "num_tokens": 137195932.0, "step": 3612 }, { "epoch": 0.45961073654751305, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.178217887878418, "learning_rate": 5e-05, "loss": 0.7986, "mean_token_accuracy": 0.8670822381973267, "num_tokens": 137229263.0, "step": 3613 }, { "epoch": 0.4597379468261035, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.709516525268555, "learning_rate": 5e-05, "loss": 0.9893, "mean_token_accuracy": 0.8587201833724976, "num_tokens": 137265250.0, "step": 3614 }, { "epoch": 0.45986515710469406, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.066201210021973, "learning_rate": 5e-05, "loss": 0.7798, "mean_token_accuracy": 0.868926465511322, "num_tokens": 137307612.0, "step": 3615 }, { "epoch": 0.4599923673832846, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.86082649230957, "learning_rate": 5e-05, "loss": 0.9599, "mean_token_accuracy": 0.8695890307426453, "num_tokens": 137336075.0, "step": 3616 }, { "epoch": 0.46011957766187506, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.402766704559326, "learning_rate": 5e-05, "loss": 0.8901, "mean_token_accuracy": 0.8566994071006775, "num_tokens": 137374571.0, "step": 3617 }, { "epoch": 0.4602467879404656, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.4764986038208, "learning_rate": 5e-05, "loss": 1.003, "mean_token_accuracy": 0.8503531217575073, "num_tokens": 137409275.0, "step": 3618 }, { "epoch": 0.4603739982190561, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.676270961761475, "learning_rate": 5e-05, "loss": 0.8943, "mean_token_accuracy": 0.8586080074310303, "num_tokens": 137447665.0, "step": 3619 }, { "epoch": 0.4605012084976466, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.85983419418335, "learning_rate": 5e-05, "loss": 0.8943, "mean_token_accuracy": 0.8656418919563293, "num_tokens": 137484256.0, "step": 3620 }, { "epoch": 0.4606284187762371, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.6252875328063965, "learning_rate": 5e-05, "loss": 0.8497, "mean_token_accuracy": 0.8603026270866394, "num_tokens": 137523868.0, "step": 3621 }, { "epoch": 0.46075562905482764, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.561244964599609, "learning_rate": 5e-05, "loss": 0.9309, "mean_token_accuracy": 0.834313690662384, "num_tokens": 137564707.0, "step": 3622 }, { "epoch": 0.4608828393334181, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.29628324508667, "learning_rate": 5e-05, "loss": 0.7783, "mean_token_accuracy": 0.8647546768188477, "num_tokens": 137600347.0, "step": 3623 }, { "epoch": 0.46101004961200864, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.776792526245117, "learning_rate": 5e-05, "loss": 0.8751, "mean_token_accuracy": 0.8531202673912048, "num_tokens": 137642874.0, "step": 3624 }, { "epoch": 0.46113725989059917, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.8913254737854, "learning_rate": 5e-05, "loss": 0.8174, "mean_token_accuracy": 0.8496416807174683, "num_tokens": 137685253.0, "step": 3625 }, { "epoch": 0.46126447016918964, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.055061340332031, "learning_rate": 5e-05, "loss": 0.8226, "mean_token_accuracy": 0.8749722242355347, "num_tokens": 137722180.0, "step": 3626 }, { "epoch": 0.4613916804477802, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.760969638824463, "learning_rate": 5e-05, "loss": 0.7476, "mean_token_accuracy": 0.8626807332038879, "num_tokens": 137756665.0, "step": 3627 }, { "epoch": 0.4615188907263707, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.081579208374023, "learning_rate": 5e-05, "loss": 0.872, "mean_token_accuracy": 0.8661252856254578, "num_tokens": 137795928.0, "step": 3628 }, { "epoch": 0.4616461010049612, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.837382793426514, "learning_rate": 5e-05, "loss": 0.7996, "mean_token_accuracy": 0.8437047004699707, "num_tokens": 137833695.0, "step": 3629 }, { "epoch": 0.4617733112835517, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.01729965209961, "learning_rate": 5e-05, "loss": 0.8631, "mean_token_accuracy": 0.8617411255836487, "num_tokens": 137871894.0, "step": 3630 }, { "epoch": 0.46190052156214223, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.95941686630249, "learning_rate": 5e-05, "loss": 0.7818, "mean_token_accuracy": 0.8513420224189758, "num_tokens": 137907818.0, "step": 3631 }, { "epoch": 0.46202773184073276, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.832658767700195, "learning_rate": 5e-05, "loss": 0.8084, "mean_token_accuracy": 0.8698705434799194, "num_tokens": 137943506.0, "step": 3632 }, { "epoch": 0.46215494211932323, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.133308410644531, "learning_rate": 5e-05, "loss": 0.8262, "mean_token_accuracy": 0.850740909576416, "num_tokens": 137975113.0, "step": 3633 }, { "epoch": 0.46228215239791376, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.508169651031494, "learning_rate": 5e-05, "loss": 0.8049, "mean_token_accuracy": 0.8614029884338379, "num_tokens": 138006503.0, "step": 3634 }, { "epoch": 0.4624093626765043, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.309609413146973, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.866130530834198, "num_tokens": 138046667.0, "step": 3635 }, { "epoch": 0.46253657295509476, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.255538463592529, "learning_rate": 5e-05, "loss": 0.8105, "mean_token_accuracy": 0.8457473516464233, "num_tokens": 138084821.0, "step": 3636 }, { "epoch": 0.4626637832336853, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.377511501312256, "learning_rate": 5e-05, "loss": 0.7798, "mean_token_accuracy": 0.8561556339263916, "num_tokens": 138123994.0, "step": 3637 }, { "epoch": 0.4627909935122758, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.140263557434082, "learning_rate": 5e-05, "loss": 0.7779, "mean_token_accuracy": 0.8541449904441833, "num_tokens": 138161225.0, "step": 3638 }, { "epoch": 0.4629182037908663, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.404889106750488, "learning_rate": 5e-05, "loss": 0.7317, "mean_token_accuracy": 0.8662059307098389, "num_tokens": 138197441.0, "step": 3639 }, { "epoch": 0.4630454140694568, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.050013542175293, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.860630214214325, "num_tokens": 138229663.0, "step": 3640 }, { "epoch": 0.46317262434804735, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.440496444702148, "learning_rate": 5e-05, "loss": 0.7868, "mean_token_accuracy": 0.8559531569480896, "num_tokens": 138267955.0, "step": 3641 }, { "epoch": 0.4632998346266378, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.97850227355957, "learning_rate": 5e-05, "loss": 0.6954, "mean_token_accuracy": 0.8695465326309204, "num_tokens": 138303968.0, "step": 3642 }, { "epoch": 0.46342704490522835, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.491354942321777, "learning_rate": 5e-05, "loss": 0.7744, "mean_token_accuracy": 0.8577567338943481, "num_tokens": 138344667.0, "step": 3643 }, { "epoch": 0.4635542551838189, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.918745994567871, "learning_rate": 5e-05, "loss": 0.6812, "mean_token_accuracy": 0.8684900403022766, "num_tokens": 138387513.0, "step": 3644 }, { "epoch": 0.46368146546240935, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.509884834289551, "learning_rate": 5e-05, "loss": 0.7066, "mean_token_accuracy": 0.874695897102356, "num_tokens": 138428070.0, "step": 3645 }, { "epoch": 0.4638086757409999, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.8877692222595215, "learning_rate": 5e-05, "loss": 0.711, "mean_token_accuracy": 0.8595530390739441, "num_tokens": 138462303.0, "step": 3646 }, { "epoch": 0.4639358860195904, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.530725479125977, "learning_rate": 5e-05, "loss": 0.8383, "mean_token_accuracy": 0.8372507095336914, "num_tokens": 138502099.0, "step": 3647 }, { "epoch": 0.4640630962981809, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.851591110229492, "learning_rate": 5e-05, "loss": 0.739, "mean_token_accuracy": 0.8513199687004089, "num_tokens": 138539400.0, "step": 3648 }, { "epoch": 0.4641903065767714, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.5667500495910645, "learning_rate": 5e-05, "loss": 0.7687, "mean_token_accuracy": 0.8567859530448914, "num_tokens": 138573082.0, "step": 3649 }, { "epoch": 0.46431751685536193, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.7890167236328125, "learning_rate": 5e-05, "loss": 0.6872, "mean_token_accuracy": 0.863999605178833, "num_tokens": 138613693.0, "step": 3650 }, { "epoch": 0.4644447271339524, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.655033588409424, "learning_rate": 5e-05, "loss": 0.7209, "mean_token_accuracy": 0.8722492456436157, "num_tokens": 138655138.0, "step": 3651 }, { "epoch": 0.46457193741254293, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.718845367431641, "learning_rate": 5e-05, "loss": 0.8097, "mean_token_accuracy": 0.8267413973808289, "num_tokens": 138697117.0, "step": 3652 }, { "epoch": 0.46469914769113346, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.721280097961426, "learning_rate": 5e-05, "loss": 0.7895, "mean_token_accuracy": 0.8534817695617676, "num_tokens": 138732723.0, "step": 3653 }, { "epoch": 0.46482635796972394, "ewc_loss": 0.28125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.612314224243164, "learning_rate": 5e-05, "loss": 0.6842, "mean_token_accuracy": 0.8598802089691162, "num_tokens": 138767680.0, "step": 3654 }, { "epoch": 0.46495356824831446, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.89333438873291, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8656973838806152, "num_tokens": 138803119.0, "step": 3655 }, { "epoch": 0.465080778526905, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.428924083709717, "learning_rate": 5e-05, "loss": 0.6645, "mean_token_accuracy": 0.8647369146347046, "num_tokens": 138845032.0, "step": 3656 }, { "epoch": 0.46520798880549546, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.184203147888184, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8641165494918823, "num_tokens": 138886521.0, "step": 3657 }, { "epoch": 0.465335199084086, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.17071008682251, "learning_rate": 5e-05, "loss": 0.6802, "mean_token_accuracy": 0.8591393232345581, "num_tokens": 138927586.0, "step": 3658 }, { "epoch": 0.4654624093626765, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.565011978149414, "learning_rate": 5e-05, "loss": 0.874, "mean_token_accuracy": 0.8538404703140259, "num_tokens": 138967382.0, "step": 3659 }, { "epoch": 0.465589619641267, "ewc_loss": 0.28125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.048236846923828, "learning_rate": 5e-05, "loss": 0.7465, "mean_token_accuracy": 0.8426817655563354, "num_tokens": 139003498.0, "step": 3660 }, { "epoch": 0.4657168299198575, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.743087768554688, "learning_rate": 5e-05, "loss": 0.9302, "mean_token_accuracy": 0.8478453159332275, "num_tokens": 139046724.0, "step": 3661 }, { "epoch": 0.46584404019844805, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.41690731048584, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.8566362857818604, "num_tokens": 139091513.0, "step": 3662 }, { "epoch": 0.4659712504770385, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.253806114196777, "learning_rate": 5e-05, "loss": 0.9673, "mean_token_accuracy": 0.8332910537719727, "num_tokens": 139130082.0, "step": 3663 }, { "epoch": 0.46609846075562905, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.297583103179932, "learning_rate": 5e-05, "loss": 0.8148, "mean_token_accuracy": 0.860664963722229, "num_tokens": 139172974.0, "step": 3664 }, { "epoch": 0.4662256710342196, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.997869491577148, "learning_rate": 5e-05, "loss": 0.7332, "mean_token_accuracy": 0.8664792776107788, "num_tokens": 139212164.0, "step": 3665 }, { "epoch": 0.46635288131281005, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.406070709228516, "learning_rate": 5e-05, "loss": 0.9773, "mean_token_accuracy": 0.8291934728622437, "num_tokens": 139252960.0, "step": 3666 }, { "epoch": 0.4664800915914006, "ewc_loss": 0.265625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00023746490478515625, "grad_norm": 5.805099010467529, "learning_rate": 5e-05, "loss": 0.7293, "mean_token_accuracy": 0.8500063419342041, "num_tokens": 139287078.0, "step": 3667 }, { "epoch": 0.4666073018699911, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.660462379455566, "learning_rate": 5e-05, "loss": 1.0623, "mean_token_accuracy": 0.8441118001937866, "num_tokens": 139321545.0, "step": 3668 }, { "epoch": 0.4667345121485816, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 5.580060958862305, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8404929637908936, "num_tokens": 139360829.0, "step": 3669 }, { "epoch": 0.4668617224271721, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.517374992370605, "learning_rate": 5e-05, "loss": 1.1604, "mean_token_accuracy": 0.8577855825424194, "num_tokens": 139396347.0, "step": 3670 }, { "epoch": 0.46698893270576264, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 7.507402420043945, "learning_rate": 5e-05, "loss": 1.0127, "mean_token_accuracy": 0.8607569932937622, "num_tokens": 139430367.0, "step": 3671 }, { "epoch": 0.4671161429843531, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.558372497558594, "learning_rate": 5e-05, "loss": 1.0557, "mean_token_accuracy": 0.8488781452178955, "num_tokens": 139468233.0, "step": 3672 }, { "epoch": 0.46724335326294364, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 8.828961372375488, "learning_rate": 5e-05, "loss": 1.1542, "mean_token_accuracy": 0.8542059659957886, "num_tokens": 139502377.0, "step": 3673 }, { "epoch": 0.46737056354153417, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.574825763702393, "learning_rate": 5e-05, "loss": 0.8884, "mean_token_accuracy": 0.8505038022994995, "num_tokens": 139540503.0, "step": 3674 }, { "epoch": 0.46749777382012464, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 9.84327220916748, "learning_rate": 5e-05, "loss": 1.1333, "mean_token_accuracy": 0.8432080745697021, "num_tokens": 139578063.0, "step": 3675 }, { "epoch": 0.46762498409871517, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.126312255859375, "learning_rate": 5e-05, "loss": 0.8415, "mean_token_accuracy": 0.8718155026435852, "num_tokens": 139615451.0, "step": 3676 }, { "epoch": 0.4677521943773057, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.187122344970703, "learning_rate": 5e-05, "loss": 1.1262, "mean_token_accuracy": 0.8455914855003357, "num_tokens": 139657779.0, "step": 3677 }, { "epoch": 0.46787940465589617, "ewc_loss": 0.625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.088173866271973, "learning_rate": 5e-05, "loss": 1.0452, "mean_token_accuracy": 0.8592689633369446, "num_tokens": 139698166.0, "step": 3678 }, { "epoch": 0.4680066149344867, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.227601051330566, "learning_rate": 5e-05, "loss": 1.0448, "mean_token_accuracy": 0.8407789468765259, "num_tokens": 139736970.0, "step": 3679 }, { "epoch": 0.4681338252130772, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.250734329223633, "learning_rate": 5e-05, "loss": 1.0156, "mean_token_accuracy": 0.8563511371612549, "num_tokens": 139771570.0, "step": 3680 }, { "epoch": 0.46826103549166775, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.247943878173828, "learning_rate": 5e-05, "loss": 0.9955, "mean_token_accuracy": 0.8356469869613647, "num_tokens": 139811927.0, "step": 3681 }, { "epoch": 0.4683882457702582, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.03669548034668, "learning_rate": 5e-05, "loss": 0.9643, "mean_token_accuracy": 0.8463739156723022, "num_tokens": 139852903.0, "step": 3682 }, { "epoch": 0.46851545604884876, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.818663597106934, "learning_rate": 5e-05, "loss": 0.9184, "mean_token_accuracy": 0.8467278480529785, "num_tokens": 139895375.0, "step": 3683 }, { "epoch": 0.4686426663274393, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.619561195373535, "learning_rate": 5e-05, "loss": 0.9383, "mean_token_accuracy": 0.8348298072814941, "num_tokens": 139927878.0, "step": 3684 }, { "epoch": 0.46876987660602976, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.816877841949463, "learning_rate": 5e-05, "loss": 0.8657, "mean_token_accuracy": 0.8542329668998718, "num_tokens": 139959864.0, "step": 3685 }, { "epoch": 0.4688970868846203, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.076302528381348, "learning_rate": 5e-05, "loss": 0.8242, "mean_token_accuracy": 0.8460428714752197, "num_tokens": 140003815.0, "step": 3686 }, { "epoch": 0.4690242971632108, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.094073295593262, "learning_rate": 5e-05, "loss": 0.9129, "mean_token_accuracy": 0.8421758413314819, "num_tokens": 140043126.0, "step": 3687 }, { "epoch": 0.4691515074418013, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.723249912261963, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8542940020561218, "num_tokens": 140078466.0, "step": 3688 }, { "epoch": 0.4692787177203918, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.247604370117188, "learning_rate": 5e-05, "loss": 0.8992, "mean_token_accuracy": 0.847126305103302, "num_tokens": 140114733.0, "step": 3689 }, { "epoch": 0.46940592799898234, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.681463241577148, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.8554630279541016, "num_tokens": 140157017.0, "step": 3690 }, { "epoch": 0.4695331382775728, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.170112609863281, "learning_rate": 5e-05, "loss": 0.8797, "mean_token_accuracy": 0.8526979684829712, "num_tokens": 140197408.0, "step": 3691 }, { "epoch": 0.46966034855616334, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.80470609664917, "learning_rate": 5e-05, "loss": 0.8099, "mean_token_accuracy": 0.8483430743217468, "num_tokens": 140232458.0, "step": 3692 }, { "epoch": 0.46978755883475387, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.901248931884766, "learning_rate": 5e-05, "loss": 0.8554, "mean_token_accuracy": 0.8541083335876465, "num_tokens": 140273790.0, "step": 3693 }, { "epoch": 0.46991476911334434, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.048152446746826, "learning_rate": 5e-05, "loss": 0.7382, "mean_token_accuracy": 0.8718461394309998, "num_tokens": 140313339.0, "step": 3694 }, { "epoch": 0.47004197939193487, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.4684600830078125, "learning_rate": 5e-05, "loss": 0.8056, "mean_token_accuracy": 0.8536513447761536, "num_tokens": 140352572.0, "step": 3695 }, { "epoch": 0.4701691896705254, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.3600921630859375, "learning_rate": 5e-05, "loss": 0.8303, "mean_token_accuracy": 0.8468636274337769, "num_tokens": 140385221.0, "step": 3696 }, { "epoch": 0.4702963999491159, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.104324817657471, "learning_rate": 5e-05, "loss": 0.7069, "mean_token_accuracy": 0.8719738721847534, "num_tokens": 140427041.0, "step": 3697 }, { "epoch": 0.4704236102277064, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.559365749359131, "learning_rate": 5e-05, "loss": 0.7379, "mean_token_accuracy": 0.8732492923736572, "num_tokens": 140460956.0, "step": 3698 }, { "epoch": 0.47055082050629693, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.988227844238281, "learning_rate": 5e-05, "loss": 0.7081, "mean_token_accuracy": 0.8715844750404358, "num_tokens": 140497644.0, "step": 3699 }, { "epoch": 0.4706780307848874, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.502808570861816, "learning_rate": 5e-05, "loss": 0.7671, "mean_token_accuracy": 0.8562541007995605, "num_tokens": 140532764.0, "step": 3700 }, { "epoch": 0.47080524106347793, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.019204139709473, "learning_rate": 5e-05, "loss": 0.7448, "mean_token_accuracy": 0.8560078144073486, "num_tokens": 140564705.0, "step": 3701 }, { "epoch": 0.47093245134206846, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.392674446105957, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8526690602302551, "num_tokens": 140606959.0, "step": 3702 }, { "epoch": 0.47105966162065893, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.128347396850586, "learning_rate": 5e-05, "loss": 0.7228, "mean_token_accuracy": 0.8645248413085938, "num_tokens": 140646034.0, "step": 3703 }, { "epoch": 0.47118687189924946, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.227022171020508, "learning_rate": 5e-05, "loss": 0.7418, "mean_token_accuracy": 0.8545873165130615, "num_tokens": 140679948.0, "step": 3704 }, { "epoch": 0.47131408217784, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.307779312133789, "learning_rate": 5e-05, "loss": 0.7844, "mean_token_accuracy": 0.8471012115478516, "num_tokens": 140713071.0, "step": 3705 }, { "epoch": 0.47144129245643046, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.9997639656066895, "learning_rate": 5e-05, "loss": 0.7078, "mean_token_accuracy": 0.8636990785598755, "num_tokens": 140744243.0, "step": 3706 }, { "epoch": 0.471568502735021, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.513210296630859, "learning_rate": 5e-05, "loss": 0.6969, "mean_token_accuracy": 0.8730939030647278, "num_tokens": 140776716.0, "step": 3707 }, { "epoch": 0.4716957130136115, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.790196418762207, "learning_rate": 5e-05, "loss": 0.6633, "mean_token_accuracy": 0.8756940960884094, "num_tokens": 140808874.0, "step": 3708 }, { "epoch": 0.471822923292202, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.765761375427246, "learning_rate": 5e-05, "loss": 0.7508, "mean_token_accuracy": 0.8655929565429688, "num_tokens": 140846198.0, "step": 3709 }, { "epoch": 0.4719501335707925, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.453122138977051, "learning_rate": 5e-05, "loss": 0.6879, "mean_token_accuracy": 0.8604624271392822, "num_tokens": 140889228.0, "step": 3710 }, { "epoch": 0.47207734384938305, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.252341270446777, "learning_rate": 5e-05, "loss": 0.798, "mean_token_accuracy": 0.8595463037490845, "num_tokens": 140924549.0, "step": 3711 }, { "epoch": 0.4722045541279735, "ewc_loss": 0.248046875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00022029876708984375, "grad_norm": 5.9473443031311035, "learning_rate": 5e-05, "loss": 0.6427, "mean_token_accuracy": 0.8683819770812988, "num_tokens": 140960995.0, "step": 3712 }, { "epoch": 0.47233176440656405, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 9.065147399902344, "learning_rate": 5e-05, "loss": 0.9173, "mean_token_accuracy": 0.8575650453567505, "num_tokens": 141000899.0, "step": 3713 }, { "epoch": 0.4724589746851546, "ewc_loss": 0.248046875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00022029876708984375, "grad_norm": 5.401330947875977, "learning_rate": 5e-05, "loss": 0.6448, "mean_token_accuracy": 0.8640870451927185, "num_tokens": 141037736.0, "step": 3714 }, { "epoch": 0.47258618496374505, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 10.215373992919922, "learning_rate": 5e-05, "loss": 1.0944, "mean_token_accuracy": 0.8392114639282227, "num_tokens": 141068916.0, "step": 3715 }, { "epoch": 0.4727133952423356, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.054024696350098, "learning_rate": 5e-05, "loss": 0.845, "mean_token_accuracy": 0.8411722183227539, "num_tokens": 141102353.0, "step": 3716 }, { "epoch": 0.4728406055209261, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 10.788516998291016, "learning_rate": 5e-05, "loss": 1.1568, "mean_token_accuracy": 0.867100715637207, "num_tokens": 141142653.0, "step": 3717 }, { "epoch": 0.4729678157995166, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 8.334572792053223, "learning_rate": 5e-05, "loss": 1.1546, "mean_token_accuracy": 0.8589942455291748, "num_tokens": 141180091.0, "step": 3718 }, { "epoch": 0.4730950260781071, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.293196678161621, "learning_rate": 5e-05, "loss": 0.9536, "mean_token_accuracy": 0.8727772235870361, "num_tokens": 141218674.0, "step": 3719 }, { "epoch": 0.47322223635669763, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 9.194892883300781, "learning_rate": 5e-05, "loss": 1.1584, "mean_token_accuracy": 0.8696517944335938, "num_tokens": 141258680.0, "step": 3720 }, { "epoch": 0.4733494466352881, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.440357208251953, "learning_rate": 5e-05, "loss": 0.8588, "mean_token_accuracy": 0.8518975973129272, "num_tokens": 141297883.0, "step": 3721 }, { "epoch": 0.47347665691387864, "ewc_loss": 0.765625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 10.432744979858398, "learning_rate": 5e-05, "loss": 1.1861, "mean_token_accuracy": 0.8595967292785645, "num_tokens": 141336549.0, "step": 3722 }, { "epoch": 0.47360386719246916, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 7.551230430603027, "learning_rate": 5e-05, "loss": 0.9822, "mean_token_accuracy": 0.857426106929779, "num_tokens": 141368697.0, "step": 3723 }, { "epoch": 0.47373107747105964, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.900764465332031, "learning_rate": 5e-05, "loss": 1.1305, "mean_token_accuracy": 0.863001823425293, "num_tokens": 141399349.0, "step": 3724 }, { "epoch": 0.47385828774965016, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000732421875, "grad_norm": 8.819323539733887, "learning_rate": 5e-05, "loss": 1.1419, "mean_token_accuracy": 0.8715493083000183, "num_tokens": 141435198.0, "step": 3725 }, { "epoch": 0.4739854980282407, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.011482238769531, "learning_rate": 5e-05, "loss": 1.1016, "mean_token_accuracy": 0.8435364961624146, "num_tokens": 141470659.0, "step": 3726 }, { "epoch": 0.47411270830683117, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 8.642212867736816, "learning_rate": 5e-05, "loss": 1.0989, "mean_token_accuracy": 0.8558225631713867, "num_tokens": 141514774.0, "step": 3727 }, { "epoch": 0.4742399185854217, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.248518943786621, "learning_rate": 5e-05, "loss": 0.9889, "mean_token_accuracy": 0.8615992069244385, "num_tokens": 141551071.0, "step": 3728 }, { "epoch": 0.4743671288640122, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.096558570861816, "learning_rate": 5e-05, "loss": 0.9679, "mean_token_accuracy": 0.8591262102127075, "num_tokens": 141590201.0, "step": 3729 }, { "epoch": 0.4744943391426027, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.708288192749023, "learning_rate": 5e-05, "loss": 0.9439, "mean_token_accuracy": 0.8651869297027588, "num_tokens": 141637580.0, "step": 3730 }, { "epoch": 0.4746215494211932, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.538364410400391, "learning_rate": 5e-05, "loss": 0.9257, "mean_token_accuracy": 0.8442406058311462, "num_tokens": 141674092.0, "step": 3731 }, { "epoch": 0.47474875969978375, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.536233901977539, "learning_rate": 5e-05, "loss": 0.9899, "mean_token_accuracy": 0.8407498002052307, "num_tokens": 141709416.0, "step": 3732 }, { "epoch": 0.4748759699783743, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.057104110717773, "learning_rate": 5e-05, "loss": 0.8143, "mean_token_accuracy": 0.856969952583313, "num_tokens": 141742445.0, "step": 3733 }, { "epoch": 0.47500318025696475, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.65069580078125, "learning_rate": 5e-05, "loss": 0.8819, "mean_token_accuracy": 0.8682807087898254, "num_tokens": 141778224.0, "step": 3734 }, { "epoch": 0.4751303905355553, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.7941789627075195, "learning_rate": 5e-05, "loss": 0.8365, "mean_token_accuracy": 0.8490501642227173, "num_tokens": 141818239.0, "step": 3735 }, { "epoch": 0.4752576008141458, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.783084869384766, "learning_rate": 5e-05, "loss": 0.9907, "mean_token_accuracy": 0.8419493436813354, "num_tokens": 141853624.0, "step": 3736 }, { "epoch": 0.4753848110927363, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.811164379119873, "learning_rate": 5e-05, "loss": 0.8088, "mean_token_accuracy": 0.8564634919166565, "num_tokens": 141893620.0, "step": 3737 }, { "epoch": 0.4755120213713268, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.776763916015625, "learning_rate": 5e-05, "loss": 0.9274, "mean_token_accuracy": 0.8651565313339233, "num_tokens": 141930504.0, "step": 3738 }, { "epoch": 0.47563923164991734, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.04880952835083, "learning_rate": 5e-05, "loss": 0.8699, "mean_token_accuracy": 0.8508183360099792, "num_tokens": 141968195.0, "step": 3739 }, { "epoch": 0.4757664419285078, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.441240310668945, "learning_rate": 5e-05, "loss": 0.9191, "mean_token_accuracy": 0.8589887619018555, "num_tokens": 142010143.0, "step": 3740 }, { "epoch": 0.47589365220709834, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.410788536071777, "learning_rate": 5e-05, "loss": 0.9204, "mean_token_accuracy": 0.8468919992446899, "num_tokens": 142051170.0, "step": 3741 }, { "epoch": 0.47602086248568887, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.741967678070068, "learning_rate": 5e-05, "loss": 0.9057, "mean_token_accuracy": 0.8383806347846985, "num_tokens": 142084327.0, "step": 3742 }, { "epoch": 0.47614807276427934, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.706485748291016, "learning_rate": 5e-05, "loss": 0.9024, "mean_token_accuracy": 0.8476631045341492, "num_tokens": 142125512.0, "step": 3743 }, { "epoch": 0.47627528304286987, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.143517971038818, "learning_rate": 5e-05, "loss": 0.7711, "mean_token_accuracy": 0.8594943284988403, "num_tokens": 142159117.0, "step": 3744 }, { "epoch": 0.4764024933214604, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.860163688659668, "learning_rate": 5e-05, "loss": 0.8502, "mean_token_accuracy": 0.8573682904243469, "num_tokens": 142194580.0, "step": 3745 }, { "epoch": 0.47652970360005087, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.880423069000244, "learning_rate": 5e-05, "loss": 0.7181, "mean_token_accuracy": 0.8728057742118835, "num_tokens": 142230945.0, "step": 3746 }, { "epoch": 0.4766569138786414, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.900006294250488, "learning_rate": 5e-05, "loss": 0.8031, "mean_token_accuracy": 0.8671044111251831, "num_tokens": 142264792.0, "step": 3747 }, { "epoch": 0.4767841241572319, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.815488815307617, "learning_rate": 5e-05, "loss": 0.7091, "mean_token_accuracy": 0.8670855760574341, "num_tokens": 142302210.0, "step": 3748 }, { "epoch": 0.4769113344358224, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.898322582244873, "learning_rate": 5e-05, "loss": 0.8547, "mean_token_accuracy": 0.8502370119094849, "num_tokens": 142345646.0, "step": 3749 }, { "epoch": 0.4770385447144129, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.816104412078857, "learning_rate": 5e-05, "loss": 0.8279, "mean_token_accuracy": 0.8343217372894287, "num_tokens": 142387097.0, "step": 3750 }, { "epoch": 0.47716575499300345, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.828343391418457, "learning_rate": 5e-05, "loss": 0.8817, "mean_token_accuracy": 0.8365120887756348, "num_tokens": 142424087.0, "step": 3751 }, { "epoch": 0.4772929652715939, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.854216575622559, "learning_rate": 5e-05, "loss": 0.7261, "mean_token_accuracy": 0.8673887252807617, "num_tokens": 142458837.0, "step": 3752 }, { "epoch": 0.47742017555018446, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.793531894683838, "learning_rate": 5e-05, "loss": 0.7654, "mean_token_accuracy": 0.8658376336097717, "num_tokens": 142495661.0, "step": 3753 }, { "epoch": 0.477547385828775, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.808110237121582, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8469200730323792, "num_tokens": 142532439.0, "step": 3754 }, { "epoch": 0.47767459610736546, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.889881134033203, "learning_rate": 5e-05, "loss": 0.8467, "mean_token_accuracy": 0.8467521667480469, "num_tokens": 142574032.0, "step": 3755 }, { "epoch": 0.477801806385956, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.654627799987793, "learning_rate": 5e-05, "loss": 0.713, "mean_token_accuracy": 0.8622050881385803, "num_tokens": 142607438.0, "step": 3756 }, { "epoch": 0.4779290166645465, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.100064277648926, "learning_rate": 5e-05, "loss": 0.8677, "mean_token_accuracy": 0.846904456615448, "num_tokens": 142642527.0, "step": 3757 }, { "epoch": 0.478056226943137, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.6900153160095215, "learning_rate": 5e-05, "loss": 0.736, "mean_token_accuracy": 0.8578978776931763, "num_tokens": 142677447.0, "step": 3758 }, { "epoch": 0.4781834372217275, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.8592095375061035, "learning_rate": 5e-05, "loss": 0.8436, "mean_token_accuracy": 0.8497065901756287, "num_tokens": 142714429.0, "step": 3759 }, { "epoch": 0.47831064750031804, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.8463454246521, "learning_rate": 5e-05, "loss": 0.7465, "mean_token_accuracy": 0.8550108671188354, "num_tokens": 142752742.0, "step": 3760 }, { "epoch": 0.4784378577789085, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.809036731719971, "learning_rate": 5e-05, "loss": 0.8064, "mean_token_accuracy": 0.8589887619018555, "num_tokens": 142788925.0, "step": 3761 }, { "epoch": 0.47856506805749904, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.7845306396484375, "learning_rate": 5e-05, "loss": 0.6869, "mean_token_accuracy": 0.8663969039916992, "num_tokens": 142823215.0, "step": 3762 }, { "epoch": 0.47869227833608957, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.888970851898193, "learning_rate": 5e-05, "loss": 0.8438, "mean_token_accuracy": 0.847305417060852, "num_tokens": 142860346.0, "step": 3763 }, { "epoch": 0.47881948861468004, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.7175188064575195, "learning_rate": 5e-05, "loss": 0.749, "mean_token_accuracy": 0.8553495407104492, "num_tokens": 142895814.0, "step": 3764 }, { "epoch": 0.4789466988932706, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.927399158477783, "learning_rate": 5e-05, "loss": 0.8923, "mean_token_accuracy": 0.8333227634429932, "num_tokens": 142934379.0, "step": 3765 }, { "epoch": 0.4790739091718611, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.737833499908447, "learning_rate": 5e-05, "loss": 0.8002, "mean_token_accuracy": 0.8362212777137756, "num_tokens": 142970512.0, "step": 3766 }, { "epoch": 0.4792011194504516, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.861708164215088, "learning_rate": 5e-05, "loss": 0.8586, "mean_token_accuracy": 0.844088077545166, "num_tokens": 143008635.0, "step": 3767 }, { "epoch": 0.4793283297290421, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.855603218078613, "learning_rate": 5e-05, "loss": 0.7225, "mean_token_accuracy": 0.8624412417411804, "num_tokens": 143047638.0, "step": 3768 }, { "epoch": 0.47945554000763263, "ewc_loss": 0.375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.661670207977295, "learning_rate": 5e-05, "loss": 0.8298, "mean_token_accuracy": 0.8504899740219116, "num_tokens": 143082470.0, "step": 3769 }, { "epoch": 0.4795827502862231, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.0596604347229, "learning_rate": 5e-05, "loss": 0.7425, "mean_token_accuracy": 0.863152265548706, "num_tokens": 143126869.0, "step": 3770 }, { "epoch": 0.47970996056481363, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.3437676429748535, "learning_rate": 5e-05, "loss": 0.7048, "mean_token_accuracy": 0.8764507174491882, "num_tokens": 143161371.0, "step": 3771 }, { "epoch": 0.47983717084340416, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.352499961853027, "learning_rate": 5e-05, "loss": 0.7115, "mean_token_accuracy": 0.8702161312103271, "num_tokens": 143199896.0, "step": 3772 }, { "epoch": 0.47996438112199463, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.983055591583252, "learning_rate": 5e-05, "loss": 0.762, "mean_token_accuracy": 0.8499564528465271, "num_tokens": 143236577.0, "step": 3773 }, { "epoch": 0.48009159140058516, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.646968841552734, "learning_rate": 5e-05, "loss": 0.7777, "mean_token_accuracy": 0.8564017415046692, "num_tokens": 143279780.0, "step": 3774 }, { "epoch": 0.4802188016791757, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.643834590911865, "learning_rate": 5e-05, "loss": 0.7057, "mean_token_accuracy": 0.8630471229553223, "num_tokens": 143321946.0, "step": 3775 }, { "epoch": 0.48034601195776616, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.999574661254883, "learning_rate": 5e-05, "loss": 0.7748, "mean_token_accuracy": 0.8660323619842529, "num_tokens": 143359278.0, "step": 3776 }, { "epoch": 0.4804732222363567, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.304769039154053, "learning_rate": 5e-05, "loss": 0.7198, "mean_token_accuracy": 0.8511204123497009, "num_tokens": 143398968.0, "step": 3777 }, { "epoch": 0.4806004325149472, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.42845630645752, "learning_rate": 5e-05, "loss": 0.8624, "mean_token_accuracy": 0.8504524230957031, "num_tokens": 143438020.0, "step": 3778 }, { "epoch": 0.4807276427935377, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00024127960205078125, "grad_norm": 5.9998087882995605, "learning_rate": 5e-05, "loss": 0.6773, "mean_token_accuracy": 0.861251711845398, "num_tokens": 143478813.0, "step": 3779 }, { "epoch": 0.4808548530721282, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.90080738067627, "learning_rate": 5e-05, "loss": 0.9474, "mean_token_accuracy": 0.8499771356582642, "num_tokens": 143519766.0, "step": 3780 }, { "epoch": 0.48098206335071875, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.016856670379639, "learning_rate": 5e-05, "loss": 0.6982, "mean_token_accuracy": 0.8634183406829834, "num_tokens": 143562570.0, "step": 3781 }, { "epoch": 0.4811092736293093, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.973246574401855, "learning_rate": 5e-05, "loss": 1.0293, "mean_token_accuracy": 0.8352166414260864, "num_tokens": 143597203.0, "step": 3782 }, { "epoch": 0.48123648390789975, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.761783123016357, "learning_rate": 5e-05, "loss": 0.8286, "mean_token_accuracy": 0.8550950288772583, "num_tokens": 143635979.0, "step": 3783 }, { "epoch": 0.4813636941864903, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.080609321594238, "learning_rate": 5e-05, "loss": 0.9421, "mean_token_accuracy": 0.8466809988021851, "num_tokens": 143679915.0, "step": 3784 }, { "epoch": 0.4814909044650808, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.817409038543701, "learning_rate": 5e-05, "loss": 0.9552, "mean_token_accuracy": 0.8406878113746643, "num_tokens": 143716275.0, "step": 3785 }, { "epoch": 0.4816181147436713, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.661899089813232, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8495386242866516, "num_tokens": 143752029.0, "step": 3786 }, { "epoch": 0.4817453250222618, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.833697319030762, "learning_rate": 5e-05, "loss": 0.9321, "mean_token_accuracy": 0.8634226322174072, "num_tokens": 143793139.0, "step": 3787 }, { "epoch": 0.48187253530085233, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 5.964715957641602, "learning_rate": 5e-05, "loss": 0.6848, "mean_token_accuracy": 0.8671403527259827, "num_tokens": 143832790.0, "step": 3788 }, { "epoch": 0.4819997455794428, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.381053924560547, "learning_rate": 5e-05, "loss": 1.0149, "mean_token_accuracy": 0.8540676832199097, "num_tokens": 143876046.0, "step": 3789 }, { "epoch": 0.48212695585803333, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.51116943359375, "learning_rate": 5e-05, "loss": 0.8344, "mean_token_accuracy": 0.8493345975875854, "num_tokens": 143917019.0, "step": 3790 }, { "epoch": 0.48225416613662386, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.879693984985352, "learning_rate": 5e-05, "loss": 0.9696, "mean_token_accuracy": 0.8629825115203857, "num_tokens": 143957789.0, "step": 3791 }, { "epoch": 0.48238137641521434, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 7.583591461181641, "learning_rate": 5e-05, "loss": 0.9281, "mean_token_accuracy": 0.8632313013076782, "num_tokens": 143993797.0, "step": 3792 }, { "epoch": 0.48250858669380486, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.62034797668457, "learning_rate": 5e-05, "loss": 0.921, "mean_token_accuracy": 0.8394657373428345, "num_tokens": 144033583.0, "step": 3793 }, { "epoch": 0.4826357969723954, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.203484535217285, "learning_rate": 5e-05, "loss": 0.9292, "mean_token_accuracy": 0.8594626188278198, "num_tokens": 144074358.0, "step": 3794 }, { "epoch": 0.48276300725098586, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.95381498336792, "learning_rate": 5e-05, "loss": 0.8328, "mean_token_accuracy": 0.8493508696556091, "num_tokens": 144114480.0, "step": 3795 }, { "epoch": 0.4828902175295764, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.115554809570312, "learning_rate": 5e-05, "loss": 0.8879, "mean_token_accuracy": 0.8644914627075195, "num_tokens": 144152306.0, "step": 3796 }, { "epoch": 0.4830174278081669, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.1622209548950195, "learning_rate": 5e-05, "loss": 0.8141, "mean_token_accuracy": 0.8574672937393188, "num_tokens": 144191134.0, "step": 3797 }, { "epoch": 0.4831446380867574, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.474222660064697, "learning_rate": 5e-05, "loss": 0.8339, "mean_token_accuracy": 0.8498303890228271, "num_tokens": 144230165.0, "step": 3798 }, { "epoch": 0.4832718483653479, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.589652061462402, "learning_rate": 5e-05, "loss": 0.8364, "mean_token_accuracy": 0.8590918183326721, "num_tokens": 144265035.0, "step": 3799 }, { "epoch": 0.48339905864393845, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.973516941070557, "learning_rate": 5e-05, "loss": 0.705, "mean_token_accuracy": 0.8750581741333008, "num_tokens": 144304692.0, "step": 3800 }, { "epoch": 0.4835262689225289, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.794638633728027, "learning_rate": 5e-05, "loss": 0.8414, "mean_token_accuracy": 0.8518197536468506, "num_tokens": 144345288.0, "step": 3801 }, { "epoch": 0.48365347920111945, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.793892860412598, "learning_rate": 5e-05, "loss": 0.7708, "mean_token_accuracy": 0.8476587533950806, "num_tokens": 144377427.0, "step": 3802 }, { "epoch": 0.48378068947971, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.932088851928711, "learning_rate": 5e-05, "loss": 0.7994, "mean_token_accuracy": 0.8662214279174805, "num_tokens": 144412405.0, "step": 3803 }, { "epoch": 0.48390789975830045, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.592499256134033, "learning_rate": 5e-05, "loss": 0.6981, "mean_token_accuracy": 0.8683470487594604, "num_tokens": 144448093.0, "step": 3804 }, { "epoch": 0.484035110036891, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.240713119506836, "learning_rate": 5e-05, "loss": 0.8706, "mean_token_accuracy": 0.8542138338088989, "num_tokens": 144483280.0, "step": 3805 }, { "epoch": 0.4841623203154815, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.313353061676025, "learning_rate": 5e-05, "loss": 0.6632, "mean_token_accuracy": 0.8765109777450562, "num_tokens": 144520827.0, "step": 3806 }, { "epoch": 0.484289530594072, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.629899978637695, "learning_rate": 5e-05, "loss": 0.8766, "mean_token_accuracy": 0.8659749031066895, "num_tokens": 144562314.0, "step": 3807 }, { "epoch": 0.4844167408726625, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.171137809753418, "learning_rate": 5e-05, "loss": 0.7109, "mean_token_accuracy": 0.8631710410118103, "num_tokens": 144599101.0, "step": 3808 }, { "epoch": 0.48454395115125304, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.925947189331055, "learning_rate": 5e-05, "loss": 0.961, "mean_token_accuracy": 0.8601303100585938, "num_tokens": 144641308.0, "step": 3809 }, { "epoch": 0.4846711614298435, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.425217628479004, "learning_rate": 5e-05, "loss": 0.82, "mean_token_accuracy": 0.8466397523880005, "num_tokens": 144677267.0, "step": 3810 }, { "epoch": 0.48479837170843404, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.857349395751953, "learning_rate": 5e-05, "loss": 1.0145, "mean_token_accuracy": 0.8465979695320129, "num_tokens": 144713904.0, "step": 3811 }, { "epoch": 0.48492558198702457, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 6.9773054122924805, "learning_rate": 5e-05, "loss": 0.8883, "mean_token_accuracy": 0.8551427125930786, "num_tokens": 144750523.0, "step": 3812 }, { "epoch": 0.48505279226561504, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.330073356628418, "learning_rate": 5e-05, "loss": 0.95, "mean_token_accuracy": 0.8577431440353394, "num_tokens": 144791569.0, "step": 3813 }, { "epoch": 0.48518000254420557, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.42529821395874, "learning_rate": 5e-05, "loss": 0.8816, "mean_token_accuracy": 0.8640706539154053, "num_tokens": 144827919.0, "step": 3814 }, { "epoch": 0.4853072128227961, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.718006610870361, "learning_rate": 5e-05, "loss": 0.8237, "mean_token_accuracy": 0.8736726641654968, "num_tokens": 144866859.0, "step": 3815 }, { "epoch": 0.48543442310138657, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.488040924072266, "learning_rate": 5e-05, "loss": 0.8411, "mean_token_accuracy": 0.8714884519577026, "num_tokens": 144903088.0, "step": 3816 }, { "epoch": 0.4855616333799771, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.5113139152526855, "learning_rate": 5e-05, "loss": 0.7966, "mean_token_accuracy": 0.8710721135139465, "num_tokens": 144938485.0, "step": 3817 }, { "epoch": 0.4856888436585676, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.216524600982666, "learning_rate": 5e-05, "loss": 0.8965, "mean_token_accuracy": 0.8392012119293213, "num_tokens": 144974296.0, "step": 3818 }, { "epoch": 0.4858160539371581, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.692996025085449, "learning_rate": 5e-05, "loss": 0.8419, "mean_token_accuracy": 0.8606107234954834, "num_tokens": 145012391.0, "step": 3819 }, { "epoch": 0.4859432642157486, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.920892715454102, "learning_rate": 5e-05, "loss": 0.8125, "mean_token_accuracy": 0.8392304182052612, "num_tokens": 145045409.0, "step": 3820 }, { "epoch": 0.48607047449433916, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.837029933929443, "learning_rate": 5e-05, "loss": 0.8254, "mean_token_accuracy": 0.8656877279281616, "num_tokens": 145081162.0, "step": 3821 }, { "epoch": 0.48619768477292963, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.809319019317627, "learning_rate": 5e-05, "loss": 0.7347, "mean_token_accuracy": 0.8642580509185791, "num_tokens": 145117045.0, "step": 3822 }, { "epoch": 0.48632489505152016, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.835629463195801, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.8652064204216003, "num_tokens": 145153931.0, "step": 3823 }, { "epoch": 0.4864521053301107, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.826658725738525, "learning_rate": 5e-05, "loss": 0.7393, "mean_token_accuracy": 0.8627464771270752, "num_tokens": 145193812.0, "step": 3824 }, { "epoch": 0.48657931560870116, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.779961585998535, "learning_rate": 5e-05, "loss": 0.8718, "mean_token_accuracy": 0.8504459857940674, "num_tokens": 145231814.0, "step": 3825 }, { "epoch": 0.4867065258872917, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.872837066650391, "learning_rate": 5e-05, "loss": 0.7606, "mean_token_accuracy": 0.8526005744934082, "num_tokens": 145279821.0, "step": 3826 }, { "epoch": 0.4868337361658822, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.713489532470703, "learning_rate": 5e-05, "loss": 0.8451, "mean_token_accuracy": 0.8495218753814697, "num_tokens": 145318972.0, "step": 3827 }, { "epoch": 0.4869609464444727, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.926059246063232, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.859799325466156, "num_tokens": 145359479.0, "step": 3828 }, { "epoch": 0.4870881567230632, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.635653972625732, "learning_rate": 5e-05, "loss": 0.8376, "mean_token_accuracy": 0.8471252918243408, "num_tokens": 145403119.0, "step": 3829 }, { "epoch": 0.48721536700165374, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.969893455505371, "learning_rate": 5e-05, "loss": 0.7713, "mean_token_accuracy": 0.8483090400695801, "num_tokens": 145439715.0, "step": 3830 }, { "epoch": 0.48734257728024427, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.549063682556152, "learning_rate": 5e-05, "loss": 0.7716, "mean_token_accuracy": 0.8651077747344971, "num_tokens": 145480715.0, "step": 3831 }, { "epoch": 0.48746978755883474, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.0136027336120605, "learning_rate": 5e-05, "loss": 0.7743, "mean_token_accuracy": 0.8537802696228027, "num_tokens": 145523501.0, "step": 3832 }, { "epoch": 0.48759699783742527, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.483271598815918, "learning_rate": 5e-05, "loss": 0.8444, "mean_token_accuracy": 0.8328295350074768, "num_tokens": 145563387.0, "step": 3833 }, { "epoch": 0.4877242081160158, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.0269317626953125, "learning_rate": 5e-05, "loss": 0.8018, "mean_token_accuracy": 0.8441964983940125, "num_tokens": 145607818.0, "step": 3834 }, { "epoch": 0.4878514183946063, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.430131912231445, "learning_rate": 5e-05, "loss": 0.8093, "mean_token_accuracy": 0.8444353938102722, "num_tokens": 145646115.0, "step": 3835 }, { "epoch": 0.4879786286731968, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.048057556152344, "learning_rate": 5e-05, "loss": 0.7154, "mean_token_accuracy": 0.86245197057724, "num_tokens": 145681555.0, "step": 3836 }, { "epoch": 0.48810583895178733, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.385975360870361, "learning_rate": 5e-05, "loss": 0.7223, "mean_token_accuracy": 0.867271363735199, "num_tokens": 145717459.0, "step": 3837 }, { "epoch": 0.4882330492303778, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.0465521812438965, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8511524796485901, "num_tokens": 145753143.0, "step": 3838 }, { "epoch": 0.48836025950896833, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.348146915435791, "learning_rate": 5e-05, "loss": 0.7858, "mean_token_accuracy": 0.8483687043190002, "num_tokens": 145790345.0, "step": 3839 }, { "epoch": 0.48848746978755886, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.057884693145752, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8514741659164429, "num_tokens": 145832900.0, "step": 3840 }, { "epoch": 0.48861468006614933, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.3216729164123535, "learning_rate": 5e-05, "loss": 0.7316, "mean_token_accuracy": 0.8619570136070251, "num_tokens": 145870309.0, "step": 3841 }, { "epoch": 0.48874189034473986, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.062961578369141, "learning_rate": 5e-05, "loss": 0.7253, "mean_token_accuracy": 0.8626749515533447, "num_tokens": 145910276.0, "step": 3842 }, { "epoch": 0.4888691006233304, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.290605068206787, "learning_rate": 5e-05, "loss": 0.7119, "mean_token_accuracy": 0.8661909103393555, "num_tokens": 145952945.0, "step": 3843 }, { "epoch": 0.48899631090192086, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.061069488525391, "learning_rate": 5e-05, "loss": 0.7232, "mean_token_accuracy": 0.8532006144523621, "num_tokens": 145987339.0, "step": 3844 }, { "epoch": 0.4891235211805114, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.290041446685791, "learning_rate": 5e-05, "loss": 0.7364, "mean_token_accuracy": 0.8589521646499634, "num_tokens": 146027447.0, "step": 3845 }, { "epoch": 0.4892507314591019, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 7.013787269592285, "learning_rate": 5e-05, "loss": 0.7086, "mean_token_accuracy": 0.8620282411575317, "num_tokens": 146064192.0, "step": 3846 }, { "epoch": 0.4893779417376924, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.347706317901611, "learning_rate": 5e-05, "loss": 0.7395, "mean_token_accuracy": 0.8596657514572144, "num_tokens": 146103468.0, "step": 3847 }, { "epoch": 0.4895051520162829, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.924579620361328, "learning_rate": 5e-05, "loss": 0.6828, "mean_token_accuracy": 0.8675087094306946, "num_tokens": 146142499.0, "step": 3848 }, { "epoch": 0.48963236229487345, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.471388816833496, "learning_rate": 5e-05, "loss": 0.7513, "mean_token_accuracy": 0.853925883769989, "num_tokens": 146177538.0, "step": 3849 }, { "epoch": 0.4897595725734639, "ewc_loss": 0.28125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.752877235412598, "learning_rate": 5e-05, "loss": 0.6854, "mean_token_accuracy": 0.8663297891616821, "num_tokens": 146218259.0, "step": 3850 }, { "epoch": 0.48988678285205445, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.713188171386719, "learning_rate": 5e-05, "loss": 0.7619, "mean_token_accuracy": 0.8591150641441345, "num_tokens": 146256630.0, "step": 3851 }, { "epoch": 0.490013993130645, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 6.445531368255615, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.8407304286956787, "num_tokens": 146290891.0, "step": 3852 }, { "epoch": 0.49014120340923545, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.178022384643555, "learning_rate": 5e-05, "loss": 0.8111, "mean_token_accuracy": 0.8518276810646057, "num_tokens": 146327148.0, "step": 3853 }, { "epoch": 0.490268413687826, "ewc_loss": 0.2412109375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00021266937255859375, "grad_norm": 5.91344690322876, "learning_rate": 5e-05, "loss": 0.6494, "mean_token_accuracy": 0.8598862886428833, "num_tokens": 146367932.0, "step": 3854 }, { "epoch": 0.4903956239664165, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 9.018823623657227, "learning_rate": 5e-05, "loss": 0.874, "mean_token_accuracy": 0.8642740845680237, "num_tokens": 146407639.0, "step": 3855 }, { "epoch": 0.490522834245007, "ewc_loss": 0.2373046875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00020885467529296875, "grad_norm": 5.342635631561279, "learning_rate": 5e-05, "loss": 0.6847, "mean_token_accuracy": 0.8491911292076111, "num_tokens": 146442974.0, "step": 3856 }, { "epoch": 0.4906500445235975, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.187165260314941, "learning_rate": 5e-05, "loss": 1.0315, "mean_token_accuracy": 0.865294337272644, "num_tokens": 146482510.0, "step": 3857 }, { "epoch": 0.49077725480218803, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 5.949626445770264, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.8717616200447083, "num_tokens": 146522616.0, "step": 3858 }, { "epoch": 0.4909044650807785, "ewc_loss": 0.765625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 10.67564868927002, "learning_rate": 5e-05, "loss": 1.1538, "mean_token_accuracy": 0.8664812445640564, "num_tokens": 146558226.0, "step": 3859 }, { "epoch": 0.49103167535936904, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 8.302165031433105, "learning_rate": 5e-05, "loss": 1.1653, "mean_token_accuracy": 0.8536262512207031, "num_tokens": 146601873.0, "step": 3860 }, { "epoch": 0.49115888563795956, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.894903182983398, "learning_rate": 5e-05, "loss": 0.9463, "mean_token_accuracy": 0.8593011498451233, "num_tokens": 146645284.0, "step": 3861 }, { "epoch": 0.49128609591655004, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 9.478113174438477, "learning_rate": 5e-05, "loss": 1.2063, "mean_token_accuracy": 0.8602756261825562, "num_tokens": 146680374.0, "step": 3862 }, { "epoch": 0.49141330619514056, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.9812798500061035, "learning_rate": 5e-05, "loss": 0.8107, "mean_token_accuracy": 0.8518326282501221, "num_tokens": 146721915.0, "step": 3863 }, { "epoch": 0.4915405164737311, "ewc_loss": 0.90625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00087738037109375, "grad_norm": 11.582493782043457, "learning_rate": 5e-05, "loss": 1.2962, "mean_token_accuracy": 0.8683161735534668, "num_tokens": 146758981.0, "step": 3864 }, { "epoch": 0.49166772675232157, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 8.822515487670898, "learning_rate": 5e-05, "loss": 1.1657, "mean_token_accuracy": 0.853705644607544, "num_tokens": 146795606.0, "step": 3865 }, { "epoch": 0.4917949370309121, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 11.280485153198242, "learning_rate": 5e-05, "loss": 1.074, "mean_token_accuracy": 0.8687444925308228, "num_tokens": 146829681.0, "step": 3866 }, { "epoch": 0.4919221473095026, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 9.31602668762207, "learning_rate": 5e-05, "loss": 1.2156, "mean_token_accuracy": 0.8733264207839966, "num_tokens": 146874142.0, "step": 3867 }, { "epoch": 0.4920493575880931, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 11.016276359558105, "learning_rate": 5e-05, "loss": 1.2349, "mean_token_accuracy": 0.8257796764373779, "num_tokens": 146911414.0, "step": 3868 }, { "epoch": 0.4921765678666836, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.254560470581055, "learning_rate": 5e-05, "loss": 1.0742, "mean_token_accuracy": 0.8401503562927246, "num_tokens": 146956376.0, "step": 3869 }, { "epoch": 0.49230377814527415, "ewc_loss": 0.78125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00075531005859375, "grad_norm": 11.44902229309082, "learning_rate": 5e-05, "loss": 1.2461, "mean_token_accuracy": 0.8446018695831299, "num_tokens": 147001539.0, "step": 3870 }, { "epoch": 0.4924309884238646, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.15372371673584, "learning_rate": 5e-05, "loss": 1.0064, "mean_token_accuracy": 0.8461579084396362, "num_tokens": 147043782.0, "step": 3871 }, { "epoch": 0.49255819870245515, "ewc_loss": 0.90625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00087738037109375, "grad_norm": 12.365311622619629, "learning_rate": 5e-05, "loss": 1.3821, "mean_token_accuracy": 0.83869469165802, "num_tokens": 147077431.0, "step": 3872 }, { "epoch": 0.4926854089810457, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 10.092117309570312, "learning_rate": 5e-05, "loss": 1.1515, "mean_token_accuracy": 0.8683633208274841, "num_tokens": 147113325.0, "step": 3873 }, { "epoch": 0.49281261925963615, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 11.350399017333984, "learning_rate": 5e-05, "loss": 1.2136, "mean_token_accuracy": 0.8383145928382874, "num_tokens": 147158363.0, "step": 3874 }, { "epoch": 0.4929398295382267, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.17457103729248, "learning_rate": 5e-05, "loss": 1.1993, "mean_token_accuracy": 0.8468407392501831, "num_tokens": 147195868.0, "step": 3875 }, { "epoch": 0.4930670398168172, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.484929084777832, "learning_rate": 5e-05, "loss": 1.1547, "mean_token_accuracy": 0.8361626267433167, "num_tokens": 147237782.0, "step": 3876 }, { "epoch": 0.4931942500954077, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.421929359436035, "learning_rate": 5e-05, "loss": 1.0613, "mean_token_accuracy": 0.8478819131851196, "num_tokens": 147274709.0, "step": 3877 }, { "epoch": 0.4933214603739982, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.967914581298828, "learning_rate": 5e-05, "loss": 1.0281, "mean_token_accuracy": 0.8628612756729126, "num_tokens": 147312261.0, "step": 3878 }, { "epoch": 0.49344867065258874, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.75665283203125, "learning_rate": 5e-05, "loss": 0.9464, "mean_token_accuracy": 0.8582640886306763, "num_tokens": 147352290.0, "step": 3879 }, { "epoch": 0.4935758809311792, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.774215698242188, "learning_rate": 5e-05, "loss": 1.0231, "mean_token_accuracy": 0.8594875335693359, "num_tokens": 147390558.0, "step": 3880 }, { "epoch": 0.49370309120976974, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.130695343017578, "learning_rate": 5e-05, "loss": 0.9053, "mean_token_accuracy": 0.8508975505828857, "num_tokens": 147429597.0, "step": 3881 }, { "epoch": 0.49383030148836027, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.090019226074219, "learning_rate": 5e-05, "loss": 1.0594, "mean_token_accuracy": 0.8504244089126587, "num_tokens": 147472302.0, "step": 3882 }, { "epoch": 0.4939575117669508, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.939301013946533, "learning_rate": 5e-05, "loss": 0.8268, "mean_token_accuracy": 0.869330883026123, "num_tokens": 147502846.0, "step": 3883 }, { "epoch": 0.49408472204554127, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.657686233520508, "learning_rate": 5e-05, "loss": 1.0893, "mean_token_accuracy": 0.8645284175872803, "num_tokens": 147542351.0, "step": 3884 }, { "epoch": 0.4942119323241318, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.608898162841797, "learning_rate": 5e-05, "loss": 0.9257, "mean_token_accuracy": 0.8639818429946899, "num_tokens": 147576912.0, "step": 3885 }, { "epoch": 0.4943391426027223, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.497407913208008, "learning_rate": 5e-05, "loss": 1.1158, "mean_token_accuracy": 0.8538333773612976, "num_tokens": 147614687.0, "step": 3886 }, { "epoch": 0.4944663528813128, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.217376708984375, "learning_rate": 5e-05, "loss": 0.9935, "mean_token_accuracy": 0.8655147552490234, "num_tokens": 147649048.0, "step": 3887 }, { "epoch": 0.4945935631599033, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.510140419006348, "learning_rate": 5e-05, "loss": 1.0234, "mean_token_accuracy": 0.848867654800415, "num_tokens": 147689337.0, "step": 3888 }, { "epoch": 0.49472077343849385, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.877246856689453, "learning_rate": 5e-05, "loss": 0.9481, "mean_token_accuracy": 0.8592828512191772, "num_tokens": 147725040.0, "step": 3889 }, { "epoch": 0.4948479837170843, "ewc_loss": 0.5, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.732096672058105, "learning_rate": 5e-05, "loss": 0.9211, "mean_token_accuracy": 0.8581271171569824, "num_tokens": 147761442.0, "step": 3890 }, { "epoch": 0.49497519399567486, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.189253807067871, "learning_rate": 5e-05, "loss": 0.8567, "mean_token_accuracy": 0.853313148021698, "num_tokens": 147795421.0, "step": 3891 }, { "epoch": 0.4951024042742654, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.518717765808105, "learning_rate": 5e-05, "loss": 0.8893, "mean_token_accuracy": 0.8582707643508911, "num_tokens": 147833591.0, "step": 3892 }, { "epoch": 0.49522961455285586, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.433895587921143, "learning_rate": 5e-05, "loss": 0.8028, "mean_token_accuracy": 0.8470838069915771, "num_tokens": 147870707.0, "step": 3893 }, { "epoch": 0.4953568248314464, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.792683601379395, "learning_rate": 5e-05, "loss": 0.904, "mean_token_accuracy": 0.8585958480834961, "num_tokens": 147905231.0, "step": 3894 }, { "epoch": 0.4954840351100369, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.983715534210205, "learning_rate": 5e-05, "loss": 0.7392, "mean_token_accuracy": 0.8606398105621338, "num_tokens": 147945890.0, "step": 3895 }, { "epoch": 0.4956112453886274, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.381336212158203, "learning_rate": 5e-05, "loss": 0.973, "mean_token_accuracy": 0.8598148822784424, "num_tokens": 147984572.0, "step": 3896 }, { "epoch": 0.4957384556672179, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.15285062789917, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8704436421394348, "num_tokens": 148020059.0, "step": 3897 }, { "epoch": 0.49586566594580844, "ewc_loss": 0.625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.785735130310059, "learning_rate": 5e-05, "loss": 0.9945, "mean_token_accuracy": 0.872094988822937, "num_tokens": 148058442.0, "step": 3898 }, { "epoch": 0.4959928762243989, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.172218322753906, "learning_rate": 5e-05, "loss": 0.8458, "mean_token_accuracy": 0.8663766384124756, "num_tokens": 148093710.0, "step": 3899 }, { "epoch": 0.49612008650298944, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.90795612335205, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.8806476593017578, "num_tokens": 148135763.0, "step": 3900 }, { "epoch": 0.49624729678157997, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.539474487304688, "learning_rate": 5e-05, "loss": 0.871, "mean_token_accuracy": 0.865877091884613, "num_tokens": 148169065.0, "step": 3901 }, { "epoch": 0.49637450706017044, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.613658905029297, "learning_rate": 5e-05, "loss": 0.8705, "mean_token_accuracy": 0.8461799621582031, "num_tokens": 148205779.0, "step": 3902 }, { "epoch": 0.496501717338761, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.40600872039795, "learning_rate": 5e-05, "loss": 0.8801, "mean_token_accuracy": 0.8428199291229248, "num_tokens": 148241893.0, "step": 3903 }, { "epoch": 0.4966289276173515, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.98982572555542, "learning_rate": 5e-05, "loss": 0.7348, "mean_token_accuracy": 0.8635683059692383, "num_tokens": 148275384.0, "step": 3904 }, { "epoch": 0.496756137895942, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.433635711669922, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8576395511627197, "num_tokens": 148311248.0, "step": 3905 }, { "epoch": 0.4968833481745325, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.726280212402344, "learning_rate": 5e-05, "loss": 0.7356, "mean_token_accuracy": 0.8568955659866333, "num_tokens": 148347717.0, "step": 3906 }, { "epoch": 0.49701055845312303, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.835309982299805, "learning_rate": 5e-05, "loss": 0.9638, "mean_token_accuracy": 0.840255618095398, "num_tokens": 148393395.0, "step": 3907 }, { "epoch": 0.4971377687317135, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.838654518127441, "learning_rate": 5e-05, "loss": 0.7244, "mean_token_accuracy": 0.867185115814209, "num_tokens": 148427843.0, "step": 3908 }, { "epoch": 0.49726497901030403, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.941121101379395, "learning_rate": 5e-05, "loss": 0.9638, "mean_token_accuracy": 0.8481407165527344, "num_tokens": 148470813.0, "step": 3909 }, { "epoch": 0.49739218928889456, "ewc_loss": 0.375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.502889156341553, "learning_rate": 5e-05, "loss": 0.8048, "mean_token_accuracy": 0.8563851118087769, "num_tokens": 148504491.0, "step": 3910 }, { "epoch": 0.49751939956748503, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.175763130187988, "learning_rate": 5e-05, "loss": 0.8821, "mean_token_accuracy": 0.8507798314094543, "num_tokens": 148545934.0, "step": 3911 }, { "epoch": 0.49764660984607556, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.881852149963379, "learning_rate": 5e-05, "loss": 0.7954, "mean_token_accuracy": 0.8655422925949097, "num_tokens": 148583521.0, "step": 3912 }, { "epoch": 0.4977738201246661, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.438239097595215, "learning_rate": 5e-05, "loss": 0.8127, "mean_token_accuracy": 0.8531104922294617, "num_tokens": 148621484.0, "step": 3913 }, { "epoch": 0.49790103040325656, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.723793983459473, "learning_rate": 5e-05, "loss": 0.8273, "mean_token_accuracy": 0.8481558561325073, "num_tokens": 148661198.0, "step": 3914 }, { "epoch": 0.4980282406818471, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.4367594718933105, "learning_rate": 5e-05, "loss": 0.7825, "mean_token_accuracy": 0.8578082323074341, "num_tokens": 148702620.0, "step": 3915 }, { "epoch": 0.4981554509604376, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.2865376472473145, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8528763055801392, "num_tokens": 148739991.0, "step": 3916 }, { "epoch": 0.4982826612390281, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.705586910247803, "learning_rate": 5e-05, "loss": 0.7267, "mean_token_accuracy": 0.8737963438034058, "num_tokens": 148780520.0, "step": 3917 }, { "epoch": 0.4984098715176186, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.857086181640625, "learning_rate": 5e-05, "loss": 0.7461, "mean_token_accuracy": 0.8484635353088379, "num_tokens": 148817604.0, "step": 3918 }, { "epoch": 0.49853708179620915, "ewc_loss": 0.375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.036307334899902, "learning_rate": 5e-05, "loss": 0.8022, "mean_token_accuracy": 0.8546915054321289, "num_tokens": 148858028.0, "step": 3919 }, { "epoch": 0.4986642920747996, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.483093738555908, "learning_rate": 5e-05, "loss": 0.6835, "mean_token_accuracy": 0.8623253107070923, "num_tokens": 148895458.0, "step": 3920 }, { "epoch": 0.49879150235339015, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.518733024597168, "learning_rate": 5e-05, "loss": 0.8588, "mean_token_accuracy": 0.8535177707672119, "num_tokens": 148933305.0, "step": 3921 }, { "epoch": 0.4989187126319807, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00023937225341796875, "grad_norm": 6.204134941101074, "learning_rate": 5e-05, "loss": 0.656, "mean_token_accuracy": 0.8683431148529053, "num_tokens": 148970587.0, "step": 3922 }, { "epoch": 0.49904592291057115, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.98326587677002, "learning_rate": 5e-05, "loss": 0.9316, "mean_token_accuracy": 0.8526427745819092, "num_tokens": 149004874.0, "step": 3923 }, { "epoch": 0.4991731331891617, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.46326208114624, "learning_rate": 5e-05, "loss": 0.6952, "mean_token_accuracy": 0.8639612197875977, "num_tokens": 149045169.0, "step": 3924 }, { "epoch": 0.4993003434677522, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.901787757873535, "learning_rate": 5e-05, "loss": 1.0303, "mean_token_accuracy": 0.8282463550567627, "num_tokens": 149081554.0, "step": 3925 }, { "epoch": 0.4994275537463427, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.200409412384033, "learning_rate": 5e-05, "loss": 0.8206, "mean_token_accuracy": 0.8497127890586853, "num_tokens": 149124500.0, "step": 3926 }, { "epoch": 0.4995547640249332, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.20024585723877, "learning_rate": 5e-05, "loss": 0.8847, "mean_token_accuracy": 0.8530287146568298, "num_tokens": 149162522.0, "step": 3927 }, { "epoch": 0.49968197430352373, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.6291279792785645, "learning_rate": 5e-05, "loss": 0.811, "mean_token_accuracy": 0.8599311113357544, "num_tokens": 149198083.0, "step": 3928 }, { "epoch": 0.4998091845821142, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.614968299865723, "learning_rate": 5e-05, "loss": 0.8383, "mean_token_accuracy": 0.8429332971572876, "num_tokens": 149234786.0, "step": 3929 }, { "epoch": 0.49993639486070474, "ewc_loss": 0.375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.536492824554443, "learning_rate": 5e-05, "loss": 0.7377, "mean_token_accuracy": 0.8752236366271973, "num_tokens": 149269859.0, "step": 3930 }, { "epoch": 0.5000636051392953, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.535834789276123, "learning_rate": 5e-05, "loss": 0.8485, "mean_token_accuracy": 0.8367629051208496, "num_tokens": 149311209.0, "step": 3931 }, { "epoch": 0.5001908154178858, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.3003010749816895, "learning_rate": 5e-05, "loss": 0.7707, "mean_token_accuracy": 0.8524670004844666, "num_tokens": 149351283.0, "step": 3932 }, { "epoch": 0.5003180256964763, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.523755073547363, "learning_rate": 5e-05, "loss": 0.7681, "mean_token_accuracy": 0.8576830625534058, "num_tokens": 149391068.0, "step": 3933 }, { "epoch": 0.5004452359750667, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.295042514801025, "learning_rate": 5e-05, "loss": 0.7766, "mean_token_accuracy": 0.8489475250244141, "num_tokens": 149435078.0, "step": 3934 }, { "epoch": 0.5005724462536573, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.252079963684082, "learning_rate": 5e-05, "loss": 0.8014, "mean_token_accuracy": 0.8445948958396912, "num_tokens": 149481336.0, "step": 3935 }, { "epoch": 0.5006996565322478, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.581052780151367, "learning_rate": 5e-05, "loss": 0.7606, "mean_token_accuracy": 0.8592007160186768, "num_tokens": 149523278.0, "step": 3936 }, { "epoch": 0.5008268668108383, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.840733528137207, "learning_rate": 5e-05, "loss": 0.6842, "mean_token_accuracy": 0.8758722543716431, "num_tokens": 149563011.0, "step": 3937 }, { "epoch": 0.5009540770894289, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.962706565856934, "learning_rate": 5e-05, "loss": 0.7955, "mean_token_accuracy": 0.8559319376945496, "num_tokens": 149606609.0, "step": 3938 }, { "epoch": 0.5010812873680194, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.5339884757995605, "learning_rate": 5e-05, "loss": 0.6661, "mean_token_accuracy": 0.8681092262268066, "num_tokens": 149647012.0, "step": 3939 }, { "epoch": 0.5012084976466098, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.252110481262207, "learning_rate": 5e-05, "loss": 0.8077, "mean_token_accuracy": 0.8616220951080322, "num_tokens": 149680153.0, "step": 3940 }, { "epoch": 0.5013357079252003, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.408256530761719, "learning_rate": 5e-05, "loss": 0.6845, "mean_token_accuracy": 0.8642097115516663, "num_tokens": 149717080.0, "step": 3941 }, { "epoch": 0.5014629182037909, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.396129608154297, "learning_rate": 5e-05, "loss": 0.8475, "mean_token_accuracy": 0.8621698617935181, "num_tokens": 149755934.0, "step": 3942 }, { "epoch": 0.5015901284823814, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.480138301849365, "learning_rate": 5e-05, "loss": 0.7035, "mean_token_accuracy": 0.8590478301048279, "num_tokens": 149792714.0, "step": 3943 }, { "epoch": 0.5017173387609719, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.336936950683594, "learning_rate": 5e-05, "loss": 0.8808, "mean_token_accuracy": 0.8499019145965576, "num_tokens": 149828647.0, "step": 3944 }, { "epoch": 0.5018445490395624, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.7391276359558105, "learning_rate": 5e-05, "loss": 0.7388, "mean_token_accuracy": 0.8580365180969238, "num_tokens": 149864303.0, "step": 3945 }, { "epoch": 0.5019717593181529, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.035974502563477, "learning_rate": 5e-05, "loss": 0.8776, "mean_token_accuracy": 0.8439244627952576, "num_tokens": 149902154.0, "step": 3946 }, { "epoch": 0.5020989695967434, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.056051731109619, "learning_rate": 5e-05, "loss": 0.7136, "mean_token_accuracy": 0.8655462861061096, "num_tokens": 149930902.0, "step": 3947 }, { "epoch": 0.5022261798753339, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.667234420776367, "learning_rate": 5e-05, "loss": 0.811, "mean_token_accuracy": 0.8566522598266602, "num_tokens": 149972033.0, "step": 3948 }, { "epoch": 0.5023533901539244, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.235809803009033, "learning_rate": 5e-05, "loss": 0.7298, "mean_token_accuracy": 0.8676491975784302, "num_tokens": 150007278.0, "step": 3949 }, { "epoch": 0.502480600432515, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.491835117340088, "learning_rate": 5e-05, "loss": 0.7871, "mean_token_accuracy": 0.8545984029769897, "num_tokens": 150054512.0, "step": 3950 }, { "epoch": 0.5026078107111055, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.205653190612793, "learning_rate": 5e-05, "loss": 0.7616, "mean_token_accuracy": 0.85356605052948, "num_tokens": 150094614.0, "step": 3951 }, { "epoch": 0.5027350209896959, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.52871561050415, "learning_rate": 5e-05, "loss": 0.7676, "mean_token_accuracy": 0.8588463664054871, "num_tokens": 150133011.0, "step": 3952 }, { "epoch": 0.5028622312682864, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.013885974884033, "learning_rate": 5e-05, "loss": 0.7467, "mean_token_accuracy": 0.8555452823638916, "num_tokens": 150170041.0, "step": 3953 }, { "epoch": 0.502989441546877, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.738324165344238, "learning_rate": 5e-05, "loss": 0.7816, "mean_token_accuracy": 0.8555334210395813, "num_tokens": 150206034.0, "step": 3954 }, { "epoch": 0.5031166518254675, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.7376532554626465, "learning_rate": 5e-05, "loss": 0.7213, "mean_token_accuracy": 0.8566697835922241, "num_tokens": 150241601.0, "step": 3955 }, { "epoch": 0.503243862104058, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.053609848022461, "learning_rate": 5e-05, "loss": 0.7754, "mean_token_accuracy": 0.8674764633178711, "num_tokens": 150279219.0, "step": 3956 }, { "epoch": 0.5033710723826486, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.456376552581787, "learning_rate": 5e-05, "loss": 0.7109, "mean_token_accuracy": 0.85612553358078, "num_tokens": 150323160.0, "step": 3957 }, { "epoch": 0.5034982826612391, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.381887435913086, "learning_rate": 5e-05, "loss": 0.8356, "mean_token_accuracy": 0.8598313331604004, "num_tokens": 150363788.0, "step": 3958 }, { "epoch": 0.5036254929398295, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.317542552947998, "learning_rate": 5e-05, "loss": 0.7228, "mean_token_accuracy": 0.8517136573791504, "num_tokens": 150395481.0, "step": 3959 }, { "epoch": 0.50375270321842, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.537490844726562, "learning_rate": 5e-05, "loss": 0.8508, "mean_token_accuracy": 0.863257884979248, "num_tokens": 150425726.0, "step": 3960 }, { "epoch": 0.5038799134970106, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.553652763366699, "learning_rate": 5e-05, "loss": 0.7674, "mean_token_accuracy": 0.8520521521568298, "num_tokens": 150463073.0, "step": 3961 }, { "epoch": 0.5040071237756011, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.157472610473633, "learning_rate": 5e-05, "loss": 0.8822, "mean_token_accuracy": 0.8445399403572083, "num_tokens": 150502938.0, "step": 3962 }, { "epoch": 0.5041343340541916, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.2093071937561035, "learning_rate": 5e-05, "loss": 0.7847, "mean_token_accuracy": 0.8537712097167969, "num_tokens": 150538601.0, "step": 3963 }, { "epoch": 0.5042615443327821, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.257690906524658, "learning_rate": 5e-05, "loss": 0.7834, "mean_token_accuracy": 0.8541643023490906, "num_tokens": 150573745.0, "step": 3964 }, { "epoch": 0.5043887546113726, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.967545986175537, "learning_rate": 5e-05, "loss": 0.8619, "mean_token_accuracy": 0.844470202922821, "num_tokens": 150613029.0, "step": 3965 }, { "epoch": 0.5045159648899631, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.382265090942383, "learning_rate": 5e-05, "loss": 0.691, "mean_token_accuracy": 0.8653520345687866, "num_tokens": 150649773.0, "step": 3966 }, { "epoch": 0.5046431751685536, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.735715866088867, "learning_rate": 5e-05, "loss": 0.8754, "mean_token_accuracy": 0.860586941242218, "num_tokens": 150694764.0, "step": 3967 }, { "epoch": 0.5047703854471441, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 5.855828285217285, "learning_rate": 5e-05, "loss": 0.7078, "mean_token_accuracy": 0.8504902124404907, "num_tokens": 150735233.0, "step": 3968 }, { "epoch": 0.5048975957257347, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.424222946166992, "learning_rate": 5e-05, "loss": 0.9223, "mean_token_accuracy": 0.8771514296531677, "num_tokens": 150769596.0, "step": 3969 }, { "epoch": 0.5050248060043252, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.11698055267334, "learning_rate": 5e-05, "loss": 0.7571, "mean_token_accuracy": 0.8502544164657593, "num_tokens": 150806096.0, "step": 3970 }, { "epoch": 0.5051520162829156, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.433370590209961, "learning_rate": 5e-05, "loss": 1.0136, "mean_token_accuracy": 0.8563297986984253, "num_tokens": 150844721.0, "step": 3971 }, { "epoch": 0.5052792265615061, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.382039546966553, "learning_rate": 5e-05, "loss": 0.8914, "mean_token_accuracy": 0.8608800172805786, "num_tokens": 150885711.0, "step": 3972 }, { "epoch": 0.5054064368400967, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.155115127563477, "learning_rate": 5e-05, "loss": 0.9308, "mean_token_accuracy": 0.8550628423690796, "num_tokens": 150924993.0, "step": 3973 }, { "epoch": 0.5055336471186872, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.443465232849121, "learning_rate": 5e-05, "loss": 0.9352, "mean_token_accuracy": 0.8678324818611145, "num_tokens": 150956139.0, "step": 3974 }, { "epoch": 0.5056608573972777, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.724009037017822, "learning_rate": 5e-05, "loss": 0.8496, "mean_token_accuracy": 0.8396763801574707, "num_tokens": 150995816.0, "step": 3975 }, { "epoch": 0.5057880676758683, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.01779556274414, "learning_rate": 5e-05, "loss": 1.0138, "mean_token_accuracy": 0.8410779237747192, "num_tokens": 151035191.0, "step": 3976 }, { "epoch": 0.5059152779544587, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.749554634094238, "learning_rate": 5e-05, "loss": 0.8503, "mean_token_accuracy": 0.8393685817718506, "num_tokens": 151069625.0, "step": 3977 }, { "epoch": 0.5060424882330492, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.405652046203613, "learning_rate": 5e-05, "loss": 0.9636, "mean_token_accuracy": 0.8464674353599548, "num_tokens": 151105999.0, "step": 3978 }, { "epoch": 0.5061696985116397, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.584484577178955, "learning_rate": 5e-05, "loss": 0.8679, "mean_token_accuracy": 0.8600490093231201, "num_tokens": 151148712.0, "step": 3979 }, { "epoch": 0.5062969087902303, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.219644069671631, "learning_rate": 5e-05, "loss": 0.7456, "mean_token_accuracy": 0.8715112209320068, "num_tokens": 151182842.0, "step": 3980 }, { "epoch": 0.5064241190688208, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.242144584655762, "learning_rate": 5e-05, "loss": 0.8972, "mean_token_accuracy": 0.8561188578605652, "num_tokens": 151223731.0, "step": 3981 }, { "epoch": 0.5065513293474113, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.707656383514404, "learning_rate": 5e-05, "loss": 0.7582, "mean_token_accuracy": 0.8534850478172302, "num_tokens": 151259711.0, "step": 3982 }, { "epoch": 0.5066785396260017, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.348043441772461, "learning_rate": 5e-05, "loss": 0.9263, "mean_token_accuracy": 0.8495147228240967, "num_tokens": 151301486.0, "step": 3983 }, { "epoch": 0.5068057499045923, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.8158416748046875, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8564876317977905, "num_tokens": 151339068.0, "step": 3984 }, { "epoch": 0.5069329601831828, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.121177673339844, "learning_rate": 5e-05, "loss": 0.9055, "mean_token_accuracy": 0.8468742370605469, "num_tokens": 151380694.0, "step": 3985 }, { "epoch": 0.5070601704617733, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.012855529785156, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.8609060049057007, "num_tokens": 151416522.0, "step": 3986 }, { "epoch": 0.5071873807403638, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.8922858238220215, "learning_rate": 5e-05, "loss": 0.8903, "mean_token_accuracy": 0.8407642841339111, "num_tokens": 151453786.0, "step": 3987 }, { "epoch": 0.5073145910189544, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.993671894073486, "learning_rate": 5e-05, "loss": 0.7822, "mean_token_accuracy": 0.8561809062957764, "num_tokens": 151492155.0, "step": 3988 }, { "epoch": 0.5074418012975448, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.9263153076171875, "learning_rate": 5e-05, "loss": 0.8066, "mean_token_accuracy": 0.8649019002914429, "num_tokens": 151534979.0, "step": 3989 }, { "epoch": 0.5075690115761353, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.765254497528076, "learning_rate": 5e-05, "loss": 0.7788, "mean_token_accuracy": 0.8510993123054504, "num_tokens": 151573670.0, "step": 3990 }, { "epoch": 0.5076962218547258, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.172649383544922, "learning_rate": 5e-05, "loss": 0.77, "mean_token_accuracy": 0.8830465078353882, "num_tokens": 151605232.0, "step": 3991 }, { "epoch": 0.5078234321333164, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.466771125793457, "learning_rate": 5e-05, "loss": 0.7902, "mean_token_accuracy": 0.8402187824249268, "num_tokens": 151648696.0, "step": 3992 }, { "epoch": 0.5079506424119069, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.531280517578125, "learning_rate": 5e-05, "loss": 0.873, "mean_token_accuracy": 0.8655576705932617, "num_tokens": 151684203.0, "step": 3993 }, { "epoch": 0.5080778526904974, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.313719749450684, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8509069681167603, "num_tokens": 151716749.0, "step": 3994 }, { "epoch": 0.5082050629690879, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.757735252380371, "learning_rate": 5e-05, "loss": 0.974, "mean_token_accuracy": 0.8455085754394531, "num_tokens": 151752567.0, "step": 3995 }, { "epoch": 0.5083322732476784, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.462209224700928, "learning_rate": 5e-05, "loss": 0.6876, "mean_token_accuracy": 0.8778824806213379, "num_tokens": 151794201.0, "step": 3996 }, { "epoch": 0.5084594835262689, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.691902160644531, "learning_rate": 5e-05, "loss": 0.9542, "mean_token_accuracy": 0.860090970993042, "num_tokens": 151833731.0, "step": 3997 }, { "epoch": 0.5085866938048594, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 6.919907569885254, "learning_rate": 5e-05, "loss": 0.8151, "mean_token_accuracy": 0.8584925532341003, "num_tokens": 151872738.0, "step": 3998 }, { "epoch": 0.50871390408345, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.182764053344727, "learning_rate": 5e-05, "loss": 0.8801, "mean_token_accuracy": 0.8661578893661499, "num_tokens": 151916801.0, "step": 3999 }, { "epoch": 0.5088411143620405, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.413613796234131, "learning_rate": 5e-05, "loss": 0.8881, "mean_token_accuracy": 0.8542905449867249, "num_tokens": 151949661.0, "step": 4000 }, { "epoch": 0.5089683246406309, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.5219221115112305, "learning_rate": 5e-05, "loss": 0.8525, "mean_token_accuracy": 0.8493940830230713, "num_tokens": 151986719.0, "step": 4001 }, { "epoch": 0.5090955349192214, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.681314945220947, "learning_rate": 5e-05, "loss": 0.8783, "mean_token_accuracy": 0.8533626198768616, "num_tokens": 152023921.0, "step": 4002 }, { "epoch": 0.509222745197812, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.118129253387451, "learning_rate": 5e-05, "loss": 0.7582, "mean_token_accuracy": 0.8654708862304688, "num_tokens": 152061246.0, "step": 4003 }, { "epoch": 0.5093499554764025, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.726723670959473, "learning_rate": 5e-05, "loss": 0.8554, "mean_token_accuracy": 0.8480329513549805, "num_tokens": 152103299.0, "step": 4004 }, { "epoch": 0.509477165754993, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.0474419593811035, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.845890998840332, "num_tokens": 152148576.0, "step": 4005 }, { "epoch": 0.5096043760335836, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.643147945404053, "learning_rate": 5e-05, "loss": 0.8131, "mean_token_accuracy": 0.8547725677490234, "num_tokens": 152189221.0, "step": 4006 }, { "epoch": 0.5097315863121741, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.024421215057373, "learning_rate": 5e-05, "loss": 0.7689, "mean_token_accuracy": 0.8520537614822388, "num_tokens": 152232732.0, "step": 4007 }, { "epoch": 0.5098587965907645, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.641073226928711, "learning_rate": 5e-05, "loss": 0.7812, "mean_token_accuracy": 0.8615847826004028, "num_tokens": 152268715.0, "step": 4008 }, { "epoch": 0.509986006869355, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.937160968780518, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8514917492866516, "num_tokens": 152311141.0, "step": 4009 }, { "epoch": 0.5101132171479456, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.741173267364502, "learning_rate": 5e-05, "loss": 0.7601, "mean_token_accuracy": 0.8725900053977966, "num_tokens": 152348892.0, "step": 4010 }, { "epoch": 0.5102404274265361, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.7885847091674805, "learning_rate": 5e-05, "loss": 0.7262, "mean_token_accuracy": 0.8609790205955505, "num_tokens": 152383300.0, "step": 4011 }, { "epoch": 0.5103676377051266, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.882287502288818, "learning_rate": 5e-05, "loss": 0.8273, "mean_token_accuracy": 0.853249192237854, "num_tokens": 152424404.0, "step": 4012 }, { "epoch": 0.5104948479837171, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.650116443634033, "learning_rate": 5e-05, "loss": 0.6772, "mean_token_accuracy": 0.8755818605422974, "num_tokens": 152465226.0, "step": 4013 }, { "epoch": 0.5106220582623076, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.09453296661377, "learning_rate": 5e-05, "loss": 0.8578, "mean_token_accuracy": 0.8493664264678955, "num_tokens": 152507819.0, "step": 4014 }, { "epoch": 0.5107492685408981, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.466670989990234, "learning_rate": 5e-05, "loss": 0.6578, "mean_token_accuracy": 0.878494381904602, "num_tokens": 152548194.0, "step": 4015 }, { "epoch": 0.5108764788194886, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.339978218078613, "learning_rate": 5e-05, "loss": 0.8393, "mean_token_accuracy": 0.8585996627807617, "num_tokens": 152580255.0, "step": 4016 }, { "epoch": 0.5110036890980791, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.345107078552246, "learning_rate": 5e-05, "loss": 0.7243, "mean_token_accuracy": 0.8552565574645996, "num_tokens": 152619813.0, "step": 4017 }, { "epoch": 0.5111308993766697, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.52883243560791, "learning_rate": 5e-05, "loss": 0.925, "mean_token_accuracy": 0.8502062559127808, "num_tokens": 152655552.0, "step": 4018 }, { "epoch": 0.5112581096552602, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.450984001159668, "learning_rate": 5e-05, "loss": 0.7878, "mean_token_accuracy": 0.8446900248527527, "num_tokens": 152693695.0, "step": 4019 }, { "epoch": 0.5113853199338506, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.443646430969238, "learning_rate": 5e-05, "loss": 0.9173, "mean_token_accuracy": 0.854846179485321, "num_tokens": 152732917.0, "step": 4020 }, { "epoch": 0.5115125302124411, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.842155456542969, "learning_rate": 5e-05, "loss": 0.7516, "mean_token_accuracy": 0.8657494783401489, "num_tokens": 152768052.0, "step": 4021 }, { "epoch": 0.5116397404910317, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.900570392608643, "learning_rate": 5e-05, "loss": 0.9131, "mean_token_accuracy": 0.8434872031211853, "num_tokens": 152808106.0, "step": 4022 }, { "epoch": 0.5117669507696222, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.378754138946533, "learning_rate": 5e-05, "loss": 0.7575, "mean_token_accuracy": 0.8795372843742371, "num_tokens": 152849989.0, "step": 4023 }, { "epoch": 0.5118941610482127, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.232779502868652, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8622446060180664, "num_tokens": 152886599.0, "step": 4024 }, { "epoch": 0.5120213713268033, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.787139415740967, "learning_rate": 5e-05, "loss": 0.8694, "mean_token_accuracy": 0.851227343082428, "num_tokens": 152926383.0, "step": 4025 }, { "epoch": 0.5121485816053937, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.779253959655762, "learning_rate": 5e-05, "loss": 0.757, "mean_token_accuracy": 0.852180004119873, "num_tokens": 152966141.0, "step": 4026 }, { "epoch": 0.5122757918839842, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.991886615753174, "learning_rate": 5e-05, "loss": 0.8215, "mean_token_accuracy": 0.8624032139778137, "num_tokens": 152998217.0, "step": 4027 }, { "epoch": 0.5124030021625747, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.652388095855713, "learning_rate": 5e-05, "loss": 0.7951, "mean_token_accuracy": 0.8426429629325867, "num_tokens": 153032258.0, "step": 4028 }, { "epoch": 0.5125302124411653, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.003290176391602, "learning_rate": 5e-05, "loss": 0.8104, "mean_token_accuracy": 0.861113965511322, "num_tokens": 153075762.0, "step": 4029 }, { "epoch": 0.5126574227197558, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.701444149017334, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.852760910987854, "num_tokens": 153112571.0, "step": 4030 }, { "epoch": 0.5127846329983463, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.922887325286865, "learning_rate": 5e-05, "loss": 0.7919, "mean_token_accuracy": 0.8669931292533875, "num_tokens": 153153348.0, "step": 4031 }, { "epoch": 0.5129118432769367, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.7816596031188965, "learning_rate": 5e-05, "loss": 0.7012, "mean_token_accuracy": 0.868025004863739, "num_tokens": 153189616.0, "step": 4032 }, { "epoch": 0.5130390535555273, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.8468098640441895, "learning_rate": 5e-05, "loss": 0.886, "mean_token_accuracy": 0.8400037288665771, "num_tokens": 153230879.0, "step": 4033 }, { "epoch": 0.5131662638341178, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.844109058380127, "learning_rate": 5e-05, "loss": 0.7647, "mean_token_accuracy": 0.8520430326461792, "num_tokens": 153275234.0, "step": 4034 }, { "epoch": 0.5132934741127083, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.768515586853027, "learning_rate": 5e-05, "loss": 0.7779, "mean_token_accuracy": 0.8692141175270081, "num_tokens": 153312606.0, "step": 4035 }, { "epoch": 0.5134206843912988, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.901995658874512, "learning_rate": 5e-05, "loss": 0.7857, "mean_token_accuracy": 0.848098874092102, "num_tokens": 153352370.0, "step": 4036 }, { "epoch": 0.5135478946698894, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.6990203857421875, "learning_rate": 5e-05, "loss": 0.7962, "mean_token_accuracy": 0.8598631024360657, "num_tokens": 153393030.0, "step": 4037 }, { "epoch": 0.5136751049484798, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.929161071777344, "learning_rate": 5e-05, "loss": 0.7634, "mean_token_accuracy": 0.8557729721069336, "num_tokens": 153423831.0, "step": 4038 }, { "epoch": 0.5138023152270703, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.659064769744873, "learning_rate": 5e-05, "loss": 0.7668, "mean_token_accuracy": 0.8660770654678345, "num_tokens": 153458388.0, "step": 4039 }, { "epoch": 0.5139295255056608, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.887670516967773, "learning_rate": 5e-05, "loss": 0.774, "mean_token_accuracy": 0.8491395115852356, "num_tokens": 153503486.0, "step": 4040 }, { "epoch": 0.5140567357842514, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.743659019470215, "learning_rate": 5e-05, "loss": 0.7956, "mean_token_accuracy": 0.8619245290756226, "num_tokens": 153541474.0, "step": 4041 }, { "epoch": 0.5141839460628419, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.746405124664307, "learning_rate": 5e-05, "loss": 0.7716, "mean_token_accuracy": 0.8476462960243225, "num_tokens": 153583518.0, "step": 4042 }, { "epoch": 0.5143111563414324, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.937383651733398, "learning_rate": 5e-05, "loss": 0.8238, "mean_token_accuracy": 0.8555048704147339, "num_tokens": 153617552.0, "step": 4043 }, { "epoch": 0.5144383666200228, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.522826194763184, "learning_rate": 5e-05, "loss": 0.7686, "mean_token_accuracy": 0.8470203280448914, "num_tokens": 153655616.0, "step": 4044 }, { "epoch": 0.5145655768986134, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.223857879638672, "learning_rate": 5e-05, "loss": 0.8174, "mean_token_accuracy": 0.8629083633422852, "num_tokens": 153692835.0, "step": 4045 }, { "epoch": 0.5146927871772039, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.3289971351623535, "learning_rate": 5e-05, "loss": 0.7355, "mean_token_accuracy": 0.8504538536071777, "num_tokens": 153720867.0, "step": 4046 }, { "epoch": 0.5148199974557944, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.48898696899414, "learning_rate": 5e-05, "loss": 0.871, "mean_token_accuracy": 0.8621888756752014, "num_tokens": 153758798.0, "step": 4047 }, { "epoch": 0.514947207734385, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.320956707000732, "learning_rate": 5e-05, "loss": 0.7918, "mean_token_accuracy": 0.8432066440582275, "num_tokens": 153798257.0, "step": 4048 }, { "epoch": 0.5150744180129755, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.508349418640137, "learning_rate": 5e-05, "loss": 0.8597, "mean_token_accuracy": 0.8685798645019531, "num_tokens": 153839241.0, "step": 4049 }, { "epoch": 0.5152016282915659, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.6977338790893555, "learning_rate": 5e-05, "loss": 0.7711, "mean_token_accuracy": 0.856311023235321, "num_tokens": 153882522.0, "step": 4050 }, { "epoch": 0.5153288385701564, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.9813947677612305, "learning_rate": 5e-05, "loss": 0.8207, "mean_token_accuracy": 0.8702115416526794, "num_tokens": 153912173.0, "step": 4051 }, { "epoch": 0.515456048848747, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.3838701248168945, "learning_rate": 5e-05, "loss": 0.9517, "mean_token_accuracy": 0.8213858604431152, "num_tokens": 153955070.0, "step": 4052 }, { "epoch": 0.5155832591273375, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.005866527557373, "learning_rate": 5e-05, "loss": 0.7253, "mean_token_accuracy": 0.8628147840499878, "num_tokens": 153995526.0, "step": 4053 }, { "epoch": 0.515710469405928, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.192486763000488, "learning_rate": 5e-05, "loss": 0.8806, "mean_token_accuracy": 0.8560307621955872, "num_tokens": 154035476.0, "step": 4054 }, { "epoch": 0.5158376796845185, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.158524513244629, "learning_rate": 5e-05, "loss": 0.7407, "mean_token_accuracy": 0.848816990852356, "num_tokens": 154079108.0, "step": 4055 }, { "epoch": 0.5159648899631091, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.971793174743652, "learning_rate": 5e-05, "loss": 1.0251, "mean_token_accuracy": 0.8323732614517212, "num_tokens": 154118484.0, "step": 4056 }, { "epoch": 0.5160921002416995, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.791428565979004, "learning_rate": 5e-05, "loss": 0.7063, "mean_token_accuracy": 0.8577790260314941, "num_tokens": 154163981.0, "step": 4057 }, { "epoch": 0.51621931052029, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.573442459106445, "learning_rate": 5e-05, "loss": 1.0706, "mean_token_accuracy": 0.8445737957954407, "num_tokens": 154205035.0, "step": 4058 }, { "epoch": 0.5163465207988805, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 6.406707763671875, "learning_rate": 5e-05, "loss": 0.7896, "mean_token_accuracy": 0.8663422465324402, "num_tokens": 154241243.0, "step": 4059 }, { "epoch": 0.5164737310774711, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.289995193481445, "learning_rate": 5e-05, "loss": 1.0846, "mean_token_accuracy": 0.8423185348510742, "num_tokens": 154289182.0, "step": 4060 }, { "epoch": 0.5166009413560616, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 7.777474403381348, "learning_rate": 5e-05, "loss": 1.0461, "mean_token_accuracy": 0.8475742340087891, "num_tokens": 154326735.0, "step": 4061 }, { "epoch": 0.5167281516346521, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.709322452545166, "learning_rate": 5e-05, "loss": 0.8821, "mean_token_accuracy": 0.8583512306213379, "num_tokens": 154367795.0, "step": 4062 }, { "epoch": 0.5168553619132426, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.817276954650879, "learning_rate": 5e-05, "loss": 0.9728, "mean_token_accuracy": 0.8729907274246216, "num_tokens": 154401949.0, "step": 4063 }, { "epoch": 0.5169825721918331, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.410702705383301, "learning_rate": 5e-05, "loss": 0.7476, "mean_token_accuracy": 0.866432785987854, "num_tokens": 154442693.0, "step": 4064 }, { "epoch": 0.5171097824704236, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.511093139648438, "learning_rate": 5e-05, "loss": 1.0265, "mean_token_accuracy": 0.857321560382843, "num_tokens": 154489219.0, "step": 4065 }, { "epoch": 0.5172369927490141, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 6.8317790031433105, "learning_rate": 5e-05, "loss": 0.8505, "mean_token_accuracy": 0.8714422583580017, "num_tokens": 154526639.0, "step": 4066 }, { "epoch": 0.5173642030276047, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.936997413635254, "learning_rate": 5e-05, "loss": 0.9634, "mean_token_accuracy": 0.870013952255249, "num_tokens": 154557339.0, "step": 4067 }, { "epoch": 0.5174914133061952, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.889796257019043, "learning_rate": 5e-05, "loss": 1.0268, "mean_token_accuracy": 0.8532334566116333, "num_tokens": 154592738.0, "step": 4068 }, { "epoch": 0.5176186235847856, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.528153896331787, "learning_rate": 5e-05, "loss": 0.8598, "mean_token_accuracy": 0.855941891670227, "num_tokens": 154631633.0, "step": 4069 }, { "epoch": 0.5177458338633761, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.519011497497559, "learning_rate": 5e-05, "loss": 1.0755, "mean_token_accuracy": 0.8329548239707947, "num_tokens": 154676274.0, "step": 4070 }, { "epoch": 0.5178730441419667, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.881289482116699, "learning_rate": 5e-05, "loss": 0.854, "mean_token_accuracy": 0.8421705961227417, "num_tokens": 154716379.0, "step": 4071 }, { "epoch": 0.5180002544205572, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.572912216186523, "learning_rate": 5e-05, "loss": 0.985, "mean_token_accuracy": 0.8516497611999512, "num_tokens": 154757038.0, "step": 4072 }, { "epoch": 0.5181274646991477, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.124060153961182, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.8808494806289673, "num_tokens": 154795684.0, "step": 4073 }, { "epoch": 0.5182546749777382, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.979952335357666, "learning_rate": 5e-05, "loss": 0.8646, "mean_token_accuracy": 0.8604214191436768, "num_tokens": 154834858.0, "step": 4074 }, { "epoch": 0.5183818852563287, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.525359153747559, "learning_rate": 5e-05, "loss": 0.8914, "mean_token_accuracy": 0.8620239496231079, "num_tokens": 154868919.0, "step": 4075 }, { "epoch": 0.5185090955349192, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.356587886810303, "learning_rate": 5e-05, "loss": 0.8251, "mean_token_accuracy": 0.8551913499832153, "num_tokens": 154910511.0, "step": 4076 }, { "epoch": 0.5186363058135097, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.737310886383057, "learning_rate": 5e-05, "loss": 0.9292, "mean_token_accuracy": 0.8426764607429504, "num_tokens": 154955345.0, "step": 4077 }, { "epoch": 0.5187635160921003, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.044550895690918, "learning_rate": 5e-05, "loss": 0.768, "mean_token_accuracy": 0.8635468482971191, "num_tokens": 154992541.0, "step": 4078 }, { "epoch": 0.5188907263706908, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.860079765319824, "learning_rate": 5e-05, "loss": 0.8548, "mean_token_accuracy": 0.8563357591629028, "num_tokens": 155033592.0, "step": 4079 }, { "epoch": 0.5190179366492813, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.885834217071533, "learning_rate": 5e-05, "loss": 0.7626, "mean_token_accuracy": 0.8595434427261353, "num_tokens": 155080034.0, "step": 4080 }, { "epoch": 0.5191451469278717, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.9561567306518555, "learning_rate": 5e-05, "loss": 0.8737, "mean_token_accuracy": 0.8453023433685303, "num_tokens": 155116207.0, "step": 4081 }, { "epoch": 0.5192723572064623, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.771986484527588, "learning_rate": 5e-05, "loss": 0.7633, "mean_token_accuracy": 0.8547642827033997, "num_tokens": 155155787.0, "step": 4082 }, { "epoch": 0.5193995674850528, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.028569221496582, "learning_rate": 5e-05, "loss": 0.8369, "mean_token_accuracy": 0.8638007640838623, "num_tokens": 155194021.0, "step": 4083 }, { "epoch": 0.5195267777636433, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.675289154052734, "learning_rate": 5e-05, "loss": 0.7377, "mean_token_accuracy": 0.8631017208099365, "num_tokens": 155228685.0, "step": 4084 }, { "epoch": 0.5196539880422338, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.105101585388184, "learning_rate": 5e-05, "loss": 0.8641, "mean_token_accuracy": 0.8536517024040222, "num_tokens": 155269146.0, "step": 4085 }, { "epoch": 0.5197811983208244, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.565490245819092, "learning_rate": 5e-05, "loss": 0.7504, "mean_token_accuracy": 0.858140230178833, "num_tokens": 155309080.0, "step": 4086 }, { "epoch": 0.5199084085994148, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.271990776062012, "learning_rate": 5e-05, "loss": 0.8569, "mean_token_accuracy": 0.8592943549156189, "num_tokens": 155348661.0, "step": 4087 }, { "epoch": 0.5200356188780053, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.423954010009766, "learning_rate": 5e-05, "loss": 0.7642, "mean_token_accuracy": 0.8519243001937866, "num_tokens": 155384984.0, "step": 4088 }, { "epoch": 0.5201628291565958, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.500483512878418, "learning_rate": 5e-05, "loss": 0.9179, "mean_token_accuracy": 0.8512042760848999, "num_tokens": 155421914.0, "step": 4089 }, { "epoch": 0.5202900394351864, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.336005687713623, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.8624095916748047, "num_tokens": 155466460.0, "step": 4090 }, { "epoch": 0.5204172497137769, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.655824661254883, "learning_rate": 5e-05, "loss": 0.9282, "mean_token_accuracy": 0.8615597486495972, "num_tokens": 155504034.0, "step": 4091 }, { "epoch": 0.5205444599923674, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.490427017211914, "learning_rate": 5e-05, "loss": 0.8196, "mean_token_accuracy": 0.8474298715591431, "num_tokens": 155542521.0, "step": 4092 }, { "epoch": 0.5206716702709578, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.553869247436523, "learning_rate": 5e-05, "loss": 0.925, "mean_token_accuracy": 0.86162269115448, "num_tokens": 155574170.0, "step": 4093 }, { "epoch": 0.5207988805495484, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 6.901769638061523, "learning_rate": 5e-05, "loss": 0.8249, "mean_token_accuracy": 0.8592934012413025, "num_tokens": 155604314.0, "step": 4094 }, { "epoch": 0.5209260908281389, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.05104923248291, "learning_rate": 5e-05, "loss": 0.8857, "mean_token_accuracy": 0.860964298248291, "num_tokens": 155643259.0, "step": 4095 }, { "epoch": 0.5210533011067294, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.345092296600342, "learning_rate": 5e-05, "loss": 0.8794, "mean_token_accuracy": 0.853614091873169, "num_tokens": 155682522.0, "step": 4096 }, { "epoch": 0.52118051138532, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.442180156707764, "learning_rate": 5e-05, "loss": 0.7881, "mean_token_accuracy": 0.8663352727890015, "num_tokens": 155714021.0, "step": 4097 }, { "epoch": 0.5213077216639105, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.595961570739746, "learning_rate": 5e-05, "loss": 0.917, "mean_token_accuracy": 0.8384258151054382, "num_tokens": 155754064.0, "step": 4098 }, { "epoch": 0.5214349319425009, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.144395351409912, "learning_rate": 5e-05, "loss": 0.7774, "mean_token_accuracy": 0.8600846529006958, "num_tokens": 155788428.0, "step": 4099 }, { "epoch": 0.5215621422210914, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.572038650512695, "learning_rate": 5e-05, "loss": 0.8214, "mean_token_accuracy": 0.8541961312294006, "num_tokens": 155826608.0, "step": 4100 }, { "epoch": 0.521689352499682, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.173849582672119, "learning_rate": 5e-05, "loss": 0.7816, "mean_token_accuracy": 0.8575356006622314, "num_tokens": 155870314.0, "step": 4101 }, { "epoch": 0.5218165627782725, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.370567321777344, "learning_rate": 5e-05, "loss": 0.8198, "mean_token_accuracy": 0.8459553718566895, "num_tokens": 155907742.0, "step": 4102 }, { "epoch": 0.521943773056863, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.304165840148926, "learning_rate": 5e-05, "loss": 0.7236, "mean_token_accuracy": 0.8734663724899292, "num_tokens": 155949837.0, "step": 4103 }, { "epoch": 0.5220709833354535, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.1891679763793945, "learning_rate": 5e-05, "loss": 0.781, "mean_token_accuracy": 0.853909969329834, "num_tokens": 155990407.0, "step": 4104 }, { "epoch": 0.522198193614044, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.423360824584961, "learning_rate": 5e-05, "loss": 0.7844, "mean_token_accuracy": 0.8527351021766663, "num_tokens": 156027157.0, "step": 4105 }, { "epoch": 0.5223254038926345, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.008159160614014, "learning_rate": 5e-05, "loss": 0.7579, "mean_token_accuracy": 0.8558470010757446, "num_tokens": 156065074.0, "step": 4106 }, { "epoch": 0.522452614171225, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.586211204528809, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.8651731014251709, "num_tokens": 156103399.0, "step": 4107 }, { "epoch": 0.5225798244498155, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.808464527130127, "learning_rate": 5e-05, "loss": 0.713, "mean_token_accuracy": 0.8609093427658081, "num_tokens": 156143825.0, "step": 4108 }, { "epoch": 0.5227070347284061, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.784628391265869, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8622793555259705, "num_tokens": 156183841.0, "step": 4109 }, { "epoch": 0.5228342450069966, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.6056365966796875, "learning_rate": 5e-05, "loss": 0.7004, "mean_token_accuracy": 0.8637933731079102, "num_tokens": 156225288.0, "step": 4110 }, { "epoch": 0.5229614552855871, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.064971923828125, "learning_rate": 5e-05, "loss": 0.7655, "mean_token_accuracy": 0.872037947177887, "num_tokens": 156257463.0, "step": 4111 }, { "epoch": 0.5230886655641775, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.321126461029053, "learning_rate": 5e-05, "loss": 0.7147, "mean_token_accuracy": 0.8563073873519897, "num_tokens": 156293059.0, "step": 4112 }, { "epoch": 0.5232158758427681, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.486974716186523, "learning_rate": 5e-05, "loss": 0.9377, "mean_token_accuracy": 0.8351144194602966, "num_tokens": 156338605.0, "step": 4113 }, { "epoch": 0.5233430861213586, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.034854888916016, "learning_rate": 5e-05, "loss": 0.7177, "mean_token_accuracy": 0.8559713363647461, "num_tokens": 156375472.0, "step": 4114 }, { "epoch": 0.5234702963999491, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.97146224975586, "learning_rate": 5e-05, "loss": 0.939, "mean_token_accuracy": 0.8589562773704529, "num_tokens": 156412290.0, "step": 4115 }, { "epoch": 0.5235975066785397, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.060102939605713, "learning_rate": 5e-05, "loss": 0.7384, "mean_token_accuracy": 0.854920506477356, "num_tokens": 156452268.0, "step": 4116 }, { "epoch": 0.5237247169571302, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.066805839538574, "learning_rate": 5e-05, "loss": 1.0155, "mean_token_accuracy": 0.850611686706543, "num_tokens": 156500659.0, "step": 4117 }, { "epoch": 0.5238519272357206, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 6.847766876220703, "learning_rate": 5e-05, "loss": 0.8399, "mean_token_accuracy": 0.8617968559265137, "num_tokens": 156534529.0, "step": 4118 }, { "epoch": 0.5239791375143111, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.201739311218262, "learning_rate": 5e-05, "loss": 0.8781, "mean_token_accuracy": 0.8710964918136597, "num_tokens": 156574459.0, "step": 4119 }, { "epoch": 0.5241063477929017, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.8757524490356445, "learning_rate": 5e-05, "loss": 0.9397, "mean_token_accuracy": 0.8594356179237366, "num_tokens": 156611626.0, "step": 4120 }, { "epoch": 0.5242335580714922, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.7674994468688965, "learning_rate": 5e-05, "loss": 0.7523, "mean_token_accuracy": 0.8619349002838135, "num_tokens": 156649454.0, "step": 4121 }, { "epoch": 0.5243607683500827, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.838604927062988, "learning_rate": 5e-05, "loss": 0.9709, "mean_token_accuracy": 0.8579789400100708, "num_tokens": 156687263.0, "step": 4122 }, { "epoch": 0.5244879786286732, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.0372314453125, "learning_rate": 5e-05, "loss": 0.7396, "mean_token_accuracy": 0.8531297445297241, "num_tokens": 156720384.0, "step": 4123 }, { "epoch": 0.5246151889072637, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.45285415649414, "learning_rate": 5e-05, "loss": 1.0104, "mean_token_accuracy": 0.86250901222229, "num_tokens": 156758365.0, "step": 4124 }, { "epoch": 0.5247423991858542, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 6.522892951965332, "learning_rate": 5e-05, "loss": 0.8162, "mean_token_accuracy": 0.8576145172119141, "num_tokens": 156790411.0, "step": 4125 }, { "epoch": 0.5248696094644447, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.147048950195312, "learning_rate": 5e-05, "loss": 1.0223, "mean_token_accuracy": 0.8595659732818604, "num_tokens": 156829176.0, "step": 4126 }, { "epoch": 0.5249968197430352, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 7.6546735763549805, "learning_rate": 5e-05, "loss": 0.9603, "mean_token_accuracy": 0.8664564490318298, "num_tokens": 156871932.0, "step": 4127 }, { "epoch": 0.5251240300216258, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.896585941314697, "learning_rate": 5e-05, "loss": 0.8522, "mean_token_accuracy": 0.8704778552055359, "num_tokens": 156905810.0, "step": 4128 }, { "epoch": 0.5252512403002163, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.292171478271484, "learning_rate": 5e-05, "loss": 0.9437, "mean_token_accuracy": 0.8654971122741699, "num_tokens": 156947218.0, "step": 4129 }, { "epoch": 0.5253784505788067, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.023155689239502, "learning_rate": 5e-05, "loss": 0.843, "mean_token_accuracy": 0.847865104675293, "num_tokens": 156978943.0, "step": 4130 }, { "epoch": 0.5255056608573972, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.265986442565918, "learning_rate": 5e-05, "loss": 0.8851, "mean_token_accuracy": 0.8688219785690308, "num_tokens": 157015015.0, "step": 4131 }, { "epoch": 0.5256328711359878, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.183195114135742, "learning_rate": 5e-05, "loss": 0.844, "mean_token_accuracy": 0.856781542301178, "num_tokens": 157051789.0, "step": 4132 }, { "epoch": 0.5257600814145783, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.545896053314209, "learning_rate": 5e-05, "loss": 0.7996, "mean_token_accuracy": 0.86670982837677, "num_tokens": 157090568.0, "step": 4133 }, { "epoch": 0.5258872916931688, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.7109270095825195, "learning_rate": 5e-05, "loss": 0.8443, "mean_token_accuracy": 0.8626033067703247, "num_tokens": 157125255.0, "step": 4134 }, { "epoch": 0.5260145019717594, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.884599208831787, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8566638827323914, "num_tokens": 157164917.0, "step": 4135 }, { "epoch": 0.5261417122503498, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.07929801940918, "learning_rate": 5e-05, "loss": 0.8917, "mean_token_accuracy": 0.850811243057251, "num_tokens": 157205186.0, "step": 4136 }, { "epoch": 0.5262689225289403, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.590962886810303, "learning_rate": 5e-05, "loss": 0.7771, "mean_token_accuracy": 0.8476600050926208, "num_tokens": 157241511.0, "step": 4137 }, { "epoch": 0.5263961328075308, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.381855964660645, "learning_rate": 5e-05, "loss": 0.8815, "mean_token_accuracy": 0.8636853694915771, "num_tokens": 157283381.0, "step": 4138 }, { "epoch": 0.5265233430861214, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.401671409606934, "learning_rate": 5e-05, "loss": 0.7524, "mean_token_accuracy": 0.8552525043487549, "num_tokens": 157321315.0, "step": 4139 }, { "epoch": 0.5266505533647119, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.723893165588379, "learning_rate": 5e-05, "loss": 0.9772, "mean_token_accuracy": 0.8459796905517578, "num_tokens": 157359205.0, "step": 4140 }, { "epoch": 0.5267777636433024, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.401331901550293, "learning_rate": 5e-05, "loss": 0.7381, "mean_token_accuracy": 0.8656347393989563, "num_tokens": 157398889.0, "step": 4141 }, { "epoch": 0.5269049739218928, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.878389358520508, "learning_rate": 5e-05, "loss": 1.0179, "mean_token_accuracy": 0.8458461165428162, "num_tokens": 157440831.0, "step": 4142 }, { "epoch": 0.5270321842004834, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 6.712782859802246, "learning_rate": 5e-05, "loss": 0.8105, "mean_token_accuracy": 0.8599182963371277, "num_tokens": 157476275.0, "step": 4143 }, { "epoch": 0.5271593944790739, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.721561431884766, "learning_rate": 5e-05, "loss": 0.9263, "mean_token_accuracy": 0.8715234994888306, "num_tokens": 157508830.0, "step": 4144 }, { "epoch": 0.5272866047576644, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.07133150100708, "learning_rate": 5e-05, "loss": 0.853, "mean_token_accuracy": 0.8648102283477783, "num_tokens": 157539077.0, "step": 4145 }, { "epoch": 0.527413815036255, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.409695625305176, "learning_rate": 5e-05, "loss": 0.9199, "mean_token_accuracy": 0.8701356053352356, "num_tokens": 157583501.0, "step": 4146 }, { "epoch": 0.5275410253148455, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.190561294555664, "learning_rate": 5e-05, "loss": 0.8802, "mean_token_accuracy": 0.8560622334480286, "num_tokens": 157629778.0, "step": 4147 }, { "epoch": 0.5276682355934359, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.252532958984375, "learning_rate": 5e-05, "loss": 0.9359, "mean_token_accuracy": 0.8623949885368347, "num_tokens": 157669375.0, "step": 4148 }, { "epoch": 0.5277954458720264, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.091026306152344, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.8686795830726624, "num_tokens": 157707400.0, "step": 4149 }, { "epoch": 0.527922656150617, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.26152515411377, "learning_rate": 5e-05, "loss": 0.9464, "mean_token_accuracy": 0.8577283620834351, "num_tokens": 157747039.0, "step": 4150 }, { "epoch": 0.5280498664292075, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.0267558097839355, "learning_rate": 5e-05, "loss": 0.8232, "mean_token_accuracy": 0.8595923781394958, "num_tokens": 157782790.0, "step": 4151 }, { "epoch": 0.528177076707798, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.190669059753418, "learning_rate": 5e-05, "loss": 0.9195, "mean_token_accuracy": 0.86293625831604, "num_tokens": 157832002.0, "step": 4152 }, { "epoch": 0.5283042869863885, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.108063220977783, "learning_rate": 5e-05, "loss": 0.8658, "mean_token_accuracy": 0.8473689556121826, "num_tokens": 157869875.0, "step": 4153 }, { "epoch": 0.528431497264979, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.9676337242126465, "learning_rate": 5e-05, "loss": 0.9163, "mean_token_accuracy": 0.8507710099220276, "num_tokens": 157908910.0, "step": 4154 }, { "epoch": 0.5285587075435695, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.2288970947265625, "learning_rate": 5e-05, "loss": 0.8538, "mean_token_accuracy": 0.8457940816879272, "num_tokens": 157944366.0, "step": 4155 }, { "epoch": 0.52868591782216, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.65762186050415, "learning_rate": 5e-05, "loss": 0.8394, "mean_token_accuracy": 0.8591282367706299, "num_tokens": 157982053.0, "step": 4156 }, { "epoch": 0.5288131281007505, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.304675579071045, "learning_rate": 5e-05, "loss": 0.8232, "mean_token_accuracy": 0.857417106628418, "num_tokens": 158022654.0, "step": 4157 }, { "epoch": 0.5289403383793411, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.477712631225586, "learning_rate": 5e-05, "loss": 0.8984, "mean_token_accuracy": 0.8362981677055359, "num_tokens": 158060508.0, "step": 4158 }, { "epoch": 0.5290675486579316, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.263794422149658, "learning_rate": 5e-05, "loss": 0.8312, "mean_token_accuracy": 0.8451735973358154, "num_tokens": 158098956.0, "step": 4159 }, { "epoch": 0.5291947589365221, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.447278022766113, "learning_rate": 5e-05, "loss": 0.8191, "mean_token_accuracy": 0.853690505027771, "num_tokens": 158134781.0, "step": 4160 }, { "epoch": 0.5293219692151125, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.163719177246094, "learning_rate": 5e-05, "loss": 0.7515, "mean_token_accuracy": 0.8652333617210388, "num_tokens": 158177495.0, "step": 4161 }, { "epoch": 0.5294491794937031, "ewc_loss": 0.375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.502112865447998, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.8545121550559998, "num_tokens": 158216672.0, "step": 4162 }, { "epoch": 0.5295763897722936, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.06863260269165, "learning_rate": 5e-05, "loss": 0.7858, "mean_token_accuracy": 0.8490918874740601, "num_tokens": 158255397.0, "step": 4163 }, { "epoch": 0.5297036000508841, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.513385772705078, "learning_rate": 5e-05, "loss": 0.7983, "mean_token_accuracy": 0.8508066534996033, "num_tokens": 158292230.0, "step": 4164 }, { "epoch": 0.5298308103294747, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.034524917602539, "learning_rate": 5e-05, "loss": 0.7498, "mean_token_accuracy": 0.8586378693580627, "num_tokens": 158329884.0, "step": 4165 }, { "epoch": 0.5299580206080652, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.521745204925537, "learning_rate": 5e-05, "loss": 0.8845, "mean_token_accuracy": 0.8319998383522034, "num_tokens": 158367290.0, "step": 4166 }, { "epoch": 0.5300852308866556, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.976828575134277, "learning_rate": 5e-05, "loss": 0.6803, "mean_token_accuracy": 0.8797556757926941, "num_tokens": 158407999.0, "step": 4167 }, { "epoch": 0.5302124411652461, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.582764148712158, "learning_rate": 5e-05, "loss": 0.7339, "mean_token_accuracy": 0.8708019256591797, "num_tokens": 158444283.0, "step": 4168 }, { "epoch": 0.5303396514438367, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.878602981567383, "learning_rate": 5e-05, "loss": 0.7369, "mean_token_accuracy": 0.8567537069320679, "num_tokens": 158484807.0, "step": 4169 }, { "epoch": 0.5304668617224272, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.685451984405518, "learning_rate": 5e-05, "loss": 0.7677, "mean_token_accuracy": 0.863113522529602, "num_tokens": 158518863.0, "step": 4170 }, { "epoch": 0.5305940720010177, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.7718634605407715, "learning_rate": 5e-05, "loss": 0.7046, "mean_token_accuracy": 0.8654902577400208, "num_tokens": 158555057.0, "step": 4171 }, { "epoch": 0.5307212822796082, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.809266090393066, "learning_rate": 5e-05, "loss": 0.7968, "mean_token_accuracy": 0.8615887761116028, "num_tokens": 158596032.0, "step": 4172 }, { "epoch": 0.5308484925581987, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.664396286010742, "learning_rate": 5e-05, "loss": 0.7617, "mean_token_accuracy": 0.8479045629501343, "num_tokens": 158631098.0, "step": 4173 }, { "epoch": 0.5309757028367892, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.961155891418457, "learning_rate": 5e-05, "loss": 0.8386, "mean_token_accuracy": 0.8484193086624146, "num_tokens": 158664579.0, "step": 4174 }, { "epoch": 0.5311029131153797, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.54703426361084, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.8486076593399048, "num_tokens": 158701220.0, "step": 4175 }, { "epoch": 0.5312301233939702, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.10054874420166, "learning_rate": 5e-05, "loss": 0.8044, "mean_token_accuracy": 0.8634630441665649, "num_tokens": 158744235.0, "step": 4176 }, { "epoch": 0.5313573336725608, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.517666339874268, "learning_rate": 5e-05, "loss": 0.7126, "mean_token_accuracy": 0.8607034087181091, "num_tokens": 158782602.0, "step": 4177 }, { "epoch": 0.5314845439511513, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.115583419799805, "learning_rate": 5e-05, "loss": 0.823, "mean_token_accuracy": 0.861808180809021, "num_tokens": 158813345.0, "step": 4178 }, { "epoch": 0.5316117542297417, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.660606861114502, "learning_rate": 5e-05, "loss": 0.7748, "mean_token_accuracy": 0.8460896015167236, "num_tokens": 158847204.0, "step": 4179 }, { "epoch": 0.5317389645083322, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.87807035446167, "learning_rate": 5e-05, "loss": 0.8298, "mean_token_accuracy": 0.8542304039001465, "num_tokens": 158889481.0, "step": 4180 }, { "epoch": 0.5318661747869228, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.027468204498291, "learning_rate": 5e-05, "loss": 0.7296, "mean_token_accuracy": 0.8679835796356201, "num_tokens": 158929267.0, "step": 4181 }, { "epoch": 0.5319933850655133, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.336122035980225, "learning_rate": 5e-05, "loss": 0.8046, "mean_token_accuracy": 0.8478074073791504, "num_tokens": 158974526.0, "step": 4182 }, { "epoch": 0.5321205953441038, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.544242858886719, "learning_rate": 5e-05, "loss": 0.7857, "mean_token_accuracy": 0.8638842105865479, "num_tokens": 159015428.0, "step": 4183 }, { "epoch": 0.5322478056226944, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.6804094314575195, "learning_rate": 5e-05, "loss": 0.7669, "mean_token_accuracy": 0.8456400036811829, "num_tokens": 159050492.0, "step": 4184 }, { "epoch": 0.5323750159012848, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.179697036743164, "learning_rate": 5e-05, "loss": 0.8612, "mean_token_accuracy": 0.8481405973434448, "num_tokens": 159085644.0, "step": 4185 }, { "epoch": 0.5325022261798753, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 6.047764301300049, "learning_rate": 5e-05, "loss": 0.7467, "mean_token_accuracy": 0.8440791368484497, "num_tokens": 159123093.0, "step": 4186 }, { "epoch": 0.5326294364584658, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.98310375213623, "learning_rate": 5e-05, "loss": 0.8998, "mean_token_accuracy": 0.8569084405899048, "num_tokens": 159157688.0, "step": 4187 }, { "epoch": 0.5327566467370564, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002307891845703125, "grad_norm": 5.57338285446167, "learning_rate": 5e-05, "loss": 0.7033, "mean_token_accuracy": 0.8507645726203918, "num_tokens": 159194473.0, "step": 4188 }, { "epoch": 0.5328838570156469, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.858820915222168, "learning_rate": 5e-05, "loss": 1.041, "mean_token_accuracy": 0.8526826500892639, "num_tokens": 159230314.0, "step": 4189 }, { "epoch": 0.5330110672942374, "ewc_loss": 0.375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.177366256713867, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8654417991638184, "num_tokens": 159266579.0, "step": 4190 }, { "epoch": 0.5331382775728278, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.90422248840332, "learning_rate": 5e-05, "loss": 1.0551, "mean_token_accuracy": 0.8686505556106567, "num_tokens": 159306502.0, "step": 4191 }, { "epoch": 0.5332654878514184, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 7.979825019836426, "learning_rate": 5e-05, "loss": 1.0807, "mean_token_accuracy": 0.8416813611984253, "num_tokens": 159342512.0, "step": 4192 }, { "epoch": 0.5333926981300089, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.110697746276855, "learning_rate": 5e-05, "loss": 0.8909, "mean_token_accuracy": 0.8605074286460876, "num_tokens": 159378408.0, "step": 4193 }, { "epoch": 0.5335199084085994, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.279997825622559, "learning_rate": 5e-05, "loss": 1.138, "mean_token_accuracy": 0.8376201391220093, "num_tokens": 159415843.0, "step": 4194 }, { "epoch": 0.53364711868719, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.314448356628418, "learning_rate": 5e-05, "loss": 0.759, "mean_token_accuracy": 0.8564043045043945, "num_tokens": 159446880.0, "step": 4195 }, { "epoch": 0.5337743289657805, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 11.003655433654785, "learning_rate": 5e-05, "loss": 1.2157, "mean_token_accuracy": 0.8604477643966675, "num_tokens": 159491960.0, "step": 4196 }, { "epoch": 0.5339015392443709, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 7.581974506378174, "learning_rate": 5e-05, "loss": 0.9846, "mean_token_accuracy": 0.8636323809623718, "num_tokens": 159529157.0, "step": 4197 }, { "epoch": 0.5340287495229614, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 11.142603874206543, "learning_rate": 5e-05, "loss": 1.182, "mean_token_accuracy": 0.852137565612793, "num_tokens": 159563589.0, "step": 4198 }, { "epoch": 0.534155959801552, "ewc_loss": 0.828125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 8.988777160644531, "learning_rate": 5e-05, "loss": 1.185, "mean_token_accuracy": 0.8758095502853394, "num_tokens": 159601789.0, "step": 4199 }, { "epoch": 0.5342831700801425, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.2875337600708, "learning_rate": 5e-05, "loss": 1.1399, "mean_token_accuracy": 0.8355709314346313, "num_tokens": 159637357.0, "step": 4200 }, { "epoch": 0.534410380358733, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 8.468331336975098, "learning_rate": 5e-05, "loss": 1.0035, "mean_token_accuracy": 0.8714315891265869, "num_tokens": 159678919.0, "step": 4201 }, { "epoch": 0.5345375906373235, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.83357048034668, "learning_rate": 5e-05, "loss": 1.0925, "mean_token_accuracy": 0.8531627058982849, "num_tokens": 159714109.0, "step": 4202 }, { "epoch": 0.534664800915914, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 7.8398308753967285, "learning_rate": 5e-05, "loss": 0.9117, "mean_token_accuracy": 0.8656899333000183, "num_tokens": 159752171.0, "step": 4203 }, { "epoch": 0.5347920111945045, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 9.827249526977539, "learning_rate": 5e-05, "loss": 1.1212, "mean_token_accuracy": 0.8481687307357788, "num_tokens": 159791251.0, "step": 4204 }, { "epoch": 0.534919221473095, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.646814823150635, "learning_rate": 5e-05, "loss": 0.89, "mean_token_accuracy": 0.8666427135467529, "num_tokens": 159819925.0, "step": 4205 }, { "epoch": 0.5350464317516855, "ewc_loss": 0.6875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.037636756896973, "learning_rate": 5e-05, "loss": 1.1343, "mean_token_accuracy": 0.8529175519943237, "num_tokens": 159859788.0, "step": 4206 }, { "epoch": 0.5351736420302761, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.024648666381836, "learning_rate": 5e-05, "loss": 1.0198, "mean_token_accuracy": 0.8500133156776428, "num_tokens": 159897202.0, "step": 4207 }, { "epoch": 0.5353008523088666, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.825227737426758, "learning_rate": 5e-05, "loss": 1.109, "mean_token_accuracy": 0.8449136018753052, "num_tokens": 159935840.0, "step": 4208 }, { "epoch": 0.5354280625874571, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.270692825317383, "learning_rate": 5e-05, "loss": 0.9848, "mean_token_accuracy": 0.8680187463760376, "num_tokens": 159975120.0, "step": 4209 }, { "epoch": 0.5355552728660475, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.330822944641113, "learning_rate": 5e-05, "loss": 1.0086, "mean_token_accuracy": 0.8631565570831299, "num_tokens": 160020935.0, "step": 4210 }, { "epoch": 0.5356824831446381, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.151185035705566, "learning_rate": 5e-05, "loss": 0.9219, "mean_token_accuracy": 0.8720352649688721, "num_tokens": 160065029.0, "step": 4211 }, { "epoch": 0.5358096934232286, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.942021369934082, "learning_rate": 5e-05, "loss": 1.0406, "mean_token_accuracy": 0.8490230441093445, "num_tokens": 160104865.0, "step": 4212 }, { "epoch": 0.5359369037018191, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.841390609741211, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8661936521530151, "num_tokens": 160145477.0, "step": 4213 }, { "epoch": 0.5360641139804097, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.812015533447266, "learning_rate": 5e-05, "loss": 1.0275, "mean_token_accuracy": 0.8492098450660706, "num_tokens": 160184906.0, "step": 4214 }, { "epoch": 0.5361913242590002, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.507906436920166, "learning_rate": 5e-05, "loss": 0.8861, "mean_token_accuracy": 0.8551257848739624, "num_tokens": 160225011.0, "step": 4215 }, { "epoch": 0.5363185345375906, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.970054626464844, "learning_rate": 5e-05, "loss": 1.009, "mean_token_accuracy": 0.8588523268699646, "num_tokens": 160263193.0, "step": 4216 }, { "epoch": 0.5364457448161811, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.38015079498291, "learning_rate": 5e-05, "loss": 0.7955, "mean_token_accuracy": 0.8830986022949219, "num_tokens": 160299552.0, "step": 4217 }, { "epoch": 0.5365729550947717, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.114028930664062, "learning_rate": 5e-05, "loss": 0.9945, "mean_token_accuracy": 0.8711846470832825, "num_tokens": 160334609.0, "step": 4218 }, { "epoch": 0.5367001653733622, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.621081829071045, "learning_rate": 5e-05, "loss": 0.8815, "mean_token_accuracy": 0.8629351258277893, "num_tokens": 160374295.0, "step": 4219 }, { "epoch": 0.5368273756519527, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.83803653717041, "learning_rate": 5e-05, "loss": 1.058, "mean_token_accuracy": 0.8423805832862854, "num_tokens": 160404420.0, "step": 4220 }, { "epoch": 0.5369545859305432, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.839684009552002, "learning_rate": 5e-05, "loss": 0.9038, "mean_token_accuracy": 0.8570626974105835, "num_tokens": 160439278.0, "step": 4221 }, { "epoch": 0.5370817962091337, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.310035705566406, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.8757996559143066, "num_tokens": 160475225.0, "step": 4222 }, { "epoch": 0.5372090064877242, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.691398620605469, "learning_rate": 5e-05, "loss": 0.8831, "mean_token_accuracy": 0.8609527945518494, "num_tokens": 160516073.0, "step": 4223 }, { "epoch": 0.5373362167663147, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.100775718688965, "learning_rate": 5e-05, "loss": 0.9521, "mean_token_accuracy": 0.8525090217590332, "num_tokens": 160550938.0, "step": 4224 }, { "epoch": 0.5374634270449052, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.339495658874512, "learning_rate": 5e-05, "loss": 0.7945, "mean_token_accuracy": 0.8669451475143433, "num_tokens": 160590382.0, "step": 4225 }, { "epoch": 0.5375906373234958, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.211523056030273, "learning_rate": 5e-05, "loss": 0.8846, "mean_token_accuracy": 0.8712641000747681, "num_tokens": 160624830.0, "step": 4226 }, { "epoch": 0.5377178476020863, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.06343936920166, "learning_rate": 5e-05, "loss": 0.7978, "mean_token_accuracy": 0.8545351624488831, "num_tokens": 160663042.0, "step": 4227 }, { "epoch": 0.5378450578806767, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.392027854919434, "learning_rate": 5e-05, "loss": 0.9574, "mean_token_accuracy": 0.8495533466339111, "num_tokens": 160700502.0, "step": 4228 }, { "epoch": 0.5379722681592672, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.010751724243164, "learning_rate": 5e-05, "loss": 0.8451, "mean_token_accuracy": 0.8435162305831909, "num_tokens": 160736746.0, "step": 4229 }, { "epoch": 0.5380994784378578, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.40501880645752, "learning_rate": 5e-05, "loss": 0.8976, "mean_token_accuracy": 0.8679007291793823, "num_tokens": 160771430.0, "step": 4230 }, { "epoch": 0.5382266887164483, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.16764497756958, "learning_rate": 5e-05, "loss": 0.8122, "mean_token_accuracy": 0.8579075336456299, "num_tokens": 160807047.0, "step": 4231 }, { "epoch": 0.5383538989950388, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.189678192138672, "learning_rate": 5e-05, "loss": 0.9754, "mean_token_accuracy": 0.8403456211090088, "num_tokens": 160844600.0, "step": 4232 }, { "epoch": 0.5384811092736294, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.3241424560546875, "learning_rate": 5e-05, "loss": 0.8112, "mean_token_accuracy": 0.8623435497283936, "num_tokens": 160883132.0, "step": 4233 }, { "epoch": 0.5386083195522198, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.863650321960449, "learning_rate": 5e-05, "loss": 0.8676, "mean_token_accuracy": 0.8590587377548218, "num_tokens": 160922408.0, "step": 4234 }, { "epoch": 0.5387355298308103, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.4005126953125, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.8626280426979065, "num_tokens": 160960563.0, "step": 4235 }, { "epoch": 0.5388627401094008, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.609588146209717, "learning_rate": 5e-05, "loss": 0.837, "mean_token_accuracy": 0.858052670955658, "num_tokens": 160995594.0, "step": 4236 }, { "epoch": 0.5389899503879914, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.3465094566345215, "learning_rate": 5e-05, "loss": 0.7965, "mean_token_accuracy": 0.8544384241104126, "num_tokens": 161028341.0, "step": 4237 }, { "epoch": 0.5391171606665819, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.5476861000061035, "learning_rate": 5e-05, "loss": 0.7983, "mean_token_accuracy": 0.8610652685165405, "num_tokens": 161064549.0, "step": 4238 }, { "epoch": 0.5392443709451724, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.2249860763549805, "learning_rate": 5e-05, "loss": 0.8171, "mean_token_accuracy": 0.8442600965499878, "num_tokens": 161105170.0, "step": 4239 }, { "epoch": 0.5393715812237628, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.566052436828613, "learning_rate": 5e-05, "loss": 0.8622, "mean_token_accuracy": 0.8390673398971558, "num_tokens": 161141018.0, "step": 4240 }, { "epoch": 0.5394987915023534, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.132272243499756, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.843055248260498, "num_tokens": 161175363.0, "step": 4241 }, { "epoch": 0.5396260017809439, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.575129985809326, "learning_rate": 5e-05, "loss": 0.7756, "mean_token_accuracy": 0.8641501069068909, "num_tokens": 161210920.0, "step": 4242 }, { "epoch": 0.5397532120595344, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.0639567375183105, "learning_rate": 5e-05, "loss": 0.7079, "mean_token_accuracy": 0.8697836399078369, "num_tokens": 161248743.0, "step": 4243 }, { "epoch": 0.5398804223381249, "ewc_loss": 0.375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.609375476837158, "learning_rate": 5e-05, "loss": 0.7733, "mean_token_accuracy": 0.8637025952339172, "num_tokens": 161282897.0, "step": 4244 }, { "epoch": 0.5400076326167155, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.980951309204102, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.8614423274993896, "num_tokens": 161323103.0, "step": 4245 }, { "epoch": 0.5401348428953059, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.670799255371094, "learning_rate": 5e-05, "loss": 0.7917, "mean_token_accuracy": 0.857801079750061, "num_tokens": 161359612.0, "step": 4246 }, { "epoch": 0.5402620531738964, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.918583393096924, "learning_rate": 5e-05, "loss": 0.7005, "mean_token_accuracy": 0.8699446320533752, "num_tokens": 161396961.0, "step": 4247 }, { "epoch": 0.540389263452487, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.714695453643799, "learning_rate": 5e-05, "loss": 0.7997, "mean_token_accuracy": 0.8535956144332886, "num_tokens": 161437380.0, "step": 4248 }, { "epoch": 0.5405164737310775, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.9094462394714355, "learning_rate": 5e-05, "loss": 0.7353, "mean_token_accuracy": 0.8552917838096619, "num_tokens": 161480237.0, "step": 4249 }, { "epoch": 0.540643684009668, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.6702775955200195, "learning_rate": 5e-05, "loss": 0.7268, "mean_token_accuracy": 0.8734939098358154, "num_tokens": 161519447.0, "step": 4250 }, { "epoch": 0.5407708942882585, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.963776111602783, "learning_rate": 5e-05, "loss": 0.7544, "mean_token_accuracy": 0.8510767817497253, "num_tokens": 161556588.0, "step": 4251 }, { "epoch": 0.540898104566849, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.5759711265563965, "learning_rate": 5e-05, "loss": 0.7585, "mean_token_accuracy": 0.8610366582870483, "num_tokens": 161591542.0, "step": 4252 }, { "epoch": 0.5410253148454395, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.073797702789307, "learning_rate": 5e-05, "loss": 0.7371, "mean_token_accuracy": 0.8604691028594971, "num_tokens": 161630453.0, "step": 4253 }, { "epoch": 0.54115252512403, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.424747467041016, "learning_rate": 5e-05, "loss": 0.6805, "mean_token_accuracy": 0.8824748992919922, "num_tokens": 161660927.0, "step": 4254 }, { "epoch": 0.5412797354026205, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.216148853302002, "learning_rate": 5e-05, "loss": 0.7423, "mean_token_accuracy": 0.8610657453536987, "num_tokens": 161703284.0, "step": 4255 }, { "epoch": 0.5414069456812111, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.210017681121826, "learning_rate": 5e-05, "loss": 0.7783, "mean_token_accuracy": 0.8456649780273438, "num_tokens": 161739212.0, "step": 4256 }, { "epoch": 0.5415341559598016, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.4377336502075195, "learning_rate": 5e-05, "loss": 0.7938, "mean_token_accuracy": 0.8569372296333313, "num_tokens": 161777635.0, "step": 4257 }, { "epoch": 0.5416613662383921, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.9225921630859375, "learning_rate": 5e-05, "loss": 0.7411, "mean_token_accuracy": 0.8530436158180237, "num_tokens": 161816019.0, "step": 4258 }, { "epoch": 0.5417885765169825, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.747560501098633, "learning_rate": 5e-05, "loss": 0.7559, "mean_token_accuracy": 0.8607214689254761, "num_tokens": 161849690.0, "step": 4259 }, { "epoch": 0.5419157867955731, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.558018684387207, "learning_rate": 5e-05, "loss": 0.6401, "mean_token_accuracy": 0.873105525970459, "num_tokens": 161889596.0, "step": 4260 }, { "epoch": 0.5420429970741636, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.196281433105469, "learning_rate": 5e-05, "loss": 0.869, "mean_token_accuracy": 0.8398512601852417, "num_tokens": 161929451.0, "step": 4261 }, { "epoch": 0.5421702073527541, "ewc_loss": 0.251953125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 6.095766544342041, "learning_rate": 5e-05, "loss": 0.7057, "mean_token_accuracy": 0.8470345735549927, "num_tokens": 161966422.0, "step": 4262 }, { "epoch": 0.5422974176313446, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.874967575073242, "learning_rate": 5e-05, "loss": 0.8877, "mean_token_accuracy": 0.8549643754959106, "num_tokens": 162005192.0, "step": 4263 }, { "epoch": 0.5424246279099352, "ewc_loss": 0.244140625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 5.623005390167236, "learning_rate": 5e-05, "loss": 0.6895, "mean_token_accuracy": 0.8547810316085815, "num_tokens": 162051039.0, "step": 4264 }, { "epoch": 0.5425518381885256, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.685701370239258, "learning_rate": 5e-05, "loss": 0.9668, "mean_token_accuracy": 0.8643805384635925, "num_tokens": 162086869.0, "step": 4265 }, { "epoch": 0.5426790484671161, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 5.956997394561768, "learning_rate": 5e-05, "loss": 0.7245, "mean_token_accuracy": 0.8613786101341248, "num_tokens": 162122442.0, "step": 4266 }, { "epoch": 0.5428062587457066, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.738265991210938, "learning_rate": 5e-05, "loss": 1.0388, "mean_token_accuracy": 0.8621677160263062, "num_tokens": 162153057.0, "step": 4267 }, { "epoch": 0.5429334690242972, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.620900630950928, "learning_rate": 5e-05, "loss": 1.028, "mean_token_accuracy": 0.843510091304779, "num_tokens": 162192086.0, "step": 4268 }, { "epoch": 0.5430606793028877, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.992059707641602, "learning_rate": 5e-05, "loss": 0.8629, "mean_token_accuracy": 0.8694721460342407, "num_tokens": 162227206.0, "step": 4269 }, { "epoch": 0.5431878895814782, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.184916496276855, "learning_rate": 5e-05, "loss": 1.0516, "mean_token_accuracy": 0.8526970744132996, "num_tokens": 162263490.0, "step": 4270 }, { "epoch": 0.5433150998600687, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 5.9406046867370605, "learning_rate": 5e-05, "loss": 0.6886, "mean_token_accuracy": 0.8711227178573608, "num_tokens": 162302750.0, "step": 4271 }, { "epoch": 0.5434423101386592, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000774383544921875, "grad_norm": 11.454913139343262, "learning_rate": 5e-05, "loss": 1.1759, "mean_token_accuracy": 0.8750000596046448, "num_tokens": 162339613.0, "step": 4272 }, { "epoch": 0.5435695204172497, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 7.561429977416992, "learning_rate": 5e-05, "loss": 1.0079, "mean_token_accuracy": 0.8528839349746704, "num_tokens": 162380163.0, "step": 4273 }, { "epoch": 0.5436967306958402, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 12.034544944763184, "learning_rate": 5e-05, "loss": 1.2746, "mean_token_accuracy": 0.854232907295227, "num_tokens": 162413306.0, "step": 4274 }, { "epoch": 0.5438239409744308, "ewc_loss": 0.90234375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0008697509765625, "grad_norm": 9.834628105163574, "learning_rate": 5e-05, "loss": 1.2947, "mean_token_accuracy": 0.8668617606163025, "num_tokens": 162446861.0, "step": 4275 }, { "epoch": 0.5439511512530213, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 11.018003463745117, "learning_rate": 5e-05, "loss": 1.1434, "mean_token_accuracy": 0.8522383570671082, "num_tokens": 162491414.0, "step": 4276 }, { "epoch": 0.5440783615316117, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 8.97916316986084, "learning_rate": 5e-05, "loss": 1.0726, "mean_token_accuracy": 0.865207850933075, "num_tokens": 162535265.0, "step": 4277 }, { "epoch": 0.5442055718102022, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 10.797446250915527, "learning_rate": 5e-05, "loss": 1.2024, "mean_token_accuracy": 0.8479692935943604, "num_tokens": 162577198.0, "step": 4278 }, { "epoch": 0.5443327820887928, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.461697578430176, "learning_rate": 5e-05, "loss": 0.9718, "mean_token_accuracy": 0.8629506826400757, "num_tokens": 162609557.0, "step": 4279 }, { "epoch": 0.5444599923673833, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.573858261108398, "learning_rate": 5e-05, "loss": 1.1416, "mean_token_accuracy": 0.8619215488433838, "num_tokens": 162642943.0, "step": 4280 }, { "epoch": 0.5445872026459738, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.669727325439453, "learning_rate": 5e-05, "loss": 1.0209, "mean_token_accuracy": 0.8599357604980469, "num_tokens": 162676444.0, "step": 4281 }, { "epoch": 0.5447144129245644, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.163633346557617, "learning_rate": 5e-05, "loss": 1.0937, "mean_token_accuracy": 0.856020987033844, "num_tokens": 162715445.0, "step": 4282 }, { "epoch": 0.5448416232031548, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.702818870544434, "learning_rate": 5e-05, "loss": 1.0787, "mean_token_accuracy": 0.8431117534637451, "num_tokens": 162751306.0, "step": 4283 }, { "epoch": 0.5449688334817453, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.642288208007812, "learning_rate": 5e-05, "loss": 1.1278, "mean_token_accuracy": 0.8297793865203857, "num_tokens": 162793587.0, "step": 4284 }, { "epoch": 0.5450960437603358, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.286438941955566, "learning_rate": 5e-05, "loss": 0.9985, "mean_token_accuracy": 0.8433554172515869, "num_tokens": 162831000.0, "step": 4285 }, { "epoch": 0.5452232540389264, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.43129825592041, "learning_rate": 5e-05, "loss": 0.9447, "mean_token_accuracy": 0.8826227784156799, "num_tokens": 162863913.0, "step": 4286 }, { "epoch": 0.5453504643175169, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.776084899902344, "learning_rate": 5e-05, "loss": 0.8659, "mean_token_accuracy": 0.8639990091323853, "num_tokens": 162899135.0, "step": 4287 }, { "epoch": 0.5454776745961074, "ewc_loss": 0.625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.699288368225098, "learning_rate": 5e-05, "loss": 1.0078, "mean_token_accuracy": 0.870373010635376, "num_tokens": 162932534.0, "step": 4288 }, { "epoch": 0.5456048848746978, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.7077717781066895, "learning_rate": 5e-05, "loss": 0.8819, "mean_token_accuracy": 0.8606346845626831, "num_tokens": 162973199.0, "step": 4289 }, { "epoch": 0.5457320951532884, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 9.963109970092773, "learning_rate": 5e-05, "loss": 1.1249, "mean_token_accuracy": 0.8530721068382263, "num_tokens": 163007018.0, "step": 4290 }, { "epoch": 0.5458593054318789, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.230260848999023, "learning_rate": 5e-05, "loss": 0.9732, "mean_token_accuracy": 0.8532204627990723, "num_tokens": 163044381.0, "step": 4291 }, { "epoch": 0.5459865157104694, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.536766052246094, "learning_rate": 5e-05, "loss": 1.0484, "mean_token_accuracy": 0.8587960004806519, "num_tokens": 163079606.0, "step": 4292 }, { "epoch": 0.5461137259890599, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.465943336486816, "learning_rate": 5e-05, "loss": 0.9944, "mean_token_accuracy": 0.8527586460113525, "num_tokens": 163120421.0, "step": 4293 }, { "epoch": 0.5462409362676505, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.783673286437988, "learning_rate": 5e-05, "loss": 1.0373, "mean_token_accuracy": 0.8417140245437622, "num_tokens": 163164761.0, "step": 4294 }, { "epoch": 0.5463681465462409, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.178290367126465, "learning_rate": 5e-05, "loss": 1.0569, "mean_token_accuracy": 0.8143379092216492, "num_tokens": 163212432.0, "step": 4295 }, { "epoch": 0.5464953568248314, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.36707592010498, "learning_rate": 5e-05, "loss": 0.8761, "mean_token_accuracy": 0.875845193862915, "num_tokens": 163251333.0, "step": 4296 }, { "epoch": 0.5466225671034219, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.626028060913086, "learning_rate": 5e-05, "loss": 0.8342, "mean_token_accuracy": 0.857134222984314, "num_tokens": 163286428.0, "step": 4297 }, { "epoch": 0.5467497773820125, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.4772310256958, "learning_rate": 5e-05, "loss": 0.9067, "mean_token_accuracy": 0.8685730695724487, "num_tokens": 163318974.0, "step": 4298 }, { "epoch": 0.546876987660603, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.114159107208252, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.8509745597839355, "num_tokens": 163358621.0, "step": 4299 }, { "epoch": 0.5470041979391935, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.87306022644043, "learning_rate": 5e-05, "loss": 0.9457, "mean_token_accuracy": 0.8668107986450195, "num_tokens": 163401062.0, "step": 4300 }, { "epoch": 0.5471314082177839, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.020175457000732, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.8533298969268799, "num_tokens": 163436194.0, "step": 4301 }, { "epoch": 0.5472586184963745, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.139564514160156, "learning_rate": 5e-05, "loss": 1.0295, "mean_token_accuracy": 0.8522804975509644, "num_tokens": 163476403.0, "step": 4302 }, { "epoch": 0.547385828774965, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.457050800323486, "learning_rate": 5e-05, "loss": 0.8471, "mean_token_accuracy": 0.8562477827072144, "num_tokens": 163515822.0, "step": 4303 }, { "epoch": 0.5475130390535555, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.800519943237305, "learning_rate": 5e-05, "loss": 1.0274, "mean_token_accuracy": 0.846840500831604, "num_tokens": 163553611.0, "step": 4304 }, { "epoch": 0.5476402493321461, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.906150817871094, "learning_rate": 5e-05, "loss": 0.8654, "mean_token_accuracy": 0.8657885789871216, "num_tokens": 163588591.0, "step": 4305 }, { "epoch": 0.5477674596107366, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.997568130493164, "learning_rate": 5e-05, "loss": 0.8898, "mean_token_accuracy": 0.8590989708900452, "num_tokens": 163630857.0, "step": 4306 }, { "epoch": 0.5478946698893271, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.883424282073975, "learning_rate": 5e-05, "loss": 0.8764, "mean_token_accuracy": 0.8476751446723938, "num_tokens": 163666684.0, "step": 4307 }, { "epoch": 0.5480218801679175, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.544836044311523, "learning_rate": 5e-05, "loss": 0.8266, "mean_token_accuracy": 0.8551739454269409, "num_tokens": 163707409.0, "step": 4308 }, { "epoch": 0.5481490904465081, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.595982074737549, "learning_rate": 5e-05, "loss": 0.7801, "mean_token_accuracy": 0.8643690347671509, "num_tokens": 163745190.0, "step": 4309 }, { "epoch": 0.5482763007250986, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.588764190673828, "learning_rate": 5e-05, "loss": 0.8263, "mean_token_accuracy": 0.8565894961357117, "num_tokens": 163781538.0, "step": 4310 }, { "epoch": 0.5484035110036891, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.261351585388184, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.857498049736023, "num_tokens": 163816412.0, "step": 4311 }, { "epoch": 0.5485307212822796, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.700974464416504, "learning_rate": 5e-05, "loss": 0.7872, "mean_token_accuracy": 0.8652351498603821, "num_tokens": 163853202.0, "step": 4312 }, { "epoch": 0.5486579315608702, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.075394630432129, "learning_rate": 5e-05, "loss": 0.7631, "mean_token_accuracy": 0.8556053638458252, "num_tokens": 163893669.0, "step": 4313 }, { "epoch": 0.5487851418394606, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.781752586364746, "learning_rate": 5e-05, "loss": 0.7965, "mean_token_accuracy": 0.8610734939575195, "num_tokens": 163932372.0, "step": 4314 }, { "epoch": 0.5489123521180511, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.9321489334106445, "learning_rate": 5e-05, "loss": 0.7006, "mean_token_accuracy": 0.8704794049263, "num_tokens": 163969431.0, "step": 4315 }, { "epoch": 0.5490395623966416, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.936434745788574, "learning_rate": 5e-05, "loss": 0.8281, "mean_token_accuracy": 0.8558279871940613, "num_tokens": 164006776.0, "step": 4316 }, { "epoch": 0.5491667726752322, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.733921051025391, "learning_rate": 5e-05, "loss": 0.6877, "mean_token_accuracy": 0.8699795603752136, "num_tokens": 164042178.0, "step": 4317 }, { "epoch": 0.5492939829538227, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.160752296447754, "learning_rate": 5e-05, "loss": 0.8603, "mean_token_accuracy": 0.8498224020004272, "num_tokens": 164089028.0, "step": 4318 }, { "epoch": 0.5494211932324132, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.687386989593506, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.8557530641555786, "num_tokens": 164130311.0, "step": 4319 }, { "epoch": 0.5495484035110036, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.141265869140625, "learning_rate": 5e-05, "loss": 0.8477, "mean_token_accuracy": 0.8564313054084778, "num_tokens": 164165398.0, "step": 4320 }, { "epoch": 0.5496756137895942, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.840444564819336, "learning_rate": 5e-05, "loss": 0.7144, "mean_token_accuracy": 0.8674964904785156, "num_tokens": 164202724.0, "step": 4321 }, { "epoch": 0.5498028240681847, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.963745594024658, "learning_rate": 5e-05, "loss": 0.8788, "mean_token_accuracy": 0.8415509462356567, "num_tokens": 164247666.0, "step": 4322 }, { "epoch": 0.5499300343467752, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.065786361694336, "learning_rate": 5e-05, "loss": 0.747, "mean_token_accuracy": 0.8593059182167053, "num_tokens": 164292323.0, "step": 4323 }, { "epoch": 0.5500572446253658, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.626032829284668, "learning_rate": 5e-05, "loss": 0.742, "mean_token_accuracy": 0.8738803863525391, "num_tokens": 164331160.0, "step": 4324 }, { "epoch": 0.5501844549039563, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.3891754150390625, "learning_rate": 5e-05, "loss": 0.7148, "mean_token_accuracy": 0.8736847639083862, "num_tokens": 164365457.0, "step": 4325 }, { "epoch": 0.5503116651825467, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.164352893829346, "learning_rate": 5e-05, "loss": 0.7786, "mean_token_accuracy": 0.8512712121009827, "num_tokens": 164403731.0, "step": 4326 }, { "epoch": 0.5504388754611372, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.799011707305908, "learning_rate": 5e-05, "loss": 0.7757, "mean_token_accuracy": 0.8634619116783142, "num_tokens": 164436616.0, "step": 4327 }, { "epoch": 0.5505660857397278, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.667544364929199, "learning_rate": 5e-05, "loss": 0.6944, "mean_token_accuracy": 0.8629319667816162, "num_tokens": 164472453.0, "step": 4328 }, { "epoch": 0.5506932960183183, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.27036190032959, "learning_rate": 5e-05, "loss": 0.8345, "mean_token_accuracy": 0.8514729142189026, "num_tokens": 164506551.0, "step": 4329 }, { "epoch": 0.5508205062969088, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.2784552574157715, "learning_rate": 5e-05, "loss": 0.6453, "mean_token_accuracy": 0.872829258441925, "num_tokens": 164542638.0, "step": 4330 }, { "epoch": 0.5509477165754993, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.742535591125488, "learning_rate": 5e-05, "loss": 0.8822, "mean_token_accuracy": 0.859276533126831, "num_tokens": 164580071.0, "step": 4331 }, { "epoch": 0.5510749268540898, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.124238967895508, "learning_rate": 5e-05, "loss": 0.7488, "mean_token_accuracy": 0.8460108041763306, "num_tokens": 164617894.0, "step": 4332 }, { "epoch": 0.5512021371326803, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.94631290435791, "learning_rate": 5e-05, "loss": 1.0104, "mean_token_accuracy": 0.8362700343132019, "num_tokens": 164660509.0, "step": 4333 }, { "epoch": 0.5513293474112708, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.59662389755249, "learning_rate": 5e-05, "loss": 0.7063, "mean_token_accuracy": 0.8691402673721313, "num_tokens": 164699683.0, "step": 4334 }, { "epoch": 0.5514565576898613, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.385688781738281, "learning_rate": 5e-05, "loss": 0.9195, "mean_token_accuracy": 0.8569539785385132, "num_tokens": 164733134.0, "step": 4335 }, { "epoch": 0.5515837679684519, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.47589635848999, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8514367938041687, "num_tokens": 164771911.0, "step": 4336 }, { "epoch": 0.5517109782470424, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.274225234985352, "learning_rate": 5e-05, "loss": 0.8126, "mean_token_accuracy": 0.8490637540817261, "num_tokens": 164810579.0, "step": 4337 }, { "epoch": 0.5518381885256328, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.17477798461914, "learning_rate": 5e-05, "loss": 0.8103, "mean_token_accuracy": 0.8663775324821472, "num_tokens": 164850215.0, "step": 4338 }, { "epoch": 0.5519653988042234, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.649105548858643, "learning_rate": 5e-05, "loss": 0.7485, "mean_token_accuracy": 0.8540962934494019, "num_tokens": 164889523.0, "step": 4339 }, { "epoch": 0.5520926090828139, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.330082893371582, "learning_rate": 5e-05, "loss": 0.8045, "mean_token_accuracy": 0.8747062087059021, "num_tokens": 164925946.0, "step": 4340 }, { "epoch": 0.5522198193614044, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.822788238525391, "learning_rate": 5e-05, "loss": 0.7332, "mean_token_accuracy": 0.8642294406890869, "num_tokens": 164961954.0, "step": 4341 }, { "epoch": 0.5523470296399949, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.8058295249938965, "learning_rate": 5e-05, "loss": 0.804, "mean_token_accuracy": 0.8607394695281982, "num_tokens": 165008013.0, "step": 4342 }, { "epoch": 0.5524742399185855, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.387604713439941, "learning_rate": 5e-05, "loss": 0.7647, "mean_token_accuracy": 0.8636578917503357, "num_tokens": 165037723.0, "step": 4343 }, { "epoch": 0.5526014501971759, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.132899761199951, "learning_rate": 5e-05, "loss": 0.7376, "mean_token_accuracy": 0.8626008629798889, "num_tokens": 165071432.0, "step": 4344 }, { "epoch": 0.5527286604757664, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.833754539489746, "learning_rate": 5e-05, "loss": 0.773, "mean_token_accuracy": 0.8661438822746277, "num_tokens": 165118784.0, "step": 4345 }, { "epoch": 0.5528558707543569, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.767727851867676, "learning_rate": 5e-05, "loss": 0.7847, "mean_token_accuracy": 0.8422722816467285, "num_tokens": 165163281.0, "step": 4346 }, { "epoch": 0.5529830810329475, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.082331657409668, "learning_rate": 5e-05, "loss": 0.8293, "mean_token_accuracy": 0.8515236377716064, "num_tokens": 165195615.0, "step": 4347 }, { "epoch": 0.553110291311538, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.631657600402832, "learning_rate": 5e-05, "loss": 0.73, "mean_token_accuracy": 0.8558381795883179, "num_tokens": 165233710.0, "step": 4348 }, { "epoch": 0.5532375015901285, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.161874771118164, "learning_rate": 5e-05, "loss": 0.8277, "mean_token_accuracy": 0.8569117188453674, "num_tokens": 165275788.0, "step": 4349 }, { "epoch": 0.5533647118687189, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.645996570587158, "learning_rate": 5e-05, "loss": 0.6957, "mean_token_accuracy": 0.8656612038612366, "num_tokens": 165316627.0, "step": 4350 }, { "epoch": 0.5534919221473095, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.132208824157715, "learning_rate": 5e-05, "loss": 0.8351, "mean_token_accuracy": 0.8593490123748779, "num_tokens": 165356643.0, "step": 4351 }, { "epoch": 0.5536191324259, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.722946643829346, "learning_rate": 5e-05, "loss": 0.737, "mean_token_accuracy": 0.854806661605835, "num_tokens": 165396897.0, "step": 4352 }, { "epoch": 0.5537463427044905, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.085522651672363, "learning_rate": 5e-05, "loss": 0.814, "mean_token_accuracy": 0.863127589225769, "num_tokens": 165434409.0, "step": 4353 }, { "epoch": 0.553873552983081, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.6963419914245605, "learning_rate": 5e-05, "loss": 0.7212, "mean_token_accuracy": 0.8577893972396851, "num_tokens": 165474674.0, "step": 4354 }, { "epoch": 0.5540007632616716, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.246675491333008, "learning_rate": 5e-05, "loss": 0.8931, "mean_token_accuracy": 0.846602201461792, "num_tokens": 165511865.0, "step": 4355 }, { "epoch": 0.554127973540262, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.467289447784424, "learning_rate": 5e-05, "loss": 0.7355, "mean_token_accuracy": 0.8575713038444519, "num_tokens": 165548589.0, "step": 4356 }, { "epoch": 0.5542551838188525, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.62820816040039, "learning_rate": 5e-05, "loss": 0.8893, "mean_token_accuracy": 0.8631480932235718, "num_tokens": 165589553.0, "step": 4357 }, { "epoch": 0.554382394097443, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.2535552978515625, "learning_rate": 5e-05, "loss": 0.8198, "mean_token_accuracy": 0.8336840867996216, "num_tokens": 165625499.0, "step": 4358 }, { "epoch": 0.5545096043760336, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.007843017578125, "learning_rate": 5e-05, "loss": 0.9493, "mean_token_accuracy": 0.8595025539398193, "num_tokens": 165663967.0, "step": 4359 }, { "epoch": 0.5546368146546241, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.484464645385742, "learning_rate": 5e-05, "loss": 0.7895, "mean_token_accuracy": 0.8525159358978271, "num_tokens": 165706382.0, "step": 4360 }, { "epoch": 0.5547640249332146, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.855034828186035, "learning_rate": 5e-05, "loss": 0.9679, "mean_token_accuracy": 0.8620368242263794, "num_tokens": 165742473.0, "step": 4361 }, { "epoch": 0.5548912352118052, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.290090084075928, "learning_rate": 5e-05, "loss": 0.8841, "mean_token_accuracy": 0.8518080711364746, "num_tokens": 165776329.0, "step": 4362 }, { "epoch": 0.5550184454903956, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.860146522521973, "learning_rate": 5e-05, "loss": 0.9038, "mean_token_accuracy": 0.8557736873626709, "num_tokens": 165812887.0, "step": 4363 }, { "epoch": 0.5551456557689861, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.051078796386719, "learning_rate": 5e-05, "loss": 0.8829, "mean_token_accuracy": 0.8701050281524658, "num_tokens": 165848933.0, "step": 4364 }, { "epoch": 0.5552728660475766, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.818044185638428, "learning_rate": 5e-05, "loss": 0.8092, "mean_token_accuracy": 0.8449541330337524, "num_tokens": 165887158.0, "step": 4365 }, { "epoch": 0.5554000763261672, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.42622184753418, "learning_rate": 5e-05, "loss": 0.9126, "mean_token_accuracy": 0.8566995859146118, "num_tokens": 165928214.0, "step": 4366 }, { "epoch": 0.5555272866047577, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.599022388458252, "learning_rate": 5e-05, "loss": 0.7504, "mean_token_accuracy": 0.8548080921173096, "num_tokens": 165964838.0, "step": 4367 }, { "epoch": 0.5556544968833482, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.231134414672852, "learning_rate": 5e-05, "loss": 0.9098, "mean_token_accuracy": 0.8487163782119751, "num_tokens": 165996097.0, "step": 4368 }, { "epoch": 0.5557817071619386, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.939143657684326, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.8562508821487427, "num_tokens": 166028660.0, "step": 4369 }, { "epoch": 0.5559089174405292, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.764851093292236, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.856239914894104, "num_tokens": 166065259.0, "step": 4370 }, { "epoch": 0.5560361277191197, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.25437068939209, "learning_rate": 5e-05, "loss": 0.7938, "mean_token_accuracy": 0.8536113500595093, "num_tokens": 166105757.0, "step": 4371 }, { "epoch": 0.5561633379977102, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.459899425506592, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.8422691226005554, "num_tokens": 166142362.0, "step": 4372 }, { "epoch": 0.5562905482763008, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.340686321258545, "learning_rate": 5e-05, "loss": 0.7769, "mean_token_accuracy": 0.8574252128601074, "num_tokens": 166179437.0, "step": 4373 }, { "epoch": 0.5564177585548913, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.391238212585449, "learning_rate": 5e-05, "loss": 0.7806, "mean_token_accuracy": 0.8501865863800049, "num_tokens": 166218555.0, "step": 4374 }, { "epoch": 0.5565449688334817, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.284265041351318, "learning_rate": 5e-05, "loss": 0.739, "mean_token_accuracy": 0.866670548915863, "num_tokens": 166252297.0, "step": 4375 }, { "epoch": 0.5566721791120722, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.399982452392578, "learning_rate": 5e-05, "loss": 0.7542, "mean_token_accuracy": 0.864238977432251, "num_tokens": 166289597.0, "step": 4376 }, { "epoch": 0.5567993893906628, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.218873500823975, "learning_rate": 5e-05, "loss": 0.7808, "mean_token_accuracy": 0.8478168845176697, "num_tokens": 166328655.0, "step": 4377 }, { "epoch": 0.5569265996692533, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.430053234100342, "learning_rate": 5e-05, "loss": 0.7572, "mean_token_accuracy": 0.855678915977478, "num_tokens": 166371825.0, "step": 4378 }, { "epoch": 0.5570538099478438, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.158357620239258, "learning_rate": 5e-05, "loss": 0.7394, "mean_token_accuracy": 0.8573664426803589, "num_tokens": 166407084.0, "step": 4379 }, { "epoch": 0.5571810202264343, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.477636814117432, "learning_rate": 5e-05, "loss": 0.7352, "mean_token_accuracy": 0.8641002178192139, "num_tokens": 166442272.0, "step": 4380 }, { "epoch": 0.5573082305050248, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.062032699584961, "learning_rate": 5e-05, "loss": 0.7615, "mean_token_accuracy": 0.8481942415237427, "num_tokens": 166475015.0, "step": 4381 }, { "epoch": 0.5574354407836153, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.582324504852295, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8599258661270142, "num_tokens": 166512421.0, "step": 4382 }, { "epoch": 0.5575626510622058, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.878372669219971, "learning_rate": 5e-05, "loss": 0.7442, "mean_token_accuracy": 0.8515611290931702, "num_tokens": 166554248.0, "step": 4383 }, { "epoch": 0.5576898613407963, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.797218322753906, "learning_rate": 5e-05, "loss": 0.7162, "mean_token_accuracy": 0.8752235174179077, "num_tokens": 166590664.0, "step": 4384 }, { "epoch": 0.5578170716193869, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.612132549285889, "learning_rate": 5e-05, "loss": 0.7159, "mean_token_accuracy": 0.8575066328048706, "num_tokens": 166625615.0, "step": 4385 }, { "epoch": 0.5579442818979774, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.196126937866211, "learning_rate": 5e-05, "loss": 0.8067, "mean_token_accuracy": 0.859883189201355, "num_tokens": 166662699.0, "step": 4386 }, { "epoch": 0.5580714921765678, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 6.181864261627197, "learning_rate": 5e-05, "loss": 0.6596, "mean_token_accuracy": 0.8640484809875488, "num_tokens": 166701218.0, "step": 4387 }, { "epoch": 0.5581987024551583, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.853734970092773, "learning_rate": 5e-05, "loss": 0.8964, "mean_token_accuracy": 0.855229377746582, "num_tokens": 166739699.0, "step": 4388 }, { "epoch": 0.5583259127337489, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.00022983551025390625, "grad_norm": 5.712014198303223, "learning_rate": 5e-05, "loss": 0.6605, "mean_token_accuracy": 0.8668562173843384, "num_tokens": 166782275.0, "step": 4389 }, { "epoch": 0.5584531230123394, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.702856063842773, "learning_rate": 5e-05, "loss": 0.9928, "mean_token_accuracy": 0.8618407249450684, "num_tokens": 166820535.0, "step": 4390 }, { "epoch": 0.5585803332909299, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 5.898224830627441, "learning_rate": 5e-05, "loss": 0.7764, "mean_token_accuracy": 0.8447031378746033, "num_tokens": 166865450.0, "step": 4391 }, { "epoch": 0.5587075435695205, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.043025970458984, "learning_rate": 5e-05, "loss": 1.0794, "mean_token_accuracy": 0.8629891872406006, "num_tokens": 166905838.0, "step": 4392 }, { "epoch": 0.5588347538481109, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 7.5340776443481445, "learning_rate": 5e-05, "loss": 1.046, "mean_token_accuracy": 0.8433986306190491, "num_tokens": 166943437.0, "step": 4393 }, { "epoch": 0.5589619641267014, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.614099502563477, "learning_rate": 5e-05, "loss": 0.9454, "mean_token_accuracy": 0.8644267320632935, "num_tokens": 166980863.0, "step": 4394 }, { "epoch": 0.5590891744052919, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 8.99466609954834, "learning_rate": 5e-05, "loss": 1.0328, "mean_token_accuracy": 0.874383270740509, "num_tokens": 167017861.0, "step": 4395 }, { "epoch": 0.5592163846838825, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.373900890350342, "learning_rate": 5e-05, "loss": 0.7027, "mean_token_accuracy": 0.8693073987960815, "num_tokens": 167051779.0, "step": 4396 }, { "epoch": 0.559343594962473, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 11.098978996276855, "learning_rate": 5e-05, "loss": 1.235, "mean_token_accuracy": 0.8577629327774048, "num_tokens": 167095205.0, "step": 4397 }, { "epoch": 0.5594708052410635, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.048304557800293, "learning_rate": 5e-05, "loss": 0.947, "mean_token_accuracy": 0.8535435199737549, "num_tokens": 167130594.0, "step": 4398 }, { "epoch": 0.5595980155196539, "ewc_loss": 0.91796875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000885009765625, "grad_norm": 12.27709674835205, "learning_rate": 5e-05, "loss": 1.3451, "mean_token_accuracy": 0.8580167889595032, "num_tokens": 167170816.0, "step": 4399 }, { "epoch": 0.5597252257982445, "ewc_loss": 0.98828125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000957489013671875, "grad_norm": 9.910165786743164, "learning_rate": 5e-05, "loss": 1.3716, "mean_token_accuracy": 0.8689570426940918, "num_tokens": 167205301.0, "step": 4400 }, { "epoch": 0.559852436076835, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 11.064743995666504, "learning_rate": 5e-05, "loss": 1.1426, "mean_token_accuracy": 0.8420851826667786, "num_tokens": 167240356.0, "step": 4401 }, { "epoch": 0.5599796463554255, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 9.199785232543945, "learning_rate": 5e-05, "loss": 1.158, "mean_token_accuracy": 0.8650510907173157, "num_tokens": 167269853.0, "step": 4402 }, { "epoch": 0.560106856634016, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.875102043151855, "learning_rate": 5e-05, "loss": 1.167, "mean_token_accuracy": 0.8621547818183899, "num_tokens": 167315398.0, "step": 4403 }, { "epoch": 0.5602340669126066, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.401251792907715, "learning_rate": 5e-05, "loss": 0.9828, "mean_token_accuracy": 0.8609694242477417, "num_tokens": 167352715.0, "step": 4404 }, { "epoch": 0.560361277191197, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 10.793880462646484, "learning_rate": 5e-05, "loss": 1.2414, "mean_token_accuracy": 0.8446601629257202, "num_tokens": 167391646.0, "step": 4405 }, { "epoch": 0.5604884874697875, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.519368171691895, "learning_rate": 5e-05, "loss": 1.0201, "mean_token_accuracy": 0.8540598154067993, "num_tokens": 167431676.0, "step": 4406 }, { "epoch": 0.560615697748378, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 10.704174995422363, "learning_rate": 5e-05, "loss": 1.1808, "mean_token_accuracy": 0.8528557419776917, "num_tokens": 167469121.0, "step": 4407 }, { "epoch": 0.5607429080269686, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 8.930603981018066, "learning_rate": 5e-05, "loss": 1.0443, "mean_token_accuracy": 0.8628817796707153, "num_tokens": 167511721.0, "step": 4408 }, { "epoch": 0.5608701183055591, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.148558616638184, "learning_rate": 5e-05, "loss": 1.0614, "mean_token_accuracy": 0.8671724200248718, "num_tokens": 167546965.0, "step": 4409 }, { "epoch": 0.5609973285841496, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.773003578186035, "learning_rate": 5e-05, "loss": 1.0561, "mean_token_accuracy": 0.8470146059989929, "num_tokens": 167584968.0, "step": 4410 }, { "epoch": 0.5611245388627402, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.655641555786133, "learning_rate": 5e-05, "loss": 1.097, "mean_token_accuracy": 0.847339391708374, "num_tokens": 167623124.0, "step": 4411 }, { "epoch": 0.5612517491413306, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.315150260925293, "learning_rate": 5e-05, "loss": 0.959, "mean_token_accuracy": 0.8544470071792603, "num_tokens": 167664877.0, "step": 4412 }, { "epoch": 0.5613789594199211, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.485617637634277, "learning_rate": 5e-05, "loss": 1.0436, "mean_token_accuracy": 0.8523045778274536, "num_tokens": 167706154.0, "step": 4413 }, { "epoch": 0.5615061696985116, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.900905132293701, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.8539905548095703, "num_tokens": 167740081.0, "step": 4414 }, { "epoch": 0.5616333799771022, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.591012001037598, "learning_rate": 5e-05, "loss": 1.0989, "mean_token_accuracy": 0.840861439704895, "num_tokens": 167783324.0, "step": 4415 }, { "epoch": 0.5617605902556927, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.819351673126221, "learning_rate": 5e-05, "loss": 0.8778, "mean_token_accuracy": 0.8621177673339844, "num_tokens": 167823216.0, "step": 4416 }, { "epoch": 0.5618878005342832, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.702195167541504, "learning_rate": 5e-05, "loss": 1.108, "mean_token_accuracy": 0.8411528468132019, "num_tokens": 167857452.0, "step": 4417 }, { "epoch": 0.5620150108128736, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.035025596618652, "learning_rate": 5e-05, "loss": 0.9065, "mean_token_accuracy": 0.8640613555908203, "num_tokens": 167896402.0, "step": 4418 }, { "epoch": 0.5621422210914642, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.521306991577148, "learning_rate": 5e-05, "loss": 1.0227, "mean_token_accuracy": 0.8603354096412659, "num_tokens": 167936441.0, "step": 4419 }, { "epoch": 0.5622694313700547, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.246415138244629, "learning_rate": 5e-05, "loss": 0.9184, "mean_token_accuracy": 0.869135856628418, "num_tokens": 167970786.0, "step": 4420 }, { "epoch": 0.5623966416486452, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.019746780395508, "learning_rate": 5e-05, "loss": 1.0126, "mean_token_accuracy": 0.8558434247970581, "num_tokens": 168011004.0, "step": 4421 }, { "epoch": 0.5625238519272358, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.137590408325195, "learning_rate": 5e-05, "loss": 0.9075, "mean_token_accuracy": 0.8566056489944458, "num_tokens": 168053418.0, "step": 4422 }, { "epoch": 0.5626510622058263, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.594801902770996, "learning_rate": 5e-05, "loss": 0.9528, "mean_token_accuracy": 0.8642875552177429, "num_tokens": 168088156.0, "step": 4423 }, { "epoch": 0.5627782724844167, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.727437496185303, "learning_rate": 5e-05, "loss": 0.8843, "mean_token_accuracy": 0.8487915992736816, "num_tokens": 168132329.0, "step": 4424 }, { "epoch": 0.5629054827630072, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.603474617004395, "learning_rate": 5e-05, "loss": 0.9164, "mean_token_accuracy": 0.8636940121650696, "num_tokens": 168167882.0, "step": 4425 }, { "epoch": 0.5630326930415978, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.330595016479492, "learning_rate": 5e-05, "loss": 0.8004, "mean_token_accuracy": 0.858507513999939, "num_tokens": 168203265.0, "step": 4426 }, { "epoch": 0.5631599033201883, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.832561492919922, "learning_rate": 5e-05, "loss": 0.9299, "mean_token_accuracy": 0.8717752695083618, "num_tokens": 168241410.0, "step": 4427 }, { "epoch": 0.5632871135987788, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.264069080352783, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.8636711835861206, "num_tokens": 168285412.0, "step": 4428 }, { "epoch": 0.5634143238773693, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.918939590454102, "learning_rate": 5e-05, "loss": 0.9941, "mean_token_accuracy": 0.8531888723373413, "num_tokens": 168324628.0, "step": 4429 }, { "epoch": 0.5635415341559598, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.516183376312256, "learning_rate": 5e-05, "loss": 0.8737, "mean_token_accuracy": 0.8525387048721313, "num_tokens": 168359759.0, "step": 4430 }, { "epoch": 0.5636687444345503, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.607608795166016, "learning_rate": 5e-05, "loss": 0.9763, "mean_token_accuracy": 0.8551783561706543, "num_tokens": 168393904.0, "step": 4431 }, { "epoch": 0.5637959547131408, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.711358070373535, "learning_rate": 5e-05, "loss": 0.8762, "mean_token_accuracy": 0.8536192178726196, "num_tokens": 168429718.0, "step": 4432 }, { "epoch": 0.5639231649917313, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.12839126586914, "learning_rate": 5e-05, "loss": 0.86, "mean_token_accuracy": 0.868831217288971, "num_tokens": 168462950.0, "step": 4433 }, { "epoch": 0.5640503752703219, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.64568567276001, "learning_rate": 5e-05, "loss": 0.8422, "mean_token_accuracy": 0.8578431606292725, "num_tokens": 168504095.0, "step": 4434 }, { "epoch": 0.5641775855489124, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.835786819458008, "learning_rate": 5e-05, "loss": 0.8647, "mean_token_accuracy": 0.854515552520752, "num_tokens": 168539931.0, "step": 4435 }, { "epoch": 0.5643047958275028, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.433169364929199, "learning_rate": 5e-05, "loss": 0.7791, "mean_token_accuracy": 0.8625872135162354, "num_tokens": 168573064.0, "step": 4436 }, { "epoch": 0.5644320061060933, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.829277992248535, "learning_rate": 5e-05, "loss": 0.832, "mean_token_accuracy": 0.8575108051300049, "num_tokens": 168607997.0, "step": 4437 }, { "epoch": 0.5645592163846839, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.197879314422607, "learning_rate": 5e-05, "loss": 0.7062, "mean_token_accuracy": 0.8779924511909485, "num_tokens": 168647590.0, "step": 4438 }, { "epoch": 0.5646864266632744, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.905572414398193, "learning_rate": 5e-05, "loss": 0.8146, "mean_token_accuracy": 0.8628157377243042, "num_tokens": 168681347.0, "step": 4439 }, { "epoch": 0.5648136369418649, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.0592875480651855, "learning_rate": 5e-05, "loss": 0.7451, "mean_token_accuracy": 0.8580354452133179, "num_tokens": 168722830.0, "step": 4440 }, { "epoch": 0.5649408472204555, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.926438808441162, "learning_rate": 5e-05, "loss": 0.8845, "mean_token_accuracy": 0.8435341119766235, "num_tokens": 168762381.0, "step": 4441 }, { "epoch": 0.5650680574990459, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.027803897857666, "learning_rate": 5e-05, "loss": 0.7326, "mean_token_accuracy": 0.8635132312774658, "num_tokens": 168802832.0, "step": 4442 }, { "epoch": 0.5651952677776364, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.89748477935791, "learning_rate": 5e-05, "loss": 0.8214, "mean_token_accuracy": 0.8560521006584167, "num_tokens": 168837780.0, "step": 4443 }, { "epoch": 0.5653224780562269, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.030342102050781, "learning_rate": 5e-05, "loss": 0.7924, "mean_token_accuracy": 0.8474279046058655, "num_tokens": 168871830.0, "step": 4444 }, { "epoch": 0.5654496883348175, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.846498966217041, "learning_rate": 5e-05, "loss": 0.7893, "mean_token_accuracy": 0.8637276887893677, "num_tokens": 168912477.0, "step": 4445 }, { "epoch": 0.565576898613408, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.083507061004639, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8491979837417603, "num_tokens": 168948976.0, "step": 4446 }, { "epoch": 0.5657041088919985, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.712058067321777, "learning_rate": 5e-05, "loss": 0.8084, "mean_token_accuracy": 0.8587614297866821, "num_tokens": 168992454.0, "step": 4447 }, { "epoch": 0.5658313191705889, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.210038185119629, "learning_rate": 5e-05, "loss": 0.7576, "mean_token_accuracy": 0.8540128469467163, "num_tokens": 169023135.0, "step": 4448 }, { "epoch": 0.5659585294491795, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.4980034828186035, "learning_rate": 5e-05, "loss": 0.7616, "mean_token_accuracy": 0.8595092296600342, "num_tokens": 169062066.0, "step": 4449 }, { "epoch": 0.56608573972777, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.380030155181885, "learning_rate": 5e-05, "loss": 0.8114, "mean_token_accuracy": 0.8388140201568604, "num_tokens": 169100317.0, "step": 4450 }, { "epoch": 0.5662129500063605, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.270941734313965, "learning_rate": 5e-05, "loss": 0.7263, "mean_token_accuracy": 0.8646400570869446, "num_tokens": 169128474.0, "step": 4451 }, { "epoch": 0.566340160284951, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.528509140014648, "learning_rate": 5e-05, "loss": 0.7226, "mean_token_accuracy": 0.8733450174331665, "num_tokens": 169164905.0, "step": 4452 }, { "epoch": 0.5664673705635416, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.057754039764404, "learning_rate": 5e-05, "loss": 0.7406, "mean_token_accuracy": 0.8550623059272766, "num_tokens": 169201153.0, "step": 4453 }, { "epoch": 0.566594580842132, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.713846206665039, "learning_rate": 5e-05, "loss": 0.7237, "mean_token_accuracy": 0.8741206526756287, "num_tokens": 169235325.0, "step": 4454 }, { "epoch": 0.5667217911207225, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.849856853485107, "learning_rate": 5e-05, "loss": 0.6652, "mean_token_accuracy": 0.8751114010810852, "num_tokens": 169276690.0, "step": 4455 }, { "epoch": 0.566849001399313, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.921571731567383, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.8445316553115845, "num_tokens": 169315751.0, "step": 4456 }, { "epoch": 0.5669762116779036, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.671864032745361, "learning_rate": 5e-05, "loss": 0.6981, "mean_token_accuracy": 0.8595298528671265, "num_tokens": 169354626.0, "step": 4457 }, { "epoch": 0.5671034219564941, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.107439994812012, "learning_rate": 5e-05, "loss": 0.7981, "mean_token_accuracy": 0.8566209077835083, "num_tokens": 169390255.0, "step": 4458 }, { "epoch": 0.5672306322350846, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.554145336151123, "learning_rate": 5e-05, "loss": 0.7237, "mean_token_accuracy": 0.8509107232093811, "num_tokens": 169431912.0, "step": 4459 }, { "epoch": 0.5673578425136752, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.239398956298828, "learning_rate": 5e-05, "loss": 0.7342, "mean_token_accuracy": 0.8824562430381775, "num_tokens": 169467803.0, "step": 4460 }, { "epoch": 0.5674850527922656, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.556570529937744, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.8498936295509338, "num_tokens": 169512252.0, "step": 4461 }, { "epoch": 0.5676122630708561, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.175053596496582, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.8622766733169556, "num_tokens": 169547632.0, "step": 4462 }, { "epoch": 0.5677394733494466, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.791983604431152, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.8571575880050659, "num_tokens": 169589355.0, "step": 4463 }, { "epoch": 0.5678666836280372, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.836359024047852, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8663265109062195, "num_tokens": 169627441.0, "step": 4464 }, { "epoch": 0.5679938939066277, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.176873683929443, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.8579099178314209, "num_tokens": 169659752.0, "step": 4465 }, { "epoch": 0.5681211041852182, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.335609436035156, "learning_rate": 5e-05, "loss": 0.7047, "mean_token_accuracy": 0.8758987188339233, "num_tokens": 169697197.0, "step": 4466 }, { "epoch": 0.5682483144638086, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.623819351196289, "learning_rate": 5e-05, "loss": 0.8312, "mean_token_accuracy": 0.839949369430542, "num_tokens": 169732523.0, "step": 4467 }, { "epoch": 0.5683755247423992, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.862447261810303, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8542154431343079, "num_tokens": 169771494.0, "step": 4468 }, { "epoch": 0.5685027350209897, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.023394584655762, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.8748663663864136, "num_tokens": 169809835.0, "step": 4469 }, { "epoch": 0.5686299452995802, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.505510330200195, "learning_rate": 5e-05, "loss": 0.7236, "mean_token_accuracy": 0.8509891033172607, "num_tokens": 169853276.0, "step": 4470 }, { "epoch": 0.5687571555781707, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.342988967895508, "learning_rate": 5e-05, "loss": 0.8692, "mean_token_accuracy": 0.84051913022995, "num_tokens": 169892106.0, "step": 4471 }, { "epoch": 0.5688843658567613, "ewc_loss": 0.28125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.322781562805176, "learning_rate": 5e-05, "loss": 0.7233, "mean_token_accuracy": 0.8503698110580444, "num_tokens": 169933635.0, "step": 4472 }, { "epoch": 0.5690115761353517, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.553010940551758, "learning_rate": 5e-05, "loss": 0.8494, "mean_token_accuracy": 0.8589494228363037, "num_tokens": 169971694.0, "step": 4473 }, { "epoch": 0.5691387864139422, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.365405082702637, "learning_rate": 5e-05, "loss": 0.7676, "mean_token_accuracy": 0.8416644930839539, "num_tokens": 170013285.0, "step": 4474 }, { "epoch": 0.5692659966925327, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.461658477783203, "learning_rate": 5e-05, "loss": 0.8693, "mean_token_accuracy": 0.8592759370803833, "num_tokens": 170052831.0, "step": 4475 }, { "epoch": 0.5693932069711233, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.727001667022705, "learning_rate": 5e-05, "loss": 0.785, "mean_token_accuracy": 0.8443938493728638, "num_tokens": 170089859.0, "step": 4476 }, { "epoch": 0.5695204172497138, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.969381332397461, "learning_rate": 5e-05, "loss": 0.8615, "mean_token_accuracy": 0.8531666994094849, "num_tokens": 170126494.0, "step": 4477 }, { "epoch": 0.5696476275283043, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.266047954559326, "learning_rate": 5e-05, "loss": 0.8013, "mean_token_accuracy": 0.8508814573287964, "num_tokens": 170167744.0, "step": 4478 }, { "epoch": 0.5697748378068948, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.361526966094971, "learning_rate": 5e-05, "loss": 0.7887, "mean_token_accuracy": 0.8573641777038574, "num_tokens": 170203988.0, "step": 4479 }, { "epoch": 0.5699020480854853, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.654759407043457, "learning_rate": 5e-05, "loss": 0.8704, "mean_token_accuracy": 0.8338449597358704, "num_tokens": 170243231.0, "step": 4480 }, { "epoch": 0.5700292583640758, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.048882007598877, "learning_rate": 5e-05, "loss": 0.8113, "mean_token_accuracy": 0.837040364742279, "num_tokens": 170280637.0, "step": 4481 }, { "epoch": 0.5701564686426663, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.673508167266846, "learning_rate": 5e-05, "loss": 0.8153, "mean_token_accuracy": 0.8484810590744019, "num_tokens": 170317106.0, "step": 4482 }, { "epoch": 0.5702836789212569, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.155209064483643, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8540109395980835, "num_tokens": 170353374.0, "step": 4483 }, { "epoch": 0.5704108891998474, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.333243370056152, "learning_rate": 5e-05, "loss": 0.7331, "mean_token_accuracy": 0.8620015978813171, "num_tokens": 170386968.0, "step": 4484 }, { "epoch": 0.5705380994784378, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.568719387054443, "learning_rate": 5e-05, "loss": 0.7691, "mean_token_accuracy": 0.862846851348877, "num_tokens": 170425386.0, "step": 4485 }, { "epoch": 0.5706653097570283, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.770198822021484, "learning_rate": 5e-05, "loss": 0.704, "mean_token_accuracy": 0.8605186343193054, "num_tokens": 170467828.0, "step": 4486 }, { "epoch": 0.5707925200356189, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.227496147155762, "learning_rate": 5e-05, "loss": 0.8225, "mean_token_accuracy": 0.8549895286560059, "num_tokens": 170506333.0, "step": 4487 }, { "epoch": 0.5709197303142094, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0002269744873046875, "grad_norm": 6.108334064483643, "learning_rate": 5e-05, "loss": 0.7114, "mean_token_accuracy": 0.8532013893127441, "num_tokens": 170542352.0, "step": 4488 }, { "epoch": 0.5710469405927999, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 9.12329387664795, "learning_rate": 5e-05, "loss": 0.8595, "mean_token_accuracy": 0.8719872236251831, "num_tokens": 170574251.0, "step": 4489 }, { "epoch": 0.5711741508713905, "ewc_loss": 0.25390625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00022411346435546875, "grad_norm": 5.566866874694824, "learning_rate": 5e-05, "loss": 0.6321, "mean_token_accuracy": 0.8693010210990906, "num_tokens": 170610294.0, "step": 4490 }, { "epoch": 0.5713013611499809, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 10.114973068237305, "learning_rate": 5e-05, "loss": 1.0181, "mean_token_accuracy": 0.8606503009796143, "num_tokens": 170649245.0, "step": 4491 }, { "epoch": 0.5714285714285714, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.240152835845947, "learning_rate": 5e-05, "loss": 0.7903, "mean_token_accuracy": 0.8543486595153809, "num_tokens": 170692106.0, "step": 4492 }, { "epoch": 0.5715557817071619, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.160874366760254, "learning_rate": 5e-05, "loss": 1.0882, "mean_token_accuracy": 0.8649082779884338, "num_tokens": 170730319.0, "step": 4493 }, { "epoch": 0.5716829919857525, "ewc_loss": 0.625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.18183708190918, "learning_rate": 5e-05, "loss": 1.0228, "mean_token_accuracy": 0.8654907941818237, "num_tokens": 170763805.0, "step": 4494 }, { "epoch": 0.571810202264343, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.261957168579102, "learning_rate": 5e-05, "loss": 0.9845, "mean_token_accuracy": 0.8404126167297363, "num_tokens": 170804966.0, "step": 4495 }, { "epoch": 0.5719374125429335, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 9.506820678710938, "learning_rate": 5e-05, "loss": 1.0884, "mean_token_accuracy": 0.86115562915802, "num_tokens": 170845652.0, "step": 4496 }, { "epoch": 0.5720646228215239, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.416847229003906, "learning_rate": 5e-05, "loss": 0.748, "mean_token_accuracy": 0.8636907339096069, "num_tokens": 170883449.0, "step": 4497 }, { "epoch": 0.5721918331001145, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 11.313288688659668, "learning_rate": 5e-05, "loss": 1.2693, "mean_token_accuracy": 0.850987434387207, "num_tokens": 170920187.0, "step": 4498 }, { "epoch": 0.572319043378705, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 7.8521904945373535, "learning_rate": 5e-05, "loss": 1.0588, "mean_token_accuracy": 0.8512998819351196, "num_tokens": 170952650.0, "step": 4499 }, { "epoch": 0.5724462536572955, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 11.40205192565918, "learning_rate": 5e-05, "loss": 1.2092, "mean_token_accuracy": 0.851777195930481, "num_tokens": 170989130.0, "step": 4500 }, { "epoch": 0.572573463935886, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 9.188223838806152, "learning_rate": 5e-05, "loss": 1.2312, "mean_token_accuracy": 0.8593697547912598, "num_tokens": 171030713.0, "step": 4501 }, { "epoch": 0.5727006742144766, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.552056312561035, "learning_rate": 5e-05, "loss": 1.0788, "mean_token_accuracy": 0.858754575252533, "num_tokens": 171066363.0, "step": 4502 }, { "epoch": 0.572827884493067, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 8.696499824523926, "learning_rate": 5e-05, "loss": 1.1218, "mean_token_accuracy": 0.8499876260757446, "num_tokens": 171108290.0, "step": 4503 }, { "epoch": 0.5729550947716575, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.97942066192627, "learning_rate": 5e-05, "loss": 1.0025, "mean_token_accuracy": 0.8825385570526123, "num_tokens": 171141990.0, "step": 4504 }, { "epoch": 0.573082305050248, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.065715789794922, "learning_rate": 5e-05, "loss": 0.9519, "mean_token_accuracy": 0.8574634790420532, "num_tokens": 171180032.0, "step": 4505 }, { "epoch": 0.5732095153288386, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.760897636413574, "learning_rate": 5e-05, "loss": 1.1196, "mean_token_accuracy": 0.8411468863487244, "num_tokens": 171218918.0, "step": 4506 }, { "epoch": 0.5733367256074291, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.669412136077881, "learning_rate": 5e-05, "loss": 0.9209, "mean_token_accuracy": 0.8572332859039307, "num_tokens": 171261028.0, "step": 4507 }, { "epoch": 0.5734639358860196, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.932626724243164, "learning_rate": 5e-05, "loss": 1.0661, "mean_token_accuracy": 0.8602153062820435, "num_tokens": 171301477.0, "step": 4508 }, { "epoch": 0.5735911461646102, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.7530903816223145, "learning_rate": 5e-05, "loss": 0.8951, "mean_token_accuracy": 0.8718185424804688, "num_tokens": 171338937.0, "step": 4509 }, { "epoch": 0.5737183564432006, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.002185821533203, "learning_rate": 5e-05, "loss": 1.1019, "mean_token_accuracy": 0.8567546606063843, "num_tokens": 171375058.0, "step": 4510 }, { "epoch": 0.5738455667217911, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.11665153503418, "learning_rate": 5e-05, "loss": 1.0033, "mean_token_accuracy": 0.8494603037834167, "num_tokens": 171410282.0, "step": 4511 }, { "epoch": 0.5739727770003816, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.672860145568848, "learning_rate": 5e-05, "loss": 1.0749, "mean_token_accuracy": 0.8570355176925659, "num_tokens": 171442864.0, "step": 4512 }, { "epoch": 0.5740999872789722, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.302814483642578, "learning_rate": 5e-05, "loss": 0.9688, "mean_token_accuracy": 0.8691976070404053, "num_tokens": 171476267.0, "step": 4513 }, { "epoch": 0.5742271975575627, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.082266807556152, "learning_rate": 5e-05, "loss": 1.0104, "mean_token_accuracy": 0.8554660677909851, "num_tokens": 171513173.0, "step": 4514 }, { "epoch": 0.5743544078361532, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.156662940979004, "learning_rate": 5e-05, "loss": 1.0296, "mean_token_accuracy": 0.8357137441635132, "num_tokens": 171555071.0, "step": 4515 }, { "epoch": 0.5744816181147436, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.627636909484863, "learning_rate": 5e-05, "loss": 0.9685, "mean_token_accuracy": 0.8581850528717041, "num_tokens": 171594263.0, "step": 4516 }, { "epoch": 0.5746088283933342, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.716484546661377, "learning_rate": 5e-05, "loss": 0.9004, "mean_token_accuracy": 0.8501855134963989, "num_tokens": 171635238.0, "step": 4517 }, { "epoch": 0.5747360386719247, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.695182800292969, "learning_rate": 5e-05, "loss": 0.995, "mean_token_accuracy": 0.8520261645317078, "num_tokens": 171674537.0, "step": 4518 }, { "epoch": 0.5748632489505152, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.225024700164795, "learning_rate": 5e-05, "loss": 0.7881, "mean_token_accuracy": 0.8699740767478943, "num_tokens": 171718606.0, "step": 4519 }, { "epoch": 0.5749904592291057, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.129083633422852, "learning_rate": 5e-05, "loss": 1.0041, "mean_token_accuracy": 0.8624448180198669, "num_tokens": 171757092.0, "step": 4520 }, { "epoch": 0.5751176695076963, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.238993167877197, "learning_rate": 5e-05, "loss": 0.8301, "mean_token_accuracy": 0.864611804485321, "num_tokens": 171799606.0, "step": 4521 }, { "epoch": 0.5752448797862867, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.2326021194458, "learning_rate": 5e-05, "loss": 1.0011, "mean_token_accuracy": 0.8661481738090515, "num_tokens": 171834686.0, "step": 4522 }, { "epoch": 0.5753720900648772, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.771945953369141, "learning_rate": 5e-05, "loss": 0.9841, "mean_token_accuracy": 0.8382259607315063, "num_tokens": 171870443.0, "step": 4523 }, { "epoch": 0.5754993003434677, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.641325950622559, "learning_rate": 5e-05, "loss": 0.9757, "mean_token_accuracy": 0.8634423613548279, "num_tokens": 171907615.0, "step": 4524 }, { "epoch": 0.5756265106220583, "ewc_loss": 0.5, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.012207984924316, "learning_rate": 5e-05, "loss": 0.9483, "mean_token_accuracy": 0.8467020988464355, "num_tokens": 171943120.0, "step": 4525 }, { "epoch": 0.5757537209006488, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.974485874176025, "learning_rate": 5e-05, "loss": 0.8921, "mean_token_accuracy": 0.8665676116943359, "num_tokens": 171980637.0, "step": 4526 }, { "epoch": 0.5758809311792393, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.770724296569824, "learning_rate": 5e-05, "loss": 0.8516, "mean_token_accuracy": 0.8598337769508362, "num_tokens": 172014232.0, "step": 4527 }, { "epoch": 0.5760081414578297, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.848046779632568, "learning_rate": 5e-05, "loss": 0.871, "mean_token_accuracy": 0.8570210933685303, "num_tokens": 172049639.0, "step": 4528 }, { "epoch": 0.5761353517364203, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.298639297485352, "learning_rate": 5e-05, "loss": 0.7987, "mean_token_accuracy": 0.8606448173522949, "num_tokens": 172086482.0, "step": 4529 }, { "epoch": 0.5762625620150108, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.145483016967773, "learning_rate": 5e-05, "loss": 0.8729, "mean_token_accuracy": 0.8658716678619385, "num_tokens": 172117220.0, "step": 4530 }, { "epoch": 0.5763897722936013, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.933990955352783, "learning_rate": 5e-05, "loss": 0.7455, "mean_token_accuracy": 0.8620692491531372, "num_tokens": 172149925.0, "step": 4531 }, { "epoch": 0.5765169825721919, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.418160438537598, "learning_rate": 5e-05, "loss": 0.8794, "mean_token_accuracy": 0.8700484037399292, "num_tokens": 172186903.0, "step": 4532 }, { "epoch": 0.5766441928507824, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.860116958618164, "learning_rate": 5e-05, "loss": 0.7746, "mean_token_accuracy": 0.8519264459609985, "num_tokens": 172226573.0, "step": 4533 }, { "epoch": 0.5767714031293728, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.477215766906738, "learning_rate": 5e-05, "loss": 0.9289, "mean_token_accuracy": 0.864516019821167, "num_tokens": 172263639.0, "step": 4534 }, { "epoch": 0.5768986134079633, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.030975341796875, "learning_rate": 5e-05, "loss": 0.7483, "mean_token_accuracy": 0.8681329488754272, "num_tokens": 172297740.0, "step": 4535 }, { "epoch": 0.5770258236865539, "ewc_loss": 0.5, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.31496524810791, "learning_rate": 5e-05, "loss": 0.9516, "mean_token_accuracy": 0.8490519523620605, "num_tokens": 172338243.0, "step": 4536 }, { "epoch": 0.5771530339651444, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.28603982925415, "learning_rate": 5e-05, "loss": 0.8154, "mean_token_accuracy": 0.8586705923080444, "num_tokens": 172383201.0, "step": 4537 }, { "epoch": 0.5772802442437349, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.920507431030273, "learning_rate": 5e-05, "loss": 0.8523, "mean_token_accuracy": 0.868330717086792, "num_tokens": 172420587.0, "step": 4538 }, { "epoch": 0.5774074545223254, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.524797439575195, "learning_rate": 5e-05, "loss": 0.8595, "mean_token_accuracy": 0.8427653312683105, "num_tokens": 172459194.0, "step": 4539 }, { "epoch": 0.5775346648009159, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.48999547958374, "learning_rate": 5e-05, "loss": 0.8557, "mean_token_accuracy": 0.8397396206855774, "num_tokens": 172492463.0, "step": 4540 }, { "epoch": 0.5776618750795064, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.616848945617676, "learning_rate": 5e-05, "loss": 0.838, "mean_token_accuracy": 0.8464614152908325, "num_tokens": 172531675.0, "step": 4541 }, { "epoch": 0.5777890853580969, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.314270973205566, "learning_rate": 5e-05, "loss": 0.772, "mean_token_accuracy": 0.8579803705215454, "num_tokens": 172571666.0, "step": 4542 }, { "epoch": 0.5779162956366874, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.568183898925781, "learning_rate": 5e-05, "loss": 0.7802, "mean_token_accuracy": 0.858396053314209, "num_tokens": 172608132.0, "step": 4543 }, { "epoch": 0.578043505915278, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.3027544021606445, "learning_rate": 5e-05, "loss": 0.7632, "mean_token_accuracy": 0.8608382940292358, "num_tokens": 172650881.0, "step": 4544 }, { "epoch": 0.5781707161938685, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.488598346710205, "learning_rate": 5e-05, "loss": 0.7634, "mean_token_accuracy": 0.8620232939720154, "num_tokens": 172684632.0, "step": 4545 }, { "epoch": 0.5782979264724589, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.3011860847473145, "learning_rate": 5e-05, "loss": 0.7237, "mean_token_accuracy": 0.8666527271270752, "num_tokens": 172724804.0, "step": 4546 }, { "epoch": 0.5784251367510495, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.439486503601074, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.8563501834869385, "num_tokens": 172762687.0, "step": 4547 }, { "epoch": 0.57855234702964, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.32381010055542, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.867506742477417, "num_tokens": 172802885.0, "step": 4548 }, { "epoch": 0.5786795573082305, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.356634616851807, "learning_rate": 5e-05, "loss": 0.738, "mean_token_accuracy": 0.8610882759094238, "num_tokens": 172836158.0, "step": 4549 }, { "epoch": 0.578806767586821, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.39307165145874, "learning_rate": 5e-05, "loss": 0.7569, "mean_token_accuracy": 0.8620181679725647, "num_tokens": 172872319.0, "step": 4550 }, { "epoch": 0.5789339778654116, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.2414398193359375, "learning_rate": 5e-05, "loss": 0.7258, "mean_token_accuracy": 0.863086462020874, "num_tokens": 172906803.0, "step": 4551 }, { "epoch": 0.579061188144002, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.504779815673828, "learning_rate": 5e-05, "loss": 0.756, "mean_token_accuracy": 0.860466718673706, "num_tokens": 172944095.0, "step": 4552 }, { "epoch": 0.5791883984225925, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.095651626586914, "learning_rate": 5e-05, "loss": 0.6999, "mean_token_accuracy": 0.8654924631118774, "num_tokens": 172980026.0, "step": 4553 }, { "epoch": 0.579315608701183, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.6608405113220215, "learning_rate": 5e-05, "loss": 0.7799, "mean_token_accuracy": 0.8528336882591248, "num_tokens": 173019558.0, "step": 4554 }, { "epoch": 0.5794428189797736, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.884025573730469, "learning_rate": 5e-05, "loss": 0.6931, "mean_token_accuracy": 0.8621759414672852, "num_tokens": 173059067.0, "step": 4555 }, { "epoch": 0.5795700292583641, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.9156599044799805, "learning_rate": 5e-05, "loss": 0.7567, "mean_token_accuracy": 0.8679276704788208, "num_tokens": 173096277.0, "step": 4556 }, { "epoch": 0.5796972395369546, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.61036491394043, "learning_rate": 5e-05, "loss": 0.7431, "mean_token_accuracy": 0.8463706374168396, "num_tokens": 173134746.0, "step": 4557 }, { "epoch": 0.5798244498155452, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.293479919433594, "learning_rate": 5e-05, "loss": 0.8939, "mean_token_accuracy": 0.8372205495834351, "num_tokens": 173164786.0, "step": 4558 }, { "epoch": 0.5799516600941356, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 0.00023365020751953125, "grad_norm": 6.24292516708374, "learning_rate": 5e-05, "loss": 0.7084, "mean_token_accuracy": 0.8515609502792358, "num_tokens": 173202948.0, "step": 4559 }, { "epoch": 0.5800788703727261, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.841787338256836, "learning_rate": 5e-05, "loss": 0.8879, "mean_token_accuracy": 0.8519605994224548, "num_tokens": 173235253.0, "step": 4560 }, { "epoch": 0.5802060806513166, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.00023746490478515625, "grad_norm": 5.9203877449035645, "learning_rate": 5e-05, "loss": 0.7014, "mean_token_accuracy": 0.8547629117965698, "num_tokens": 173279878.0, "step": 4561 }, { "epoch": 0.5803332909299072, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.357746124267578, "learning_rate": 5e-05, "loss": 0.9607, "mean_token_accuracy": 0.8593100309371948, "num_tokens": 173316719.0, "step": 4562 }, { "epoch": 0.5804605012084977, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.238920211791992, "learning_rate": 5e-05, "loss": 0.7216, "mean_token_accuracy": 0.8635035157203674, "num_tokens": 173347077.0, "step": 4563 }, { "epoch": 0.5805877114870882, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.11876106262207, "learning_rate": 5e-05, "loss": 1.0302, "mean_token_accuracy": 0.848871111869812, "num_tokens": 173386891.0, "step": 4564 }, { "epoch": 0.5807149217656786, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.400564670562744, "learning_rate": 5e-05, "loss": 0.8791, "mean_token_accuracy": 0.8600181341171265, "num_tokens": 173426919.0, "step": 4565 }, { "epoch": 0.5808421320442692, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.638335227966309, "learning_rate": 5e-05, "loss": 0.8352, "mean_token_accuracy": 0.8648024797439575, "num_tokens": 173461380.0, "step": 4566 }, { "epoch": 0.5809693423228597, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.686726570129395, "learning_rate": 5e-05, "loss": 0.9735, "mean_token_accuracy": 0.8598212599754333, "num_tokens": 173504680.0, "step": 4567 }, { "epoch": 0.5810965526014502, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.093233108520508, "learning_rate": 5e-05, "loss": 0.7214, "mean_token_accuracy": 0.8593423962593079, "num_tokens": 173541229.0, "step": 4568 }, { "epoch": 0.5812237628800407, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.979475021362305, "learning_rate": 5e-05, "loss": 1.0363, "mean_token_accuracy": 0.8552137613296509, "num_tokens": 173573805.0, "step": 4569 }, { "epoch": 0.5813509731586313, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.111267566680908, "learning_rate": 5e-05, "loss": 0.803, "mean_token_accuracy": 0.8476337194442749, "num_tokens": 173613899.0, "step": 4570 }, { "epoch": 0.5814781834372217, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 10.415005683898926, "learning_rate": 5e-05, "loss": 1.1083, "mean_token_accuracy": 0.8627682328224182, "num_tokens": 173650590.0, "step": 4571 }, { "epoch": 0.5816053937158122, "ewc_loss": 0.625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.021940231323242, "learning_rate": 5e-05, "loss": 1.0468, "mean_token_accuracy": 0.8576375246047974, "num_tokens": 173684067.0, "step": 4572 }, { "epoch": 0.5817326039944027, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.950597763061523, "learning_rate": 5e-05, "loss": 0.9498, "mean_token_accuracy": 0.8620770573616028, "num_tokens": 173721846.0, "step": 4573 }, { "epoch": 0.5818598142729933, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.050750732421875, "learning_rate": 5e-05, "loss": 1.1258, "mean_token_accuracy": 0.8405452966690063, "num_tokens": 173754616.0, "step": 4574 }, { "epoch": 0.5819870245515838, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.281224727630615, "learning_rate": 5e-05, "loss": 0.8355, "mean_token_accuracy": 0.854034960269928, "num_tokens": 173790208.0, "step": 4575 }, { "epoch": 0.5821142348301743, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.713549613952637, "learning_rate": 5e-05, "loss": 1.0963, "mean_token_accuracy": 0.8464478254318237, "num_tokens": 173833132.0, "step": 4576 }, { "epoch": 0.5822414451087647, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.389181137084961, "learning_rate": 5e-05, "loss": 0.9205, "mean_token_accuracy": 0.8517894744873047, "num_tokens": 173870526.0, "step": 4577 }, { "epoch": 0.5823686553873553, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.12082290649414, "learning_rate": 5e-05, "loss": 0.9795, "mean_token_accuracy": 0.8525575995445251, "num_tokens": 173901561.0, "step": 4578 }, { "epoch": 0.5824958656659458, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.142654418945312, "learning_rate": 5e-05, "loss": 1.0327, "mean_token_accuracy": 0.8470792770385742, "num_tokens": 173941197.0, "step": 4579 }, { "epoch": 0.5826230759445363, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.744997024536133, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8537312746047974, "num_tokens": 173981755.0, "step": 4580 }, { "epoch": 0.5827502862231269, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.691633224487305, "learning_rate": 5e-05, "loss": 0.9744, "mean_token_accuracy": 0.8530336022377014, "num_tokens": 174015831.0, "step": 4581 }, { "epoch": 0.5828774965017174, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.143921375274658, "learning_rate": 5e-05, "loss": 0.7889, "mean_token_accuracy": 0.8635170459747314, "num_tokens": 174051623.0, "step": 4582 }, { "epoch": 0.5830047067803078, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.704286575317383, "learning_rate": 5e-05, "loss": 0.9794, "mean_token_accuracy": 0.8454302549362183, "num_tokens": 174086422.0, "step": 4583 }, { "epoch": 0.5831319170588983, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.307811260223389, "learning_rate": 5e-05, "loss": 0.8733, "mean_token_accuracy": 0.8517037034034729, "num_tokens": 174128557.0, "step": 4584 }, { "epoch": 0.5832591273374889, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.189591407775879, "learning_rate": 5e-05, "loss": 0.8962, "mean_token_accuracy": 0.8489063382148743, "num_tokens": 174169704.0, "step": 4585 }, { "epoch": 0.5833863376160794, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.550539970397949, "learning_rate": 5e-05, "loss": 0.8971, "mean_token_accuracy": 0.8453223705291748, "num_tokens": 174213975.0, "step": 4586 }, { "epoch": 0.5835135478946699, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.635919094085693, "learning_rate": 5e-05, "loss": 0.8177, "mean_token_accuracy": 0.8551976084709167, "num_tokens": 174254425.0, "step": 4587 }, { "epoch": 0.5836407581732604, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.733160495758057, "learning_rate": 5e-05, "loss": 0.8728, "mean_token_accuracy": 0.8485555648803711, "num_tokens": 174292411.0, "step": 4588 }, { "epoch": 0.5837679684518509, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.239723205566406, "learning_rate": 5e-05, "loss": 0.8223, "mean_token_accuracy": 0.8392279148101807, "num_tokens": 174335103.0, "step": 4589 }, { "epoch": 0.5838951787304414, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.929209232330322, "learning_rate": 5e-05, "loss": 0.8741, "mean_token_accuracy": 0.8468156456947327, "num_tokens": 174377984.0, "step": 4590 }, { "epoch": 0.5840223890090319, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.021903991699219, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8559961318969727, "num_tokens": 174408211.0, "step": 4591 }, { "epoch": 0.5841495992876224, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.027478218078613, "learning_rate": 5e-05, "loss": 0.8531, "mean_token_accuracy": 0.8435291051864624, "num_tokens": 174444137.0, "step": 4592 }, { "epoch": 0.584276809566213, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.898774147033691, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.8501672148704529, "num_tokens": 174488382.0, "step": 4593 }, { "epoch": 0.5844040198448035, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.068459510803223, "learning_rate": 5e-05, "loss": 0.7833, "mean_token_accuracy": 0.8715547323226929, "num_tokens": 174524436.0, "step": 4594 }, { "epoch": 0.5845312301233939, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.815807819366455, "learning_rate": 5e-05, "loss": 0.7884, "mean_token_accuracy": 0.8474233746528625, "num_tokens": 174563830.0, "step": 4595 }, { "epoch": 0.5846584404019844, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.12955379486084, "learning_rate": 5e-05, "loss": 0.8266, "mean_token_accuracy": 0.8572548627853394, "num_tokens": 174598600.0, "step": 4596 }, { "epoch": 0.584785650680575, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.7201104164123535, "learning_rate": 5e-05, "loss": 0.7119, "mean_token_accuracy": 0.865107536315918, "num_tokens": 174631234.0, "step": 4597 }, { "epoch": 0.5849128609591655, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.206344604492188, "learning_rate": 5e-05, "loss": 0.8411, "mean_token_accuracy": 0.8554649949073792, "num_tokens": 174673933.0, "step": 4598 }, { "epoch": 0.585040071237756, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.650259017944336, "learning_rate": 5e-05, "loss": 0.7275, "mean_token_accuracy": 0.8597908616065979, "num_tokens": 174715801.0, "step": 4599 }, { "epoch": 0.5851672815163466, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.284721374511719, "learning_rate": 5e-05, "loss": 0.8448, "mean_token_accuracy": 0.8566228151321411, "num_tokens": 174757309.0, "step": 4600 }, { "epoch": 0.585294491794937, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.587479591369629, "learning_rate": 5e-05, "loss": 0.6808, "mean_token_accuracy": 0.8724812269210815, "num_tokens": 174787942.0, "step": 4601 }, { "epoch": 0.5854217020735275, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.404706954956055, "learning_rate": 5e-05, "loss": 0.8564, "mean_token_accuracy": 0.8598477840423584, "num_tokens": 174827483.0, "step": 4602 }, { "epoch": 0.585548912352118, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.5147600173950195, "learning_rate": 5e-05, "loss": 0.6878, "mean_token_accuracy": 0.870688259601593, "num_tokens": 174862847.0, "step": 4603 }, { "epoch": 0.5856761226307086, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.539292335510254, "learning_rate": 5e-05, "loss": 0.9108, "mean_token_accuracy": 0.8482244610786438, "num_tokens": 174903327.0, "step": 4604 }, { "epoch": 0.5858033329092991, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.576538562774658, "learning_rate": 5e-05, "loss": 0.7887, "mean_token_accuracy": 0.8442513346672058, "num_tokens": 174939278.0, "step": 4605 }, { "epoch": 0.5859305431878896, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.478184700012207, "learning_rate": 5e-05, "loss": 0.8892, "mean_token_accuracy": 0.8546916842460632, "num_tokens": 174972671.0, "step": 4606 }, { "epoch": 0.5860577534664801, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.855212688446045, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8489279747009277, "num_tokens": 175009753.0, "step": 4607 }, { "epoch": 0.5861849637450706, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.10246467590332, "learning_rate": 5e-05, "loss": 0.8203, "mean_token_accuracy": 0.8738625645637512, "num_tokens": 175044271.0, "step": 4608 }, { "epoch": 0.5863121740236611, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.260595321655273, "learning_rate": 5e-05, "loss": 0.8327, "mean_token_accuracy": 0.8511670827865601, "num_tokens": 175086132.0, "step": 4609 }, { "epoch": 0.5864393843022516, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.554121494293213, "learning_rate": 5e-05, "loss": 0.7534, "mean_token_accuracy": 0.8717277646064758, "num_tokens": 175114902.0, "step": 4610 }, { "epoch": 0.5865665945808421, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.646333694458008, "learning_rate": 5e-05, "loss": 0.8289, "mean_token_accuracy": 0.8525241017341614, "num_tokens": 175152274.0, "step": 4611 }, { "epoch": 0.5866938048594327, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.108682632446289, "learning_rate": 5e-05, "loss": 0.7244, "mean_token_accuracy": 0.863762378692627, "num_tokens": 175198449.0, "step": 4612 }, { "epoch": 0.5868210151380232, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.8668532371521, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8718843460083008, "num_tokens": 175236851.0, "step": 4613 }, { "epoch": 0.5869482254166136, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.918745040893555, "learning_rate": 5e-05, "loss": 0.7448, "mean_token_accuracy": 0.8590831756591797, "num_tokens": 175275694.0, "step": 4614 }, { "epoch": 0.5870754356952042, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.894956588745117, "learning_rate": 5e-05, "loss": 0.8584, "mean_token_accuracy": 0.8432300090789795, "num_tokens": 175321484.0, "step": 4615 }, { "epoch": 0.5872026459737947, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.9219255447387695, "learning_rate": 5e-05, "loss": 0.7313, "mean_token_accuracy": 0.8598594665527344, "num_tokens": 175358387.0, "step": 4616 }, { "epoch": 0.5873298562523852, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.827362537384033, "learning_rate": 5e-05, "loss": 0.8175, "mean_token_accuracy": 0.8511019945144653, "num_tokens": 175402363.0, "step": 4617 }, { "epoch": 0.5874570665309757, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.990665912628174, "learning_rate": 5e-05, "loss": 0.7569, "mean_token_accuracy": 0.8553102016448975, "num_tokens": 175440655.0, "step": 4618 }, { "epoch": 0.5875842768095663, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.744622707366943, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8558158874511719, "num_tokens": 175476041.0, "step": 4619 }, { "epoch": 0.5877114870881567, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.020070552825928, "learning_rate": 5e-05, "loss": 0.6864, "mean_token_accuracy": 0.8758276104927063, "num_tokens": 175514673.0, "step": 4620 }, { "epoch": 0.5878386973667472, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.717034339904785, "learning_rate": 5e-05, "loss": 0.7539, "mean_token_accuracy": 0.8689507842063904, "num_tokens": 175555396.0, "step": 4621 }, { "epoch": 0.5879659076453377, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.994016170501709, "learning_rate": 5e-05, "loss": 0.6827, "mean_token_accuracy": 0.8748500347137451, "num_tokens": 175597588.0, "step": 4622 }, { "epoch": 0.5880931179239283, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.768189907073975, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.8637259006500244, "num_tokens": 175636911.0, "step": 4623 }, { "epoch": 0.5882203282025188, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.8700690269470215, "learning_rate": 5e-05, "loss": 0.7795, "mean_token_accuracy": 0.8441405892372131, "num_tokens": 175671778.0, "step": 4624 }, { "epoch": 0.5883475384811093, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.926755905151367, "learning_rate": 5e-05, "loss": 0.7356, "mean_token_accuracy": 0.8774340748786926, "num_tokens": 175711218.0, "step": 4625 }, { "epoch": 0.5884747487596997, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.684873580932617, "learning_rate": 5e-05, "loss": 0.7334, "mean_token_accuracy": 0.8597806692123413, "num_tokens": 175744630.0, "step": 4626 }, { "epoch": 0.5886019590382903, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.169686317443848, "learning_rate": 5e-05, "loss": 0.8163, "mean_token_accuracy": 0.8603342175483704, "num_tokens": 175783342.0, "step": 4627 }, { "epoch": 0.5887291693168808, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.435861587524414, "learning_rate": 5e-05, "loss": 0.7294, "mean_token_accuracy": 0.8472483158111572, "num_tokens": 175820361.0, "step": 4628 }, { "epoch": 0.5888563795954713, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.54094409942627, "learning_rate": 5e-05, "loss": 0.8725, "mean_token_accuracy": 0.8539093732833862, "num_tokens": 175860388.0, "step": 4629 }, { "epoch": 0.5889835898740619, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.184401988983154, "learning_rate": 5e-05, "loss": 0.7323, "mean_token_accuracy": 0.8559415340423584, "num_tokens": 175895725.0, "step": 4630 }, { "epoch": 0.5891108001526524, "ewc_loss": 0.5, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.959364891052246, "learning_rate": 5e-05, "loss": 0.9152, "mean_token_accuracy": 0.8625723719596863, "num_tokens": 175933535.0, "step": 4631 }, { "epoch": 0.5892380104312428, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.199132442474365, "learning_rate": 5e-05, "loss": 0.7629, "mean_token_accuracy": 0.8479706645011902, "num_tokens": 175968681.0, "step": 4632 }, { "epoch": 0.5893652207098333, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.989177703857422, "learning_rate": 5e-05, "loss": 0.9882, "mean_token_accuracy": 0.8517845869064331, "num_tokens": 176006088.0, "step": 4633 }, { "epoch": 0.5894924309884239, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.886131286621094, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8741785287857056, "num_tokens": 176040855.0, "step": 4634 }, { "epoch": 0.5896196412670144, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.12866497039795, "learning_rate": 5e-05, "loss": 0.8968, "mean_token_accuracy": 0.858090877532959, "num_tokens": 176080015.0, "step": 4635 }, { "epoch": 0.5897468515456049, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.891629219055176, "learning_rate": 5e-05, "loss": 0.9356, "mean_token_accuracy": 0.8488786816596985, "num_tokens": 176120266.0, "step": 4636 }, { "epoch": 0.5898740618241954, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.768078327178955, "learning_rate": 5e-05, "loss": 0.7418, "mean_token_accuracy": 0.863660991191864, "num_tokens": 176150436.0, "step": 4637 }, { "epoch": 0.5900012721027859, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.884784698486328, "learning_rate": 5e-05, "loss": 0.9773, "mean_token_accuracy": 0.849665105342865, "num_tokens": 176185340.0, "step": 4638 }, { "epoch": 0.5901284823813764, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.943964958190918, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8514576554298401, "num_tokens": 176225372.0, "step": 4639 }, { "epoch": 0.5902556926599669, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.666610717773438, "learning_rate": 5e-05, "loss": 1.0069, "mean_token_accuracy": 0.8593809008598328, "num_tokens": 176267114.0, "step": 4640 }, { "epoch": 0.5903829029385574, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.296868324279785, "learning_rate": 5e-05, "loss": 0.7478, "mean_token_accuracy": 0.8736222982406616, "num_tokens": 176302937.0, "step": 4641 }, { "epoch": 0.590510113217148, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.593667984008789, "learning_rate": 5e-05, "loss": 1.0104, "mean_token_accuracy": 0.8693947792053223, "num_tokens": 176340307.0, "step": 4642 }, { "epoch": 0.5906373234957385, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 7.669976711273193, "learning_rate": 5e-05, "loss": 1.0341, "mean_token_accuracy": 0.8503019213676453, "num_tokens": 176385420.0, "step": 4643 }, { "epoch": 0.5907645337743289, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.241399765014648, "learning_rate": 5e-05, "loss": 0.9293, "mean_token_accuracy": 0.8553601503372192, "num_tokens": 176424975.0, "step": 4644 }, { "epoch": 0.5908917440529194, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.619622230529785, "learning_rate": 5e-05, "loss": 1.0756, "mean_token_accuracy": 0.8416705131530762, "num_tokens": 176465471.0, "step": 4645 }, { "epoch": 0.59101895433151, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.911545276641846, "learning_rate": 5e-05, "loss": 0.7967, "mean_token_accuracy": 0.8608155250549316, "num_tokens": 176493717.0, "step": 4646 }, { "epoch": 0.5911461646101005, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.999736785888672, "learning_rate": 5e-05, "loss": 1.0018, "mean_token_accuracy": 0.8570924401283264, "num_tokens": 176533636.0, "step": 4647 }, { "epoch": 0.591273374888691, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.019469738006592, "learning_rate": 5e-05, "loss": 0.8744, "mean_token_accuracy": 0.8567187786102295, "num_tokens": 176573248.0, "step": 4648 }, { "epoch": 0.5914005851672816, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.253805160522461, "learning_rate": 5e-05, "loss": 0.8789, "mean_token_accuracy": 0.8725660443305969, "num_tokens": 176614607.0, "step": 4649 }, { "epoch": 0.591527795445872, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.807767868041992, "learning_rate": 5e-05, "loss": 0.9427, "mean_token_accuracy": 0.8535177707672119, "num_tokens": 176657482.0, "step": 4650 }, { "epoch": 0.5916550057244625, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.140472412109375, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.8692992329597473, "num_tokens": 176694448.0, "step": 4651 }, { "epoch": 0.591782216003053, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.392874717712402, "learning_rate": 5e-05, "loss": 0.9676, "mean_token_accuracy": 0.8531920909881592, "num_tokens": 176726899.0, "step": 4652 }, { "epoch": 0.5919094262816436, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.696022987365723, "learning_rate": 5e-05, "loss": 0.7516, "mean_token_accuracy": 0.8524695038795471, "num_tokens": 176767495.0, "step": 4653 }, { "epoch": 0.5920366365602341, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.610688209533691, "learning_rate": 5e-05, "loss": 0.9469, "mean_token_accuracy": 0.8550319671630859, "num_tokens": 176802804.0, "step": 4654 }, { "epoch": 0.5921638468388246, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.740586757659912, "learning_rate": 5e-05, "loss": 0.7564, "mean_token_accuracy": 0.8687875270843506, "num_tokens": 176847224.0, "step": 4655 }, { "epoch": 0.592291057117415, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.602214813232422, "learning_rate": 5e-05, "loss": 0.9604, "mean_token_accuracy": 0.8556678891181946, "num_tokens": 176884423.0, "step": 4656 }, { "epoch": 0.5924182673960056, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 6.935647010803223, "learning_rate": 5e-05, "loss": 0.8278, "mean_token_accuracy": 0.8679838180541992, "num_tokens": 176917776.0, "step": 4657 }, { "epoch": 0.5925454776745961, "ewc_loss": 0.5, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.415923118591309, "learning_rate": 5e-05, "loss": 0.8898, "mean_token_accuracy": 0.8672152161598206, "num_tokens": 176951387.0, "step": 4658 }, { "epoch": 0.5926726879531866, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.024649620056152, "learning_rate": 5e-05, "loss": 0.8636, "mean_token_accuracy": 0.8503232002258301, "num_tokens": 176986533.0, "step": 4659 }, { "epoch": 0.5927998982317771, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0003108978271484375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.322556495666504, "learning_rate": 5e-05, "loss": 0.8831, "mean_token_accuracy": 0.8693690299987793, "num_tokens": 177021541.0, "step": 4660 }, { "epoch": 0.5929271085103677, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.892982482910156, "learning_rate": 5e-05, "loss": 0.8056, "mean_token_accuracy": 0.8610882759094238, "num_tokens": 177062933.0, "step": 4661 }, { "epoch": 0.5930543187889582, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.506782531738281, "learning_rate": 5e-05, "loss": 0.9824, "mean_token_accuracy": 0.8401644229888916, "num_tokens": 177102549.0, "step": 4662 }, { "epoch": 0.5931815290675486, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.625840187072754, "learning_rate": 5e-05, "loss": 0.773, "mean_token_accuracy": 0.8589091300964355, "num_tokens": 177139618.0, "step": 4663 }, { "epoch": 0.5933087393461391, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.8598051071167, "learning_rate": 5e-05, "loss": 0.9826, "mean_token_accuracy": 0.8514413833618164, "num_tokens": 177171912.0, "step": 4664 }, { "epoch": 0.5934359496247297, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.644283771514893, "learning_rate": 5e-05, "loss": 0.8047, "mean_token_accuracy": 0.8589516878128052, "num_tokens": 177214279.0, "step": 4665 }, { "epoch": 0.5935631599033202, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.929085731506348, "learning_rate": 5e-05, "loss": 1.0154, "mean_token_accuracy": 0.8557500839233398, "num_tokens": 177247277.0, "step": 4666 }, { "epoch": 0.5936903701819107, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.092325687408447, "learning_rate": 5e-05, "loss": 0.8749, "mean_token_accuracy": 0.8580178022384644, "num_tokens": 177287097.0, "step": 4667 }, { "epoch": 0.5938175804605013, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.465865135192871, "learning_rate": 5e-05, "loss": 0.9281, "mean_token_accuracy": 0.8719705939292908, "num_tokens": 177324400.0, "step": 4668 }, { "epoch": 0.5939447907390917, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.605315208435059, "learning_rate": 5e-05, "loss": 0.9057, "mean_token_accuracy": 0.8666649460792542, "num_tokens": 177362962.0, "step": 4669 }, { "epoch": 0.5940720010176822, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.728373050689697, "learning_rate": 5e-05, "loss": 0.8726, "mean_token_accuracy": 0.8588641285896301, "num_tokens": 177402708.0, "step": 4670 }, { "epoch": 0.5941992112962727, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.803506851196289, "learning_rate": 5e-05, "loss": 0.8635, "mean_token_accuracy": 0.8682195544242859, "num_tokens": 177433957.0, "step": 4671 }, { "epoch": 0.5943264215748633, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.382521152496338, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.863540768623352, "num_tokens": 177465265.0, "step": 4672 }, { "epoch": 0.5944536318534538, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.6312384605407715, "learning_rate": 5e-05, "loss": 0.8697, "mean_token_accuracy": 0.851032018661499, "num_tokens": 177495448.0, "step": 4673 }, { "epoch": 0.5945808421320443, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.450900554656982, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8636584281921387, "num_tokens": 177531582.0, "step": 4674 }, { "epoch": 0.5947080524106347, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.338944435119629, "learning_rate": 5e-05, "loss": 0.7906, "mean_token_accuracy": 0.8579477071762085, "num_tokens": 177565039.0, "step": 4675 }, { "epoch": 0.5948352626892253, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.6096954345703125, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.858356773853302, "num_tokens": 177601872.0, "step": 4676 }, { "epoch": 0.5949624729678158, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.167484283447266, "learning_rate": 5e-05, "loss": 0.8128, "mean_token_accuracy": 0.8449338674545288, "num_tokens": 177645693.0, "step": 4677 }, { "epoch": 0.5950896832464063, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.630370616912842, "learning_rate": 5e-05, "loss": 0.7817, "mean_token_accuracy": 0.8645004034042358, "num_tokens": 177685334.0, "step": 4678 }, { "epoch": 0.5952168935249968, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.13924503326416, "learning_rate": 5e-05, "loss": 0.8947, "mean_token_accuracy": 0.8194774389266968, "num_tokens": 177727851.0, "step": 4679 }, { "epoch": 0.5953441038035874, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.612329483032227, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.854549765586853, "num_tokens": 177769115.0, "step": 4680 }, { "epoch": 0.5954713140821778, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.132816314697266, "learning_rate": 5e-05, "loss": 0.7616, "mean_token_accuracy": 0.8528835773468018, "num_tokens": 177809094.0, "step": 4681 }, { "epoch": 0.5955985243607683, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.598166465759277, "learning_rate": 5e-05, "loss": 0.7775, "mean_token_accuracy": 0.858676552772522, "num_tokens": 177842409.0, "step": 4682 }, { "epoch": 0.5957257346393589, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.0589985847473145, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8530054092407227, "num_tokens": 177882525.0, "step": 4683 }, { "epoch": 0.5958529449179494, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.712869167327881, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.8418722152709961, "num_tokens": 177924293.0, "step": 4684 }, { "epoch": 0.5959801551965399, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.8894362449646, "learning_rate": 5e-05, "loss": 0.7287, "mean_token_accuracy": 0.8610364198684692, "num_tokens": 177964716.0, "step": 4685 }, { "epoch": 0.5961073654751304, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.934178352355957, "learning_rate": 5e-05, "loss": 0.7379, "mean_token_accuracy": 0.8795621991157532, "num_tokens": 177999029.0, "step": 4686 }, { "epoch": 0.5962345757537209, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.642243385314941, "learning_rate": 5e-05, "loss": 0.686, "mean_token_accuracy": 0.8718322515487671, "num_tokens": 178033872.0, "step": 4687 }, { "epoch": 0.5963617860323114, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.254997253417969, "learning_rate": 5e-05, "loss": 0.8173, "mean_token_accuracy": 0.8614046573638916, "num_tokens": 178077835.0, "step": 4688 }, { "epoch": 0.5964889963109019, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.362181663513184, "learning_rate": 5e-05, "loss": 0.6793, "mean_token_accuracy": 0.8667416572570801, "num_tokens": 178119205.0, "step": 4689 }, { "epoch": 0.5966162065894924, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.67491340637207, "learning_rate": 5e-05, "loss": 0.9227, "mean_token_accuracy": 0.8495374917984009, "num_tokens": 178157923.0, "step": 4690 }, { "epoch": 0.596743416868083, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.190631866455078, "learning_rate": 5e-05, "loss": 0.6589, "mean_token_accuracy": 0.8742042183876038, "num_tokens": 178200465.0, "step": 4691 }, { "epoch": 0.5968706271466735, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.982733726501465, "learning_rate": 5e-05, "loss": 1.0257, "mean_token_accuracy": 0.8301867246627808, "num_tokens": 178234767.0, "step": 4692 }, { "epoch": 0.5969978374252639, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.497652530670166, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8506709337234497, "num_tokens": 178269569.0, "step": 4693 }, { "epoch": 0.5971250477038544, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.614354133605957, "learning_rate": 5e-05, "loss": 0.9484, "mean_token_accuracy": 0.8587422370910645, "num_tokens": 178300789.0, "step": 4694 }, { "epoch": 0.597252257982445, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.3633246421813965, "learning_rate": 5e-05, "loss": 0.8388, "mean_token_accuracy": 0.8659806251525879, "num_tokens": 178333341.0, "step": 4695 }, { "epoch": 0.5973794682610355, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.43189811706543, "learning_rate": 5e-05, "loss": 0.8667, "mean_token_accuracy": 0.8433554172515869, "num_tokens": 178368743.0, "step": 4696 }, { "epoch": 0.597506678539626, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.31330680847168, "learning_rate": 5e-05, "loss": 0.8327, "mean_token_accuracy": 0.8785678148269653, "num_tokens": 178407432.0, "step": 4697 }, { "epoch": 0.5976338888182166, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.299612522125244, "learning_rate": 5e-05, "loss": 0.7426, "mean_token_accuracy": 0.8574658632278442, "num_tokens": 178440245.0, "step": 4698 }, { "epoch": 0.597761099096807, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.181722640991211, "learning_rate": 5e-05, "loss": 0.9574, "mean_token_accuracy": 0.8622146844863892, "num_tokens": 178484446.0, "step": 4699 }, { "epoch": 0.5978883093753975, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 5.930997371673584, "learning_rate": 5e-05, "loss": 0.7192, "mean_token_accuracy": 0.8612545728683472, "num_tokens": 178522390.0, "step": 4700 }, { "epoch": 0.598015519653988, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.716650009155273, "learning_rate": 5e-05, "loss": 1.0116, "mean_token_accuracy": 0.8639688491821289, "num_tokens": 178561671.0, "step": 4701 }, { "epoch": 0.5981427299325786, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 6.750526428222656, "learning_rate": 5e-05, "loss": 0.9299, "mean_token_accuracy": 0.8416141867637634, "num_tokens": 178605289.0, "step": 4702 }, { "epoch": 0.5982699402111691, "ewc_loss": 0.625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.265037536621094, "learning_rate": 5e-05, "loss": 1.0316, "mean_token_accuracy": 0.8656664490699768, "num_tokens": 178647947.0, "step": 4703 }, { "epoch": 0.5983971504897596, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.054329872131348, "learning_rate": 5e-05, "loss": 0.9799, "mean_token_accuracy": 0.871160626411438, "num_tokens": 178686779.0, "step": 4704 }, { "epoch": 0.59852436076835, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.717531204223633, "learning_rate": 5e-05, "loss": 0.9096, "mean_token_accuracy": 0.8487846851348877, "num_tokens": 178730517.0, "step": 4705 }, { "epoch": 0.5986515710469406, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.779106140136719, "learning_rate": 5e-05, "loss": 0.9948, "mean_token_accuracy": 0.859940767288208, "num_tokens": 178762521.0, "step": 4706 }, { "epoch": 0.5987787813255311, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 6.825253009796143, "learning_rate": 5e-05, "loss": 0.8377, "mean_token_accuracy": 0.8510474562644958, "num_tokens": 178804816.0, "step": 4707 }, { "epoch": 0.5989059916041216, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.725767135620117, "learning_rate": 5e-05, "loss": 1.0262, "mean_token_accuracy": 0.8448686003684998, "num_tokens": 178845413.0, "step": 4708 }, { "epoch": 0.5990332018827121, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.267243385314941, "learning_rate": 5e-05, "loss": 0.8779, "mean_token_accuracy": 0.8577749729156494, "num_tokens": 178886887.0, "step": 4709 }, { "epoch": 0.5991604121613027, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.6944050788879395, "learning_rate": 5e-05, "loss": 0.8306, "mean_token_accuracy": 0.8620907068252563, "num_tokens": 178922067.0, "step": 4710 }, { "epoch": 0.5992876224398932, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.031789779663086, "learning_rate": 5e-05, "loss": 0.9778, "mean_token_accuracy": 0.8377388715744019, "num_tokens": 178955645.0, "step": 4711 }, { "epoch": 0.5994148327184836, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.825447082519531, "learning_rate": 5e-05, "loss": 0.7447, "mean_token_accuracy": 0.8594324588775635, "num_tokens": 178992389.0, "step": 4712 }, { "epoch": 0.5995420429970741, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.474119186401367, "learning_rate": 5e-05, "loss": 0.9398, "mean_token_accuracy": 0.847061038017273, "num_tokens": 179031140.0, "step": 4713 }, { "epoch": 0.5996692532756647, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.643188953399658, "learning_rate": 5e-05, "loss": 0.7059, "mean_token_accuracy": 0.8748267889022827, "num_tokens": 179065486.0, "step": 4714 }, { "epoch": 0.5997964635542552, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.574335098266602, "learning_rate": 5e-05, "loss": 0.9373, "mean_token_accuracy": 0.8553124666213989, "num_tokens": 179104957.0, "step": 4715 }, { "epoch": 0.5999236738328457, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.762452125549316, "learning_rate": 5e-05, "loss": 0.7959, "mean_token_accuracy": 0.8590492010116577, "num_tokens": 179146022.0, "step": 4716 }, { "epoch": 0.6000508841114363, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.515826225280762, "learning_rate": 5e-05, "loss": 0.8993, "mean_token_accuracy": 0.8684883713722229, "num_tokens": 179184691.0, "step": 4717 }, { "epoch": 0.6001780943900267, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 6.86432409286499, "learning_rate": 5e-05, "loss": 0.7841, "mean_token_accuracy": 0.862938642501831, "num_tokens": 179223734.0, "step": 4718 }, { "epoch": 0.6003053046686172, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.466045379638672, "learning_rate": 5e-05, "loss": 0.8978, "mean_token_accuracy": 0.8669483661651611, "num_tokens": 179265997.0, "step": 4719 }, { "epoch": 0.6004325149472077, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 6.872827053070068, "learning_rate": 5e-05, "loss": 0.8215, "mean_token_accuracy": 0.8536187410354614, "num_tokens": 179302734.0, "step": 4720 }, { "epoch": 0.6005597252257983, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.531895637512207, "learning_rate": 5e-05, "loss": 0.9715, "mean_token_accuracy": 0.8538182377815247, "num_tokens": 179343779.0, "step": 4721 }, { "epoch": 0.6006869355043888, "ewc_loss": 0.375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.745888710021973, "learning_rate": 5e-05, "loss": 0.7784, "mean_token_accuracy": 0.8684127926826477, "num_tokens": 179384713.0, "step": 4722 }, { "epoch": 0.6008141457829793, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.769527435302734, "learning_rate": 5e-05, "loss": 0.9458, "mean_token_accuracy": 0.8644396066665649, "num_tokens": 179418005.0, "step": 4723 }, { "epoch": 0.6009413560615697, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.7321624755859375, "learning_rate": 5e-05, "loss": 0.8323, "mean_token_accuracy": 0.8520849347114563, "num_tokens": 179459437.0, "step": 4724 }, { "epoch": 0.6010685663401603, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.876351356506348, "learning_rate": 5e-05, "loss": 0.9819, "mean_token_accuracy": 0.86183762550354, "num_tokens": 179490534.0, "step": 4725 }, { "epoch": 0.6011957766187508, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.0533881187438965, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.8559896349906921, "num_tokens": 179533141.0, "step": 4726 }, { "epoch": 0.6013229868973413, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.552695274353027, "learning_rate": 5e-05, "loss": 0.9999, "mean_token_accuracy": 0.8496330380439758, "num_tokens": 179565438.0, "step": 4727 }, { "epoch": 0.6014501971759318, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.5043768882751465, "learning_rate": 5e-05, "loss": 0.9635, "mean_token_accuracy": 0.8412008285522461, "num_tokens": 179605184.0, "step": 4728 }, { "epoch": 0.6015774074545224, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.889937400817871, "learning_rate": 5e-05, "loss": 0.8936, "mean_token_accuracy": 0.8588794469833374, "num_tokens": 179643355.0, "step": 4729 }, { "epoch": 0.6017046177331128, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.721707820892334, "learning_rate": 5e-05, "loss": 0.8938, "mean_token_accuracy": 0.8625073432922363, "num_tokens": 179682471.0, "step": 4730 }, { "epoch": 0.6018318280117033, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.4889349937438965, "learning_rate": 5e-05, "loss": 0.8673, "mean_token_accuracy": 0.8501465916633606, "num_tokens": 179722290.0, "step": 4731 }, { "epoch": 0.6019590382902938, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.562554359436035, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.8457024693489075, "num_tokens": 179762409.0, "step": 4732 }, { "epoch": 0.6020862485688844, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.543806552886963, "learning_rate": 5e-05, "loss": 0.8511, "mean_token_accuracy": 0.849989652633667, "num_tokens": 179802561.0, "step": 4733 }, { "epoch": 0.6022134588474749, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.258451461791992, "learning_rate": 5e-05, "loss": 0.7908, "mean_token_accuracy": 0.860504686832428, "num_tokens": 179842923.0, "step": 4734 }, { "epoch": 0.6023406691260654, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.700602054595947, "learning_rate": 5e-05, "loss": 0.8373, "mean_token_accuracy": 0.8569011688232422, "num_tokens": 179882430.0, "step": 4735 }, { "epoch": 0.6024678794046558, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.11952543258667, "learning_rate": 5e-05, "loss": 0.792, "mean_token_accuracy": 0.8511117696762085, "num_tokens": 179923974.0, "step": 4736 }, { "epoch": 0.6025950896832464, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.684117317199707, "learning_rate": 5e-05, "loss": 0.7763, "mean_token_accuracy": 0.866858720779419, "num_tokens": 179966468.0, "step": 4737 }, { "epoch": 0.6027222999618369, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.109505653381348, "learning_rate": 5e-05, "loss": 0.713, "mean_token_accuracy": 0.873537540435791, "num_tokens": 180005356.0, "step": 4738 }, { "epoch": 0.6028495102404274, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.6636810302734375, "learning_rate": 5e-05, "loss": 0.8066, "mean_token_accuracy": 0.8570243120193481, "num_tokens": 180044765.0, "step": 4739 }, { "epoch": 0.602976720519018, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.059957504272461, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.8520022630691528, "num_tokens": 180082311.0, "step": 4740 }, { "epoch": 0.6031039307976085, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.718275547027588, "learning_rate": 5e-05, "loss": 0.7788, "mean_token_accuracy": 0.864392876625061, "num_tokens": 180122889.0, "step": 4741 }, { "epoch": 0.6032311410761989, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.956814765930176, "learning_rate": 5e-05, "loss": 0.7206, "mean_token_accuracy": 0.8656229376792908, "num_tokens": 180157425.0, "step": 4742 }, { "epoch": 0.6033583513547894, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.833791255950928, "learning_rate": 5e-05, "loss": 0.7878, "mean_token_accuracy": 0.8663907051086426, "num_tokens": 180196099.0, "step": 4743 }, { "epoch": 0.60348556163338, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.827183723449707, "learning_rate": 5e-05, "loss": 0.7879, "mean_token_accuracy": 0.842003345489502, "num_tokens": 180238036.0, "step": 4744 }, { "epoch": 0.6036127719119705, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.974983215332031, "learning_rate": 5e-05, "loss": 0.8041, "mean_token_accuracy": 0.8606762886047363, "num_tokens": 180276193.0, "step": 4745 }, { "epoch": 0.603739982190561, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.702920913696289, "learning_rate": 5e-05, "loss": 0.7464, "mean_token_accuracy": 0.8561383485794067, "num_tokens": 180313470.0, "step": 4746 }, { "epoch": 0.6038671924691515, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.149349212646484, "learning_rate": 5e-05, "loss": 0.8288, "mean_token_accuracy": 0.8627453446388245, "num_tokens": 180354588.0, "step": 4747 }, { "epoch": 0.603994402747742, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.6097283363342285, "learning_rate": 5e-05, "loss": 0.7294, "mean_token_accuracy": 0.8589255213737488, "num_tokens": 180395792.0, "step": 4748 }, { "epoch": 0.6041216130263325, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.257061958312988, "learning_rate": 5e-05, "loss": 0.8804, "mean_token_accuracy": 0.8503142595291138, "num_tokens": 180436823.0, "step": 4749 }, { "epoch": 0.604248823304923, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.6460723876953125, "learning_rate": 5e-05, "loss": 0.6991, "mean_token_accuracy": 0.8680707216262817, "num_tokens": 180472643.0, "step": 4750 }, { "epoch": 0.6043760335835135, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.20671558380127, "learning_rate": 5e-05, "loss": 0.8774, "mean_token_accuracy": 0.8496249318122864, "num_tokens": 180511788.0, "step": 4751 }, { "epoch": 0.6045032438621041, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.889532089233398, "learning_rate": 5e-05, "loss": 0.7241, "mean_token_accuracy": 0.8705579042434692, "num_tokens": 180547871.0, "step": 4752 }, { "epoch": 0.6046304541406946, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.819974899291992, "learning_rate": 5e-05, "loss": 0.8533, "mean_token_accuracy": 0.8466994166374207, "num_tokens": 180580521.0, "step": 4753 }, { "epoch": 0.604757664419285, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.3341827392578125, "learning_rate": 5e-05, "loss": 0.8149, "mean_token_accuracy": 0.8507398366928101, "num_tokens": 180617726.0, "step": 4754 }, { "epoch": 0.6048848746978756, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.203855991363525, "learning_rate": 5e-05, "loss": 0.767, "mean_token_accuracy": 0.8566429615020752, "num_tokens": 180657705.0, "step": 4755 }, { "epoch": 0.6050120849764661, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.851191520690918, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.8649930953979492, "num_tokens": 180693582.0, "step": 4756 }, { "epoch": 0.6051392952550566, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.603133678436279, "learning_rate": 5e-05, "loss": 0.7383, "mean_token_accuracy": 0.8514085412025452, "num_tokens": 180731159.0, "step": 4757 }, { "epoch": 0.6052665055336471, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.376435279846191, "learning_rate": 5e-05, "loss": 0.8753, "mean_token_accuracy": 0.84784996509552, "num_tokens": 180769901.0, "step": 4758 }, { "epoch": 0.6053937158122377, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.153039455413818, "learning_rate": 5e-05, "loss": 0.6916, "mean_token_accuracy": 0.8622981309890747, "num_tokens": 180811763.0, "step": 4759 }, { "epoch": 0.6055209260908282, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.95216178894043, "learning_rate": 5e-05, "loss": 0.9057, "mean_token_accuracy": 0.8616077899932861, "num_tokens": 180846073.0, "step": 4760 }, { "epoch": 0.6056481363694186, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.917778968811035, "learning_rate": 5e-05, "loss": 0.6821, "mean_token_accuracy": 0.8681234121322632, "num_tokens": 180883609.0, "step": 4761 }, { "epoch": 0.6057753466480091, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.4326171875, "learning_rate": 5e-05, "loss": 1.0072, "mean_token_accuracy": 0.8558927774429321, "num_tokens": 180921492.0, "step": 4762 }, { "epoch": 0.6059025569265997, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.44887638092041, "learning_rate": 5e-05, "loss": 0.8302, "mean_token_accuracy": 0.8451970219612122, "num_tokens": 180960298.0, "step": 4763 }, { "epoch": 0.6060297672051902, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.075286865234375, "learning_rate": 5e-05, "loss": 1.0306, "mean_token_accuracy": 0.8509979248046875, "num_tokens": 180997002.0, "step": 4764 }, { "epoch": 0.6061569774837807, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.691103935241699, "learning_rate": 5e-05, "loss": 1.0169, "mean_token_accuracy": 0.8384508490562439, "num_tokens": 181033575.0, "step": 4765 }, { "epoch": 0.6062841877623713, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.54528284072876, "learning_rate": 5e-05, "loss": 0.852, "mean_token_accuracy": 0.8595147132873535, "num_tokens": 181068584.0, "step": 4766 }, { "epoch": 0.6064113980409617, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.751713752746582, "learning_rate": 5e-05, "loss": 0.9756, "mean_token_accuracy": 0.8638426065444946, "num_tokens": 181103017.0, "step": 4767 }, { "epoch": 0.6065386083195522, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.338290214538574, "learning_rate": 5e-05, "loss": 0.7381, "mean_token_accuracy": 0.8602004051208496, "num_tokens": 181137590.0, "step": 4768 }, { "epoch": 0.6066658185981427, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.33779525756836, "learning_rate": 5e-05, "loss": 1.0094, "mean_token_accuracy": 0.8584287166595459, "num_tokens": 181177581.0, "step": 4769 }, { "epoch": 0.6067930288767333, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.623913288116455, "learning_rate": 5e-05, "loss": 0.7937, "mean_token_accuracy": 0.8685564994812012, "num_tokens": 181212582.0, "step": 4770 }, { "epoch": 0.6069202391553238, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.760380744934082, "learning_rate": 5e-05, "loss": 1.0014, "mean_token_accuracy": 0.8503799438476562, "num_tokens": 181246373.0, "step": 4771 }, { "epoch": 0.6070474494339143, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003147125244140625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.661901950836182, "learning_rate": 5e-05, "loss": 0.886, "mean_token_accuracy": 0.8725805282592773, "num_tokens": 181280657.0, "step": 4772 }, { "epoch": 0.6071746597125047, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.437966823577881, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8683645129203796, "num_tokens": 181322165.0, "step": 4773 }, { "epoch": 0.6073018699910953, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.382219314575195, "learning_rate": 5e-05, "loss": 0.9346, "mean_token_accuracy": 0.8578678965568542, "num_tokens": 181358612.0, "step": 4774 }, { "epoch": 0.6074290802696858, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.796590328216553, "learning_rate": 5e-05, "loss": 0.7313, "mean_token_accuracy": 0.8720365762710571, "num_tokens": 181392928.0, "step": 4775 }, { "epoch": 0.6075562905482763, "ewc_loss": 0.5, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.461316108703613, "learning_rate": 5e-05, "loss": 0.9821, "mean_token_accuracy": 0.84173983335495, "num_tokens": 181431127.0, "step": 4776 }, { "epoch": 0.6076835008268668, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.011904239654541, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.860255241394043, "num_tokens": 181469830.0, "step": 4777 }, { "epoch": 0.6078107111054574, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.9433159828186035, "learning_rate": 5e-05, "loss": 0.8591, "mean_token_accuracy": 0.8581430912017822, "num_tokens": 181505590.0, "step": 4778 }, { "epoch": 0.6079379213840478, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.4105730056762695, "learning_rate": 5e-05, "loss": 0.8001, "mean_token_accuracy": 0.8744838833808899, "num_tokens": 181541675.0, "step": 4779 }, { "epoch": 0.6080651316626383, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.479634761810303, "learning_rate": 5e-05, "loss": 0.8381, "mean_token_accuracy": 0.8462634086608887, "num_tokens": 181579890.0, "step": 4780 }, { "epoch": 0.6081923419412288, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.500826835632324, "learning_rate": 5e-05, "loss": 0.8282, "mean_token_accuracy": 0.8571070432662964, "num_tokens": 181613821.0, "step": 4781 }, { "epoch": 0.6083195522198194, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.398380756378174, "learning_rate": 5e-05, "loss": 0.8245, "mean_token_accuracy": 0.8475167751312256, "num_tokens": 181656899.0, "step": 4782 }, { "epoch": 0.6084467624984099, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.357576847076416, "learning_rate": 5e-05, "loss": 0.8242, "mean_token_accuracy": 0.8426529169082642, "num_tokens": 181690044.0, "step": 4783 }, { "epoch": 0.6085739727770004, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.470208168029785, "learning_rate": 5e-05, "loss": 0.765, "mean_token_accuracy": 0.8629763126373291, "num_tokens": 181725038.0, "step": 4784 }, { "epoch": 0.6087011830555908, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.184927463531494, "learning_rate": 5e-05, "loss": 0.7584, "mean_token_accuracy": 0.8590682744979858, "num_tokens": 181767914.0, "step": 4785 }, { "epoch": 0.6088283933341814, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.576681137084961, "learning_rate": 5e-05, "loss": 0.7451, "mean_token_accuracy": 0.870063304901123, "num_tokens": 181798955.0, "step": 4786 }, { "epoch": 0.6089556036127719, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.040951728820801, "learning_rate": 5e-05, "loss": 0.7906, "mean_token_accuracy": 0.8454208374023438, "num_tokens": 181831967.0, "step": 4787 }, { "epoch": 0.6090828138913624, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.694444179534912, "learning_rate": 5e-05, "loss": 0.789, "mean_token_accuracy": 0.8568768501281738, "num_tokens": 181865717.0, "step": 4788 }, { "epoch": 0.609210024169953, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.8659210205078125, "learning_rate": 5e-05, "loss": 0.7909, "mean_token_accuracy": 0.8429605960845947, "num_tokens": 181902847.0, "step": 4789 }, { "epoch": 0.6093372344485435, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.8999552726745605, "learning_rate": 5e-05, "loss": 0.7579, "mean_token_accuracy": 0.869974672794342, "num_tokens": 181937129.0, "step": 4790 }, { "epoch": 0.6094644447271339, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.617472171783447, "learning_rate": 5e-05, "loss": 0.7158, "mean_token_accuracy": 0.858529269695282, "num_tokens": 181979576.0, "step": 4791 }, { "epoch": 0.6095916550057244, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.250727653503418, "learning_rate": 5e-05, "loss": 0.8253, "mean_token_accuracy": 0.858709990978241, "num_tokens": 182021510.0, "step": 4792 }, { "epoch": 0.609718865284315, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.233248710632324, "learning_rate": 5e-05, "loss": 0.7086, "mean_token_accuracy": 0.8601611256599426, "num_tokens": 182061178.0, "step": 4793 }, { "epoch": 0.6098460755629055, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.871919631958008, "learning_rate": 5e-05, "loss": 0.9542, "mean_token_accuracy": 0.8414534330368042, "num_tokens": 182100575.0, "step": 4794 }, { "epoch": 0.609973285841496, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.795929908752441, "learning_rate": 5e-05, "loss": 0.7008, "mean_token_accuracy": 0.8574408292770386, "num_tokens": 182136142.0, "step": 4795 }, { "epoch": 0.6101004961200865, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.682507514953613, "learning_rate": 5e-05, "loss": 1.0441, "mean_token_accuracy": 0.8474797010421753, "num_tokens": 182173108.0, "step": 4796 }, { "epoch": 0.610227706398677, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 5.986840724945068, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.847248911857605, "num_tokens": 182209931.0, "step": 4797 }, { "epoch": 0.6103549166772675, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.057765007019043, "learning_rate": 5e-05, "loss": 1.0901, "mean_token_accuracy": 0.861478328704834, "num_tokens": 182244128.0, "step": 4798 }, { "epoch": 0.610482126955858, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 7.455881118774414, "learning_rate": 5e-05, "loss": 0.9913, "mean_token_accuracy": 0.8588030338287354, "num_tokens": 182278112.0, "step": 4799 }, { "epoch": 0.6106093372344485, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.988831520080566, "learning_rate": 5e-05, "loss": 0.9926, "mean_token_accuracy": 0.8633993864059448, "num_tokens": 182313753.0, "step": 4800 }, { "epoch": 0.6107365475130391, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 8.625079154968262, "learning_rate": 5e-05, "loss": 1.0017, "mean_token_accuracy": 0.878600537776947, "num_tokens": 182348758.0, "step": 4801 }, { "epoch": 0.6108637577916296, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.108417510986328, "learning_rate": 5e-05, "loss": 0.8261, "mean_token_accuracy": 0.8571261763572693, "num_tokens": 182381084.0, "step": 4802 }, { "epoch": 0.61099096807022, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.63242244720459, "learning_rate": 5e-05, "loss": 1.0618, "mean_token_accuracy": 0.8598340153694153, "num_tokens": 182420530.0, "step": 4803 }, { "epoch": 0.6111181783488105, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 6.536776065826416, "learning_rate": 5e-05, "loss": 0.8312, "mean_token_accuracy": 0.8566259145736694, "num_tokens": 182462286.0, "step": 4804 }, { "epoch": 0.6112453886274011, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.087845802307129, "learning_rate": 5e-05, "loss": 1.0404, "mean_token_accuracy": 0.8724648356437683, "num_tokens": 182497209.0, "step": 4805 }, { "epoch": 0.6113725989059916, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 7.675617218017578, "learning_rate": 5e-05, "loss": 0.9777, "mean_token_accuracy": 0.8712960481643677, "num_tokens": 182530609.0, "step": 4806 }, { "epoch": 0.6114998091845821, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.141498565673828, "learning_rate": 5e-05, "loss": 0.9767, "mean_token_accuracy": 0.8662359118461609, "num_tokens": 182566698.0, "step": 4807 }, { "epoch": 0.6116270194631727, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 8.463953971862793, "learning_rate": 5e-05, "loss": 1.0578, "mean_token_accuracy": 0.8646489381790161, "num_tokens": 182609524.0, "step": 4808 }, { "epoch": 0.6117542297417632, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.775339126586914, "learning_rate": 5e-05, "loss": 0.86, "mean_token_accuracy": 0.862430214881897, "num_tokens": 182649339.0, "step": 4809 }, { "epoch": 0.6118814400203536, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.747658729553223, "learning_rate": 5e-05, "loss": 1.0216, "mean_token_accuracy": 0.8592093586921692, "num_tokens": 182684142.0, "step": 4810 }, { "epoch": 0.6120086502989441, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.280059814453125, "learning_rate": 5e-05, "loss": 0.8564, "mean_token_accuracy": 0.8571460247039795, "num_tokens": 182722857.0, "step": 4811 }, { "epoch": 0.6121358605775347, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.584333419799805, "learning_rate": 5e-05, "loss": 0.976, "mean_token_accuracy": 0.8522025942802429, "num_tokens": 182760196.0, "step": 4812 }, { "epoch": 0.6122630708561252, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.4052300453186035, "learning_rate": 5e-05, "loss": 0.8994, "mean_token_accuracy": 0.8539907932281494, "num_tokens": 182795807.0, "step": 4813 }, { "epoch": 0.6123902811347157, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.024481773376465, "learning_rate": 5e-05, "loss": 0.9062, "mean_token_accuracy": 0.8509371876716614, "num_tokens": 182837956.0, "step": 4814 }, { "epoch": 0.6125174914133062, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.611616134643555, "learning_rate": 5e-05, "loss": 0.8822, "mean_token_accuracy": 0.864324152469635, "num_tokens": 182873374.0, "step": 4815 }, { "epoch": 0.6126447016918967, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.502591133117676, "learning_rate": 5e-05, "loss": 0.7762, "mean_token_accuracy": 0.8719177842140198, "num_tokens": 182908018.0, "step": 4816 }, { "epoch": 0.6127719119704872, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.747856616973877, "learning_rate": 5e-05, "loss": 0.8803, "mean_token_accuracy": 0.8569881916046143, "num_tokens": 182950314.0, "step": 4817 }, { "epoch": 0.6128991222490777, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.128276824951172, "learning_rate": 5e-05, "loss": 0.7834, "mean_token_accuracy": 0.8567724823951721, "num_tokens": 182988434.0, "step": 4818 }, { "epoch": 0.6130263325276682, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.02885627746582, "learning_rate": 5e-05, "loss": 0.8317, "mean_token_accuracy": 0.864587664604187, "num_tokens": 183026355.0, "step": 4819 }, { "epoch": 0.6131535428062588, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.750888824462891, "learning_rate": 5e-05, "loss": 0.7516, "mean_token_accuracy": 0.8592735528945923, "num_tokens": 183063031.0, "step": 4820 }, { "epoch": 0.6132807530848493, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.44547176361084, "learning_rate": 5e-05, "loss": 0.9423, "mean_token_accuracy": 0.8465575575828552, "num_tokens": 183102082.0, "step": 4821 }, { "epoch": 0.6134079633634397, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.501034736633301, "learning_rate": 5e-05, "loss": 0.7416, "mean_token_accuracy": 0.8574236631393433, "num_tokens": 183137961.0, "step": 4822 }, { "epoch": 0.6135351736420303, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.716655731201172, "learning_rate": 5e-05, "loss": 0.8723, "mean_token_accuracy": 0.8721221089363098, "num_tokens": 183173779.0, "step": 4823 }, { "epoch": 0.6136623839206208, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.530728816986084, "learning_rate": 5e-05, "loss": 0.7968, "mean_token_accuracy": 0.8510032892227173, "num_tokens": 183215044.0, "step": 4824 }, { "epoch": 0.6137895941992113, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.781023979187012, "learning_rate": 5e-05, "loss": 0.9112, "mean_token_accuracy": 0.8646222949028015, "num_tokens": 183248793.0, "step": 4825 }, { "epoch": 0.6139168044778018, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.79129695892334, "learning_rate": 5e-05, "loss": 0.8299, "mean_token_accuracy": 0.8600476384162903, "num_tokens": 183291180.0, "step": 4826 }, { "epoch": 0.6140440147563924, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.564748764038086, "learning_rate": 5e-05, "loss": 0.9189, "mean_token_accuracy": 0.860227644443512, "num_tokens": 183326214.0, "step": 4827 }, { "epoch": 0.6141712250349828, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.111402988433838, "learning_rate": 5e-05, "loss": 0.8422, "mean_token_accuracy": 0.8629077672958374, "num_tokens": 183362403.0, "step": 4828 }, { "epoch": 0.6142984353135733, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.172599792480469, "learning_rate": 5e-05, "loss": 0.871, "mean_token_accuracy": 0.8659325838088989, "num_tokens": 183407419.0, "step": 4829 }, { "epoch": 0.6144256455921638, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.292776107788086, "learning_rate": 5e-05, "loss": 0.8568, "mean_token_accuracy": 0.8586201071739197, "num_tokens": 183443140.0, "step": 4830 }, { "epoch": 0.6145528558707544, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.894942760467529, "learning_rate": 5e-05, "loss": 0.9087, "mean_token_accuracy": 0.8435279130935669, "num_tokens": 183485986.0, "step": 4831 }, { "epoch": 0.6146800661493449, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.212152004241943, "learning_rate": 5e-05, "loss": 0.8815, "mean_token_accuracy": 0.8437258005142212, "num_tokens": 183525302.0, "step": 4832 }, { "epoch": 0.6148072764279354, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.92722749710083, "learning_rate": 5e-05, "loss": 0.8597, "mean_token_accuracy": 0.8542961478233337, "num_tokens": 183561575.0, "step": 4833 }, { "epoch": 0.6149344867065258, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.9731645584106445, "learning_rate": 5e-05, "loss": 0.7524, "mean_token_accuracy": 0.8674494624137878, "num_tokens": 183596465.0, "step": 4834 }, { "epoch": 0.6150616969851164, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.106603622436523, "learning_rate": 5e-05, "loss": 0.9125, "mean_token_accuracy": 0.8364286422729492, "num_tokens": 183631733.0, "step": 4835 }, { "epoch": 0.6151889072637069, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.795437812805176, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8546903133392334, "num_tokens": 183676138.0, "step": 4836 }, { "epoch": 0.6153161175422974, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.196331024169922, "learning_rate": 5e-05, "loss": 0.8574, "mean_token_accuracy": 0.8581234216690063, "num_tokens": 183711004.0, "step": 4837 }, { "epoch": 0.615443327820888, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.830452919006348, "learning_rate": 5e-05, "loss": 0.6916, "mean_token_accuracy": 0.880584180355072, "num_tokens": 183744945.0, "step": 4838 }, { "epoch": 0.6155705380994785, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.0626802444458, "learning_rate": 5e-05, "loss": 0.832, "mean_token_accuracy": 0.8678909540176392, "num_tokens": 183779119.0, "step": 4839 }, { "epoch": 0.6156977483780689, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.034887790679932, "learning_rate": 5e-05, "loss": 0.7356, "mean_token_accuracy": 0.8677133321762085, "num_tokens": 183814136.0, "step": 4840 }, { "epoch": 0.6158249586566594, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.747107982635498, "learning_rate": 5e-05, "loss": 0.8301, "mean_token_accuracy": 0.8546411991119385, "num_tokens": 183850957.0, "step": 4841 }, { "epoch": 0.61595216893525, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.292956829071045, "learning_rate": 5e-05, "loss": 0.7516, "mean_token_accuracy": 0.8698574304580688, "num_tokens": 183890049.0, "step": 4842 }, { "epoch": 0.6160793792138405, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.423657417297363, "learning_rate": 5e-05, "loss": 0.7715, "mean_token_accuracy": 0.8598037958145142, "num_tokens": 183933055.0, "step": 4843 }, { "epoch": 0.616206589492431, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.455361843109131, "learning_rate": 5e-05, "loss": 0.7827, "mean_token_accuracy": 0.8596498966217041, "num_tokens": 183975059.0, "step": 4844 }, { "epoch": 0.6163337997710215, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.234421730041504, "learning_rate": 5e-05, "loss": 0.788, "mean_token_accuracy": 0.8518651723861694, "num_tokens": 184015130.0, "step": 4845 }, { "epoch": 0.616461010049612, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.526589393615723, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.848142683506012, "num_tokens": 184048416.0, "step": 4846 }, { "epoch": 0.6165882203282025, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.160205364227295, "learning_rate": 5e-05, "loss": 0.7802, "mean_token_accuracy": 0.8514209985733032, "num_tokens": 184081591.0, "step": 4847 }, { "epoch": 0.616715430606793, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.503539562225342, "learning_rate": 5e-05, "loss": 0.6978, "mean_token_accuracy": 0.8802722096443176, "num_tokens": 184118078.0, "step": 4848 }, { "epoch": 0.6168426408853835, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.190954685211182, "learning_rate": 5e-05, "loss": 0.7265, "mean_token_accuracy": 0.8629655838012695, "num_tokens": 184154860.0, "step": 4849 }, { "epoch": 0.6169698511639741, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.404580593109131, "learning_rate": 5e-05, "loss": 0.721, "mean_token_accuracy": 0.8713619709014893, "num_tokens": 184186425.0, "step": 4850 }, { "epoch": 0.6170970614425646, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.280263900756836, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8529447317123413, "num_tokens": 184224891.0, "step": 4851 }, { "epoch": 0.617224271721155, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.283077716827393, "learning_rate": 5e-05, "loss": 0.6993, "mean_token_accuracy": 0.8711921572685242, "num_tokens": 184266648.0, "step": 4852 }, { "epoch": 0.6173514819997455, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.377437591552734, "learning_rate": 5e-05, "loss": 0.8057, "mean_token_accuracy": 0.8465842008590698, "num_tokens": 184301453.0, "step": 4853 }, { "epoch": 0.6174786922783361, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.1450276374816895, "learning_rate": 5e-05, "loss": 0.748, "mean_token_accuracy": 0.8584058880805969, "num_tokens": 184344271.0, "step": 4854 }, { "epoch": 0.6176059025569266, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.5160722732543945, "learning_rate": 5e-05, "loss": 0.7302, "mean_token_accuracy": 0.863526463508606, "num_tokens": 184378212.0, "step": 4855 }, { "epoch": 0.6177331128355171, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.9790239334106445, "learning_rate": 5e-05, "loss": 0.6942, "mean_token_accuracy": 0.8652300834655762, "num_tokens": 184414177.0, "step": 4856 }, { "epoch": 0.6178603231141077, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.706366062164307, "learning_rate": 5e-05, "loss": 0.8004, "mean_token_accuracy": 0.8478667736053467, "num_tokens": 184455088.0, "step": 4857 }, { "epoch": 0.6179875333926982, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.756825923919678, "learning_rate": 5e-05, "loss": 0.7509, "mean_token_accuracy": 0.8482093811035156, "num_tokens": 184496574.0, "step": 4858 }, { "epoch": 0.6181147436712886, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 8.006704330444336, "learning_rate": 5e-05, "loss": 0.8138, "mean_token_accuracy": 0.8533707857131958, "num_tokens": 184542303.0, "step": 4859 }, { "epoch": 0.6182419539498791, "ewc_loss": 0.265625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 6.385822296142578, "learning_rate": 5e-05, "loss": 0.6956, "mean_token_accuracy": 0.8592872619628906, "num_tokens": 184583337.0, "step": 4860 }, { "epoch": 0.6183691642284697, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.564215660095215, "learning_rate": 5e-05, "loss": 0.8585, "mean_token_accuracy": 0.8561937212944031, "num_tokens": 184622424.0, "step": 4861 }, { "epoch": 0.6184963745070602, "ewc_loss": 0.24609375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00021457672119140625, "grad_norm": 5.828160762786865, "learning_rate": 5e-05, "loss": 0.7036, "mean_token_accuracy": 0.8435248136520386, "num_tokens": 184660137.0, "step": 4862 }, { "epoch": 0.6186235847856507, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.472979545593262, "learning_rate": 5e-05, "loss": 0.9557, "mean_token_accuracy": 0.856183648109436, "num_tokens": 184698353.0, "step": 4863 }, { "epoch": 0.6187507950642412, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00023555755615234375, "grad_norm": 5.4926934242248535, "learning_rate": 5e-05, "loss": 0.7161, "mean_token_accuracy": 0.8503255844116211, "num_tokens": 184738734.0, "step": 4864 }, { "epoch": 0.6188780053428317, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.401398658752441, "learning_rate": 5e-05, "loss": 1.1092, "mean_token_accuracy": 0.8494561314582825, "num_tokens": 184774525.0, "step": 4865 }, { "epoch": 0.6190052156214222, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 6.717979431152344, "learning_rate": 5e-05, "loss": 0.8664, "mean_token_accuracy": 0.8620281219482422, "num_tokens": 184809028.0, "step": 4866 }, { "epoch": 0.6191324259000127, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.101729393005371, "learning_rate": 5e-05, "loss": 1.0287, "mean_token_accuracy": 0.8785601854324341, "num_tokens": 184845479.0, "step": 4867 }, { "epoch": 0.6192596361786032, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 8.650525093078613, "learning_rate": 5e-05, "loss": 1.1991, "mean_token_accuracy": 0.8425025939941406, "num_tokens": 184886596.0, "step": 4868 }, { "epoch": 0.6193868464571938, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.847841739654541, "learning_rate": 5e-05, "loss": 0.8594, "mean_token_accuracy": 0.8556971549987793, "num_tokens": 184923763.0, "step": 4869 }, { "epoch": 0.6195140567357843, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000743865966796875, "grad_norm": 10.21542739868164, "learning_rate": 5e-05, "loss": 1.1797, "mean_token_accuracy": 0.8647469878196716, "num_tokens": 184959882.0, "step": 4870 }, { "epoch": 0.6196412670143747, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 6.589427947998047, "learning_rate": 5e-05, "loss": 0.7535, "mean_token_accuracy": 0.8745557069778442, "num_tokens": 184999063.0, "step": 4871 }, { "epoch": 0.6197684772929652, "ewc_loss": 0.95703125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000926971435546875, "grad_norm": 12.314062118530273, "learning_rate": 5e-05, "loss": 1.3251, "mean_token_accuracy": 0.8746472597122192, "num_tokens": 185040016.0, "step": 4872 }, { "epoch": 0.6198956875715558, "ewc_loss": 0.890625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 9.512480735778809, "learning_rate": 5e-05, "loss": 1.308, "mean_token_accuracy": 0.858135998249054, "num_tokens": 185075085.0, "step": 4873 }, { "epoch": 0.6200228978501463, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 11.452534675598145, "learning_rate": 5e-05, "loss": 1.1385, "mean_token_accuracy": 0.8580684661865234, "num_tokens": 185111621.0, "step": 4874 }, { "epoch": 0.6201501081287368, "ewc_loss": 0.8125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.424333572387695, "learning_rate": 5e-05, "loss": 1.2575, "mean_token_accuracy": 0.8520479202270508, "num_tokens": 185149681.0, "step": 4875 }, { "epoch": 0.6202773184073274, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 11.400222778320312, "learning_rate": 5e-05, "loss": 1.1578, "mean_token_accuracy": 0.8654706478118896, "num_tokens": 185186449.0, "step": 4876 }, { "epoch": 0.6204045286859178, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.507978439331055, "learning_rate": 5e-05, "loss": 0.995, "mean_token_accuracy": 0.8645012378692627, "num_tokens": 185219324.0, "step": 4877 }, { "epoch": 0.6205317389645083, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 11.747011184692383, "learning_rate": 5e-05, "loss": 1.2562, "mean_token_accuracy": 0.8566851615905762, "num_tokens": 185256049.0, "step": 4878 }, { "epoch": 0.6206589492430988, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.983440399169922, "learning_rate": 5e-05, "loss": 1.0218, "mean_token_accuracy": 0.8641409277915955, "num_tokens": 185293304.0, "step": 4879 }, { "epoch": 0.6207861595216894, "ewc_loss": 0.8125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 11.823589324951172, "learning_rate": 5e-05, "loss": 1.2173, "mean_token_accuracy": 0.862368106842041, "num_tokens": 185334007.0, "step": 4880 }, { "epoch": 0.6209133698002799, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 9.945856094360352, "learning_rate": 5e-05, "loss": 1.1808, "mean_token_accuracy": 0.8524742126464844, "num_tokens": 185372932.0, "step": 4881 }, { "epoch": 0.6210405800788704, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 11.035419464111328, "learning_rate": 5e-05, "loss": 1.0862, "mean_token_accuracy": 0.8658593893051147, "num_tokens": 185410267.0, "step": 4882 }, { "epoch": 0.6211677903574608, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 9.669973373413086, "learning_rate": 5e-05, "loss": 1.1105, "mean_token_accuracy": 0.8513056039810181, "num_tokens": 185445838.0, "step": 4883 }, { "epoch": 0.6212950006360514, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.482568740844727, "learning_rate": 5e-05, "loss": 1.0629, "mean_token_accuracy": 0.8618651628494263, "num_tokens": 185486002.0, "step": 4884 }, { "epoch": 0.6214222109146419, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.11998176574707, "learning_rate": 5e-05, "loss": 0.9983, "mean_token_accuracy": 0.8552688956260681, "num_tokens": 185524501.0, "step": 4885 }, { "epoch": 0.6215494211932324, "ewc_loss": 0.625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 10.108834266662598, "learning_rate": 5e-05, "loss": 1.0377, "mean_token_accuracy": 0.8563814163208008, "num_tokens": 185562105.0, "step": 4886 }, { "epoch": 0.621676631471823, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.610978126525879, "learning_rate": 5e-05, "loss": 0.9076, "mean_token_accuracy": 0.8670039176940918, "num_tokens": 185602821.0, "step": 4887 }, { "epoch": 0.6218038417504135, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 10.01583194732666, "learning_rate": 5e-05, "loss": 1.0572, "mean_token_accuracy": 0.8512473106384277, "num_tokens": 185643188.0, "step": 4888 }, { "epoch": 0.6219310520290039, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.252270698547363, "learning_rate": 5e-05, "loss": 0.9472, "mean_token_accuracy": 0.8448734879493713, "num_tokens": 185685770.0, "step": 4889 }, { "epoch": 0.6220582623075944, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.189414978027344, "learning_rate": 5e-05, "loss": 1.0368, "mean_token_accuracy": 0.8619552254676819, "num_tokens": 185718981.0, "step": 4890 }, { "epoch": 0.622185472586185, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.245079040527344, "learning_rate": 5e-05, "loss": 0.8819, "mean_token_accuracy": 0.8617769479751587, "num_tokens": 185752184.0, "step": 4891 }, { "epoch": 0.6223126828647755, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.366548538208008, "learning_rate": 5e-05, "loss": 1.0802, "mean_token_accuracy": 0.8572621941566467, "num_tokens": 185791405.0, "step": 4892 }, { "epoch": 0.622439893143366, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.634119033813477, "learning_rate": 5e-05, "loss": 0.9944, "mean_token_accuracy": 0.8399912118911743, "num_tokens": 185832227.0, "step": 4893 }, { "epoch": 0.6225671034219565, "ewc_loss": 0.625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 10.040812492370605, "learning_rate": 5e-05, "loss": 1.0339, "mean_token_accuracy": 0.8649473786354065, "num_tokens": 185867175.0, "step": 4894 }, { "epoch": 0.622694313700547, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.821037292480469, "learning_rate": 5e-05, "loss": 0.9535, "mean_token_accuracy": 0.8561677932739258, "num_tokens": 185907582.0, "step": 4895 }, { "epoch": 0.6228215239791375, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.363367080688477, "learning_rate": 5e-05, "loss": 1.0047, "mean_token_accuracy": 0.8487785458564758, "num_tokens": 185944625.0, "step": 4896 }, { "epoch": 0.622948734257728, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.500533103942871, "learning_rate": 5e-05, "loss": 0.8631, "mean_token_accuracy": 0.8692474961280823, "num_tokens": 185975579.0, "step": 4897 }, { "epoch": 0.6230759445363185, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.90368938446045, "learning_rate": 5e-05, "loss": 0.9181, "mean_token_accuracy": 0.8637675046920776, "num_tokens": 186016948.0, "step": 4898 }, { "epoch": 0.6232031548149091, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.9354472160339355, "learning_rate": 5e-05, "loss": 0.8474, "mean_token_accuracy": 0.8527476191520691, "num_tokens": 186060700.0, "step": 4899 }, { "epoch": 0.6233303650934996, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.89376449584961, "learning_rate": 5e-05, "loss": 0.8778, "mean_token_accuracy": 0.8738645315170288, "num_tokens": 186099624.0, "step": 4900 }, { "epoch": 0.62345757537209, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.460875988006592, "learning_rate": 5e-05, "loss": 0.7819, "mean_token_accuracy": 0.8579938411712646, "num_tokens": 186142360.0, "step": 4901 }, { "epoch": 0.6235847856506805, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.179659843444824, "learning_rate": 5e-05, "loss": 0.9224, "mean_token_accuracy": 0.8716417551040649, "num_tokens": 186182642.0, "step": 4902 }, { "epoch": 0.6237119959292711, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.392023086547852, "learning_rate": 5e-05, "loss": 0.7922, "mean_token_accuracy": 0.8549908995628357, "num_tokens": 186218867.0, "step": 4903 }, { "epoch": 0.6238392062078616, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.400005340576172, "learning_rate": 5e-05, "loss": 0.9403, "mean_token_accuracy": 0.8712649345397949, "num_tokens": 186257424.0, "step": 4904 }, { "epoch": 0.6239664164864521, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.793737411499023, "learning_rate": 5e-05, "loss": 0.819, "mean_token_accuracy": 0.8610637187957764, "num_tokens": 186298588.0, "step": 4905 }, { "epoch": 0.6240936267650427, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.046195983886719, "learning_rate": 5e-05, "loss": 0.9279, "mean_token_accuracy": 0.8679584264755249, "num_tokens": 186334638.0, "step": 4906 }, { "epoch": 0.6242208370436332, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.127896308898926, "learning_rate": 5e-05, "loss": 0.869, "mean_token_accuracy": 0.8522534370422363, "num_tokens": 186368303.0, "step": 4907 }, { "epoch": 0.6243480473222236, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.278594970703125, "learning_rate": 5e-05, "loss": 0.8346, "mean_token_accuracy": 0.8750416040420532, "num_tokens": 186406133.0, "step": 4908 }, { "epoch": 0.6244752576008141, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.939292907714844, "learning_rate": 5e-05, "loss": 0.7544, "mean_token_accuracy": 0.8781533241271973, "num_tokens": 186440763.0, "step": 4909 }, { "epoch": 0.6246024678794047, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.860611438751221, "learning_rate": 5e-05, "loss": 0.8563, "mean_token_accuracy": 0.8583351373672485, "num_tokens": 186479662.0, "step": 4910 }, { "epoch": 0.6247296781579952, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.576265811920166, "learning_rate": 5e-05, "loss": 0.8252, "mean_token_accuracy": 0.8483823537826538, "num_tokens": 186518944.0, "step": 4911 }, { "epoch": 0.6248568884365857, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.785132884979248, "learning_rate": 5e-05, "loss": 0.7772, "mean_token_accuracy": 0.8691152930259705, "num_tokens": 186555902.0, "step": 4912 }, { "epoch": 0.6249840987151762, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.3271942138671875, "learning_rate": 5e-05, "loss": 0.7011, "mean_token_accuracy": 0.8772268295288086, "num_tokens": 186591962.0, "step": 4913 }, { "epoch": 0.6251113089937667, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.771909713745117, "learning_rate": 5e-05, "loss": 0.8258, "mean_token_accuracy": 0.852940559387207, "num_tokens": 186631841.0, "step": 4914 }, { "epoch": 0.6252385192723572, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.225637912750244, "learning_rate": 5e-05, "loss": 0.7034, "mean_token_accuracy": 0.870705246925354, "num_tokens": 186677963.0, "step": 4915 }, { "epoch": 0.6253657295509477, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.724728107452393, "learning_rate": 5e-05, "loss": 0.7896, "mean_token_accuracy": 0.8578479290008545, "num_tokens": 186715658.0, "step": 4916 }, { "epoch": 0.6254929398295382, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.138880252838135, "learning_rate": 5e-05, "loss": 0.7053, "mean_token_accuracy": 0.8678444623947144, "num_tokens": 186749405.0, "step": 4917 }, { "epoch": 0.6256201501081288, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.782094955444336, "learning_rate": 5e-05, "loss": 0.8067, "mean_token_accuracy": 0.851934552192688, "num_tokens": 186785667.0, "step": 4918 }, { "epoch": 0.6257473603867193, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.009815216064453, "learning_rate": 5e-05, "loss": 0.686, "mean_token_accuracy": 0.8734481334686279, "num_tokens": 186817425.0, "step": 4919 }, { "epoch": 0.6258745706653097, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.909380912780762, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.867682933807373, "num_tokens": 186859693.0, "step": 4920 }, { "epoch": 0.6260017809439002, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.883760929107666, "learning_rate": 5e-05, "loss": 0.7032, "mean_token_accuracy": 0.868651270866394, "num_tokens": 186897774.0, "step": 4921 }, { "epoch": 0.6261289912224908, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.995078086853027, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.8618389368057251, "num_tokens": 186935530.0, "step": 4922 }, { "epoch": 0.6262562015010813, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.848010063171387, "learning_rate": 5e-05, "loss": 0.7579, "mean_token_accuracy": 0.8488521575927734, "num_tokens": 186970778.0, "step": 4923 }, { "epoch": 0.6263834117796718, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.05321216583252, "learning_rate": 5e-05, "loss": 0.8001, "mean_token_accuracy": 0.8600752353668213, "num_tokens": 187007338.0, "step": 4924 }, { "epoch": 0.6265106220582624, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.851505756378174, "learning_rate": 5e-05, "loss": 0.723, "mean_token_accuracy": 0.8603502511978149, "num_tokens": 187042614.0, "step": 4925 }, { "epoch": 0.6266378323368528, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.004643440246582, "learning_rate": 5e-05, "loss": 0.8202, "mean_token_accuracy": 0.8556376695632935, "num_tokens": 187081753.0, "step": 4926 }, { "epoch": 0.6267650426154433, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.009184837341309, "learning_rate": 5e-05, "loss": 0.7506, "mean_token_accuracy": 0.8535200357437134, "num_tokens": 187116758.0, "step": 4927 }, { "epoch": 0.6268922528940338, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.686066150665283, "learning_rate": 5e-05, "loss": 0.8041, "mean_token_accuracy": 0.8512694239616394, "num_tokens": 187152296.0, "step": 4928 }, { "epoch": 0.6270194631726244, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.390955448150635, "learning_rate": 5e-05, "loss": 0.7571, "mean_token_accuracy": 0.8536950349807739, "num_tokens": 187187619.0, "step": 4929 }, { "epoch": 0.6271466734512149, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.147698402404785, "learning_rate": 5e-05, "loss": 0.7274, "mean_token_accuracy": 0.8598731756210327, "num_tokens": 187227872.0, "step": 4930 }, { "epoch": 0.6272738837298054, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.9007649421691895, "learning_rate": 5e-05, "loss": 0.7863, "mean_token_accuracy": 0.85813307762146, "num_tokens": 187265720.0, "step": 4931 }, { "epoch": 0.6274010940083958, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.594191551208496, "learning_rate": 5e-05, "loss": 0.7832, "mean_token_accuracy": 0.8374288082122803, "num_tokens": 187307119.0, "step": 4932 }, { "epoch": 0.6275283042869864, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.48723316192627, "learning_rate": 5e-05, "loss": 0.8202, "mean_token_accuracy": 0.8653272390365601, "num_tokens": 187342804.0, "step": 4933 }, { "epoch": 0.6276555145655769, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 6.044987201690674, "learning_rate": 5e-05, "loss": 0.6677, "mean_token_accuracy": 0.8612308502197266, "num_tokens": 187381540.0, "step": 4934 }, { "epoch": 0.6277827248441674, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 9.263519287109375, "learning_rate": 5e-05, "loss": 0.9934, "mean_token_accuracy": 0.8427443504333496, "num_tokens": 187424952.0, "step": 4935 }, { "epoch": 0.627909935122758, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00024127960205078125, "grad_norm": 5.88670015335083, "learning_rate": 5e-05, "loss": 0.7222, "mean_token_accuracy": 0.8492039442062378, "num_tokens": 187463668.0, "step": 4936 }, { "epoch": 0.6280371454013485, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.644551277160645, "learning_rate": 5e-05, "loss": 1.0418, "mean_token_accuracy": 0.8552696704864502, "num_tokens": 187505962.0, "step": 4937 }, { "epoch": 0.6281643556799389, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.956524848937988, "learning_rate": 5e-05, "loss": 0.806, "mean_token_accuracy": 0.8537116646766663, "num_tokens": 187546342.0, "step": 4938 }, { "epoch": 0.6282915659585294, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.652017593383789, "learning_rate": 5e-05, "loss": 0.9361, "mean_token_accuracy": 0.8593792915344238, "num_tokens": 187587378.0, "step": 4939 }, { "epoch": 0.62841877623712, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.315667152404785, "learning_rate": 5e-05, "loss": 0.913, "mean_token_accuracy": 0.8604744672775269, "num_tokens": 187621972.0, "step": 4940 }, { "epoch": 0.6285459865157105, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.879946708679199, "learning_rate": 5e-05, "loss": 0.792, "mean_token_accuracy": 0.8478025794029236, "num_tokens": 187657205.0, "step": 4941 }, { "epoch": 0.628673196794301, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.28310775756836, "learning_rate": 5e-05, "loss": 0.9856, "mean_token_accuracy": 0.8536645174026489, "num_tokens": 187698201.0, "step": 4942 }, { "epoch": 0.6288004070728915, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.461673259735107, "learning_rate": 5e-05, "loss": 0.7079, "mean_token_accuracy": 0.8706474304199219, "num_tokens": 187734883.0, "step": 4943 }, { "epoch": 0.628927617351482, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.17948055267334, "learning_rate": 5e-05, "loss": 0.9207, "mean_token_accuracy": 0.8726745247840881, "num_tokens": 187777459.0, "step": 4944 }, { "epoch": 0.6290548276300725, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.479398727416992, "learning_rate": 5e-05, "loss": 0.8631, "mean_token_accuracy": 0.852796733379364, "num_tokens": 187819661.0, "step": 4945 }, { "epoch": 0.629182037908663, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.786966323852539, "learning_rate": 5e-05, "loss": 0.8081, "mean_token_accuracy": 0.8644613027572632, "num_tokens": 187857685.0, "step": 4946 }, { "epoch": 0.6293092481872535, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.45491886138916, "learning_rate": 5e-05, "loss": 0.9746, "mean_token_accuracy": 0.8370345234870911, "num_tokens": 187896614.0, "step": 4947 }, { "epoch": 0.6294364584658441, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.794917583465576, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8510897159576416, "num_tokens": 187925480.0, "step": 4948 }, { "epoch": 0.6295636687444346, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.782539367675781, "learning_rate": 5e-05, "loss": 0.9062, "mean_token_accuracy": 0.8610237240791321, "num_tokens": 187965045.0, "step": 4949 }, { "epoch": 0.629690879023025, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.942349433898926, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8577960729598999, "num_tokens": 188001633.0, "step": 4950 }, { "epoch": 0.6298180893016155, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.290217399597168, "learning_rate": 5e-05, "loss": 0.8883, "mean_token_accuracy": 0.8569122552871704, "num_tokens": 188039931.0, "step": 4951 }, { "epoch": 0.6299452995802061, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.467208385467529, "learning_rate": 5e-05, "loss": 0.865, "mean_token_accuracy": 0.8421260118484497, "num_tokens": 188085353.0, "step": 4952 }, { "epoch": 0.6300725098587966, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.61435604095459, "learning_rate": 5e-05, "loss": 0.7429, "mean_token_accuracy": 0.8782914280891418, "num_tokens": 188118966.0, "step": 4953 }, { "epoch": 0.6301997201373871, "ewc_loss": 0.375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.6514363288879395, "learning_rate": 5e-05, "loss": 0.7099, "mean_token_accuracy": 0.8813582062721252, "num_tokens": 188153680.0, "step": 4954 }, { "epoch": 0.6303269304159776, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.458529472351074, "learning_rate": 5e-05, "loss": 0.7297, "mean_token_accuracy": 0.8720368146896362, "num_tokens": 188189549.0, "step": 4955 }, { "epoch": 0.6304541406945681, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.4121599197387695, "learning_rate": 5e-05, "loss": 0.7936, "mean_token_accuracy": 0.8518024682998657, "num_tokens": 188230469.0, "step": 4956 }, { "epoch": 0.6305813509731586, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.690130233764648, "learning_rate": 5e-05, "loss": 0.7811, "mean_token_accuracy": 0.8616248369216919, "num_tokens": 188265074.0, "step": 4957 }, { "epoch": 0.6307085612517491, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.021694183349609, "learning_rate": 5e-05, "loss": 0.7622, "mean_token_accuracy": 0.851157546043396, "num_tokens": 188302759.0, "step": 4958 }, { "epoch": 0.6308357715303397, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.010879516601562, "learning_rate": 5e-05, "loss": 0.9025, "mean_token_accuracy": 0.8319751024246216, "num_tokens": 188337627.0, "step": 4959 }, { "epoch": 0.6309629818089302, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.696149826049805, "learning_rate": 5e-05, "loss": 0.7097, "mean_token_accuracy": 0.8590490221977234, "num_tokens": 188365094.0, "step": 4960 }, { "epoch": 0.6310901920875207, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.297686576843262, "learning_rate": 5e-05, "loss": 0.8411, "mean_token_accuracy": 0.8619047999382019, "num_tokens": 188403244.0, "step": 4961 }, { "epoch": 0.6312174023661112, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.499526023864746, "learning_rate": 5e-05, "loss": 0.6846, "mean_token_accuracy": 0.8668270111083984, "num_tokens": 188442785.0, "step": 4962 }, { "epoch": 0.6313446126447017, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.547333717346191, "learning_rate": 5e-05, "loss": 0.8985, "mean_token_accuracy": 0.8522056937217712, "num_tokens": 188479298.0, "step": 4963 }, { "epoch": 0.6314718229232922, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.33466100692749, "learning_rate": 5e-05, "loss": 0.7675, "mean_token_accuracy": 0.8422706127166748, "num_tokens": 188519304.0, "step": 4964 }, { "epoch": 0.6315990332018827, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.887953758239746, "learning_rate": 5e-05, "loss": 0.9535, "mean_token_accuracy": 0.843329906463623, "num_tokens": 188559372.0, "step": 4965 }, { "epoch": 0.6317262434804732, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.223100185394287, "learning_rate": 5e-05, "loss": 0.6744, "mean_token_accuracy": 0.8752824664115906, "num_tokens": 188595337.0, "step": 4966 }, { "epoch": 0.6318534537590638, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.21176528930664, "learning_rate": 5e-05, "loss": 0.9653, "mean_token_accuracy": 0.8618547320365906, "num_tokens": 188632988.0, "step": 4967 }, { "epoch": 0.6319806640376543, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.475218296051025, "learning_rate": 5e-05, "loss": 0.719, "mean_token_accuracy": 0.8715521097183228, "num_tokens": 188667205.0, "step": 4968 }, { "epoch": 0.6321078743162447, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.190278053283691, "learning_rate": 5e-05, "loss": 1.0217, "mean_token_accuracy": 0.8530504703521729, "num_tokens": 188707717.0, "step": 4969 }, { "epoch": 0.6322350845948352, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.173247814178467, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.8679230213165283, "num_tokens": 188742811.0, "step": 4970 }, { "epoch": 0.6323622948734258, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.56466293334961, "learning_rate": 5e-05, "loss": 0.9506, "mean_token_accuracy": 0.8595864772796631, "num_tokens": 188781004.0, "step": 4971 }, { "epoch": 0.6324895051520163, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.767878532409668, "learning_rate": 5e-05, "loss": 0.9256, "mean_token_accuracy": 0.8540748357772827, "num_tokens": 188813923.0, "step": 4972 }, { "epoch": 0.6326167154306068, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.715503692626953, "learning_rate": 5e-05, "loss": 0.8434, "mean_token_accuracy": 0.8621742725372314, "num_tokens": 188855375.0, "step": 4973 }, { "epoch": 0.6327439257091974, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.924386501312256, "learning_rate": 5e-05, "loss": 0.8442, "mean_token_accuracy": 0.875075101852417, "num_tokens": 188892807.0, "step": 4974 }, { "epoch": 0.6328711359877878, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.461408615112305, "learning_rate": 5e-05, "loss": 0.7903, "mean_token_accuracy": 0.8663785457611084, "num_tokens": 188932942.0, "step": 4975 }, { "epoch": 0.6329983462663783, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.558760643005371, "learning_rate": 5e-05, "loss": 0.8268, "mean_token_accuracy": 0.8590152263641357, "num_tokens": 188968206.0, "step": 4976 }, { "epoch": 0.6331255565449688, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.639248371124268, "learning_rate": 5e-05, "loss": 0.8401, "mean_token_accuracy": 0.8529518246650696, "num_tokens": 189001315.0, "step": 4977 }, { "epoch": 0.6332527668235594, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.281075954437256, "learning_rate": 5e-05, "loss": 0.8347, "mean_token_accuracy": 0.8405047655105591, "num_tokens": 189039729.0, "step": 4978 }, { "epoch": 0.6333799771021499, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.630064010620117, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8757584095001221, "num_tokens": 189077455.0, "step": 4979 }, { "epoch": 0.6335071873807404, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.3176188468933105, "learning_rate": 5e-05, "loss": 0.7981, "mean_token_accuracy": 0.8506487011909485, "num_tokens": 189118417.0, "step": 4980 }, { "epoch": 0.6336343976593308, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.424018383026123, "learning_rate": 5e-05, "loss": 0.7386, "mean_token_accuracy": 0.8708017468452454, "num_tokens": 189155581.0, "step": 4981 }, { "epoch": 0.6337616079379214, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.467113971710205, "learning_rate": 5e-05, "loss": 0.7801, "mean_token_accuracy": 0.8559793829917908, "num_tokens": 189191192.0, "step": 4982 }, { "epoch": 0.6338888182165119, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.267935752868652, "learning_rate": 5e-05, "loss": 0.7384, "mean_token_accuracy": 0.8654972910881042, "num_tokens": 189230475.0, "step": 4983 }, { "epoch": 0.6340160284951024, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.530435085296631, "learning_rate": 5e-05, "loss": 0.7687, "mean_token_accuracy": 0.8603132963180542, "num_tokens": 189265949.0, "step": 4984 }, { "epoch": 0.6341432387736929, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.19610595703125, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.8586509823799133, "num_tokens": 189302914.0, "step": 4985 }, { "epoch": 0.6342704490522835, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.551097393035889, "learning_rate": 5e-05, "loss": 0.7357, "mean_token_accuracy": 0.8693283200263977, "num_tokens": 189347741.0, "step": 4986 }, { "epoch": 0.6343976593308739, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.131659030914307, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8684202432632446, "num_tokens": 189389672.0, "step": 4987 }, { "epoch": 0.6345248696094644, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.626800060272217, "learning_rate": 5e-05, "loss": 0.7402, "mean_token_accuracy": 0.8696853518486023, "num_tokens": 189425418.0, "step": 4988 }, { "epoch": 0.6346520798880549, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.0003275871276855, "learning_rate": 5e-05, "loss": 0.6967, "mean_token_accuracy": 0.8690636157989502, "num_tokens": 189457042.0, "step": 4989 }, { "epoch": 0.6347792901666455, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.772007942199707, "learning_rate": 5e-05, "loss": 0.8112, "mean_token_accuracy": 0.8482078313827515, "num_tokens": 189497135.0, "step": 4990 }, { "epoch": 0.634906500445236, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.792900562286377, "learning_rate": 5e-05, "loss": 0.6554, "mean_token_accuracy": 0.8783645629882812, "num_tokens": 189534829.0, "step": 4991 }, { "epoch": 0.6350337107238265, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.094544410705566, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.8544464707374573, "num_tokens": 189570535.0, "step": 4992 }, { "epoch": 0.635160921002417, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.416566848754883, "learning_rate": 5e-05, "loss": 0.6753, "mean_token_accuracy": 0.8669369220733643, "num_tokens": 189606448.0, "step": 4993 }, { "epoch": 0.6352881312810075, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.654243469238281, "learning_rate": 5e-05, "loss": 0.8583, "mean_token_accuracy": 0.860603928565979, "num_tokens": 189650880.0, "step": 4994 }, { "epoch": 0.635415341559598, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 5.920840263366699, "learning_rate": 5e-05, "loss": 0.7282, "mean_token_accuracy": 0.8452561497688293, "num_tokens": 189690866.0, "step": 4995 }, { "epoch": 0.6355425518381885, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.495190620422363, "learning_rate": 5e-05, "loss": 1.0538, "mean_token_accuracy": 0.8324319124221802, "num_tokens": 189726891.0, "step": 4996 }, { "epoch": 0.6356697621167791, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 5.796329975128174, "learning_rate": 5e-05, "loss": 0.705, "mean_token_accuracy": 0.8601271510124207, "num_tokens": 189763893.0, "step": 4997 }, { "epoch": 0.6357969723953696, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.028697967529297, "learning_rate": 5e-05, "loss": 1.1023, "mean_token_accuracy": 0.8453966379165649, "num_tokens": 189804122.0, "step": 4998 }, { "epoch": 0.63592418267396, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.037351608276367, "learning_rate": 5e-05, "loss": 0.9099, "mean_token_accuracy": 0.8559725284576416, "num_tokens": 189836355.0, "step": 4999 }, { "epoch": 0.6360513929525505, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.16899585723877, "learning_rate": 5e-05, "loss": 1.0041, "mean_token_accuracy": 0.8634894490242004, "num_tokens": 189871839.0, "step": 5000 }, { "epoch": 0.6361786032311411, "ewc_loss": 0.625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.570243835449219, "learning_rate": 5e-05, "loss": 1.0018, "mean_token_accuracy": 0.8698960542678833, "num_tokens": 189915631.0, "step": 5001 }, { "epoch": 0.6363058135097316, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.937527179718018, "learning_rate": 5e-05, "loss": 0.7993, "mean_token_accuracy": 0.8530327081680298, "num_tokens": 189953635.0, "step": 5002 }, { "epoch": 0.6364330237883221, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.215576171875, "learning_rate": 5e-05, "loss": 1.0696, "mean_token_accuracy": 0.8622124195098877, "num_tokens": 189999338.0, "step": 5003 }, { "epoch": 0.6365602340669126, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.150599956512451, "learning_rate": 5e-05, "loss": 0.7739, "mean_token_accuracy": 0.8607760667800903, "num_tokens": 190045804.0, "step": 5004 }, { "epoch": 0.6366874443455031, "ewc_loss": 0.890625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 11.771939277648926, "learning_rate": 5e-05, "loss": 1.3174, "mean_token_accuracy": 0.8574872612953186, "num_tokens": 190083350.0, "step": 5005 }, { "epoch": 0.6368146546240936, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.0007781982421875, "grad_norm": 8.894036293029785, "learning_rate": 5e-05, "loss": 1.254, "mean_token_accuracy": 0.8509775400161743, "num_tokens": 190121910.0, "step": 5006 }, { "epoch": 0.6369418649026841, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.67748737335205, "learning_rate": 5e-05, "loss": 1.054, "mean_token_accuracy": 0.8635351657867432, "num_tokens": 190163857.0, "step": 5007 }, { "epoch": 0.6370690751812746, "ewc_loss": 0.828125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 9.378146171569824, "learning_rate": 5e-05, "loss": 1.267, "mean_token_accuracy": 0.8543591499328613, "num_tokens": 190204337.0, "step": 5008 }, { "epoch": 0.6371962854598652, "ewc_loss": 0.625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.583298683166504, "learning_rate": 5e-05, "loss": 1.0035, "mean_token_accuracy": 0.8703572154045105, "num_tokens": 190240646.0, "step": 5009 }, { "epoch": 0.6373234957384557, "ewc_loss": 0.625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.91364574432373, "learning_rate": 5e-05, "loss": 1.0484, "mean_token_accuracy": 0.8543624877929688, "num_tokens": 190273507.0, "step": 5010 }, { "epoch": 0.6374507060170462, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.707802772521973, "learning_rate": 5e-05, "loss": 0.9943, "mean_token_accuracy": 0.8593310117721558, "num_tokens": 190306876.0, "step": 5011 }, { "epoch": 0.6375779162956366, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.298872947692871, "learning_rate": 5e-05, "loss": 0.9387, "mean_token_accuracy": 0.8497155904769897, "num_tokens": 190343126.0, "step": 5012 }, { "epoch": 0.6377051265742272, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.532159805297852, "learning_rate": 5e-05, "loss": 0.9981, "mean_token_accuracy": 0.8510472774505615, "num_tokens": 190382089.0, "step": 5013 }, { "epoch": 0.6378323368528177, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.459288120269775, "learning_rate": 5e-05, "loss": 0.857, "mean_token_accuracy": 0.84599769115448, "num_tokens": 190419577.0, "step": 5014 }, { "epoch": 0.6379595471314082, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.92978572845459, "learning_rate": 5e-05, "loss": 1.0245, "mean_token_accuracy": 0.8473905324935913, "num_tokens": 190459725.0, "step": 5015 }, { "epoch": 0.6380867574099988, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 6.948433876037598, "learning_rate": 5e-05, "loss": 0.8539, "mean_token_accuracy": 0.8465820550918579, "num_tokens": 190493575.0, "step": 5016 }, { "epoch": 0.6382139676885893, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.394909858703613, "learning_rate": 5e-05, "loss": 1.0164, "mean_token_accuracy": 0.860106348991394, "num_tokens": 190530326.0, "step": 5017 }, { "epoch": 0.6383411779671797, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.0422210693359375, "learning_rate": 5e-05, "loss": 0.8596, "mean_token_accuracy": 0.8574338555335999, "num_tokens": 190565727.0, "step": 5018 }, { "epoch": 0.6384683882457702, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.5642671585083, "learning_rate": 5e-05, "loss": 1.0683, "mean_token_accuracy": 0.8477045893669128, "num_tokens": 190600163.0, "step": 5019 }, { "epoch": 0.6385955985243608, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 7.511805057525635, "learning_rate": 5e-05, "loss": 0.9661, "mean_token_accuracy": 0.8500925302505493, "num_tokens": 190640801.0, "step": 5020 }, { "epoch": 0.6387228088029513, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.244322776794434, "learning_rate": 5e-05, "loss": 0.9512, "mean_token_accuracy": 0.8786208033561707, "num_tokens": 190679886.0, "step": 5021 }, { "epoch": 0.6388500190815418, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.854743480682373, "learning_rate": 5e-05, "loss": 0.9693, "mean_token_accuracy": 0.8588370084762573, "num_tokens": 190726246.0, "step": 5022 }, { "epoch": 0.6389772293601323, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.632718086242676, "learning_rate": 5e-05, "loss": 0.963, "mean_token_accuracy": 0.8598113656044006, "num_tokens": 190765496.0, "step": 5023 }, { "epoch": 0.6391044396387228, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.80318546295166, "learning_rate": 5e-05, "loss": 0.9283, "mean_token_accuracy": 0.8544226288795471, "num_tokens": 190808770.0, "step": 5024 }, { "epoch": 0.6392316499173133, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.351821899414062, "learning_rate": 5e-05, "loss": 0.9058, "mean_token_accuracy": 0.8691912293434143, "num_tokens": 190844242.0, "step": 5025 }, { "epoch": 0.6393588601959038, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.344799041748047, "learning_rate": 5e-05, "loss": 0.8913, "mean_token_accuracy": 0.8413330316543579, "num_tokens": 190887658.0, "step": 5026 }, { "epoch": 0.6394860704744944, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.716863632202148, "learning_rate": 5e-05, "loss": 0.9599, "mean_token_accuracy": 0.8638354539871216, "num_tokens": 190922322.0, "step": 5027 }, { "epoch": 0.6396132807530849, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.912241458892822, "learning_rate": 5e-05, "loss": 0.8035, "mean_token_accuracy": 0.8529303073883057, "num_tokens": 190959085.0, "step": 5028 }, { "epoch": 0.6397404910316754, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.282419204711914, "learning_rate": 5e-05, "loss": 0.9973, "mean_token_accuracy": 0.8717101812362671, "num_tokens": 190991646.0, "step": 5029 }, { "epoch": 0.6398677013102658, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.180076599121094, "learning_rate": 5e-05, "loss": 0.867, "mean_token_accuracy": 0.8583595752716064, "num_tokens": 191039489.0, "step": 5030 }, { "epoch": 0.6399949115888564, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.24715518951416, "learning_rate": 5e-05, "loss": 1.0486, "mean_token_accuracy": 0.8564128279685974, "num_tokens": 191078166.0, "step": 5031 }, { "epoch": 0.6401221218674469, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.955399513244629, "learning_rate": 5e-05, "loss": 1.0055, "mean_token_accuracy": 0.8383290767669678, "num_tokens": 191115981.0, "step": 5032 }, { "epoch": 0.6402493321460374, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.272427558898926, "learning_rate": 5e-05, "loss": 0.9336, "mean_token_accuracy": 0.8613568544387817, "num_tokens": 191153989.0, "step": 5033 }, { "epoch": 0.6403765424246279, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.31569766998291, "learning_rate": 5e-05, "loss": 0.951, "mean_token_accuracy": 0.8609218597412109, "num_tokens": 191192799.0, "step": 5034 }, { "epoch": 0.6405037527032185, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.261753082275391, "learning_rate": 5e-05, "loss": 0.8211, "mean_token_accuracy": 0.858002781867981, "num_tokens": 191231923.0, "step": 5035 }, { "epoch": 0.6406309629818089, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.332085609436035, "learning_rate": 5e-05, "loss": 0.8975, "mean_token_accuracy": 0.8592602014541626, "num_tokens": 191271650.0, "step": 5036 }, { "epoch": 0.6407581732603994, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.051071643829346, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.8628256916999817, "num_tokens": 191300212.0, "step": 5037 }, { "epoch": 0.6408853835389899, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.000261306762695, "learning_rate": 5e-05, "loss": 0.864, "mean_token_accuracy": 0.854142427444458, "num_tokens": 191336163.0, "step": 5038 }, { "epoch": 0.6410125938175805, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.290139675140381, "learning_rate": 5e-05, "loss": 0.7872, "mean_token_accuracy": 0.8594471216201782, "num_tokens": 191374749.0, "step": 5039 }, { "epoch": 0.641139804096171, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.635773658752441, "learning_rate": 5e-05, "loss": 0.7972, "mean_token_accuracy": 0.8614825010299683, "num_tokens": 191413342.0, "step": 5040 }, { "epoch": 0.6412670143747615, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.398591995239258, "learning_rate": 5e-05, "loss": 0.7664, "mean_token_accuracy": 0.8651379346847534, "num_tokens": 191450332.0, "step": 5041 }, { "epoch": 0.6413942246533519, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.519386291503906, "learning_rate": 5e-05, "loss": 0.7633, "mean_token_accuracy": 0.8631287813186646, "num_tokens": 191486969.0, "step": 5042 }, { "epoch": 0.6415214349319425, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.364338397979736, "learning_rate": 5e-05, "loss": 0.7511, "mean_token_accuracy": 0.8641008734703064, "num_tokens": 191525074.0, "step": 5043 }, { "epoch": 0.641648645210533, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.502615451812744, "learning_rate": 5e-05, "loss": 0.7805, "mean_token_accuracy": 0.857117235660553, "num_tokens": 191564324.0, "step": 5044 }, { "epoch": 0.6417758554891235, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.319352149963379, "learning_rate": 5e-05, "loss": 0.7754, "mean_token_accuracy": 0.8533523082733154, "num_tokens": 191604081.0, "step": 5045 }, { "epoch": 0.641903065767714, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.570230007171631, "learning_rate": 5e-05, "loss": 0.8073, "mean_token_accuracy": 0.8459135293960571, "num_tokens": 191639811.0, "step": 5046 }, { "epoch": 0.6420302760463046, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.09119987487793, "learning_rate": 5e-05, "loss": 0.7102, "mean_token_accuracy": 0.8689608573913574, "num_tokens": 191680941.0, "step": 5047 }, { "epoch": 0.642157486324895, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.895950794219971, "learning_rate": 5e-05, "loss": 0.8084, "mean_token_accuracy": 0.855032742023468, "num_tokens": 191718039.0, "step": 5048 }, { "epoch": 0.6422846966034855, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.696352481842041, "learning_rate": 5e-05, "loss": 0.7239, "mean_token_accuracy": 0.8596657514572144, "num_tokens": 191759837.0, "step": 5049 }, { "epoch": 0.6424119068820761, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.366411209106445, "learning_rate": 5e-05, "loss": 0.852, "mean_token_accuracy": 0.8537776470184326, "num_tokens": 191801791.0, "step": 5050 }, { "epoch": 0.6425391171606666, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.285542011260986, "learning_rate": 5e-05, "loss": 0.6724, "mean_token_accuracy": 0.869146466255188, "num_tokens": 191838193.0, "step": 5051 }, { "epoch": 0.6426663274392571, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.963191032409668, "learning_rate": 5e-05, "loss": 0.9869, "mean_token_accuracy": 0.8319703340530396, "num_tokens": 191874811.0, "step": 5052 }, { "epoch": 0.6427935377178476, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 5.995673179626465, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.8532540798187256, "num_tokens": 191921289.0, "step": 5053 }, { "epoch": 0.6429207479964381, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.536935806274414, "learning_rate": 5e-05, "loss": 0.998, "mean_token_accuracy": 0.8611161708831787, "num_tokens": 191961796.0, "step": 5054 }, { "epoch": 0.6430479582750286, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.404025077819824, "learning_rate": 5e-05, "loss": 0.6983, "mean_token_accuracy": 0.8780204057693481, "num_tokens": 192000853.0, "step": 5055 }, { "epoch": 0.6431751685536191, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.4506254196167, "learning_rate": 5e-05, "loss": 1.0207, "mean_token_accuracy": 0.8627498149871826, "num_tokens": 192029126.0, "step": 5056 }, { "epoch": 0.6433023788322096, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.631454944610596, "learning_rate": 5e-05, "loss": 0.9597, "mean_token_accuracy": 0.8514158725738525, "num_tokens": 192070472.0, "step": 5057 }, { "epoch": 0.6434295891108002, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.148722648620605, "learning_rate": 5e-05, "loss": 0.9555, "mean_token_accuracy": 0.847787618637085, "num_tokens": 192107844.0, "step": 5058 }, { "epoch": 0.6435567993893907, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.584877014160156, "learning_rate": 5e-05, "loss": 0.9637, "mean_token_accuracy": 0.8605084419250488, "num_tokens": 192140637.0, "step": 5059 }, { "epoch": 0.6436840096679812, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.740047931671143, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.8668022155761719, "num_tokens": 192176177.0, "step": 5060 }, { "epoch": 0.6438112199465716, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.16258430480957, "learning_rate": 5e-05, "loss": 0.9822, "mean_token_accuracy": 0.8639368414878845, "num_tokens": 192218131.0, "step": 5061 }, { "epoch": 0.6439384302251622, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.575324535369873, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.862552285194397, "num_tokens": 192258720.0, "step": 5062 }, { "epoch": 0.6440656405037527, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.059648513793945, "learning_rate": 5e-05, "loss": 1.023, "mean_token_accuracy": 0.849025309085846, "num_tokens": 192303197.0, "step": 5063 }, { "epoch": 0.6441928507823432, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.2300519943237305, "learning_rate": 5e-05, "loss": 0.9431, "mean_token_accuracy": 0.8359338641166687, "num_tokens": 192340882.0, "step": 5064 }, { "epoch": 0.6443200610609338, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.324398040771484, "learning_rate": 5e-05, "loss": 0.908, "mean_token_accuracy": 0.86264967918396, "num_tokens": 192383596.0, "step": 5065 }, { "epoch": 0.6444472713395243, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.6945695877075195, "learning_rate": 5e-05, "loss": 0.9062, "mean_token_accuracy": 0.8599616289138794, "num_tokens": 192423861.0, "step": 5066 }, { "epoch": 0.6445744816181147, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.737337589263916, "learning_rate": 5e-05, "loss": 0.8086, "mean_token_accuracy": 0.8662864565849304, "num_tokens": 192456564.0, "step": 5067 }, { "epoch": 0.6447016918967052, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.673246383666992, "learning_rate": 5e-05, "loss": 0.8825, "mean_token_accuracy": 0.8582494258880615, "num_tokens": 192496917.0, "step": 5068 }, { "epoch": 0.6448289021752958, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.634398460388184, "learning_rate": 5e-05, "loss": 0.8244, "mean_token_accuracy": 0.8578594326972961, "num_tokens": 192539548.0, "step": 5069 }, { "epoch": 0.6449561124538863, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.355658531188965, "learning_rate": 5e-05, "loss": 0.8501, "mean_token_accuracy": 0.8492214679718018, "num_tokens": 192577430.0, "step": 5070 }, { "epoch": 0.6450833227324768, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.792216777801514, "learning_rate": 5e-05, "loss": 0.7807, "mean_token_accuracy": 0.8707412481307983, "num_tokens": 192617343.0, "step": 5071 }, { "epoch": 0.6452105330110673, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.053980827331543, "learning_rate": 5e-05, "loss": 0.7546, "mean_token_accuracy": 0.8601855039596558, "num_tokens": 192662274.0, "step": 5072 }, { "epoch": 0.6453377432896578, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.023637771606445, "learning_rate": 5e-05, "loss": 0.823, "mean_token_accuracy": 0.8665862083435059, "num_tokens": 192703305.0, "step": 5073 }, { "epoch": 0.6454649535682483, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.769974708557129, "learning_rate": 5e-05, "loss": 0.7568, "mean_token_accuracy": 0.8539965748786926, "num_tokens": 192738325.0, "step": 5074 }, { "epoch": 0.6455921638468388, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.290796279907227, "learning_rate": 5e-05, "loss": 0.8288, "mean_token_accuracy": 0.872398316860199, "num_tokens": 192778285.0, "step": 5075 }, { "epoch": 0.6457193741254293, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.549795627593994, "learning_rate": 5e-05, "loss": 0.7227, "mean_token_accuracy": 0.8623583912849426, "num_tokens": 192822822.0, "step": 5076 }, { "epoch": 0.6458465844040199, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.604758262634277, "learning_rate": 5e-05, "loss": 0.9073, "mean_token_accuracy": 0.8578600883483887, "num_tokens": 192857312.0, "step": 5077 }, { "epoch": 0.6459737946826104, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.455924987792969, "learning_rate": 5e-05, "loss": 0.7058, "mean_token_accuracy": 0.8660665154457092, "num_tokens": 192892363.0, "step": 5078 }, { "epoch": 0.6461010049612008, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.757135391235352, "learning_rate": 5e-05, "loss": 0.9456, "mean_token_accuracy": 0.8498675227165222, "num_tokens": 192931774.0, "step": 5079 }, { "epoch": 0.6462282152397913, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.659084796905518, "learning_rate": 5e-05, "loss": 0.7406, "mean_token_accuracy": 0.8667351603507996, "num_tokens": 192964463.0, "step": 5080 }, { "epoch": 0.6463554255183819, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.553098678588867, "learning_rate": 5e-05, "loss": 0.883, "mean_token_accuracy": 0.8704524040222168, "num_tokens": 192999807.0, "step": 5081 }, { "epoch": 0.6464826357969724, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.086048126220703, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.8556276559829712, "num_tokens": 193041408.0, "step": 5082 }, { "epoch": 0.6466098460755629, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.094307899475098, "learning_rate": 5e-05, "loss": 0.9716, "mean_token_accuracy": 0.833090603351593, "num_tokens": 193078104.0, "step": 5083 }, { "epoch": 0.6467370563541535, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.391849517822266, "learning_rate": 5e-05, "loss": 0.8907, "mean_token_accuracy": 0.8425981998443604, "num_tokens": 193118821.0, "step": 5084 }, { "epoch": 0.6468642666327439, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.701286792755127, "learning_rate": 5e-05, "loss": 0.8587, "mean_token_accuracy": 0.8556116223335266, "num_tokens": 193152633.0, "step": 5085 }, { "epoch": 0.6469914769113344, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.500570774078369, "learning_rate": 5e-05, "loss": 0.813, "mean_token_accuracy": 0.8635208606719971, "num_tokens": 193191398.0, "step": 5086 }, { "epoch": 0.6471186871899249, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.493102073669434, "learning_rate": 5e-05, "loss": 0.832, "mean_token_accuracy": 0.8532388806343079, "num_tokens": 193229612.0, "step": 5087 }, { "epoch": 0.6472458974685155, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.505016803741455, "learning_rate": 5e-05, "loss": 0.8429, "mean_token_accuracy": 0.845548152923584, "num_tokens": 193271993.0, "step": 5088 }, { "epoch": 0.647373107747106, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.399219036102295, "learning_rate": 5e-05, "loss": 0.8235, "mean_token_accuracy": 0.8502653241157532, "num_tokens": 193308146.0, "step": 5089 }, { "epoch": 0.6475003180256965, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.498758316040039, "learning_rate": 5e-05, "loss": 0.7641, "mean_token_accuracy": 0.8646090626716614, "num_tokens": 193348448.0, "step": 5090 }, { "epoch": 0.6476275283042869, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.296459197998047, "learning_rate": 5e-05, "loss": 0.7571, "mean_token_accuracy": 0.8600054979324341, "num_tokens": 193379279.0, "step": 5091 }, { "epoch": 0.6477547385828775, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.574588775634766, "learning_rate": 5e-05, "loss": 0.7617, "mean_token_accuracy": 0.8636761903762817, "num_tokens": 193415999.0, "step": 5092 }, { "epoch": 0.647881948861468, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.112642765045166, "learning_rate": 5e-05, "loss": 0.7459, "mean_token_accuracy": 0.8625471591949463, "num_tokens": 193450287.0, "step": 5093 }, { "epoch": 0.6480091591400585, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.754034996032715, "learning_rate": 5e-05, "loss": 0.7803, "mean_token_accuracy": 0.8624590635299683, "num_tokens": 193485980.0, "step": 5094 }, { "epoch": 0.648136369418649, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.894063949584961, "learning_rate": 5e-05, "loss": 0.7614, "mean_token_accuracy": 0.8555164337158203, "num_tokens": 193525179.0, "step": 5095 }, { "epoch": 0.6482635796972396, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.000653266906738, "learning_rate": 5e-05, "loss": 0.8125, "mean_token_accuracy": 0.8570358753204346, "num_tokens": 193561580.0, "step": 5096 }, { "epoch": 0.64839078997583, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.65371561050415, "learning_rate": 5e-05, "loss": 0.7673, "mean_token_accuracy": 0.8492071032524109, "num_tokens": 193598218.0, "step": 5097 }, { "epoch": 0.6485180002544205, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.282051086425781, "learning_rate": 5e-05, "loss": 0.8724, "mean_token_accuracy": 0.8499606847763062, "num_tokens": 193635256.0, "step": 5098 }, { "epoch": 0.648645210533011, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.410903453826904, "learning_rate": 5e-05, "loss": 0.7287, "mean_token_accuracy": 0.8496497869491577, "num_tokens": 193672957.0, "step": 5099 }, { "epoch": 0.6487724208116016, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.649127006530762, "learning_rate": 5e-05, "loss": 0.8572, "mean_token_accuracy": 0.8671073913574219, "num_tokens": 193710575.0, "step": 5100 }, { "epoch": 0.6488996310901921, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.235466003417969, "learning_rate": 5e-05, "loss": 0.704, "mean_token_accuracy": 0.8677288293838501, "num_tokens": 193749006.0, "step": 5101 }, { "epoch": 0.6490268413687826, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.935335159301758, "learning_rate": 5e-05, "loss": 0.9939, "mean_token_accuracy": 0.8444572687149048, "num_tokens": 193788975.0, "step": 5102 }, { "epoch": 0.649154051647373, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.421854019165039, "learning_rate": 5e-05, "loss": 0.7439, "mean_token_accuracy": 0.8596563935279846, "num_tokens": 193822898.0, "step": 5103 }, { "epoch": 0.6492812619259636, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.772817611694336, "learning_rate": 5e-05, "loss": 0.8956, "mean_token_accuracy": 0.8722812533378601, "num_tokens": 193862439.0, "step": 5104 }, { "epoch": 0.6494084722045541, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.108672618865967, "learning_rate": 5e-05, "loss": 0.8643, "mean_token_accuracy": 0.8565818071365356, "num_tokens": 193906721.0, "step": 5105 }, { "epoch": 0.6495356824831446, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.927670001983643, "learning_rate": 5e-05, "loss": 0.8626, "mean_token_accuracy": 0.8639739155769348, "num_tokens": 193940521.0, "step": 5106 }, { "epoch": 0.6496628927617352, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.8471269607543945, "learning_rate": 5e-05, "loss": 0.9104, "mean_token_accuracy": 0.8502247929573059, "num_tokens": 193979340.0, "step": 5107 }, { "epoch": 0.6497901030403257, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.9979047775268555, "learning_rate": 5e-05, "loss": 0.7646, "mean_token_accuracy": 0.8599624037742615, "num_tokens": 194018562.0, "step": 5108 }, { "epoch": 0.6499173133189162, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.300213813781738, "learning_rate": 5e-05, "loss": 0.9626, "mean_token_accuracy": 0.830195426940918, "num_tokens": 194056804.0, "step": 5109 }, { "epoch": 0.6500445235975066, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.641761302947998, "learning_rate": 5e-05, "loss": 0.7279, "mean_token_accuracy": 0.8600798845291138, "num_tokens": 194095194.0, "step": 5110 }, { "epoch": 0.6501717338760972, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.37055778503418, "learning_rate": 5e-05, "loss": 0.8993, "mean_token_accuracy": 0.8562679290771484, "num_tokens": 194135061.0, "step": 5111 }, { "epoch": 0.6502989441546877, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.737835884094238, "learning_rate": 5e-05, "loss": 0.7446, "mean_token_accuracy": 0.8606961965560913, "num_tokens": 194169017.0, "step": 5112 }, { "epoch": 0.6504261544332782, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.171783447265625, "learning_rate": 5e-05, "loss": 0.8902, "mean_token_accuracy": 0.8491619825363159, "num_tokens": 194213424.0, "step": 5113 }, { "epoch": 0.6505533647118688, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.974647521972656, "learning_rate": 5e-05, "loss": 0.7433, "mean_token_accuracy": 0.8698010444641113, "num_tokens": 194247465.0, "step": 5114 }, { "epoch": 0.6506805749904593, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.91624641418457, "learning_rate": 5e-05, "loss": 0.8131, "mean_token_accuracy": 0.8649783134460449, "num_tokens": 194282575.0, "step": 5115 }, { "epoch": 0.6508077852690497, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.128428936004639, "learning_rate": 5e-05, "loss": 0.7749, "mean_token_accuracy": 0.860538899898529, "num_tokens": 194311822.0, "step": 5116 }, { "epoch": 0.6509349955476402, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.732499599456787, "learning_rate": 5e-05, "loss": 0.8443, "mean_token_accuracy": 0.849783718585968, "num_tokens": 194346990.0, "step": 5117 }, { "epoch": 0.6510622058262308, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.2050371170043945, "learning_rate": 5e-05, "loss": 0.762, "mean_token_accuracy": 0.8607736229896545, "num_tokens": 194384836.0, "step": 5118 }, { "epoch": 0.6511894161048213, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.628637790679932, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8711967468261719, "num_tokens": 194418080.0, "step": 5119 }, { "epoch": 0.6513166263834118, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.187285423278809, "learning_rate": 5e-05, "loss": 0.8137, "mean_token_accuracy": 0.845931351184845, "num_tokens": 194460111.0, "step": 5120 }, { "epoch": 0.6514438366620023, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.634436130523682, "learning_rate": 5e-05, "loss": 0.7561, "mean_token_accuracy": 0.8752602338790894, "num_tokens": 194498157.0, "step": 5121 }, { "epoch": 0.6515710469405928, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.0790324211120605, "learning_rate": 5e-05, "loss": 0.7438, "mean_token_accuracy": 0.8625513911247253, "num_tokens": 194538928.0, "step": 5122 }, { "epoch": 0.6516982572191833, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.747315883636475, "learning_rate": 5e-05, "loss": 0.7925, "mean_token_accuracy": 0.8617401123046875, "num_tokens": 194582276.0, "step": 5123 }, { "epoch": 0.6518254674977738, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.902137756347656, "learning_rate": 5e-05, "loss": 0.7263, "mean_token_accuracy": 0.860741913318634, "num_tokens": 194620197.0, "step": 5124 }, { "epoch": 0.6519526777763643, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.946324825286865, "learning_rate": 5e-05, "loss": 0.7821, "mean_token_accuracy": 0.8669620752334595, "num_tokens": 194654634.0, "step": 5125 }, { "epoch": 0.6520798880549549, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.657942771911621, "learning_rate": 5e-05, "loss": 0.7461, "mean_token_accuracy": 0.853259801864624, "num_tokens": 194696400.0, "step": 5126 }, { "epoch": 0.6522070983335454, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.284811973571777, "learning_rate": 5e-05, "loss": 0.8158, "mean_token_accuracy": 0.8654389381408691, "num_tokens": 194738326.0, "step": 5127 }, { "epoch": 0.6523343086121358, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.316368579864502, "learning_rate": 5e-05, "loss": 0.6928, "mean_token_accuracy": 0.8626359701156616, "num_tokens": 194778550.0, "step": 5128 }, { "epoch": 0.6524615188907263, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.790307998657227, "learning_rate": 5e-05, "loss": 0.9101, "mean_token_accuracy": 0.8566689491271973, "num_tokens": 194815124.0, "step": 5129 }, { "epoch": 0.6525887291693169, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.06879186630249, "learning_rate": 5e-05, "loss": 0.719, "mean_token_accuracy": 0.8609066605567932, "num_tokens": 194851372.0, "step": 5130 }, { "epoch": 0.6527159394479074, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.207930564880371, "learning_rate": 5e-05, "loss": 0.9419, "mean_token_accuracy": 0.8709473609924316, "num_tokens": 194886705.0, "step": 5131 }, { "epoch": 0.6528431497264979, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.387430667877197, "learning_rate": 5e-05, "loss": 0.7528, "mean_token_accuracy": 0.8606477975845337, "num_tokens": 194930372.0, "step": 5132 }, { "epoch": 0.6529703600050885, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.96194076538086, "learning_rate": 5e-05, "loss": 0.971, "mean_token_accuracy": 0.8628818988800049, "num_tokens": 194971869.0, "step": 5133 }, { "epoch": 0.6530975702836789, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.401998043060303, "learning_rate": 5e-05, "loss": 0.9275, "mean_token_accuracy": 0.8531436920166016, "num_tokens": 195011271.0, "step": 5134 }, { "epoch": 0.6532247805622694, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.681641101837158, "learning_rate": 5e-05, "loss": 0.8542, "mean_token_accuracy": 0.8613409399986267, "num_tokens": 195046328.0, "step": 5135 }, { "epoch": 0.6533519908408599, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.463838577270508, "learning_rate": 5e-05, "loss": 0.9431, "mean_token_accuracy": 0.8647136092185974, "num_tokens": 195086424.0, "step": 5136 }, { "epoch": 0.6534792011194505, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.3414435386657715, "learning_rate": 5e-05, "loss": 0.7624, "mean_token_accuracy": 0.8513801097869873, "num_tokens": 195121751.0, "step": 5137 }, { "epoch": 0.653606411398041, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.38994312286377, "learning_rate": 5e-05, "loss": 0.9981, "mean_token_accuracy": 0.8620672225952148, "num_tokens": 195158231.0, "step": 5138 }, { "epoch": 0.6537336216766315, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.069411754608154, "learning_rate": 5e-05, "loss": 0.7422, "mean_token_accuracy": 0.8592246770858765, "num_tokens": 195195009.0, "step": 5139 }, { "epoch": 0.6538608319552219, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.806801795959473, "learning_rate": 5e-05, "loss": 1.0299, "mean_token_accuracy": 0.8689464330673218, "num_tokens": 195232892.0, "step": 5140 }, { "epoch": 0.6539880422338125, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 7.157569408416748, "learning_rate": 5e-05, "loss": 0.9545, "mean_token_accuracy": 0.8569689989089966, "num_tokens": 195283233.0, "step": 5141 }, { "epoch": 0.654115252512403, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.026250839233398, "learning_rate": 5e-05, "loss": 0.9626, "mean_token_accuracy": 0.8776494264602661, "num_tokens": 195319402.0, "step": 5142 }, { "epoch": 0.6542424627909935, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.288400650024414, "learning_rate": 5e-05, "loss": 1.0117, "mean_token_accuracy": 0.8661075234413147, "num_tokens": 195358330.0, "step": 5143 }, { "epoch": 0.654369673069584, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.540313720703125, "learning_rate": 5e-05, "loss": 0.8103, "mean_token_accuracy": 0.872976541519165, "num_tokens": 195398074.0, "step": 5144 }, { "epoch": 0.6544968833481746, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.738027572631836, "learning_rate": 5e-05, "loss": 0.9706, "mean_token_accuracy": 0.8729128837585449, "num_tokens": 195435684.0, "step": 5145 }, { "epoch": 0.654624093626765, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.071155071258545, "learning_rate": 5e-05, "loss": 0.7982, "mean_token_accuracy": 0.8704869747161865, "num_tokens": 195475126.0, "step": 5146 }, { "epoch": 0.6547513039053555, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.316497802734375, "learning_rate": 5e-05, "loss": 0.9586, "mean_token_accuracy": 0.8573819398880005, "num_tokens": 195516398.0, "step": 5147 }, { "epoch": 0.654878514183946, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.596981048583984, "learning_rate": 5e-05, "loss": 0.8852, "mean_token_accuracy": 0.8638536930084229, "num_tokens": 195558777.0, "step": 5148 }, { "epoch": 0.6550057244625366, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.314751148223877, "learning_rate": 5e-05, "loss": 0.821, "mean_token_accuracy": 0.8553153276443481, "num_tokens": 195598663.0, "step": 5149 }, { "epoch": 0.6551329347411271, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.200333595275879, "learning_rate": 5e-05, "loss": 0.9397, "mean_token_accuracy": 0.8479846715927124, "num_tokens": 195628594.0, "step": 5150 }, { "epoch": 0.6552601450197176, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.711517333984375, "learning_rate": 5e-05, "loss": 0.7671, "mean_token_accuracy": 0.8550848960876465, "num_tokens": 195662484.0, "step": 5151 }, { "epoch": 0.655387355298308, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.583189964294434, "learning_rate": 5e-05, "loss": 0.9617, "mean_token_accuracy": 0.8486326932907104, "num_tokens": 195697753.0, "step": 5152 }, { "epoch": 0.6555145655768986, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.534820556640625, "learning_rate": 5e-05, "loss": 0.7183, "mean_token_accuracy": 0.8710623979568481, "num_tokens": 195740931.0, "step": 5153 }, { "epoch": 0.6556417758554891, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.916385650634766, "learning_rate": 5e-05, "loss": 1.0157, "mean_token_accuracy": 0.8467945456504822, "num_tokens": 195783056.0, "step": 5154 }, { "epoch": 0.6557689861340796, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.616556644439697, "learning_rate": 5e-05, "loss": 0.8389, "mean_token_accuracy": 0.8466969728469849, "num_tokens": 195819515.0, "step": 5155 }, { "epoch": 0.6558961964126702, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.047197341918945, "learning_rate": 5e-05, "loss": 1.035, "mean_token_accuracy": 0.845366358757019, "num_tokens": 195861039.0, "step": 5156 }, { "epoch": 0.6560234066912607, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 6.96799898147583, "learning_rate": 5e-05, "loss": 0.9262, "mean_token_accuracy": 0.8438518047332764, "num_tokens": 195897818.0, "step": 5157 }, { "epoch": 0.6561506169698512, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.847244262695312, "learning_rate": 5e-05, "loss": 0.916, "mean_token_accuracy": 0.8759259581565857, "num_tokens": 195930752.0, "step": 5158 }, { "epoch": 0.6562778272484416, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.272834777832031, "learning_rate": 5e-05, "loss": 0.9186, "mean_token_accuracy": 0.8549646735191345, "num_tokens": 195965036.0, "step": 5159 }, { "epoch": 0.6564050375270322, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.575271606445312, "learning_rate": 5e-05, "loss": 0.9947, "mean_token_accuracy": 0.8590310215950012, "num_tokens": 195999293.0, "step": 5160 }, { "epoch": 0.6565322478056227, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.215770721435547, "learning_rate": 5e-05, "loss": 0.9077, "mean_token_accuracy": 0.848938524723053, "num_tokens": 196039066.0, "step": 5161 }, { "epoch": 0.6566594580842132, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.665682792663574, "learning_rate": 5e-05, "loss": 1.009, "mean_token_accuracy": 0.8523362874984741, "num_tokens": 196077805.0, "step": 5162 }, { "epoch": 0.6567866683628037, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.03028678894043, "learning_rate": 5e-05, "loss": 0.9167, "mean_token_accuracy": 0.8365849256515503, "num_tokens": 196117460.0, "step": 5163 }, { "epoch": 0.6569138786413943, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.847318649291992, "learning_rate": 5e-05, "loss": 0.9927, "mean_token_accuracy": 0.8611723780632019, "num_tokens": 196154855.0, "step": 5164 }, { "epoch": 0.6570410889199847, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.164154052734375, "learning_rate": 5e-05, "loss": 0.9284, "mean_token_accuracy": 0.8442356586456299, "num_tokens": 196197360.0, "step": 5165 }, { "epoch": 0.6571682991985752, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.624577522277832, "learning_rate": 5e-05, "loss": 1.057, "mean_token_accuracy": 0.8366886377334595, "num_tokens": 196237419.0, "step": 5166 }, { "epoch": 0.6572955094771658, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.550792694091797, "learning_rate": 5e-05, "loss": 0.9395, "mean_token_accuracy": 0.8490055203437805, "num_tokens": 196275533.0, "step": 5167 }, { "epoch": 0.6574227197557563, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.004570007324219, "learning_rate": 5e-05, "loss": 0.9047, "mean_token_accuracy": 0.8645536303520203, "num_tokens": 196312269.0, "step": 5168 }, { "epoch": 0.6575499300343468, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.738056182861328, "learning_rate": 5e-05, "loss": 0.9074, "mean_token_accuracy": 0.8532137274742126, "num_tokens": 196349338.0, "step": 5169 }, { "epoch": 0.6576771403129373, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.57559871673584, "learning_rate": 5e-05, "loss": 0.8961, "mean_token_accuracy": 0.8481410145759583, "num_tokens": 196396000.0, "step": 5170 }, { "epoch": 0.6578043505915278, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.597340106964111, "learning_rate": 5e-05, "loss": 0.871, "mean_token_accuracy": 0.8509958982467651, "num_tokens": 196428119.0, "step": 5171 }, { "epoch": 0.6579315608701183, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.596240043640137, "learning_rate": 5e-05, "loss": 0.8788, "mean_token_accuracy": 0.8461854457855225, "num_tokens": 196470466.0, "step": 5172 }, { "epoch": 0.6580587711487088, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.2934675216674805, "learning_rate": 5e-05, "loss": 0.799, "mean_token_accuracy": 0.8604331612586975, "num_tokens": 196508883.0, "step": 5173 }, { "epoch": 0.6581859814272993, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.758339881896973, "learning_rate": 5e-05, "loss": 0.8556, "mean_token_accuracy": 0.846881628036499, "num_tokens": 196552762.0, "step": 5174 }, { "epoch": 0.6583131917058899, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.1315813064575195, "learning_rate": 5e-05, "loss": 0.7095, "mean_token_accuracy": 0.8787588477134705, "num_tokens": 196581394.0, "step": 5175 }, { "epoch": 0.6584404019844804, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.739616870880127, "learning_rate": 5e-05, "loss": 0.7922, "mean_token_accuracy": 0.8659507036209106, "num_tokens": 196623237.0, "step": 5176 }, { "epoch": 0.6585676122630708, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.137954235076904, "learning_rate": 5e-05, "loss": 0.7234, "mean_token_accuracy": 0.8739670515060425, "num_tokens": 196659219.0, "step": 5177 }, { "epoch": 0.6586948225416613, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.703664302825928, "learning_rate": 5e-05, "loss": 0.8162, "mean_token_accuracy": 0.8531977534294128, "num_tokens": 196700102.0, "step": 5178 }, { "epoch": 0.6588220328202519, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.064980506896973, "learning_rate": 5e-05, "loss": 0.6703, "mean_token_accuracy": 0.8857159614562988, "num_tokens": 196740569.0, "step": 5179 }, { "epoch": 0.6589492430988424, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.810482978820801, "learning_rate": 5e-05, "loss": 0.8294, "mean_token_accuracy": 0.8531041741371155, "num_tokens": 196778634.0, "step": 5180 }, { "epoch": 0.6590764533774329, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.932565212249756, "learning_rate": 5e-05, "loss": 0.774, "mean_token_accuracy": 0.8521010875701904, "num_tokens": 196805021.0, "step": 5181 }, { "epoch": 0.6592036636560235, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.95535135269165, "learning_rate": 5e-05, "loss": 0.8147, "mean_token_accuracy": 0.8625485301017761, "num_tokens": 196839676.0, "step": 5182 }, { "epoch": 0.6593308739346139, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.7757391929626465, "learning_rate": 5e-05, "loss": 0.712, "mean_token_accuracy": 0.8663581609725952, "num_tokens": 196877564.0, "step": 5183 }, { "epoch": 0.6594580842132044, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.16189193725586, "learning_rate": 5e-05, "loss": 0.8598, "mean_token_accuracy": 0.8522039651870728, "num_tokens": 196913819.0, "step": 5184 }, { "epoch": 0.6595852944917949, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.649811744689941, "learning_rate": 5e-05, "loss": 0.7582, "mean_token_accuracy": 0.8468554019927979, "num_tokens": 196951019.0, "step": 5185 }, { "epoch": 0.6597125047703855, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.335318565368652, "learning_rate": 5e-05, "loss": 0.8697, "mean_token_accuracy": 0.8584817051887512, "num_tokens": 196995112.0, "step": 5186 }, { "epoch": 0.659839715048976, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.659419059753418, "learning_rate": 5e-05, "loss": 0.7175, "mean_token_accuracy": 0.8633155822753906, "num_tokens": 197035334.0, "step": 5187 }, { "epoch": 0.6599669253275665, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.294914245605469, "learning_rate": 5e-05, "loss": 0.9043, "mean_token_accuracy": 0.8504728078842163, "num_tokens": 197074234.0, "step": 5188 }, { "epoch": 0.6600941356061569, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.918951988220215, "learning_rate": 5e-05, "loss": 0.7831, "mean_token_accuracy": 0.8525965213775635, "num_tokens": 197110977.0, "step": 5189 }, { "epoch": 0.6602213458847475, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.927647590637207, "learning_rate": 5e-05, "loss": 0.8831, "mean_token_accuracy": 0.8485201597213745, "num_tokens": 197148634.0, "step": 5190 }, { "epoch": 0.660348556163338, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.343145847320557, "learning_rate": 5e-05, "loss": 0.7916, "mean_token_accuracy": 0.8619564771652222, "num_tokens": 197186361.0, "step": 5191 }, { "epoch": 0.6604757664419285, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.330904960632324, "learning_rate": 5e-05, "loss": 0.782, "mean_token_accuracy": 0.8579930067062378, "num_tokens": 197222642.0, "step": 5192 }, { "epoch": 0.660602976720519, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.7849836349487305, "learning_rate": 5e-05, "loss": 0.7766, "mean_token_accuracy": 0.8719885349273682, "num_tokens": 197263320.0, "step": 5193 }, { "epoch": 0.6607301869991096, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.8165974617004395, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8481350541114807, "num_tokens": 197294041.0, "step": 5194 }, { "epoch": 0.6608573972777, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.172571182250977, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8668352365493774, "num_tokens": 197324878.0, "step": 5195 }, { "epoch": 0.6609846075562905, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.4375410079956055, "learning_rate": 5e-05, "loss": 0.6743, "mean_token_accuracy": 0.8695598840713501, "num_tokens": 197368318.0, "step": 5196 }, { "epoch": 0.661111817834881, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.59868335723877, "learning_rate": 5e-05, "loss": 0.8693, "mean_token_accuracy": 0.8643335700035095, "num_tokens": 197407665.0, "step": 5197 }, { "epoch": 0.6612390281134716, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.157686233520508, "learning_rate": 5e-05, "loss": 0.7157, "mean_token_accuracy": 0.8593062162399292, "num_tokens": 197445812.0, "step": 5198 }, { "epoch": 0.6613662383920621, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003376007080078125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 9.010149955749512, "learning_rate": 5e-05, "loss": 0.9072, "mean_token_accuracy": 0.8677568435668945, "num_tokens": 197482733.0, "step": 5199 }, { "epoch": 0.6614934486706526, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.254181861877441, "learning_rate": 5e-05, "loss": 0.7298, "mean_token_accuracy": 0.8594526052474976, "num_tokens": 197520874.0, "step": 5200 }, { "epoch": 0.661620658949243, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.036519050598145, "learning_rate": 5e-05, "loss": 0.9856, "mean_token_accuracy": 0.8553597927093506, "num_tokens": 197563422.0, "step": 5201 }, { "epoch": 0.6617478692278336, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 6.977282524108887, "learning_rate": 5e-05, "loss": 0.8566, "mean_token_accuracy": 0.8522516489028931, "num_tokens": 197603632.0, "step": 5202 }, { "epoch": 0.6618750795064241, "ewc_loss": 0.5, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.222700119018555, "learning_rate": 5e-05, "loss": 0.923, "mean_token_accuracy": 0.8616701364517212, "num_tokens": 197642090.0, "step": 5203 }, { "epoch": 0.6620022897850146, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.861782550811768, "learning_rate": 5e-05, "loss": 1.0018, "mean_token_accuracy": 0.8302778005599976, "num_tokens": 197679294.0, "step": 5204 }, { "epoch": 0.6621295000636052, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.079408645629883, "learning_rate": 5e-05, "loss": 0.8627, "mean_token_accuracy": 0.8382375240325928, "num_tokens": 197721254.0, "step": 5205 }, { "epoch": 0.6622567103421957, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.366859436035156, "learning_rate": 5e-05, "loss": 0.8433, "mean_token_accuracy": 0.8785623908042908, "num_tokens": 197764656.0, "step": 5206 }, { "epoch": 0.6623839206207861, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.732151985168457, "learning_rate": 5e-05, "loss": 0.7603, "mean_token_accuracy": 0.8604800701141357, "num_tokens": 197805900.0, "step": 5207 }, { "epoch": 0.6625111308993766, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.170537948608398, "learning_rate": 5e-05, "loss": 0.8577, "mean_token_accuracy": 0.864047646522522, "num_tokens": 197842752.0, "step": 5208 }, { "epoch": 0.6626383411779672, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.123549938201904, "learning_rate": 5e-05, "loss": 0.7502, "mean_token_accuracy": 0.8694437146186829, "num_tokens": 197883531.0, "step": 5209 }, { "epoch": 0.6627655514565577, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.519504547119141, "learning_rate": 5e-05, "loss": 0.7882, "mean_token_accuracy": 0.8640854358673096, "num_tokens": 197920875.0, "step": 5210 }, { "epoch": 0.6628927617351482, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.643161773681641, "learning_rate": 5e-05, "loss": 0.813, "mean_token_accuracy": 0.8602675199508667, "num_tokens": 197955194.0, "step": 5211 }, { "epoch": 0.6630199720137387, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.9708781242370605, "learning_rate": 5e-05, "loss": 0.7107, "mean_token_accuracy": 0.8698757886886597, "num_tokens": 197991730.0, "step": 5212 }, { "epoch": 0.6631471822923293, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.988935470581055, "learning_rate": 5e-05, "loss": 0.8077, "mean_token_accuracy": 0.8632784485816956, "num_tokens": 198030032.0, "step": 5213 }, { "epoch": 0.6632743925709197, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.723329544067383, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8631120920181274, "num_tokens": 198068274.0, "step": 5214 }, { "epoch": 0.6634016028495102, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.166544914245605, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.8559134006500244, "num_tokens": 198105326.0, "step": 5215 }, { "epoch": 0.6635288131281007, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.6076836585998535, "learning_rate": 5e-05, "loss": 0.6824, "mean_token_accuracy": 0.8729187846183777, "num_tokens": 198147485.0, "step": 5216 }, { "epoch": 0.6636560234066913, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.34231185913086, "learning_rate": 5e-05, "loss": 0.8574, "mean_token_accuracy": 0.8635979890823364, "num_tokens": 198184928.0, "step": 5217 }, { "epoch": 0.6637832336852818, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.512597560882568, "learning_rate": 5e-05, "loss": 0.7278, "mean_token_accuracy": 0.8583921790122986, "num_tokens": 198224641.0, "step": 5218 }, { "epoch": 0.6639104439638723, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.501120567321777, "learning_rate": 5e-05, "loss": 0.8492, "mean_token_accuracy": 0.8697220087051392, "num_tokens": 198263821.0, "step": 5219 }, { "epoch": 0.6640376542424627, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.472877025604248, "learning_rate": 5e-05, "loss": 0.7272, "mean_token_accuracy": 0.8625103235244751, "num_tokens": 198300315.0, "step": 5220 }, { "epoch": 0.6641648645210533, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.651451110839844, "learning_rate": 5e-05, "loss": 0.9006, "mean_token_accuracy": 0.8646543025970459, "num_tokens": 198336543.0, "step": 5221 }, { "epoch": 0.6642920747996438, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.537840366363525, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.862970232963562, "num_tokens": 198374728.0, "step": 5222 }, { "epoch": 0.6644192850782343, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.712270736694336, "learning_rate": 5e-05, "loss": 0.9269, "mean_token_accuracy": 0.8644571304321289, "num_tokens": 198414504.0, "step": 5223 }, { "epoch": 0.6645464953568249, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.762599945068359, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8728685975074768, "num_tokens": 198449976.0, "step": 5224 }, { "epoch": 0.6646737056354154, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.532992362976074, "learning_rate": 5e-05, "loss": 0.9225, "mean_token_accuracy": 0.8665532469749451, "num_tokens": 198489522.0, "step": 5225 }, { "epoch": 0.6648009159140058, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.126871109008789, "learning_rate": 5e-05, "loss": 0.8191, "mean_token_accuracy": 0.8688892126083374, "num_tokens": 198524460.0, "step": 5226 }, { "epoch": 0.6649281261925963, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.11933422088623, "learning_rate": 5e-05, "loss": 0.8783, "mean_token_accuracy": 0.8664907217025757, "num_tokens": 198569139.0, "step": 5227 }, { "epoch": 0.6650553364711869, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.399679660797119, "learning_rate": 5e-05, "loss": 0.7945, "mean_token_accuracy": 0.878602147102356, "num_tokens": 198597166.0, "step": 5228 }, { "epoch": 0.6651825467497774, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.731790065765381, "learning_rate": 5e-05, "loss": 0.8517, "mean_token_accuracy": 0.8655923008918762, "num_tokens": 198631646.0, "step": 5229 }, { "epoch": 0.6653097570283679, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.412917137145996, "learning_rate": 5e-05, "loss": 0.8729, "mean_token_accuracy": 0.8491666913032532, "num_tokens": 198675743.0, "step": 5230 }, { "epoch": 0.6654369673069584, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.664252758026123, "learning_rate": 5e-05, "loss": 0.8564, "mean_token_accuracy": 0.8562994003295898, "num_tokens": 198715826.0, "step": 5231 }, { "epoch": 0.6655641775855489, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.201112747192383, "learning_rate": 5e-05, "loss": 0.7899, "mean_token_accuracy": 0.8610875606536865, "num_tokens": 198757320.0, "step": 5232 }, { "epoch": 0.6656913878641394, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.829699516296387, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.8690640926361084, "num_tokens": 198791110.0, "step": 5233 }, { "epoch": 0.6658185981427299, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.953339576721191, "learning_rate": 5e-05, "loss": 0.7378, "mean_token_accuracy": 0.8662905097007751, "num_tokens": 198828506.0, "step": 5234 }, { "epoch": 0.6659458084213205, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.98817777633667, "learning_rate": 5e-05, "loss": 0.8945, "mean_token_accuracy": 0.8440260291099548, "num_tokens": 198870759.0, "step": 5235 }, { "epoch": 0.666073018699911, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.883325576782227, "learning_rate": 5e-05, "loss": 0.7165, "mean_token_accuracy": 0.8690186142921448, "num_tokens": 198908178.0, "step": 5236 }, { "epoch": 0.6662002289785015, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.959961414337158, "learning_rate": 5e-05, "loss": 0.854, "mean_token_accuracy": 0.861784040927887, "num_tokens": 198942847.0, "step": 5237 }, { "epoch": 0.6663274392570919, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.975165843963623, "learning_rate": 5e-05, "loss": 0.7526, "mean_token_accuracy": 0.859602689743042, "num_tokens": 198981444.0, "step": 5238 }, { "epoch": 0.6664546495356825, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.813969135284424, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.870354413986206, "num_tokens": 199019159.0, "step": 5239 }, { "epoch": 0.666581859814273, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.082268714904785, "learning_rate": 5e-05, "loss": 0.7904, "mean_token_accuracy": 0.8547724485397339, "num_tokens": 199056696.0, "step": 5240 }, { "epoch": 0.6667090700928635, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.687396049499512, "learning_rate": 5e-05, "loss": 0.7528, "mean_token_accuracy": 0.8731623888015747, "num_tokens": 199092358.0, "step": 5241 }, { "epoch": 0.666836280371454, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.16241455078125, "learning_rate": 5e-05, "loss": 0.8624, "mean_token_accuracy": 0.8301709294319153, "num_tokens": 199135527.0, "step": 5242 }, { "epoch": 0.6669634906500446, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.549596309661865, "learning_rate": 5e-05, "loss": 0.7814, "mean_token_accuracy": 0.8639893531799316, "num_tokens": 199171886.0, "step": 5243 }, { "epoch": 0.667090700928635, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.2367024421691895, "learning_rate": 5e-05, "loss": 0.772, "mean_token_accuracy": 0.8571279048919678, "num_tokens": 199209807.0, "step": 5244 }, { "epoch": 0.6672179112072255, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.442485809326172, "learning_rate": 5e-05, "loss": 0.7817, "mean_token_accuracy": 0.8584474921226501, "num_tokens": 199244878.0, "step": 5245 }, { "epoch": 0.667345121485816, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.284407615661621, "learning_rate": 5e-05, "loss": 0.7772, "mean_token_accuracy": 0.8548230528831482, "num_tokens": 199279445.0, "step": 5246 }, { "epoch": 0.6674723317644066, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.371086120605469, "learning_rate": 5e-05, "loss": 0.6979, "mean_token_accuracy": 0.8765313625335693, "num_tokens": 199315504.0, "step": 5247 }, { "epoch": 0.6675995420429971, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.302145957946777, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8440689444541931, "num_tokens": 199354345.0, "step": 5248 }, { "epoch": 0.6677267523215876, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.311788558959961, "learning_rate": 5e-05, "loss": 0.8031, "mean_token_accuracy": 0.8435304760932922, "num_tokens": 199401530.0, "step": 5249 }, { "epoch": 0.667853962600178, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.336907386779785, "learning_rate": 5e-05, "loss": 0.7634, "mean_token_accuracy": 0.8542258739471436, "num_tokens": 199440428.0, "step": 5250 }, { "epoch": 0.6679811728787686, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.23690938949585, "learning_rate": 5e-05, "loss": 0.6903, "mean_token_accuracy": 0.8786017298698425, "num_tokens": 199478875.0, "step": 5251 }, { "epoch": 0.6681083831573591, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.386637210845947, "learning_rate": 5e-05, "loss": 0.8138, "mean_token_accuracy": 0.839442253112793, "num_tokens": 199517287.0, "step": 5252 }, { "epoch": 0.6682355934359496, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.183233737945557, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.8552267551422119, "num_tokens": 199553723.0, "step": 5253 }, { "epoch": 0.6683628037145402, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.427186965942383, "learning_rate": 5e-05, "loss": 0.7287, "mean_token_accuracy": 0.8635756969451904, "num_tokens": 199592059.0, "step": 5254 }, { "epoch": 0.6684900139931307, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.106022834777832, "learning_rate": 5e-05, "loss": 0.7447, "mean_token_accuracy": 0.8548768758773804, "num_tokens": 199633005.0, "step": 5255 }, { "epoch": 0.6686172242717211, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.502152442932129, "learning_rate": 5e-05, "loss": 0.8358, "mean_token_accuracy": 0.8328412771224976, "num_tokens": 199677507.0, "step": 5256 }, { "epoch": 0.6687444345503116, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.995526313781738, "learning_rate": 5e-05, "loss": 0.7084, "mean_token_accuracy": 0.8611433506011963, "num_tokens": 199716040.0, "step": 5257 }, { "epoch": 0.6688716448289022, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.641855716705322, "learning_rate": 5e-05, "loss": 0.7166, "mean_token_accuracy": 0.8636179566383362, "num_tokens": 199750438.0, "step": 5258 }, { "epoch": 0.6689988551074927, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.80348539352417, "learning_rate": 5e-05, "loss": 0.7148, "mean_token_accuracy": 0.8599581718444824, "num_tokens": 199786889.0, "step": 5259 }, { "epoch": 0.6691260653860832, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 8.244616508483887, "learning_rate": 5e-05, "loss": 0.7699, "mean_token_accuracy": 0.8639102578163147, "num_tokens": 199821673.0, "step": 5260 }, { "epoch": 0.6692532756646737, "ewc_loss": 0.240234375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 0.00020599365234375, "grad_norm": 5.867188453674316, "learning_rate": 5e-05, "loss": 0.6986, "mean_token_accuracy": 0.846085786819458, "num_tokens": 199854313.0, "step": 5261 }, { "epoch": 0.6693804859432643, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.659456253051758, "learning_rate": 5e-05, "loss": 0.9456, "mean_token_accuracy": 0.8631201982498169, "num_tokens": 199892853.0, "step": 5262 }, { "epoch": 0.6695076962218547, "ewc_loss": 0.2314453125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 0.00019741058349609375, "grad_norm": 4.98394775390625, "learning_rate": 5e-05, "loss": 0.6412, "mean_token_accuracy": 0.857530951499939, "num_tokens": 199931754.0, "step": 5263 }, { "epoch": 0.6696349065004452, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 11.710564613342285, "learning_rate": 5e-05, "loss": 1.191, "mean_token_accuracy": 0.8736414909362793, "num_tokens": 199965570.0, "step": 5264 }, { "epoch": 0.6697621167790357, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003490447998046875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 6.988916397094727, "learning_rate": 5e-05, "loss": 0.9333, "mean_token_accuracy": 0.8721303939819336, "num_tokens": 200004931.0, "step": 5265 }, { "epoch": 0.6698893270576263, "ewc_loss": 0.99609375, "ewc_loss_diag": 0.0003490447998046875, "ewc_loss_parallel": 0.0009613037109375, "grad_norm": 12.724000930786133, "learning_rate": 5e-05, "loss": 1.3927, "mean_token_accuracy": 0.8709652423858643, "num_tokens": 200040315.0, "step": 5266 }, { "epoch": 0.6700165373362168, "ewc_loss": 1.125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0010833740234375, "grad_norm": 10.315642356872559, "learning_rate": 5e-05, "loss": 1.541, "mean_token_accuracy": 0.8554674386978149, "num_tokens": 200081206.0, "step": 5267 }, { "epoch": 0.6701437476148073, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000354766845703125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 11.319714546203613, "learning_rate": 5e-05, "loss": 1.1015, "mean_token_accuracy": 0.8647074699401855, "num_tokens": 200119975.0, "step": 5268 }, { "epoch": 0.6702709578933977, "ewc_loss": 0.875, "ewc_loss_diag": 0.000354766845703125, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 9.501811027526855, "learning_rate": 5e-05, "loss": 1.3561, "mean_token_accuracy": 0.8387795686721802, "num_tokens": 200154956.0, "step": 5269 }, { "epoch": 0.6703981681719883, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000354766845703125, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 10.878071784973145, "learning_rate": 5e-05, "loss": 1.2037, "mean_token_accuracy": 0.8509786128997803, "num_tokens": 200194499.0, "step": 5270 }, { "epoch": 0.6705253784505788, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.500033378601074, "learning_rate": 5e-05, "loss": 1.0054, "mean_token_accuracy": 0.8655491471290588, "num_tokens": 200227142.0, "step": 5271 }, { "epoch": 0.6706525887291693, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.317447662353516, "learning_rate": 5e-05, "loss": 1.1721, "mean_token_accuracy": 0.8466651439666748, "num_tokens": 200262123.0, "step": 5272 }, { "epoch": 0.6707797990077599, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.1178560256958, "learning_rate": 5e-05, "loss": 0.996, "mean_token_accuracy": 0.8553128242492676, "num_tokens": 200299648.0, "step": 5273 }, { "epoch": 0.6709070092863504, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.156095504760742, "learning_rate": 5e-05, "loss": 1.1429, "mean_token_accuracy": 0.8418837785720825, "num_tokens": 200337191.0, "step": 5274 }, { "epoch": 0.6710342195649408, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.101264953613281, "learning_rate": 5e-05, "loss": 0.9594, "mean_token_accuracy": 0.8674712777137756, "num_tokens": 200371487.0, "step": 5275 }, { "epoch": 0.6711614298435313, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.951556205749512, "learning_rate": 5e-05, "loss": 1.0537, "mean_token_accuracy": 0.8626607656478882, "num_tokens": 200412634.0, "step": 5276 }, { "epoch": 0.6712886401221219, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.216675758361816, "learning_rate": 5e-05, "loss": 0.9828, "mean_token_accuracy": 0.8659252524375916, "num_tokens": 200449663.0, "step": 5277 }, { "epoch": 0.6714158504007124, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.578909873962402, "learning_rate": 5e-05, "loss": 1.0896, "mean_token_accuracy": 0.845624566078186, "num_tokens": 200494962.0, "step": 5278 }, { "epoch": 0.6715430606793029, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.111387252807617, "learning_rate": 5e-05, "loss": 0.9184, "mean_token_accuracy": 0.872700572013855, "num_tokens": 200531101.0, "step": 5279 }, { "epoch": 0.6716702709578934, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.297945022583008, "learning_rate": 5e-05, "loss": 0.9969, "mean_token_accuracy": 0.867756187915802, "num_tokens": 200568725.0, "step": 5280 }, { "epoch": 0.6717974812364839, "ewc_loss": 0.5, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.846182346343994, "learning_rate": 5e-05, "loss": 0.918, "mean_token_accuracy": 0.855860710144043, "num_tokens": 200607319.0, "step": 5281 }, { "epoch": 0.6719246915150744, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.284711837768555, "learning_rate": 5e-05, "loss": 1.025, "mean_token_accuracy": 0.8570113182067871, "num_tokens": 200654791.0, "step": 5282 }, { "epoch": 0.6720519017936649, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.618818283081055, "learning_rate": 5e-05, "loss": 0.8752, "mean_token_accuracy": 0.8645566701889038, "num_tokens": 200694175.0, "step": 5283 }, { "epoch": 0.6721791120722554, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.485611915588379, "learning_rate": 5e-05, "loss": 1.0429, "mean_token_accuracy": 0.8598500490188599, "num_tokens": 200726675.0, "step": 5284 }, { "epoch": 0.672306322350846, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.71011209487915, "learning_rate": 5e-05, "loss": 0.9389, "mean_token_accuracy": 0.8502687215805054, "num_tokens": 200759358.0, "step": 5285 }, { "epoch": 0.6724335326294365, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.505938529968262, "learning_rate": 5e-05, "loss": 1.0707, "mean_token_accuracy": 0.8626717925071716, "num_tokens": 200797735.0, "step": 5286 }, { "epoch": 0.6725607429080269, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.056740760803223, "learning_rate": 5e-05, "loss": 0.9811, "mean_token_accuracy": 0.855238676071167, "num_tokens": 200835446.0, "step": 5287 }, { "epoch": 0.6726879531866174, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.027138710021973, "learning_rate": 5e-05, "loss": 1.007, "mean_token_accuracy": 0.8653908967971802, "num_tokens": 200881020.0, "step": 5288 }, { "epoch": 0.672815163465208, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.167488098144531, "learning_rate": 5e-05, "loss": 0.9377, "mean_token_accuracy": 0.8625896573066711, "num_tokens": 200921549.0, "step": 5289 }, { "epoch": 0.6729423737437985, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.428714752197266, "learning_rate": 5e-05, "loss": 0.9943, "mean_token_accuracy": 0.848885178565979, "num_tokens": 200953991.0, "step": 5290 }, { "epoch": 0.673069584022389, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.840557098388672, "learning_rate": 5e-05, "loss": 0.8695, "mean_token_accuracy": 0.8648310899734497, "num_tokens": 200990240.0, "step": 5291 }, { "epoch": 0.6731967943009796, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.35502815246582, "learning_rate": 5e-05, "loss": 0.9022, "mean_token_accuracy": 0.8734999299049377, "num_tokens": 201022433.0, "step": 5292 }, { "epoch": 0.67332400457957, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.300247669219971, "learning_rate": 5e-05, "loss": 0.8099, "mean_token_accuracy": 0.8620227575302124, "num_tokens": 201055801.0, "step": 5293 }, { "epoch": 0.6734512148581605, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.758604049682617, "learning_rate": 5e-05, "loss": 0.996, "mean_token_accuracy": 0.8522912263870239, "num_tokens": 201102141.0, "step": 5294 }, { "epoch": 0.673578425136751, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.046977996826172, "learning_rate": 5e-05, "loss": 0.8763, "mean_token_accuracy": 0.8488882780075073, "num_tokens": 201139834.0, "step": 5295 }, { "epoch": 0.6737056354153416, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.109179496765137, "learning_rate": 5e-05, "loss": 0.9827, "mean_token_accuracy": 0.8691843152046204, "num_tokens": 201169650.0, "step": 5296 }, { "epoch": 0.6738328456939321, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.381371974945068, "learning_rate": 5e-05, "loss": 0.8523, "mean_token_accuracy": 0.8656288385391235, "num_tokens": 201208716.0, "step": 5297 }, { "epoch": 0.6739600559725226, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.875617027282715, "learning_rate": 5e-05, "loss": 1.009, "mean_token_accuracy": 0.8638319969177246, "num_tokens": 201253976.0, "step": 5298 }, { "epoch": 0.674087266251113, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.890690803527832, "learning_rate": 5e-05, "loss": 0.9502, "mean_token_accuracy": 0.8507739305496216, "num_tokens": 201291478.0, "step": 5299 }, { "epoch": 0.6742144765297036, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.081350326538086, "learning_rate": 5e-05, "loss": 0.9232, "mean_token_accuracy": 0.8586242198944092, "num_tokens": 201328824.0, "step": 5300 }, { "epoch": 0.6743416868082941, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.957835674285889, "learning_rate": 5e-05, "loss": 0.9387, "mean_token_accuracy": 0.8439329266548157, "num_tokens": 201363859.0, "step": 5301 }, { "epoch": 0.6744688970868846, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.559616565704346, "learning_rate": 5e-05, "loss": 0.83, "mean_token_accuracy": 0.8672939538955688, "num_tokens": 201402635.0, "step": 5302 }, { "epoch": 0.6745961073654752, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.634053707122803, "learning_rate": 5e-05, "loss": 0.8477, "mean_token_accuracy": 0.8524925708770752, "num_tokens": 201444203.0, "step": 5303 }, { "epoch": 0.6747233176440657, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.6916985511779785, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.8457179069519043, "num_tokens": 201486397.0, "step": 5304 }, { "epoch": 0.6748505279226561, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.175931930541992, "learning_rate": 5e-05, "loss": 0.7972, "mean_token_accuracy": 0.8545063734054565, "num_tokens": 201528262.0, "step": 5305 }, { "epoch": 0.6749777382012466, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.012578964233398, "learning_rate": 5e-05, "loss": 0.8709, "mean_token_accuracy": 0.8598288297653198, "num_tokens": 201566274.0, "step": 5306 }, { "epoch": 0.6751049484798372, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.8855767250061035, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8546012043952942, "num_tokens": 201610196.0, "step": 5307 }, { "epoch": 0.6752321587584277, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.20779037475586, "learning_rate": 5e-05, "loss": 0.8897, "mean_token_accuracy": 0.8551821708679199, "num_tokens": 201650144.0, "step": 5308 }, { "epoch": 0.6753593690370182, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.786965847015381, "learning_rate": 5e-05, "loss": 0.7483, "mean_token_accuracy": 0.8602026104927063, "num_tokens": 201689550.0, "step": 5309 }, { "epoch": 0.6754865793156087, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.343603134155273, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8753579258918762, "num_tokens": 201729807.0, "step": 5310 }, { "epoch": 0.6756137895941993, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.796054840087891, "learning_rate": 5e-05, "loss": 0.7323, "mean_token_accuracy": 0.8647469282150269, "num_tokens": 201759540.0, "step": 5311 }, { "epoch": 0.6757409998727897, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.394079208374023, "learning_rate": 5e-05, "loss": 0.9346, "mean_token_accuracy": 0.8548562526702881, "num_tokens": 201800067.0, "step": 5312 }, { "epoch": 0.6758682101513802, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.008307933807373, "learning_rate": 5e-05, "loss": 0.7596, "mean_token_accuracy": 0.862342357635498, "num_tokens": 201832688.0, "step": 5313 }, { "epoch": 0.6759954204299707, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.081585884094238, "learning_rate": 5e-05, "loss": 0.8966, "mean_token_accuracy": 0.8556356430053711, "num_tokens": 201876090.0, "step": 5314 }, { "epoch": 0.6761226307085613, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.389293670654297, "learning_rate": 5e-05, "loss": 0.8479, "mean_token_accuracy": 0.8430286645889282, "num_tokens": 201914339.0, "step": 5315 }, { "epoch": 0.6762498409871518, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.504013538360596, "learning_rate": 5e-05, "loss": 0.787, "mean_token_accuracy": 0.8696979284286499, "num_tokens": 201945746.0, "step": 5316 }, { "epoch": 0.6763770512657423, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.699489116668701, "learning_rate": 5e-05, "loss": 0.7771, "mean_token_accuracy": 0.8660873770713806, "num_tokens": 201978773.0, "step": 5317 }, { "epoch": 0.6765042615443327, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.075718879699707, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8582181930541992, "num_tokens": 202019838.0, "step": 5318 }, { "epoch": 0.6766314718229233, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.870189666748047, "learning_rate": 5e-05, "loss": 0.8177, "mean_token_accuracy": 0.8608930706977844, "num_tokens": 202060825.0, "step": 5319 }, { "epoch": 0.6767586821015138, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.865408420562744, "learning_rate": 5e-05, "loss": 0.6817, "mean_token_accuracy": 0.8770113587379456, "num_tokens": 202092656.0, "step": 5320 }, { "epoch": 0.6768858923801043, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.011826515197754, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.8443905711174011, "num_tokens": 202126982.0, "step": 5321 }, { "epoch": 0.6770131026586949, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.710035800933838, "learning_rate": 5e-05, "loss": 0.7148, "mean_token_accuracy": 0.8642966151237488, "num_tokens": 202157550.0, "step": 5322 }, { "epoch": 0.6771403129372854, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.18212604522705, "learning_rate": 5e-05, "loss": 0.8736, "mean_token_accuracy": 0.8467903137207031, "num_tokens": 202192168.0, "step": 5323 }, { "epoch": 0.6772675232158758, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0003490447998046875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.5929856300354, "learning_rate": 5e-05, "loss": 0.6863, "mean_token_accuracy": 0.871820867061615, "num_tokens": 202226887.0, "step": 5324 }, { "epoch": 0.6773947334944663, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0003490447998046875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.335334777832031, "learning_rate": 5e-05, "loss": 0.8889, "mean_token_accuracy": 0.8464884161949158, "num_tokens": 202269631.0, "step": 5325 }, { "epoch": 0.6775219437730569, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003490447998046875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.641297817230225, "learning_rate": 5e-05, "loss": 0.7404, "mean_token_accuracy": 0.8587439060211182, "num_tokens": 202306951.0, "step": 5326 }, { "epoch": 0.6776491540516474, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.237696647644043, "learning_rate": 5e-05, "loss": 0.8244, "mean_token_accuracy": 0.870287299156189, "num_tokens": 202342922.0, "step": 5327 }, { "epoch": 0.6777763643302379, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.920712471008301, "learning_rate": 5e-05, "loss": 0.7509, "mean_token_accuracy": 0.8583928942680359, "num_tokens": 202382125.0, "step": 5328 }, { "epoch": 0.6779035746088284, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.866613864898682, "learning_rate": 5e-05, "loss": 0.8267, "mean_token_accuracy": 0.8599497079849243, "num_tokens": 202424914.0, "step": 5329 }, { "epoch": 0.6780307848874189, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.295589923858643, "learning_rate": 5e-05, "loss": 0.7631, "mean_token_accuracy": 0.8616853356361389, "num_tokens": 202460828.0, "step": 5330 }, { "epoch": 0.6781579951660094, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.395079612731934, "learning_rate": 5e-05, "loss": 0.7549, "mean_token_accuracy": 0.8651224970817566, "num_tokens": 202492484.0, "step": 5331 }, { "epoch": 0.6782852054445999, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.584033966064453, "learning_rate": 5e-05, "loss": 0.8779, "mean_token_accuracy": 0.8313661813735962, "num_tokens": 202529079.0, "step": 5332 }, { "epoch": 0.6784124157231904, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.140364646911621, "learning_rate": 5e-05, "loss": 0.793, "mean_token_accuracy": 0.8426917791366577, "num_tokens": 202563024.0, "step": 5333 }, { "epoch": 0.678539626001781, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.6334228515625, "learning_rate": 5e-05, "loss": 0.7746, "mean_token_accuracy": 0.8614482283592224, "num_tokens": 202601264.0, "step": 5334 }, { "epoch": 0.6786668362803715, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.13514518737793, "learning_rate": 5e-05, "loss": 0.7587, "mean_token_accuracy": 0.8559398651123047, "num_tokens": 202642933.0, "step": 5335 }, { "epoch": 0.6787940465589619, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.490914821624756, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.8676932454109192, "num_tokens": 202674851.0, "step": 5336 }, { "epoch": 0.6789212568375524, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.323095798492432, "learning_rate": 5e-05, "loss": 0.7408, "mean_token_accuracy": 0.87176913022995, "num_tokens": 202711117.0, "step": 5337 }, { "epoch": 0.679048467116143, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.210650444030762, "learning_rate": 5e-05, "loss": 0.7133, "mean_token_accuracy": 0.8693371415138245, "num_tokens": 202749025.0, "step": 5338 }, { "epoch": 0.6791756773947335, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.620456695556641, "learning_rate": 5e-05, "loss": 0.818, "mean_token_accuracy": 0.8419337868690491, "num_tokens": 202790733.0, "step": 5339 }, { "epoch": 0.679302887673324, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.850336074829102, "learning_rate": 5e-05, "loss": 0.7294, "mean_token_accuracy": 0.8526819348335266, "num_tokens": 202825753.0, "step": 5340 }, { "epoch": 0.6794300979519146, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.036005020141602, "learning_rate": 5e-05, "loss": 0.7498, "mean_token_accuracy": 0.8725720643997192, "num_tokens": 202862069.0, "step": 5341 }, { "epoch": 0.679557308230505, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0002384185791015625, "grad_norm": 6.3831000328063965, "learning_rate": 5e-05, "loss": 0.6796, "mean_token_accuracy": 0.8614222407341003, "num_tokens": 202904526.0, "step": 5342 }, { "epoch": 0.6796845185090955, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.69154167175293, "learning_rate": 5e-05, "loss": 0.8991, "mean_token_accuracy": 0.8485054969787598, "num_tokens": 202942342.0, "step": 5343 }, { "epoch": 0.679811728787686, "ewc_loss": 0.25390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000217437744140625, "grad_norm": 5.8247857093811035, "learning_rate": 5e-05, "loss": 0.6725, "mean_token_accuracy": 0.853259801864624, "num_tokens": 202974644.0, "step": 5344 }, { "epoch": 0.6799389390662766, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.580552101135254, "learning_rate": 5e-05, "loss": 0.9338, "mean_token_accuracy": 0.8720343112945557, "num_tokens": 203011728.0, "step": 5345 }, { "epoch": 0.6800661493448671, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.736917018890381, "learning_rate": 5e-05, "loss": 0.7219, "mean_token_accuracy": 0.857248842716217, "num_tokens": 203051768.0, "step": 5346 }, { "epoch": 0.6801933596234576, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.144678115844727, "learning_rate": 5e-05, "loss": 1.1249, "mean_token_accuracy": 0.851526141166687, "num_tokens": 203094943.0, "step": 5347 }, { "epoch": 0.680320569902048, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.151841640472412, "learning_rate": 5e-05, "loss": 0.9202, "mean_token_accuracy": 0.862563967704773, "num_tokens": 203130587.0, "step": 5348 }, { "epoch": 0.6804477801806386, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.176114082336426, "learning_rate": 5e-05, "loss": 1.0226, "mean_token_accuracy": 0.8549416065216064, "num_tokens": 203170030.0, "step": 5349 }, { "epoch": 0.6805749904592291, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 8.687993049621582, "learning_rate": 5e-05, "loss": 1.063, "mean_token_accuracy": 0.8551504611968994, "num_tokens": 203209382.0, "step": 5350 }, { "epoch": 0.6807022007378196, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.939363479614258, "learning_rate": 5e-05, "loss": 0.7892, "mean_token_accuracy": 0.858238935470581, "num_tokens": 203247750.0, "step": 5351 }, { "epoch": 0.6808294110164101, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.291647911071777, "learning_rate": 5e-05, "loss": 1.1001, "mean_token_accuracy": 0.8640461564064026, "num_tokens": 203284443.0, "step": 5352 }, { "epoch": 0.6809566212950007, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 6.424667835235596, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.8712846040725708, "num_tokens": 203324525.0, "step": 5353 }, { "epoch": 0.6810838315735911, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 11.523877143859863, "learning_rate": 5e-05, "loss": 1.245, "mean_token_accuracy": 0.8636143207550049, "num_tokens": 203369633.0, "step": 5354 }, { "epoch": 0.6812110418521816, "ewc_loss": 0.765625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 8.766372680664062, "learning_rate": 5e-05, "loss": 1.1796, "mean_token_accuracy": 0.8584322929382324, "num_tokens": 203409549.0, "step": 5355 }, { "epoch": 0.6813382521307721, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.533020973205566, "learning_rate": 5e-05, "loss": 1.0947, "mean_token_accuracy": 0.850362241268158, "num_tokens": 203446856.0, "step": 5356 }, { "epoch": 0.6814654624093627, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 9.03369426727295, "learning_rate": 5e-05, "loss": 1.0951, "mean_token_accuracy": 0.8832409381866455, "num_tokens": 203478575.0, "step": 5357 }, { "epoch": 0.6815926726879532, "ewc_loss": 0.625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.618746757507324, "learning_rate": 5e-05, "loss": 1.0028, "mean_token_accuracy": 0.8724050521850586, "num_tokens": 203515617.0, "step": 5358 }, { "epoch": 0.6817198829665437, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.5003023147583, "learning_rate": 5e-05, "loss": 1.0296, "mean_token_accuracy": 0.8537684679031372, "num_tokens": 203548636.0, "step": 5359 }, { "epoch": 0.6818470932451343, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.977862358093262, "learning_rate": 5e-05, "loss": 0.9818, "mean_token_accuracy": 0.8679137825965881, "num_tokens": 203581358.0, "step": 5360 }, { "epoch": 0.6819743035237247, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.798701763153076, "learning_rate": 5e-05, "loss": 0.8397, "mean_token_accuracy": 0.8759637475013733, "num_tokens": 203615052.0, "step": 5361 }, { "epoch": 0.6821015138023152, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.990287780761719, "learning_rate": 5e-05, "loss": 1.0186, "mean_token_accuracy": 0.8551998734474182, "num_tokens": 203655743.0, "step": 5362 }, { "epoch": 0.6822287240809057, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.192450046539307, "learning_rate": 5e-05, "loss": 0.8619, "mean_token_accuracy": 0.8490679264068604, "num_tokens": 203689370.0, "step": 5363 }, { "epoch": 0.6823559343594963, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.412399291992188, "learning_rate": 5e-05, "loss": 1.0158, "mean_token_accuracy": 0.8576835989952087, "num_tokens": 203723594.0, "step": 5364 }, { "epoch": 0.6824831446380868, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.096010684967041, "learning_rate": 5e-05, "loss": 0.8971, "mean_token_accuracy": 0.843838095664978, "num_tokens": 203761746.0, "step": 5365 }, { "epoch": 0.6826103549166773, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.70763874053955, "learning_rate": 5e-05, "loss": 1.0536, "mean_token_accuracy": 0.8582665920257568, "num_tokens": 203801904.0, "step": 5366 }, { "epoch": 0.6827375651952677, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.540841102600098, "learning_rate": 5e-05, "loss": 0.9129, "mean_token_accuracy": 0.8620960712432861, "num_tokens": 203837775.0, "step": 5367 }, { "epoch": 0.6828647754738583, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.567497253417969, "learning_rate": 5e-05, "loss": 0.9945, "mean_token_accuracy": 0.87825608253479, "num_tokens": 203876456.0, "step": 5368 }, { "epoch": 0.6829919857524488, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.986738681793213, "learning_rate": 5e-05, "loss": 0.9915, "mean_token_accuracy": 0.8567301034927368, "num_tokens": 203917946.0, "step": 5369 }, { "epoch": 0.6831191960310393, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.984707832336426, "learning_rate": 5e-05, "loss": 0.9911, "mean_token_accuracy": 0.863440752029419, "num_tokens": 203959544.0, "step": 5370 }, { "epoch": 0.6832464063096299, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.093578338623047, "learning_rate": 5e-05, "loss": 0.9733, "mean_token_accuracy": 0.8495177030563354, "num_tokens": 204000882.0, "step": 5371 }, { "epoch": 0.6833736165882204, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.333189964294434, "learning_rate": 5e-05, "loss": 0.9244, "mean_token_accuracy": 0.8649264574050903, "num_tokens": 204037715.0, "step": 5372 }, { "epoch": 0.6835008268668108, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.832942008972168, "learning_rate": 5e-05, "loss": 0.8804, "mean_token_accuracy": 0.8563288450241089, "num_tokens": 204073145.0, "step": 5373 }, { "epoch": 0.6836280371454013, "ewc_loss": 0.5, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.189973831176758, "learning_rate": 5e-05, "loss": 0.9275, "mean_token_accuracy": 0.8573344945907593, "num_tokens": 204112781.0, "step": 5374 }, { "epoch": 0.6837552474239919, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.266976833343506, "learning_rate": 5e-05, "loss": 0.807, "mean_token_accuracy": 0.8578643798828125, "num_tokens": 204151128.0, "step": 5375 }, { "epoch": 0.6838824577025824, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.653099060058594, "learning_rate": 5e-05, "loss": 0.9357, "mean_token_accuracy": 0.8678790330886841, "num_tokens": 204188666.0, "step": 5376 }, { "epoch": 0.6840096679811729, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.859116077423096, "learning_rate": 5e-05, "loss": 0.7155, "mean_token_accuracy": 0.8818438053131104, "num_tokens": 204225410.0, "step": 5377 }, { "epoch": 0.6841368782597634, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.130486488342285, "learning_rate": 5e-05, "loss": 1.0754, "mean_token_accuracy": 0.8435108065605164, "num_tokens": 204271506.0, "step": 5378 }, { "epoch": 0.6842640885383539, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.127424240112305, "learning_rate": 5e-05, "loss": 0.8412, "mean_token_accuracy": 0.8580126762390137, "num_tokens": 204310569.0, "step": 5379 }, { "epoch": 0.6843912988169444, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.013701438903809, "learning_rate": 5e-05, "loss": 0.9952, "mean_token_accuracy": 0.8645769953727722, "num_tokens": 204349614.0, "step": 5380 }, { "epoch": 0.6845185090955349, "ewc_loss": 0.5, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.762208461761475, "learning_rate": 5e-05, "loss": 0.9552, "mean_token_accuracy": 0.8485807180404663, "num_tokens": 204381985.0, "step": 5381 }, { "epoch": 0.6846457193741254, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.207932472229004, "learning_rate": 5e-05, "loss": 0.9315, "mean_token_accuracy": 0.8547618985176086, "num_tokens": 204418433.0, "step": 5382 }, { "epoch": 0.684772929652716, "ewc_loss": 0.5, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.06878662109375, "learning_rate": 5e-05, "loss": 0.9047, "mean_token_accuracy": 0.860902726650238, "num_tokens": 204457544.0, "step": 5383 }, { "epoch": 0.6849001399313065, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.37977409362793, "learning_rate": 5e-05, "loss": 0.8266, "mean_token_accuracy": 0.8619682788848877, "num_tokens": 204500626.0, "step": 5384 }, { "epoch": 0.6850273502098969, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.041810989379883, "learning_rate": 5e-05, "loss": 0.9023, "mean_token_accuracy": 0.8472232222557068, "num_tokens": 204540861.0, "step": 5385 }, { "epoch": 0.6851545604884874, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.255985736846924, "learning_rate": 5e-05, "loss": 0.766, "mean_token_accuracy": 0.8711974620819092, "num_tokens": 204581555.0, "step": 5386 }, { "epoch": 0.685281770767078, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.637552738189697, "learning_rate": 5e-05, "loss": 0.7893, "mean_token_accuracy": 0.8630305528640747, "num_tokens": 204614931.0, "step": 5387 }, { "epoch": 0.6854089810456685, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.572577476501465, "learning_rate": 5e-05, "loss": 0.8102, "mean_token_accuracy": 0.8583840131759644, "num_tokens": 204649678.0, "step": 5388 }, { "epoch": 0.685536191324259, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.1967949867248535, "learning_rate": 5e-05, "loss": 0.7826, "mean_token_accuracy": 0.8541039824485779, "num_tokens": 204690617.0, "step": 5389 }, { "epoch": 0.6856634016028496, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.834824562072754, "learning_rate": 5e-05, "loss": 0.8157, "mean_token_accuracy": 0.8547468185424805, "num_tokens": 204725685.0, "step": 5390 }, { "epoch": 0.68579061188144, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.952418804168701, "learning_rate": 5e-05, "loss": 0.8102, "mean_token_accuracy": 0.8395071029663086, "num_tokens": 204768265.0, "step": 5391 }, { "epoch": 0.6859178221600305, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.017972946166992, "learning_rate": 5e-05, "loss": 0.809, "mean_token_accuracy": 0.8621400594711304, "num_tokens": 204806211.0, "step": 5392 }, { "epoch": 0.686045032438621, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.7452216148376465, "learning_rate": 5e-05, "loss": 0.739, "mean_token_accuracy": 0.8611127138137817, "num_tokens": 204844982.0, "step": 5393 }, { "epoch": 0.6861722427172116, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.294147491455078, "learning_rate": 5e-05, "loss": 0.8117, "mean_token_accuracy": 0.8735803365707397, "num_tokens": 204881676.0, "step": 5394 }, { "epoch": 0.6862994529958021, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.577861309051514, "learning_rate": 5e-05, "loss": 0.7312, "mean_token_accuracy": 0.8583310842514038, "num_tokens": 204920316.0, "step": 5395 }, { "epoch": 0.6864266632743926, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.486741065979004, "learning_rate": 5e-05, "loss": 0.9095, "mean_token_accuracy": 0.8491243124008179, "num_tokens": 204960269.0, "step": 5396 }, { "epoch": 0.686553873552983, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.615055084228516, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8587082028388977, "num_tokens": 204993238.0, "step": 5397 }, { "epoch": 0.6866810838315736, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.475579261779785, "learning_rate": 5e-05, "loss": 0.9332, "mean_token_accuracy": 0.8509188890457153, "num_tokens": 205029550.0, "step": 5398 }, { "epoch": 0.6868082941101641, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.841693878173828, "learning_rate": 5e-05, "loss": 0.7823, "mean_token_accuracy": 0.8536120057106018, "num_tokens": 205073198.0, "step": 5399 }, { "epoch": 0.6869355043887546, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.258772850036621, "learning_rate": 5e-05, "loss": 0.8706, "mean_token_accuracy": 0.8651481866836548, "num_tokens": 205107581.0, "step": 5400 }, { "epoch": 0.6870627146673451, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.1698079109191895, "learning_rate": 5e-05, "loss": 0.7909, "mean_token_accuracy": 0.8594866991043091, "num_tokens": 205144030.0, "step": 5401 }, { "epoch": 0.6871899249459357, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.794434070587158, "learning_rate": 5e-05, "loss": 0.8538, "mean_token_accuracy": 0.8560497760772705, "num_tokens": 205180583.0, "step": 5402 }, { "epoch": 0.6873171352245261, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.511248588562012, "learning_rate": 5e-05, "loss": 0.812, "mean_token_accuracy": 0.8586819171905518, "num_tokens": 205215889.0, "step": 5403 }, { "epoch": 0.6874443455031166, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.337767124176025, "learning_rate": 5e-05, "loss": 0.8308, "mean_token_accuracy": 0.8414024114608765, "num_tokens": 205254725.0, "step": 5404 }, { "epoch": 0.6875715557817071, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.70739221572876, "learning_rate": 5e-05, "loss": 0.7738, "mean_token_accuracy": 0.8677361011505127, "num_tokens": 205289533.0, "step": 5405 }, { "epoch": 0.6876987660602977, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.108071804046631, "learning_rate": 5e-05, "loss": 0.8346, "mean_token_accuracy": 0.8380139470100403, "num_tokens": 205330462.0, "step": 5406 }, { "epoch": 0.6878259763388882, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.786148548126221, "learning_rate": 5e-05, "loss": 0.8532, "mean_token_accuracy": 0.8451561331748962, "num_tokens": 205371462.0, "step": 5407 }, { "epoch": 0.6879531866174787, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.972638130187988, "learning_rate": 5e-05, "loss": 0.6643, "mean_token_accuracy": 0.8830796480178833, "num_tokens": 205404629.0, "step": 5408 }, { "epoch": 0.6880803968960693, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.92238712310791, "learning_rate": 5e-05, "loss": 0.8434, "mean_token_accuracy": 0.8493635058403015, "num_tokens": 205440600.0, "step": 5409 }, { "epoch": 0.6882076071746597, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.756059646606445, "learning_rate": 5e-05, "loss": 0.7193, "mean_token_accuracy": 0.8614683151245117, "num_tokens": 205473696.0, "step": 5410 }, { "epoch": 0.6883348174532502, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.189671516418457, "learning_rate": 5e-05, "loss": 0.8778, "mean_token_accuracy": 0.8486742973327637, "num_tokens": 205519889.0, "step": 5411 }, { "epoch": 0.6884620277318407, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.493259906768799, "learning_rate": 5e-05, "loss": 0.668, "mean_token_accuracy": 0.874845564365387, "num_tokens": 205552303.0, "step": 5412 }, { "epoch": 0.6885892380104313, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.568089485168457, "learning_rate": 5e-05, "loss": 0.9079, "mean_token_accuracy": 0.845706045627594, "num_tokens": 205588918.0, "step": 5413 }, { "epoch": 0.6887164482890218, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.3198771476745605, "learning_rate": 5e-05, "loss": 0.7377, "mean_token_accuracy": 0.8591791391372681, "num_tokens": 205624634.0, "step": 5414 }, { "epoch": 0.6888436585676123, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.823273658752441, "learning_rate": 5e-05, "loss": 0.9054, "mean_token_accuracy": 0.8649052977561951, "num_tokens": 205668700.0, "step": 5415 }, { "epoch": 0.6889708688462027, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.5068888664245605, "learning_rate": 5e-05, "loss": 0.7603, "mean_token_accuracy": 0.8550417423248291, "num_tokens": 205711051.0, "step": 5416 }, { "epoch": 0.6890980791247933, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.607190132141113, "learning_rate": 5e-05, "loss": 0.9204, "mean_token_accuracy": 0.8631242513656616, "num_tokens": 205750349.0, "step": 5417 }, { "epoch": 0.6892252894033838, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.131449222564697, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8720719814300537, "num_tokens": 205789525.0, "step": 5418 }, { "epoch": 0.6893524996819743, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.828030109405518, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.8616693019866943, "num_tokens": 205827266.0, "step": 5419 }, { "epoch": 0.6894797099605648, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.777373790740967, "learning_rate": 5e-05, "loss": 0.8358, "mean_token_accuracy": 0.8665664792060852, "num_tokens": 205866170.0, "step": 5420 }, { "epoch": 0.6896069202391554, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.04571008682251, "learning_rate": 5e-05, "loss": 0.7354, "mean_token_accuracy": 0.8675328493118286, "num_tokens": 205905613.0, "step": 5421 }, { "epoch": 0.6897341305177458, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.083850860595703, "learning_rate": 5e-05, "loss": 0.9083, "mean_token_accuracy": 0.8384582996368408, "num_tokens": 205943142.0, "step": 5422 }, { "epoch": 0.6898613407963363, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.885054111480713, "learning_rate": 5e-05, "loss": 0.7365, "mean_token_accuracy": 0.8587136268615723, "num_tokens": 205983321.0, "step": 5423 }, { "epoch": 0.6899885510749268, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.8815507888793945, "learning_rate": 5e-05, "loss": 0.8573, "mean_token_accuracy": 0.848984956741333, "num_tokens": 206019013.0, "step": 5424 }, { "epoch": 0.6901157613535174, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.21116304397583, "learning_rate": 5e-05, "loss": 0.7612, "mean_token_accuracy": 0.8596486449241638, "num_tokens": 206059958.0, "step": 5425 }, { "epoch": 0.6902429716321079, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.34816312789917, "learning_rate": 5e-05, "loss": 0.7603, "mean_token_accuracy": 0.8629193305969238, "num_tokens": 206096536.0, "step": 5426 }, { "epoch": 0.6903701819106984, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.7188262939453125, "learning_rate": 5e-05, "loss": 0.817, "mean_token_accuracy": 0.854347288608551, "num_tokens": 206138399.0, "step": 5427 }, { "epoch": 0.6904973921892888, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.788926601409912, "learning_rate": 5e-05, "loss": 0.6918, "mean_token_accuracy": 0.8734680414199829, "num_tokens": 206171666.0, "step": 5428 }, { "epoch": 0.6906246024678794, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.235398292541504, "learning_rate": 5e-05, "loss": 0.8098, "mean_token_accuracy": 0.866273045539856, "num_tokens": 206212439.0, "step": 5429 }, { "epoch": 0.6907518127464699, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.344823360443115, "learning_rate": 5e-05, "loss": 0.7064, "mean_token_accuracy": 0.8601973056793213, "num_tokens": 206245213.0, "step": 5430 }, { "epoch": 0.6908790230250604, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.78450870513916, "learning_rate": 5e-05, "loss": 0.8814, "mean_token_accuracy": 0.8605150580406189, "num_tokens": 206283875.0, "step": 5431 }, { "epoch": 0.691006233303651, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.049348831176758, "learning_rate": 5e-05, "loss": 0.7216, "mean_token_accuracy": 0.8537486791610718, "num_tokens": 206319442.0, "step": 5432 }, { "epoch": 0.6911334435822415, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.238365173339844, "learning_rate": 5e-05, "loss": 1.0161, "mean_token_accuracy": 0.8426061272621155, "num_tokens": 206358210.0, "step": 5433 }, { "epoch": 0.6912606538608319, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.296599864959717, "learning_rate": 5e-05, "loss": 0.7077, "mean_token_accuracy": 0.8708631992340088, "num_tokens": 206390900.0, "step": 5434 }, { "epoch": 0.6913878641394224, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.182665824890137, "learning_rate": 5e-05, "loss": 1.0623, "mean_token_accuracy": 0.8410050868988037, "num_tokens": 206426296.0, "step": 5435 }, { "epoch": 0.691515074418013, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.203147888183594, "learning_rate": 5e-05, "loss": 0.8717, "mean_token_accuracy": 0.8609636425971985, "num_tokens": 206457010.0, "step": 5436 }, { "epoch": 0.6916422846966035, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.199987411499023, "learning_rate": 5e-05, "loss": 0.9086, "mean_token_accuracy": 0.8660116791725159, "num_tokens": 206496324.0, "step": 5437 }, { "epoch": 0.691769494975194, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.025992393493652, "learning_rate": 5e-05, "loss": 0.9705, "mean_token_accuracy": 0.85300612449646, "num_tokens": 206530507.0, "step": 5438 }, { "epoch": 0.6918967052537845, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.161336421966553, "learning_rate": 5e-05, "loss": 0.7973, "mean_token_accuracy": 0.8644598126411438, "num_tokens": 206565123.0, "step": 5439 }, { "epoch": 0.692023915532375, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.259511947631836, "learning_rate": 5e-05, "loss": 0.8859, "mean_token_accuracy": 0.8633608222007751, "num_tokens": 206602026.0, "step": 5440 }, { "epoch": 0.6921511258109655, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.056127071380615, "learning_rate": 5e-05, "loss": 0.7998, "mean_token_accuracy": 0.8545913100242615, "num_tokens": 206643371.0, "step": 5441 }, { "epoch": 0.692278336089556, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.731118202209473, "learning_rate": 5e-05, "loss": 0.8059, "mean_token_accuracy": 0.8716260194778442, "num_tokens": 206675822.0, "step": 5442 }, { "epoch": 0.6924055463681466, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.56502628326416, "learning_rate": 5e-05, "loss": 0.7048, "mean_token_accuracy": 0.895822286605835, "num_tokens": 206703973.0, "step": 5443 }, { "epoch": 0.6925327566467371, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.094667434692383, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.8547821640968323, "num_tokens": 206740645.0, "step": 5444 }, { "epoch": 0.6926599669253276, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.94035530090332, "learning_rate": 5e-05, "loss": 0.7998, "mean_token_accuracy": 0.8695052266120911, "num_tokens": 206785620.0, "step": 5445 }, { "epoch": 0.692787177203918, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.874884128570557, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.8516243696212769, "num_tokens": 206818761.0, "step": 5446 }, { "epoch": 0.6929143874825086, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.928264617919922, "learning_rate": 5e-05, "loss": 0.7966, "mean_token_accuracy": 0.8688596487045288, "num_tokens": 206857523.0, "step": 5447 }, { "epoch": 0.6930415977610991, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.96438455581665, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8534098267555237, "num_tokens": 206893389.0, "step": 5448 }, { "epoch": 0.6931688080396896, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.763848781585693, "learning_rate": 5e-05, "loss": 0.8275, "mean_token_accuracy": 0.8542072772979736, "num_tokens": 206924997.0, "step": 5449 }, { "epoch": 0.6932960183182801, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.073604583740234, "learning_rate": 5e-05, "loss": 0.7572, "mean_token_accuracy": 0.8577167391777039, "num_tokens": 206957032.0, "step": 5450 }, { "epoch": 0.6934232285968707, "ewc_loss": 0.375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.665755748748779, "learning_rate": 5e-05, "loss": 0.7835, "mean_token_accuracy": 0.8639078736305237, "num_tokens": 206995407.0, "step": 5451 }, { "epoch": 0.6935504388754611, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.0433478355407715, "learning_rate": 5e-05, "loss": 0.7398, "mean_token_accuracy": 0.8627287149429321, "num_tokens": 207031563.0, "step": 5452 }, { "epoch": 0.6936776491540516, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.729836940765381, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.8644000291824341, "num_tokens": 207064641.0, "step": 5453 }, { "epoch": 0.6938048594326421, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.896876335144043, "learning_rate": 5e-05, "loss": 0.8031, "mean_token_accuracy": 0.8422017097473145, "num_tokens": 207099474.0, "step": 5454 }, { "epoch": 0.6939320697112327, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.963429927825928, "learning_rate": 5e-05, "loss": 0.7691, "mean_token_accuracy": 0.870085597038269, "num_tokens": 207139250.0, "step": 5455 }, { "epoch": 0.6940592799898232, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.5943284034729, "learning_rate": 5e-05, "loss": 0.7147, "mean_token_accuracy": 0.8617361187934875, "num_tokens": 207179523.0, "step": 5456 }, { "epoch": 0.6941864902684137, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.376490592956543, "learning_rate": 5e-05, "loss": 0.813, "mean_token_accuracy": 0.8661212921142578, "num_tokens": 207218143.0, "step": 5457 }, { "epoch": 0.6943137005470043, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.214004993438721, "learning_rate": 5e-05, "loss": 0.6988, "mean_token_accuracy": 0.8614311218261719, "num_tokens": 207256946.0, "step": 5458 }, { "epoch": 0.6944409108255947, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.939948081970215, "learning_rate": 5e-05, "loss": 0.8726, "mean_token_accuracy": 0.8727788925170898, "num_tokens": 207296793.0, "step": 5459 }, { "epoch": 0.6945681211041852, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 5.997866630554199, "learning_rate": 5e-05, "loss": 0.746, "mean_token_accuracy": 0.8480685353279114, "num_tokens": 207333602.0, "step": 5460 }, { "epoch": 0.6946953313827757, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.366792678833008, "learning_rate": 5e-05, "loss": 1.0229, "mean_token_accuracy": 0.8491560816764832, "num_tokens": 207375806.0, "step": 5461 }, { "epoch": 0.6948225416613663, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.503243923187256, "learning_rate": 5e-05, "loss": 0.7375, "mean_token_accuracy": 0.8811018466949463, "num_tokens": 207407856.0, "step": 5462 }, { "epoch": 0.6949497519399568, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.957595825195312, "learning_rate": 5e-05, "loss": 1.0364, "mean_token_accuracy": 0.8444712162017822, "num_tokens": 207447000.0, "step": 5463 }, { "epoch": 0.6950769622185473, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.65427303314209, "learning_rate": 5e-05, "loss": 0.9222, "mean_token_accuracy": 0.8685796856880188, "num_tokens": 207479860.0, "step": 5464 }, { "epoch": 0.6952041724971377, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.433965682983398, "learning_rate": 5e-05, "loss": 0.7839, "mean_token_accuracy": 0.8734776973724365, "num_tokens": 207515454.0, "step": 5465 }, { "epoch": 0.6953313827757283, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.822420120239258, "learning_rate": 5e-05, "loss": 0.9982, "mean_token_accuracy": 0.8547338247299194, "num_tokens": 207559259.0, "step": 5466 }, { "epoch": 0.6954585930543188, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.079928398132324, "learning_rate": 5e-05, "loss": 0.744, "mean_token_accuracy": 0.8511386513710022, "num_tokens": 207595909.0, "step": 5467 }, { "epoch": 0.6955858033329093, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.985960006713867, "learning_rate": 5e-05, "loss": 1.1414, "mean_token_accuracy": 0.8344905376434326, "num_tokens": 207636063.0, "step": 5468 }, { "epoch": 0.6957130136114998, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.274165630340576, "learning_rate": 5e-05, "loss": 0.7814, "mean_token_accuracy": 0.8631681799888611, "num_tokens": 207675990.0, "step": 5469 }, { "epoch": 0.6958402238900904, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 10.443666458129883, "learning_rate": 5e-05, "loss": 1.1464, "mean_token_accuracy": 0.8615222573280334, "num_tokens": 207708996.0, "step": 5470 }, { "epoch": 0.6959674341686808, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 8.134712219238281, "learning_rate": 5e-05, "loss": 1.1484, "mean_token_accuracy": 0.8420209884643555, "num_tokens": 207750433.0, "step": 5471 }, { "epoch": 0.6960946444472713, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.127455711364746, "learning_rate": 5e-05, "loss": 1.0243, "mean_token_accuracy": 0.8530006408691406, "num_tokens": 207792479.0, "step": 5472 }, { "epoch": 0.6962218547258618, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 8.97327709197998, "learning_rate": 5e-05, "loss": 1.1026, "mean_token_accuracy": 0.8618899583816528, "num_tokens": 207833420.0, "step": 5473 }, { "epoch": 0.6963490650044524, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.519084453582764, "learning_rate": 5e-05, "loss": 0.8404, "mean_token_accuracy": 0.871009886264801, "num_tokens": 207875933.0, "step": 5474 }, { "epoch": 0.6964762752830429, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.388931274414062, "learning_rate": 5e-05, "loss": 1.0456, "mean_token_accuracy": 0.8579353094100952, "num_tokens": 207915481.0, "step": 5475 }, { "epoch": 0.6966034855616334, "ewc_loss": 0.5, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.34169864654541, "learning_rate": 5e-05, "loss": 0.9271, "mean_token_accuracy": 0.8564621210098267, "num_tokens": 207952024.0, "step": 5476 }, { "epoch": 0.6967306958402238, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.850318908691406, "learning_rate": 5e-05, "loss": 1.0157, "mean_token_accuracy": 0.8514649868011475, "num_tokens": 207990938.0, "step": 5477 }, { "epoch": 0.6968579061188144, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 7.9465012550354, "learning_rate": 5e-05, "loss": 0.982, "mean_token_accuracy": 0.8585565090179443, "num_tokens": 208032492.0, "step": 5478 }, { "epoch": 0.6969851163974049, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.6111626625061035, "learning_rate": 5e-05, "loss": 0.8614, "mean_token_accuracy": 0.8497231006622314, "num_tokens": 208068817.0, "step": 5479 }, { "epoch": 0.6971123266759954, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.537126541137695, "learning_rate": 5e-05, "loss": 0.964, "mean_token_accuracy": 0.8630871772766113, "num_tokens": 208105090.0, "step": 5480 }, { "epoch": 0.697239536954586, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.893603324890137, "learning_rate": 5e-05, "loss": 0.8012, "mean_token_accuracy": 0.8559727072715759, "num_tokens": 208136653.0, "step": 5481 }, { "epoch": 0.6973667472331765, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.917630195617676, "learning_rate": 5e-05, "loss": 0.9668, "mean_token_accuracy": 0.8613329529762268, "num_tokens": 208175556.0, "step": 5482 }, { "epoch": 0.6974939575117669, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.849527835845947, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.8670686483383179, "num_tokens": 208214549.0, "step": 5483 }, { "epoch": 0.6976211677903574, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.021326065063477, "learning_rate": 5e-05, "loss": 0.9181, "mean_token_accuracy": 0.8769795298576355, "num_tokens": 208242582.0, "step": 5484 }, { "epoch": 0.697748378068948, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.1542582511901855, "learning_rate": 5e-05, "loss": 0.9159, "mean_token_accuracy": 0.8547681570053101, "num_tokens": 208281639.0, "step": 5485 }, { "epoch": 0.6978755883475385, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.702598571777344, "learning_rate": 5e-05, "loss": 0.8855, "mean_token_accuracy": 0.8761592507362366, "num_tokens": 208323543.0, "step": 5486 }, { "epoch": 0.698002798626129, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.388131618499756, "learning_rate": 5e-05, "loss": 0.9292, "mean_token_accuracy": 0.8569635152816772, "num_tokens": 208359881.0, "step": 5487 }, { "epoch": 0.6981300089047195, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.268901824951172, "learning_rate": 5e-05, "loss": 0.9262, "mean_token_accuracy": 0.8496760725975037, "num_tokens": 208404541.0, "step": 5488 }, { "epoch": 0.69825721918331, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.331538200378418, "learning_rate": 5e-05, "loss": 0.8515, "mean_token_accuracy": 0.86342453956604, "num_tokens": 208437731.0, "step": 5489 }, { "epoch": 0.6983844294619005, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.146037101745605, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.8588917255401611, "num_tokens": 208475171.0, "step": 5490 }, { "epoch": 0.698511639740491, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.0240631103515625, "learning_rate": 5e-05, "loss": 0.7882, "mean_token_accuracy": 0.8593283891677856, "num_tokens": 208511410.0, "step": 5491 }, { "epoch": 0.6986388500190815, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.33266830444336, "learning_rate": 5e-05, "loss": 0.895, "mean_token_accuracy": 0.8644319772720337, "num_tokens": 208551133.0, "step": 5492 }, { "epoch": 0.6987660602976721, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.730962753295898, "learning_rate": 5e-05, "loss": 0.7549, "mean_token_accuracy": 0.863093376159668, "num_tokens": 208593394.0, "step": 5493 }, { "epoch": 0.6988932705762626, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.603645324707031, "learning_rate": 5e-05, "loss": 0.9609, "mean_token_accuracy": 0.8485251665115356, "num_tokens": 208627160.0, "step": 5494 }, { "epoch": 0.699020480854853, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.626784324645996, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.8581066131591797, "num_tokens": 208664896.0, "step": 5495 }, { "epoch": 0.6991476911334435, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.805355072021484, "learning_rate": 5e-05, "loss": 0.9737, "mean_token_accuracy": 0.8549368381500244, "num_tokens": 208702118.0, "step": 5496 }, { "epoch": 0.6992749014120341, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 6.7888994216918945, "learning_rate": 5e-05, "loss": 0.8377, "mean_token_accuracy": 0.8528492450714111, "num_tokens": 208741243.0, "step": 5497 }, { "epoch": 0.6994021116906246, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.734519004821777, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.8697658777236938, "num_tokens": 208778914.0, "step": 5498 }, { "epoch": 0.6995293219692151, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.2299394607543945, "learning_rate": 5e-05, "loss": 0.8842, "mean_token_accuracy": 0.858381986618042, "num_tokens": 208815683.0, "step": 5499 }, { "epoch": 0.6996565322478057, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.180737495422363, "learning_rate": 5e-05, "loss": 0.9198, "mean_token_accuracy": 0.8613385558128357, "num_tokens": 208853812.0, "step": 5500 }, { "epoch": 0.6997837425263961, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.647130966186523, "learning_rate": 5e-05, "loss": 0.8608, "mean_token_accuracy": 0.8702236413955688, "num_tokens": 208894326.0, "step": 5501 }, { "epoch": 0.6999109528049866, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.499935626983643, "learning_rate": 5e-05, "loss": 0.8007, "mean_token_accuracy": 0.8688106536865234, "num_tokens": 208929644.0, "step": 5502 }, { "epoch": 0.7000381630835771, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.855690956115723, "learning_rate": 5e-05, "loss": 0.9015, "mean_token_accuracy": 0.856117844581604, "num_tokens": 208972836.0, "step": 5503 }, { "epoch": 0.7001653733621677, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.185571193695068, "learning_rate": 5e-05, "loss": 0.7885, "mean_token_accuracy": 0.8613712787628174, "num_tokens": 209016255.0, "step": 5504 }, { "epoch": 0.7002925836407582, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.83135986328125, "learning_rate": 5e-05, "loss": 0.8791, "mean_token_accuracy": 0.8483225107192993, "num_tokens": 209054576.0, "step": 5505 }, { "epoch": 0.7004197939193487, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.12325382232666, "learning_rate": 5e-05, "loss": 0.7316, "mean_token_accuracy": 0.8683516979217529, "num_tokens": 209097881.0, "step": 5506 }, { "epoch": 0.7005470041979391, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.790012836456299, "learning_rate": 5e-05, "loss": 0.852, "mean_token_accuracy": 0.8520052433013916, "num_tokens": 209137354.0, "step": 5507 }, { "epoch": 0.7006742144765297, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.0631513595581055, "learning_rate": 5e-05, "loss": 0.762, "mean_token_accuracy": 0.8539796471595764, "num_tokens": 209170505.0, "step": 5508 }, { "epoch": 0.7008014247551202, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.802926540374756, "learning_rate": 5e-05, "loss": 0.8199, "mean_token_accuracy": 0.852432370185852, "num_tokens": 209207745.0, "step": 5509 }, { "epoch": 0.7009286350337107, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.00991153717041, "learning_rate": 5e-05, "loss": 0.7675, "mean_token_accuracy": 0.855310320854187, "num_tokens": 209246674.0, "step": 5510 }, { "epoch": 0.7010558453123013, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.792601108551025, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.8613811731338501, "num_tokens": 209289019.0, "step": 5511 }, { "epoch": 0.7011830555908918, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.003636837005615, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.8588502407073975, "num_tokens": 209323053.0, "step": 5512 }, { "epoch": 0.7013102658694823, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.795597076416016, "learning_rate": 5e-05, "loss": 0.8022, "mean_token_accuracy": 0.8599229454994202, "num_tokens": 209358810.0, "step": 5513 }, { "epoch": 0.7014374761480727, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.952746868133545, "learning_rate": 5e-05, "loss": 0.7522, "mean_token_accuracy": 0.8621010184288025, "num_tokens": 209401868.0, "step": 5514 }, { "epoch": 0.7015646864266633, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.886301040649414, "learning_rate": 5e-05, "loss": 0.8247, "mean_token_accuracy": 0.8498106002807617, "num_tokens": 209435212.0, "step": 5515 }, { "epoch": 0.7016918967052538, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.808614253997803, "learning_rate": 5e-05, "loss": 0.7112, "mean_token_accuracy": 0.8671817183494568, "num_tokens": 209472357.0, "step": 5516 }, { "epoch": 0.7018191069838443, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.084959983825684, "learning_rate": 5e-05, "loss": 0.8602, "mean_token_accuracy": 0.8440256714820862, "num_tokens": 209512687.0, "step": 5517 }, { "epoch": 0.7019463172624348, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.597871780395508, "learning_rate": 5e-05, "loss": 0.7051, "mean_token_accuracy": 0.8644375801086426, "num_tokens": 209547223.0, "step": 5518 }, { "epoch": 0.7020735275410254, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.362895965576172, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.861762523651123, "num_tokens": 209583299.0, "step": 5519 }, { "epoch": 0.7022007378196158, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.4151082038879395, "learning_rate": 5e-05, "loss": 0.7119, "mean_token_accuracy": 0.8607348799705505, "num_tokens": 209618792.0, "step": 5520 }, { "epoch": 0.7023279480982063, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.655879974365234, "learning_rate": 5e-05, "loss": 0.9055, "mean_token_accuracy": 0.8532388806343079, "num_tokens": 209654324.0, "step": 5521 }, { "epoch": 0.7024551583767968, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.400331497192383, "learning_rate": 5e-05, "loss": 0.6841, "mean_token_accuracy": 0.8744546175003052, "num_tokens": 209684963.0, "step": 5522 }, { "epoch": 0.7025823686553874, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.713786125183105, "learning_rate": 5e-05, "loss": 0.8729, "mean_token_accuracy": 0.8746383786201477, "num_tokens": 209724197.0, "step": 5523 }, { "epoch": 0.7027095789339779, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.7678399085998535, "learning_rate": 5e-05, "loss": 0.7474, "mean_token_accuracy": 0.8669080138206482, "num_tokens": 209759397.0, "step": 5524 }, { "epoch": 0.7028367892125684, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.23266315460205, "learning_rate": 5e-05, "loss": 0.8938, "mean_token_accuracy": 0.8626383543014526, "num_tokens": 209802504.0, "step": 5525 }, { "epoch": 0.7029639994911588, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.417229652404785, "learning_rate": 5e-05, "loss": 0.8517, "mean_token_accuracy": 0.8631037473678589, "num_tokens": 209838798.0, "step": 5526 }, { "epoch": 0.7030912097697494, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.357429504394531, "learning_rate": 5e-05, "loss": 0.8176, "mean_token_accuracy": 0.8520700931549072, "num_tokens": 209880935.0, "step": 5527 }, { "epoch": 0.7032184200483399, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.0359468460083, "learning_rate": 5e-05, "loss": 0.8747, "mean_token_accuracy": 0.8554193377494812, "num_tokens": 209921514.0, "step": 5528 }, { "epoch": 0.7033456303269304, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.662099361419678, "learning_rate": 5e-05, "loss": 0.7057, "mean_token_accuracy": 0.8670173287391663, "num_tokens": 209960082.0, "step": 5529 }, { "epoch": 0.703472840605521, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.471233367919922, "learning_rate": 5e-05, "loss": 0.8705, "mean_token_accuracy": 0.8609326481819153, "num_tokens": 209996338.0, "step": 5530 }, { "epoch": 0.7036000508841115, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.384101390838623, "learning_rate": 5e-05, "loss": 0.7607, "mean_token_accuracy": 0.8501051664352417, "num_tokens": 210039140.0, "step": 5531 }, { "epoch": 0.7037272611627019, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.659443855285645, "learning_rate": 5e-05, "loss": 0.8806, "mean_token_accuracy": 0.863361656665802, "num_tokens": 210074166.0, "step": 5532 }, { "epoch": 0.7038544714412924, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.469415187835693, "learning_rate": 5e-05, "loss": 0.7032, "mean_token_accuracy": 0.8725537061691284, "num_tokens": 210110295.0, "step": 5533 }, { "epoch": 0.703981681719883, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.593935012817383, "learning_rate": 5e-05, "loss": 0.9587, "mean_token_accuracy": 0.8520650863647461, "num_tokens": 210151994.0, "step": 5534 }, { "epoch": 0.7041088919984735, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.8528666496276855, "learning_rate": 5e-05, "loss": 0.7785, "mean_token_accuracy": 0.8645201921463013, "num_tokens": 210184769.0, "step": 5535 }, { "epoch": 0.704236102277064, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.147640228271484, "learning_rate": 5e-05, "loss": 0.9149, "mean_token_accuracy": 0.8554804921150208, "num_tokens": 210225561.0, "step": 5536 }, { "epoch": 0.7043633125556545, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.331329822540283, "learning_rate": 5e-05, "loss": 0.8001, "mean_token_accuracy": 0.872917890548706, "num_tokens": 210265544.0, "step": 5537 }, { "epoch": 0.704490522834245, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.574856758117676, "learning_rate": 5e-05, "loss": 0.7769, "mean_token_accuracy": 0.8714328408241272, "num_tokens": 210298178.0, "step": 5538 }, { "epoch": 0.7046177331128355, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.608460903167725, "learning_rate": 5e-05, "loss": 0.863, "mean_token_accuracy": 0.8534320592880249, "num_tokens": 210342391.0, "step": 5539 }, { "epoch": 0.704744943391426, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.3076395988464355, "learning_rate": 5e-05, "loss": 0.7757, "mean_token_accuracy": 0.8653810024261475, "num_tokens": 210386673.0, "step": 5540 }, { "epoch": 0.7048721536700165, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.5269775390625, "learning_rate": 5e-05, "loss": 0.8346, "mean_token_accuracy": 0.8508540391921997, "num_tokens": 210424050.0, "step": 5541 }, { "epoch": 0.7049993639486071, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.435041904449463, "learning_rate": 5e-05, "loss": 0.7495, "mean_token_accuracy": 0.8682018518447876, "num_tokens": 210462593.0, "step": 5542 }, { "epoch": 0.7051265742271976, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.2077226638793945, "learning_rate": 5e-05, "loss": 0.7582, "mean_token_accuracy": 0.8611275553703308, "num_tokens": 210496534.0, "step": 5543 }, { "epoch": 0.705253784505788, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.720818519592285, "learning_rate": 5e-05, "loss": 0.8117, "mean_token_accuracy": 0.8550136685371399, "num_tokens": 210534945.0, "step": 5544 }, { "epoch": 0.7053809947843785, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.8492608070373535, "learning_rate": 5e-05, "loss": 0.7052, "mean_token_accuracy": 0.8717794418334961, "num_tokens": 210576742.0, "step": 5545 }, { "epoch": 0.7055082050629691, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.08078670501709, "learning_rate": 5e-05, "loss": 0.8744, "mean_token_accuracy": 0.842993974685669, "num_tokens": 210616248.0, "step": 5546 }, { "epoch": 0.7056354153415596, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.486854076385498, "learning_rate": 5e-05, "loss": 0.7465, "mean_token_accuracy": 0.8552731275558472, "num_tokens": 210658229.0, "step": 5547 }, { "epoch": 0.7057626256201501, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.540635108947754, "learning_rate": 5e-05, "loss": 0.8205, "mean_token_accuracy": 0.873440146446228, "num_tokens": 210696312.0, "step": 5548 }, { "epoch": 0.7058898358987407, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.15305233001709, "learning_rate": 5e-05, "loss": 0.7232, "mean_token_accuracy": 0.85606849193573, "num_tokens": 210736916.0, "step": 5549 }, { "epoch": 0.7060170461773311, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.052682876586914, "learning_rate": 5e-05, "loss": 1.0083, "mean_token_accuracy": 0.8419833183288574, "num_tokens": 210777779.0, "step": 5550 }, { "epoch": 0.7061442564559216, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.163386821746826, "learning_rate": 5e-05, "loss": 0.8035, "mean_token_accuracy": 0.8401846289634705, "num_tokens": 210823134.0, "step": 5551 }, { "epoch": 0.7062714667345121, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.23185920715332, "learning_rate": 5e-05, "loss": 1.0076, "mean_token_accuracy": 0.8574156761169434, "num_tokens": 210862792.0, "step": 5552 }, { "epoch": 0.7063986770131027, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 6.902549743652344, "learning_rate": 5e-05, "loss": 0.8358, "mean_token_accuracy": 0.8652832508087158, "num_tokens": 210905411.0, "step": 5553 }, { "epoch": 0.7065258872916932, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.501280784606934, "learning_rate": 5e-05, "loss": 0.9831, "mean_token_accuracy": 0.8504664897918701, "num_tokens": 210944090.0, "step": 5554 }, { "epoch": 0.7066530975702837, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.912574291229248, "learning_rate": 5e-05, "loss": 0.9315, "mean_token_accuracy": 0.8569021224975586, "num_tokens": 210977372.0, "step": 5555 }, { "epoch": 0.7067803078488741, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.112575531005859, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8511897325515747, "num_tokens": 211021800.0, "step": 5556 }, { "epoch": 0.7069075181274647, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.716014862060547, "learning_rate": 5e-05, "loss": 0.9763, "mean_token_accuracy": 0.8564985990524292, "num_tokens": 211063568.0, "step": 5557 }, { "epoch": 0.7070347284060552, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.340921401977539, "learning_rate": 5e-05, "loss": 0.7234, "mean_token_accuracy": 0.8610088229179382, "num_tokens": 211100422.0, "step": 5558 }, { "epoch": 0.7071619386846457, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.116598129272461, "learning_rate": 5e-05, "loss": 1.0188, "mean_token_accuracy": 0.8532532453536987, "num_tokens": 211138644.0, "step": 5559 }, { "epoch": 0.7072891489632362, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.566479206085205, "learning_rate": 5e-05, "loss": 0.7862, "mean_token_accuracy": 0.8574262857437134, "num_tokens": 211173033.0, "step": 5560 }, { "epoch": 0.7074163592418268, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.874686241149902, "learning_rate": 5e-05, "loss": 0.9569, "mean_token_accuracy": 0.8678581118583679, "num_tokens": 211212902.0, "step": 5561 }, { "epoch": 0.7075435695204173, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.259328365325928, "learning_rate": 5e-05, "loss": 0.9607, "mean_token_accuracy": 0.8365448117256165, "num_tokens": 211258812.0, "step": 5562 }, { "epoch": 0.7076707797990077, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.165383338928223, "learning_rate": 5e-05, "loss": 0.9232, "mean_token_accuracy": 0.8533860445022583, "num_tokens": 211296496.0, "step": 5563 }, { "epoch": 0.7077979900775982, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.669970512390137, "learning_rate": 5e-05, "loss": 0.9224, "mean_token_accuracy": 0.8562716245651245, "num_tokens": 211337081.0, "step": 5564 }, { "epoch": 0.7079252003561888, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.643641948699951, "learning_rate": 5e-05, "loss": 0.8892, "mean_token_accuracy": 0.8500206470489502, "num_tokens": 211373845.0, "step": 5565 }, { "epoch": 0.7080524106347793, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.553391456604004, "learning_rate": 5e-05, "loss": 0.8736, "mean_token_accuracy": 0.858637809753418, "num_tokens": 211412994.0, "step": 5566 }, { "epoch": 0.7081796209133698, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.732811450958252, "learning_rate": 5e-05, "loss": 0.8852, "mean_token_accuracy": 0.8499311804771423, "num_tokens": 211456678.0, "step": 5567 }, { "epoch": 0.7083068311919604, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.071501731872559, "learning_rate": 5e-05, "loss": 0.8078, "mean_token_accuracy": 0.8517582416534424, "num_tokens": 211497040.0, "step": 5568 }, { "epoch": 0.7084340414705508, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.138961791992188, "learning_rate": 5e-05, "loss": 0.8943, "mean_token_accuracy": 0.8498615026473999, "num_tokens": 211534113.0, "step": 5569 }, { "epoch": 0.7085612517491413, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.631538391113281, "learning_rate": 5e-05, "loss": 0.7238, "mean_token_accuracy": 0.862811803817749, "num_tokens": 211569518.0, "step": 5570 }, { "epoch": 0.7086884620277318, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.489324569702148, "learning_rate": 5e-05, "loss": 0.8762, "mean_token_accuracy": 0.8701713681221008, "num_tokens": 211607918.0, "step": 5571 }, { "epoch": 0.7088156723063224, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.485210418701172, "learning_rate": 5e-05, "loss": 0.7899, "mean_token_accuracy": 0.8486001491546631, "num_tokens": 211647222.0, "step": 5572 }, { "epoch": 0.7089428825849129, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.68196964263916, "learning_rate": 5e-05, "loss": 0.9103, "mean_token_accuracy": 0.8658574819564819, "num_tokens": 211684584.0, "step": 5573 }, { "epoch": 0.7090700928635034, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.608119964599609, "learning_rate": 5e-05, "loss": 0.7634, "mean_token_accuracy": 0.8631658554077148, "num_tokens": 211718968.0, "step": 5574 }, { "epoch": 0.7091973031420938, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.644152641296387, "learning_rate": 5e-05, "loss": 0.9367, "mean_token_accuracy": 0.8626313805580139, "num_tokens": 211756230.0, "step": 5575 }, { "epoch": 0.7093245134206844, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 6.987656593322754, "learning_rate": 5e-05, "loss": 0.774, "mean_token_accuracy": 0.8785884976387024, "num_tokens": 211794155.0, "step": 5576 }, { "epoch": 0.7094517236992749, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.214680671691895, "learning_rate": 5e-05, "loss": 0.9319, "mean_token_accuracy": 0.8593818545341492, "num_tokens": 211834483.0, "step": 5577 }, { "epoch": 0.7095789339778654, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.44050407409668, "learning_rate": 5e-05, "loss": 0.9108, "mean_token_accuracy": 0.8479067087173462, "num_tokens": 211873369.0, "step": 5578 }, { "epoch": 0.709706144256456, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.598819732666016, "learning_rate": 5e-05, "loss": 0.8834, "mean_token_accuracy": 0.8506064414978027, "num_tokens": 211911985.0, "step": 5579 }, { "epoch": 0.7098333545350465, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.685843467712402, "learning_rate": 5e-05, "loss": 0.844, "mean_token_accuracy": 0.863121509552002, "num_tokens": 211953739.0, "step": 5580 }, { "epoch": 0.7099605648136369, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.244058132171631, "learning_rate": 5e-05, "loss": 0.7491, "mean_token_accuracy": 0.8682122826576233, "num_tokens": 211988817.0, "step": 5581 }, { "epoch": 0.7100877750922274, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.712652683258057, "learning_rate": 5e-05, "loss": 0.875, "mean_token_accuracy": 0.8486316204071045, "num_tokens": 212030891.0, "step": 5582 }, { "epoch": 0.710214985370818, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.153841495513916, "learning_rate": 5e-05, "loss": 0.7394, "mean_token_accuracy": 0.8710892200469971, "num_tokens": 212065727.0, "step": 5583 }, { "epoch": 0.7103421956494085, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.676663398742676, "learning_rate": 5e-05, "loss": 0.8334, "mean_token_accuracy": 0.8574584126472473, "num_tokens": 212098718.0, "step": 5584 }, { "epoch": 0.710469405927999, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.064545631408691, "learning_rate": 5e-05, "loss": 0.7612, "mean_token_accuracy": 0.8583465814590454, "num_tokens": 212142597.0, "step": 5585 }, { "epoch": 0.7105966162065895, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.754368305206299, "learning_rate": 5e-05, "loss": 0.8754, "mean_token_accuracy": 0.8366985321044922, "num_tokens": 212182878.0, "step": 5586 }, { "epoch": 0.71072382648518, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.92822265625, "learning_rate": 5e-05, "loss": 0.7137, "mean_token_accuracy": 0.8650542497634888, "num_tokens": 212216759.0, "step": 5587 }, { "epoch": 0.7108510367637705, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.895251750946045, "learning_rate": 5e-05, "loss": 0.8395, "mean_token_accuracy": 0.852398157119751, "num_tokens": 212253559.0, "step": 5588 }, { "epoch": 0.710978247042361, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.763359546661377, "learning_rate": 5e-05, "loss": 0.7164, "mean_token_accuracy": 0.8634019494056702, "num_tokens": 212296821.0, "step": 5589 }, { "epoch": 0.7111054573209515, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.078786849975586, "learning_rate": 5e-05, "loss": 0.8045, "mean_token_accuracy": 0.8656520843505859, "num_tokens": 212334872.0, "step": 5590 }, { "epoch": 0.7112326675995421, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.618593692779541, "learning_rate": 5e-05, "loss": 0.7146, "mean_token_accuracy": 0.8631181120872498, "num_tokens": 212375692.0, "step": 5591 }, { "epoch": 0.7113598778781326, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.272987365722656, "learning_rate": 5e-05, "loss": 0.8629, "mean_token_accuracy": 0.8531668186187744, "num_tokens": 212407301.0, "step": 5592 }, { "epoch": 0.711487088156723, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.5130815505981445, "learning_rate": 5e-05, "loss": 0.7196, "mean_token_accuracy": 0.8610774278640747, "num_tokens": 212443831.0, "step": 5593 }, { "epoch": 0.7116142984353135, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.447444915771484, "learning_rate": 5e-05, "loss": 0.84, "mean_token_accuracy": 0.8670108318328857, "num_tokens": 212477595.0, "step": 5594 }, { "epoch": 0.7117415087139041, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.526472091674805, "learning_rate": 5e-05, "loss": 0.8009, "mean_token_accuracy": 0.8409661054611206, "num_tokens": 212513786.0, "step": 5595 }, { "epoch": 0.7118687189924946, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.442255020141602, "learning_rate": 5e-05, "loss": 0.8834, "mean_token_accuracy": 0.8578965663909912, "num_tokens": 212548111.0, "step": 5596 }, { "epoch": 0.7119959292710851, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.782813549041748, "learning_rate": 5e-05, "loss": 0.7909, "mean_token_accuracy": 0.8546187281608582, "num_tokens": 212585395.0, "step": 5597 }, { "epoch": 0.7121231395496757, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.103999137878418, "learning_rate": 5e-05, "loss": 0.8662, "mean_token_accuracy": 0.8696780204772949, "num_tokens": 212625351.0, "step": 5598 }, { "epoch": 0.7122503498282661, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.208898544311523, "learning_rate": 5e-05, "loss": 0.8192, "mean_token_accuracy": 0.8529685735702515, "num_tokens": 212657689.0, "step": 5599 }, { "epoch": 0.7123775601068566, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.542420387268066, "learning_rate": 5e-05, "loss": 0.7589, "mean_token_accuracy": 0.8755710124969482, "num_tokens": 212694741.0, "step": 5600 }, { "epoch": 0.7125047703854471, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.593786716461182, "learning_rate": 5e-05, "loss": 0.8334, "mean_token_accuracy": 0.8538247346878052, "num_tokens": 212727109.0, "step": 5601 }, { "epoch": 0.7126319806640377, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.116004467010498, "learning_rate": 5e-05, "loss": 0.8039, "mean_token_accuracy": 0.848690390586853, "num_tokens": 212763031.0, "step": 5602 }, { "epoch": 0.7127591909426282, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.75559663772583, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8634594082832336, "num_tokens": 212803601.0, "step": 5603 }, { "epoch": 0.7128864012212187, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.9969868659973145, "learning_rate": 5e-05, "loss": 0.6949, "mean_token_accuracy": 0.8758906126022339, "num_tokens": 212848331.0, "step": 5604 }, { "epoch": 0.7130136114998091, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.707024574279785, "learning_rate": 5e-05, "loss": 0.8119, "mean_token_accuracy": 0.8564894199371338, "num_tokens": 212890364.0, "step": 5605 }, { "epoch": 0.7131408217783997, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.056166172027588, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.8653373718261719, "num_tokens": 212931234.0, "step": 5606 }, { "epoch": 0.7132680320569902, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.572137355804443, "learning_rate": 5e-05, "loss": 0.7698, "mean_token_accuracy": 0.8599568009376526, "num_tokens": 212970103.0, "step": 5607 }, { "epoch": 0.7133952423355807, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.162816524505615, "learning_rate": 5e-05, "loss": 0.7673, "mean_token_accuracy": 0.854458749294281, "num_tokens": 213004895.0, "step": 5608 }, { "epoch": 0.7135224526141712, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.416431903839111, "learning_rate": 5e-05, "loss": 0.7498, "mean_token_accuracy": 0.8628032207489014, "num_tokens": 213044263.0, "step": 5609 }, { "epoch": 0.7136496628927618, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.295743465423584, "learning_rate": 5e-05, "loss": 0.761, "mean_token_accuracy": 0.8548287153244019, "num_tokens": 213081685.0, "step": 5610 }, { "epoch": 0.7137768731713523, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.254443168640137, "learning_rate": 5e-05, "loss": 0.7388, "mean_token_accuracy": 0.8628376126289368, "num_tokens": 213116986.0, "step": 5611 }, { "epoch": 0.7139040834499427, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.422109127044678, "learning_rate": 5e-05, "loss": 0.7232, "mean_token_accuracy": 0.8688755035400391, "num_tokens": 213156313.0, "step": 5612 }, { "epoch": 0.7140312937285332, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.137446880340576, "learning_rate": 5e-05, "loss": 0.7756, "mean_token_accuracy": 0.8474004864692688, "num_tokens": 213187216.0, "step": 5613 }, { "epoch": 0.7141585040071238, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.489821910858154, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8575273156166077, "num_tokens": 213223380.0, "step": 5614 }, { "epoch": 0.7142857142857143, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.033536911010742, "learning_rate": 5e-05, "loss": 0.684, "mean_token_accuracy": 0.8744962811470032, "num_tokens": 213259100.0, "step": 5615 }, { "epoch": 0.7144129245643048, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.639378070831299, "learning_rate": 5e-05, "loss": 0.7851, "mean_token_accuracy": 0.8518022298812866, "num_tokens": 213300103.0, "step": 5616 }, { "epoch": 0.7145401348428954, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.809076309204102, "learning_rate": 5e-05, "loss": 0.6808, "mean_token_accuracy": 0.8686188459396362, "num_tokens": 213338283.0, "step": 5617 }, { "epoch": 0.7146673451214858, "ewc_loss": 0.375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.933466911315918, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8652875423431396, "num_tokens": 213382211.0, "step": 5618 }, { "epoch": 0.7147945554000763, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.453083038330078, "learning_rate": 5e-05, "loss": 0.6937, "mean_token_accuracy": 0.8610833883285522, "num_tokens": 213418790.0, "step": 5619 }, { "epoch": 0.7149217656786668, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.496628761291504, "learning_rate": 5e-05, "loss": 0.8255, "mean_token_accuracy": 0.8604344129562378, "num_tokens": 213457843.0, "step": 5620 }, { "epoch": 0.7150489759572574, "ewc_loss": 0.25, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00021457672119140625, "grad_norm": 5.851799964904785, "learning_rate": 5e-05, "loss": 0.6676, "mean_token_accuracy": 0.8596894145011902, "num_tokens": 213496182.0, "step": 5621 }, { "epoch": 0.7151761862358479, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.479433059692383, "learning_rate": 5e-05, "loss": 0.8831, "mean_token_accuracy": 0.8753297328948975, "num_tokens": 213535254.0, "step": 5622 }, { "epoch": 0.7153033965144384, "ewc_loss": 0.265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0002307891845703125, "grad_norm": 5.365104675292969, "learning_rate": 5e-05, "loss": 0.6535, "mean_token_accuracy": 0.8652535676956177, "num_tokens": 213574331.0, "step": 5623 }, { "epoch": 0.7154306067930288, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 10.745351791381836, "learning_rate": 5e-05, "loss": 1.1278, "mean_token_accuracy": 0.8589590787887573, "num_tokens": 213608185.0, "step": 5624 }, { "epoch": 0.7155578170716194, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 6.699052810668945, "learning_rate": 5e-05, "loss": 0.8796, "mean_token_accuracy": 0.8667988777160645, "num_tokens": 213642032.0, "step": 5625 }, { "epoch": 0.7156850273502099, "ewc_loss": 0.78125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 10.879894256591797, "learning_rate": 5e-05, "loss": 1.1747, "mean_token_accuracy": 0.8638581037521362, "num_tokens": 213674652.0, "step": 5626 }, { "epoch": 0.7158122376288004, "ewc_loss": 0.859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000823974609375, "grad_norm": 8.968392372131348, "learning_rate": 5e-05, "loss": 1.2597, "mean_token_accuracy": 0.8654477596282959, "num_tokens": 213711112.0, "step": 5627 }, { "epoch": 0.715939447907391, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.841388702392578, "learning_rate": 5e-05, "loss": 0.9617, "mean_token_accuracy": 0.851371169090271, "num_tokens": 213750319.0, "step": 5628 }, { "epoch": 0.7160666581859815, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 9.974950790405273, "learning_rate": 5e-05, "loss": 1.2356, "mean_token_accuracy": 0.8665326833724976, "num_tokens": 213788408.0, "step": 5629 }, { "epoch": 0.7161938684645719, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.2675557136535645, "learning_rate": 5e-05, "loss": 0.9316, "mean_token_accuracy": 0.8401580452919006, "num_tokens": 213826822.0, "step": 5630 }, { "epoch": 0.7163210787431624, "ewc_loss": 0.921875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000885009765625, "grad_norm": 11.824496269226074, "learning_rate": 5e-05, "loss": 1.4072, "mean_token_accuracy": 0.8437443971633911, "num_tokens": 213864046.0, "step": 5631 }, { "epoch": 0.716448289021753, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.162737846374512, "learning_rate": 5e-05, "loss": 1.2195, "mean_token_accuracy": 0.8633904457092285, "num_tokens": 213907235.0, "step": 5632 }, { "epoch": 0.7165754993003435, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 11.343194961547852, "learning_rate": 5e-05, "loss": 1.1206, "mean_token_accuracy": 0.8615822792053223, "num_tokens": 213942126.0, "step": 5633 }, { "epoch": 0.716702709578934, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 9.359722137451172, "learning_rate": 5e-05, "loss": 1.2458, "mean_token_accuracy": 0.8610745668411255, "num_tokens": 213980486.0, "step": 5634 }, { "epoch": 0.7168299198575245, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 11.091541290283203, "learning_rate": 5e-05, "loss": 1.1072, "mean_token_accuracy": 0.870563805103302, "num_tokens": 214016991.0, "step": 5635 }, { "epoch": 0.716957130136115, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.428764343261719, "learning_rate": 5e-05, "loss": 1.0231, "mean_token_accuracy": 0.8634176254272461, "num_tokens": 214049424.0, "step": 5636 }, { "epoch": 0.7170843404147055, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 11.31379508972168, "learning_rate": 5e-05, "loss": 1.1906, "mean_token_accuracy": 0.8588545322418213, "num_tokens": 214090422.0, "step": 5637 }, { "epoch": 0.717211550693296, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.16934871673584, "learning_rate": 5e-05, "loss": 0.9458, "mean_token_accuracy": 0.8703199625015259, "num_tokens": 214123075.0, "step": 5638 }, { "epoch": 0.7173387609718865, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 12.054126739501953, "learning_rate": 5e-05, "loss": 1.3397, "mean_token_accuracy": 0.8391798734664917, "num_tokens": 214158962.0, "step": 5639 }, { "epoch": 0.7174659712504771, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 9.392500877380371, "learning_rate": 5e-05, "loss": 1.0906, "mean_token_accuracy": 0.8662840127944946, "num_tokens": 214201867.0, "step": 5640 }, { "epoch": 0.7175931815290676, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 11.872148513793945, "learning_rate": 5e-05, "loss": 1.1672, "mean_token_accuracy": 0.8702468872070312, "num_tokens": 214241300.0, "step": 5641 }, { "epoch": 0.717720391807658, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.13638687133789, "learning_rate": 5e-05, "loss": 1.1619, "mean_token_accuracy": 0.8570640087127686, "num_tokens": 214276455.0, "step": 5642 }, { "epoch": 0.7178476020862485, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 11.036436080932617, "learning_rate": 5e-05, "loss": 1.1064, "mean_token_accuracy": 0.864816427230835, "num_tokens": 214313737.0, "step": 5643 }, { "epoch": 0.7179748123648391, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.76654052734375, "learning_rate": 5e-05, "loss": 1.0533, "mean_token_accuracy": 0.8623613715171814, "num_tokens": 214351179.0, "step": 5644 }, { "epoch": 0.7181020226434296, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.402383804321289, "learning_rate": 5e-05, "loss": 1.0129, "mean_token_accuracy": 0.8743646144866943, "num_tokens": 214387606.0, "step": 5645 }, { "epoch": 0.7182292329220201, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.215143203735352, "learning_rate": 5e-05, "loss": 1.0249, "mean_token_accuracy": 0.8524777889251709, "num_tokens": 214428149.0, "step": 5646 }, { "epoch": 0.7183564432006107, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.953486442565918, "learning_rate": 5e-05, "loss": 0.9942, "mean_token_accuracy": 0.8682112097740173, "num_tokens": 214468878.0, "step": 5647 }, { "epoch": 0.7184836534792011, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.573407173156738, "learning_rate": 5e-05, "loss": 0.9477, "mean_token_accuracy": 0.8524316549301147, "num_tokens": 214508813.0, "step": 5648 }, { "epoch": 0.7186108637577916, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.947211265563965, "learning_rate": 5e-05, "loss": 1.0135, "mean_token_accuracy": 0.8628799915313721, "num_tokens": 214553896.0, "step": 5649 }, { "epoch": 0.7187380740363821, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.099715232849121, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8670780658721924, "num_tokens": 214593727.0, "step": 5650 }, { "epoch": 0.7188652843149727, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.352745056152344, "learning_rate": 5e-05, "loss": 1.047, "mean_token_accuracy": 0.8633903861045837, "num_tokens": 214627029.0, "step": 5651 }, { "epoch": 0.7189924945935632, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.252883911132812, "learning_rate": 5e-05, "loss": 0.9378, "mean_token_accuracy": 0.8439887762069702, "num_tokens": 214665484.0, "step": 5652 }, { "epoch": 0.7191197048721537, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.554076194763184, "learning_rate": 5e-05, "loss": 1.0602, "mean_token_accuracy": 0.8696029782295227, "num_tokens": 214708654.0, "step": 5653 }, { "epoch": 0.7192469151507441, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.903595924377441, "learning_rate": 5e-05, "loss": 0.9845, "mean_token_accuracy": 0.8548828363418579, "num_tokens": 214746962.0, "step": 5654 }, { "epoch": 0.7193741254293347, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.925609588623047, "learning_rate": 5e-05, "loss": 1.0156, "mean_token_accuracy": 0.8604930639266968, "num_tokens": 214784691.0, "step": 5655 }, { "epoch": 0.7195013357079252, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.950201034545898, "learning_rate": 5e-05, "loss": 0.9151, "mean_token_accuracy": 0.8704770803451538, "num_tokens": 214817182.0, "step": 5656 }, { "epoch": 0.7196285459865157, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.1098051071167, "learning_rate": 5e-05, "loss": 0.984, "mean_token_accuracy": 0.8470388650894165, "num_tokens": 214857053.0, "step": 5657 }, { "epoch": 0.7197557562651062, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.452730178833008, "learning_rate": 5e-05, "loss": 0.9773, "mean_token_accuracy": 0.8359739184379578, "num_tokens": 214894908.0, "step": 5658 }, { "epoch": 0.7198829665436968, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.700017929077148, "learning_rate": 5e-05, "loss": 0.9022, "mean_token_accuracy": 0.8590657114982605, "num_tokens": 214930235.0, "step": 5659 }, { "epoch": 0.7200101768222873, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.762937545776367, "learning_rate": 5e-05, "loss": 0.7704, "mean_token_accuracy": 0.869381844997406, "num_tokens": 214967140.0, "step": 5660 }, { "epoch": 0.7201373871008777, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.789897918701172, "learning_rate": 5e-05, "loss": 0.9033, "mean_token_accuracy": 0.858235776424408, "num_tokens": 215004039.0, "step": 5661 }, { "epoch": 0.7202645973794682, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.2592267990112305, "learning_rate": 5e-05, "loss": 0.7209, "mean_token_accuracy": 0.870630145072937, "num_tokens": 215039305.0, "step": 5662 }, { "epoch": 0.7203918076580588, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.116265296936035, "learning_rate": 5e-05, "loss": 0.9054, "mean_token_accuracy": 0.8683542013168335, "num_tokens": 215076455.0, "step": 5663 }, { "epoch": 0.7205190179366493, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.180645942687988, "learning_rate": 5e-05, "loss": 0.7968, "mean_token_accuracy": 0.8514808416366577, "num_tokens": 215116826.0, "step": 5664 }, { "epoch": 0.7206462282152398, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.458572387695312, "learning_rate": 5e-05, "loss": 1.0089, "mean_token_accuracy": 0.8506385684013367, "num_tokens": 215162302.0, "step": 5665 }, { "epoch": 0.7207734384938304, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.67922830581665, "learning_rate": 5e-05, "loss": 0.8177, "mean_token_accuracy": 0.8547206521034241, "num_tokens": 215192402.0, "step": 5666 }, { "epoch": 0.7209006487724208, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.156055450439453, "learning_rate": 5e-05, "loss": 0.9929, "mean_token_accuracy": 0.8546757102012634, "num_tokens": 215228521.0, "step": 5667 }, { "epoch": 0.7210278590510113, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.205025672912598, "learning_rate": 5e-05, "loss": 0.8466, "mean_token_accuracy": 0.8617349863052368, "num_tokens": 215264997.0, "step": 5668 }, { "epoch": 0.7211550693296018, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.161125183105469, "learning_rate": 5e-05, "loss": 0.7834, "mean_token_accuracy": 0.8823179602622986, "num_tokens": 215301997.0, "step": 5669 }, { "epoch": 0.7212822796081924, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.203022003173828, "learning_rate": 5e-05, "loss": 0.8591, "mean_token_accuracy": 0.8528770208358765, "num_tokens": 215343637.0, "step": 5670 }, { "epoch": 0.7214094898867829, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.420299053192139, "learning_rate": 5e-05, "loss": 0.7686, "mean_token_accuracy": 0.8632698655128479, "num_tokens": 215381966.0, "step": 5671 }, { "epoch": 0.7215367001653734, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.063361167907715, "learning_rate": 5e-05, "loss": 0.8181, "mean_token_accuracy": 0.859333872795105, "num_tokens": 215422930.0, "step": 5672 }, { "epoch": 0.7216639104439638, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.120845317840576, "learning_rate": 5e-05, "loss": 0.71, "mean_token_accuracy": 0.8716013431549072, "num_tokens": 215461204.0, "step": 5673 }, { "epoch": 0.7217911207225544, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.169266700744629, "learning_rate": 5e-05, "loss": 0.8119, "mean_token_accuracy": 0.8587325811386108, "num_tokens": 215493771.0, "step": 5674 }, { "epoch": 0.7219183310011449, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.8469929695129395, "learning_rate": 5e-05, "loss": 0.7271, "mean_token_accuracy": 0.860525369644165, "num_tokens": 215529398.0, "step": 5675 }, { "epoch": 0.7220455412797354, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.515172958374023, "learning_rate": 5e-05, "loss": 0.8201, "mean_token_accuracy": 0.8668439984321594, "num_tokens": 215565794.0, "step": 5676 }, { "epoch": 0.7221727515583259, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.774089336395264, "learning_rate": 5e-05, "loss": 0.7482, "mean_token_accuracy": 0.8539478182792664, "num_tokens": 215604005.0, "step": 5677 }, { "epoch": 0.7222999618369165, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.607946395874023, "learning_rate": 5e-05, "loss": 0.823, "mean_token_accuracy": 0.8714380860328674, "num_tokens": 215635325.0, "step": 5678 }, { "epoch": 0.7224271721155069, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.099930763244629, "learning_rate": 5e-05, "loss": 0.7325, "mean_token_accuracy": 0.8630872964859009, "num_tokens": 215667477.0, "step": 5679 }, { "epoch": 0.7225543823940974, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.1961088180542, "learning_rate": 5e-05, "loss": 0.865, "mean_token_accuracy": 0.8523620367050171, "num_tokens": 215707087.0, "step": 5680 }, { "epoch": 0.722681592672688, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.445777416229248, "learning_rate": 5e-05, "loss": 0.749, "mean_token_accuracy": 0.8655654788017273, "num_tokens": 215746182.0, "step": 5681 }, { "epoch": 0.7228088029512785, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.762217044830322, "learning_rate": 5e-05, "loss": 0.816, "mean_token_accuracy": 0.8570084571838379, "num_tokens": 215785941.0, "step": 5682 }, { "epoch": 0.722936013229869, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.50206995010376, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.8651015758514404, "num_tokens": 215823562.0, "step": 5683 }, { "epoch": 0.7230632235084595, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.62297248840332, "learning_rate": 5e-05, "loss": 0.8326, "mean_token_accuracy": 0.8411027789115906, "num_tokens": 215860454.0, "step": 5684 }, { "epoch": 0.72319043378705, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.352261066436768, "learning_rate": 5e-05, "loss": 0.711, "mean_token_accuracy": 0.86797034740448, "num_tokens": 215895088.0, "step": 5685 }, { "epoch": 0.7233176440656405, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.699927806854248, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.8539459705352783, "num_tokens": 215935170.0, "step": 5686 }, { "epoch": 0.723444854344231, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.0864081382751465, "learning_rate": 5e-05, "loss": 0.6932, "mean_token_accuracy": 0.8731337785720825, "num_tokens": 215977323.0, "step": 5687 }, { "epoch": 0.7235720646228215, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.9795002937316895, "learning_rate": 5e-05, "loss": 0.7564, "mean_token_accuracy": 0.8716799020767212, "num_tokens": 216016729.0, "step": 5688 }, { "epoch": 0.7236992749014121, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.7111053466796875, "learning_rate": 5e-05, "loss": 0.7163, "mean_token_accuracy": 0.8604533076286316, "num_tokens": 216058514.0, "step": 5689 }, { "epoch": 0.7238264851800026, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.430988311767578, "learning_rate": 5e-05, "loss": 0.8918, "mean_token_accuracy": 0.8404232263565063, "num_tokens": 216098633.0, "step": 5690 }, { "epoch": 0.723953695458593, "ewc_loss": 0.28125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.408821105957031, "learning_rate": 5e-05, "loss": 0.6245, "mean_token_accuracy": 0.8814383745193481, "num_tokens": 216136124.0, "step": 5691 }, { "epoch": 0.7240809057371835, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.774088859558105, "learning_rate": 5e-05, "loss": 0.8939, "mean_token_accuracy": 0.8571298718452454, "num_tokens": 216178799.0, "step": 5692 }, { "epoch": 0.7242081160157741, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.545188903808594, "learning_rate": 5e-05, "loss": 0.7677, "mean_token_accuracy": 0.8458700180053711, "num_tokens": 216215233.0, "step": 5693 }, { "epoch": 0.7243353262943646, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.588274002075195, "learning_rate": 5e-05, "loss": 0.8534, "mean_token_accuracy": 0.865699291229248, "num_tokens": 216252258.0, "step": 5694 }, { "epoch": 0.7244625365729551, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.169110298156738, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.8477398753166199, "num_tokens": 216287845.0, "step": 5695 }, { "epoch": 0.7245897468515456, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.788351535797119, "learning_rate": 5e-05, "loss": 0.7896, "mean_token_accuracy": 0.8695315718650818, "num_tokens": 216325490.0, "step": 5696 }, { "epoch": 0.7247169571301361, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.819427013397217, "learning_rate": 5e-05, "loss": 0.8124, "mean_token_accuracy": 0.852872371673584, "num_tokens": 216359353.0, "step": 5697 }, { "epoch": 0.7248441674087266, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.9833550453186035, "learning_rate": 5e-05, "loss": 0.7857, "mean_token_accuracy": 0.8512842655181885, "num_tokens": 216393620.0, "step": 5698 }, { "epoch": 0.7249713776873171, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.235426902770996, "learning_rate": 5e-05, "loss": 0.8034, "mean_token_accuracy": 0.866382360458374, "num_tokens": 216427400.0, "step": 5699 }, { "epoch": 0.7250985879659076, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.605736255645752, "learning_rate": 5e-05, "loss": 0.7249, "mean_token_accuracy": 0.8580607175827026, "num_tokens": 216463817.0, "step": 5700 }, { "epoch": 0.7252257982444982, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.498876571655273, "learning_rate": 5e-05, "loss": 0.8428, "mean_token_accuracy": 0.8584935665130615, "num_tokens": 216495596.0, "step": 5701 }, { "epoch": 0.7253530085230887, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.495453834533691, "learning_rate": 5e-05, "loss": 0.7073, "mean_token_accuracy": 0.8641142845153809, "num_tokens": 216533003.0, "step": 5702 }, { "epoch": 0.7254802188016791, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.621936798095703, "learning_rate": 5e-05, "loss": 0.877, "mean_token_accuracy": 0.8631942272186279, "num_tokens": 216569280.0, "step": 5703 }, { "epoch": 0.7256074290802697, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.63710880279541, "learning_rate": 5e-05, "loss": 0.694, "mean_token_accuracy": 0.8713048696517944, "num_tokens": 216603073.0, "step": 5704 }, { "epoch": 0.7257346393588602, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.50643539428711, "learning_rate": 5e-05, "loss": 0.8901, "mean_token_accuracy": 0.8555386662483215, "num_tokens": 216641971.0, "step": 5705 }, { "epoch": 0.7258618496374507, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.97790002822876, "learning_rate": 5e-05, "loss": 0.8061, "mean_token_accuracy": 0.8445591926574707, "num_tokens": 216683119.0, "step": 5706 }, { "epoch": 0.7259890599160412, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.108201026916504, "learning_rate": 5e-05, "loss": 0.9238, "mean_token_accuracy": 0.8361685872077942, "num_tokens": 216722139.0, "step": 5707 }, { "epoch": 0.7261162701946318, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.320013523101807, "learning_rate": 5e-05, "loss": 0.7341, "mean_token_accuracy": 0.8694870471954346, "num_tokens": 216757712.0, "step": 5708 }, { "epoch": 0.7262434804732223, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.694493293762207, "learning_rate": 5e-05, "loss": 0.8215, "mean_token_accuracy": 0.8503730297088623, "num_tokens": 216794132.0, "step": 5709 }, { "epoch": 0.7263706907518127, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.419983863830566, "learning_rate": 5e-05, "loss": 0.7691, "mean_token_accuracy": 0.8559299111366272, "num_tokens": 216833937.0, "step": 5710 }, { "epoch": 0.7264979010304032, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.619019508361816, "learning_rate": 5e-05, "loss": 0.7975, "mean_token_accuracy": 0.8579011559486389, "num_tokens": 216872979.0, "step": 5711 }, { "epoch": 0.7266251113089938, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.225964546203613, "learning_rate": 5e-05, "loss": 0.7159, "mean_token_accuracy": 0.8710979223251343, "num_tokens": 216915369.0, "step": 5712 }, { "epoch": 0.7267523215875843, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.819549560546875, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8660379648208618, "num_tokens": 216948876.0, "step": 5713 }, { "epoch": 0.7268795318661748, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.903647422790527, "learning_rate": 5e-05, "loss": 0.7322, "mean_token_accuracy": 0.8615783452987671, "num_tokens": 216988254.0, "step": 5714 }, { "epoch": 0.7270067421447653, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.114282608032227, "learning_rate": 5e-05, "loss": 0.8196, "mean_token_accuracy": 0.8575043678283691, "num_tokens": 217024792.0, "step": 5715 }, { "epoch": 0.7271339524233558, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.629000186920166, "learning_rate": 5e-05, "loss": 0.7313, "mean_token_accuracy": 0.8603408932685852, "num_tokens": 217063228.0, "step": 5716 }, { "epoch": 0.7272611627019463, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.407404899597168, "learning_rate": 5e-05, "loss": 0.9299, "mean_token_accuracy": 0.8338249921798706, "num_tokens": 217102601.0, "step": 5717 }, { "epoch": 0.7273883729805368, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.479502201080322, "learning_rate": 5e-05, "loss": 0.7012, "mean_token_accuracy": 0.8646011352539062, "num_tokens": 217137839.0, "step": 5718 }, { "epoch": 0.7275155832591274, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.582927703857422, "learning_rate": 5e-05, "loss": 0.879, "mean_token_accuracy": 0.865867018699646, "num_tokens": 217179862.0, "step": 5719 }, { "epoch": 0.7276427935377179, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.585844993591309, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.8375958800315857, "num_tokens": 217217728.0, "step": 5720 }, { "epoch": 0.7277700038163084, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.446062088012695, "learning_rate": 5e-05, "loss": 0.893, "mean_token_accuracy": 0.8579638004302979, "num_tokens": 217255410.0, "step": 5721 }, { "epoch": 0.7278972140948988, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.0002923011779785, "learning_rate": 5e-05, "loss": 0.7866, "mean_token_accuracy": 0.8511537313461304, "num_tokens": 217294365.0, "step": 5722 }, { "epoch": 0.7280244243734894, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.884387016296387, "learning_rate": 5e-05, "loss": 0.8153, "mean_token_accuracy": 0.8659490942955017, "num_tokens": 217333451.0, "step": 5723 }, { "epoch": 0.7281516346520799, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.541299343109131, "learning_rate": 5e-05, "loss": 0.8039, "mean_token_accuracy": 0.8540241718292236, "num_tokens": 217368737.0, "step": 5724 }, { "epoch": 0.7282788449306704, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.191885948181152, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.8424675464630127, "num_tokens": 217409476.0, "step": 5725 }, { "epoch": 0.7284060552092609, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.984375476837158, "learning_rate": 5e-05, "loss": 0.7936, "mean_token_accuracy": 0.8664683103561401, "num_tokens": 217446481.0, "step": 5726 }, { "epoch": 0.7285332654878515, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.748789310455322, "learning_rate": 5e-05, "loss": 0.7046, "mean_token_accuracy": 0.8677879571914673, "num_tokens": 217482647.0, "step": 5727 }, { "epoch": 0.7286604757664419, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.293519973754883, "learning_rate": 5e-05, "loss": 0.8667, "mean_token_accuracy": 0.8548676371574402, "num_tokens": 217520783.0, "step": 5728 }, { "epoch": 0.7287876860450324, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.502891540527344, "learning_rate": 5e-05, "loss": 0.7243, "mean_token_accuracy": 0.8592023849487305, "num_tokens": 217557061.0, "step": 5729 }, { "epoch": 0.7289148963236229, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.546136856079102, "learning_rate": 5e-05, "loss": 0.832, "mean_token_accuracy": 0.8687957525253296, "num_tokens": 217586833.0, "step": 5730 }, { "epoch": 0.7290421066022135, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.392950057983398, "learning_rate": 5e-05, "loss": 0.7422, "mean_token_accuracy": 0.8517037034034729, "num_tokens": 217622644.0, "step": 5731 }, { "epoch": 0.729169316880804, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.734321594238281, "learning_rate": 5e-05, "loss": 0.8948, "mean_token_accuracy": 0.8640258312225342, "num_tokens": 217659345.0, "step": 5732 }, { "epoch": 0.7292965271593945, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.536079406738281, "learning_rate": 5e-05, "loss": 0.7194, "mean_token_accuracy": 0.868312656879425, "num_tokens": 217698134.0, "step": 5733 }, { "epoch": 0.7294237374379849, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.609969139099121, "learning_rate": 5e-05, "loss": 0.9456, "mean_token_accuracy": 0.8515968322753906, "num_tokens": 217730578.0, "step": 5734 }, { "epoch": 0.7295509477165755, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.992373943328857, "learning_rate": 5e-05, "loss": 0.7328, "mean_token_accuracy": 0.8764313459396362, "num_tokens": 217768090.0, "step": 5735 }, { "epoch": 0.729678157995166, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.068949699401855, "learning_rate": 5e-05, "loss": 0.8783, "mean_token_accuracy": 0.8592038154602051, "num_tokens": 217809350.0, "step": 5736 }, { "epoch": 0.7298053682737565, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.476787090301514, "learning_rate": 5e-05, "loss": 0.8471, "mean_token_accuracy": 0.855589747428894, "num_tokens": 217852478.0, "step": 5737 }, { "epoch": 0.7299325785523471, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.5069780349731445, "learning_rate": 5e-05, "loss": 0.7862, "mean_token_accuracy": 0.8670310974121094, "num_tokens": 217885463.0, "step": 5738 }, { "epoch": 0.7300597888309376, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.64727258682251, "learning_rate": 5e-05, "loss": 0.8526, "mean_token_accuracy": 0.851112961769104, "num_tokens": 217925712.0, "step": 5739 }, { "epoch": 0.730186999109528, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.377590656280518, "learning_rate": 5e-05, "loss": 0.7835, "mean_token_accuracy": 0.8581550717353821, "num_tokens": 217969628.0, "step": 5740 }, { "epoch": 0.7303142093881185, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.433250904083252, "learning_rate": 5e-05, "loss": 0.7696, "mean_token_accuracy": 0.861677885055542, "num_tokens": 218012027.0, "step": 5741 }, { "epoch": 0.7304414196667091, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.632421970367432, "learning_rate": 5e-05, "loss": 0.8209, "mean_token_accuracy": 0.8553920984268188, "num_tokens": 218054018.0, "step": 5742 }, { "epoch": 0.7305686299452996, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.000509262084961, "learning_rate": 5e-05, "loss": 0.792, "mean_token_accuracy": 0.8475748300552368, "num_tokens": 218099181.0, "step": 5743 }, { "epoch": 0.7306958402238901, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.036733627319336, "learning_rate": 5e-05, "loss": 0.7973, "mean_token_accuracy": 0.8671689033508301, "num_tokens": 218135330.0, "step": 5744 }, { "epoch": 0.7308230505024806, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.61525297164917, "learning_rate": 5e-05, "loss": 0.7814, "mean_token_accuracy": 0.8440946340560913, "num_tokens": 218171785.0, "step": 5745 }, { "epoch": 0.7309502607810711, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.433714866638184, "learning_rate": 5e-05, "loss": 0.9349, "mean_token_accuracy": 0.8374547958374023, "num_tokens": 218204230.0, "step": 5746 }, { "epoch": 0.7310774710596616, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.363529682159424, "learning_rate": 5e-05, "loss": 0.6848, "mean_token_accuracy": 0.8732144832611084, "num_tokens": 218237755.0, "step": 5747 }, { "epoch": 0.7312046813382521, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.75589370727539, "learning_rate": 5e-05, "loss": 0.9478, "mean_token_accuracy": 0.8462013602256775, "num_tokens": 218271850.0, "step": 5748 }, { "epoch": 0.7313318916168426, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.390664577484131, "learning_rate": 5e-05, "loss": 0.7354, "mean_token_accuracy": 0.8637269139289856, "num_tokens": 218308466.0, "step": 5749 }, { "epoch": 0.7314591018954332, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.771634101867676, "learning_rate": 5e-05, "loss": 0.9771, "mean_token_accuracy": 0.8462750911712646, "num_tokens": 218346133.0, "step": 5750 }, { "epoch": 0.7315863121740237, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.857080459594727, "learning_rate": 5e-05, "loss": 0.7982, "mean_token_accuracy": 0.8577797412872314, "num_tokens": 218384341.0, "step": 5751 }, { "epoch": 0.7317135224526141, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.192452430725098, "learning_rate": 5e-05, "loss": 0.8765, "mean_token_accuracy": 0.8614358901977539, "num_tokens": 218418314.0, "step": 5752 }, { "epoch": 0.7318407327312046, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.574134349822998, "learning_rate": 5e-05, "loss": 0.8818, "mean_token_accuracy": 0.8512855172157288, "num_tokens": 218455258.0, "step": 5753 }, { "epoch": 0.7319679430097952, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.235672473907471, "learning_rate": 5e-05, "loss": 0.7919, "mean_token_accuracy": 0.8601683974266052, "num_tokens": 218498733.0, "step": 5754 }, { "epoch": 0.7320951532883857, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.175071716308594, "learning_rate": 5e-05, "loss": 0.8953, "mean_token_accuracy": 0.853430986404419, "num_tokens": 218541365.0, "step": 5755 }, { "epoch": 0.7322223635669762, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.612138748168945, "learning_rate": 5e-05, "loss": 0.7733, "mean_token_accuracy": 0.8498806953430176, "num_tokens": 218578292.0, "step": 5756 }, { "epoch": 0.7323495738455668, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.544601440429688, "learning_rate": 5e-05, "loss": 0.8989, "mean_token_accuracy": 0.8588762879371643, "num_tokens": 218617988.0, "step": 5757 }, { "epoch": 0.7324767841241572, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.4038496017456055, "learning_rate": 5e-05, "loss": 0.7601, "mean_token_accuracy": 0.8562588691711426, "num_tokens": 218657759.0, "step": 5758 }, { "epoch": 0.7326039944027477, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.779562950134277, "learning_rate": 5e-05, "loss": 0.9348, "mean_token_accuracy": 0.8601338863372803, "num_tokens": 218693619.0, "step": 5759 }, { "epoch": 0.7327312046813382, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.467809677124023, "learning_rate": 5e-05, "loss": 0.7219, "mean_token_accuracy": 0.8645172119140625, "num_tokens": 218725978.0, "step": 5760 }, { "epoch": 0.7328584149599288, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.83155632019043, "learning_rate": 5e-05, "loss": 0.9442, "mean_token_accuracy": 0.8641989231109619, "num_tokens": 218769616.0, "step": 5761 }, { "epoch": 0.7329856252385193, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.853650093078613, "learning_rate": 5e-05, "loss": 0.8008, "mean_token_accuracy": 0.8645526766777039, "num_tokens": 218806053.0, "step": 5762 }, { "epoch": 0.7331128355171098, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.487716674804688, "learning_rate": 5e-05, "loss": 0.967, "mean_token_accuracy": 0.8507277965545654, "num_tokens": 218844650.0, "step": 5763 }, { "epoch": 0.7332400457957003, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.341811656951904, "learning_rate": 5e-05, "loss": 0.8969, "mean_token_accuracy": 0.8514899015426636, "num_tokens": 218885943.0, "step": 5764 }, { "epoch": 0.7333672560742908, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.909097194671631, "learning_rate": 5e-05, "loss": 0.8855, "mean_token_accuracy": 0.8606444597244263, "num_tokens": 218923939.0, "step": 5765 }, { "epoch": 0.7334944663528813, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.5618109703063965, "learning_rate": 5e-05, "loss": 0.8547, "mean_token_accuracy": 0.859613835811615, "num_tokens": 218964252.0, "step": 5766 }, { "epoch": 0.7336216766314718, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.663967609405518, "learning_rate": 5e-05, "loss": 0.8644, "mean_token_accuracy": 0.8547470569610596, "num_tokens": 219009815.0, "step": 5767 }, { "epoch": 0.7337488869100623, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.311595916748047, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.8638483285903931, "num_tokens": 219047998.0, "step": 5768 }, { "epoch": 0.7338760971886529, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.950955867767334, "learning_rate": 5e-05, "loss": 0.9167, "mean_token_accuracy": 0.8450326919555664, "num_tokens": 219082447.0, "step": 5769 }, { "epoch": 0.7340033074672434, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.824981689453125, "learning_rate": 5e-05, "loss": 0.7893, "mean_token_accuracy": 0.8490117788314819, "num_tokens": 219119393.0, "step": 5770 }, { "epoch": 0.7341305177458338, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.407715797424316, "learning_rate": 5e-05, "loss": 0.9478, "mean_token_accuracy": 0.84995436668396, "num_tokens": 219155954.0, "step": 5771 }, { "epoch": 0.7342577280244243, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.530722618103027, "learning_rate": 5e-05, "loss": 0.7653, "mean_token_accuracy": 0.8583943843841553, "num_tokens": 219195960.0, "step": 5772 }, { "epoch": 0.7343849383030149, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.665451049804688, "learning_rate": 5e-05, "loss": 0.9307, "mean_token_accuracy": 0.857191264629364, "num_tokens": 219238323.0, "step": 5773 }, { "epoch": 0.7345121485816054, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.6367292404174805, "learning_rate": 5e-05, "loss": 0.7491, "mean_token_accuracy": 0.8638808727264404, "num_tokens": 219279044.0, "step": 5774 }, { "epoch": 0.7346393588601959, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.549993515014648, "learning_rate": 5e-05, "loss": 0.9089, "mean_token_accuracy": 0.8671178221702576, "num_tokens": 219312635.0, "step": 5775 }, { "epoch": 0.7347665691387865, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.1017889976501465, "learning_rate": 5e-05, "loss": 0.7863, "mean_token_accuracy": 0.8754062652587891, "num_tokens": 219349424.0, "step": 5776 }, { "epoch": 0.7348937794173769, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.955742359161377, "learning_rate": 5e-05, "loss": 0.8954, "mean_token_accuracy": 0.8565497994422913, "num_tokens": 219389230.0, "step": 5777 }, { "epoch": 0.7350209896959674, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.624505043029785, "learning_rate": 5e-05, "loss": 0.8845, "mean_token_accuracy": 0.8594577312469482, "num_tokens": 219424082.0, "step": 5778 }, { "epoch": 0.7351481999745579, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.263136386871338, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8608083128929138, "num_tokens": 219463324.0, "step": 5779 }, { "epoch": 0.7352754102531485, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.919723987579346, "learning_rate": 5e-05, "loss": 0.7972, "mean_token_accuracy": 0.8790339827537537, "num_tokens": 219500653.0, "step": 5780 }, { "epoch": 0.735402620531739, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.945699214935303, "learning_rate": 5e-05, "loss": 0.725, "mean_token_accuracy": 0.8692567348480225, "num_tokens": 219538906.0, "step": 5781 }, { "epoch": 0.7355298308103295, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.957813262939453, "learning_rate": 5e-05, "loss": 0.899, "mean_token_accuracy": 0.8371938467025757, "num_tokens": 219578950.0, "step": 5782 }, { "epoch": 0.7356570410889199, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.901393890380859, "learning_rate": 5e-05, "loss": 0.6978, "mean_token_accuracy": 0.875161349773407, "num_tokens": 219614298.0, "step": 5783 }, { "epoch": 0.7357842513675105, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.9572954177856445, "learning_rate": 5e-05, "loss": 0.8511, "mean_token_accuracy": 0.8526661396026611, "num_tokens": 219655521.0, "step": 5784 }, { "epoch": 0.735911461646101, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.862696647644043, "learning_rate": 5e-05, "loss": 0.7317, "mean_token_accuracy": 0.8634722232818604, "num_tokens": 219689112.0, "step": 5785 }, { "epoch": 0.7360386719246915, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.020456314086914, "learning_rate": 5e-05, "loss": 0.9122, "mean_token_accuracy": 0.8358211517333984, "num_tokens": 219727432.0, "step": 5786 }, { "epoch": 0.736165882203282, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.768728733062744, "learning_rate": 5e-05, "loss": 0.7462, "mean_token_accuracy": 0.8572795391082764, "num_tokens": 219767231.0, "step": 5787 }, { "epoch": 0.7362930924818726, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.137712478637695, "learning_rate": 5e-05, "loss": 0.8112, "mean_token_accuracy": 0.8672721982002258, "num_tokens": 219798999.0, "step": 5788 }, { "epoch": 0.736420302760463, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.684180736541748, "learning_rate": 5e-05, "loss": 0.7032, "mean_token_accuracy": 0.8701701164245605, "num_tokens": 219844867.0, "step": 5789 }, { "epoch": 0.7365475130390535, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.259773254394531, "learning_rate": 5e-05, "loss": 0.8599, "mean_token_accuracy": 0.8589051365852356, "num_tokens": 219882817.0, "step": 5790 }, { "epoch": 0.736674723317644, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.699109077453613, "learning_rate": 5e-05, "loss": 0.7718, "mean_token_accuracy": 0.8452678918838501, "num_tokens": 219916776.0, "step": 5791 }, { "epoch": 0.7368019335962346, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.219164848327637, "learning_rate": 5e-05, "loss": 0.807, "mean_token_accuracy": 0.8785039186477661, "num_tokens": 219953471.0, "step": 5792 }, { "epoch": 0.7369291438748251, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.843839168548584, "learning_rate": 5e-05, "loss": 0.788, "mean_token_accuracy": 0.8521132469177246, "num_tokens": 219983313.0, "step": 5793 }, { "epoch": 0.7370563541534156, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.01978874206543, "learning_rate": 5e-05, "loss": 0.9248, "mean_token_accuracy": 0.8402260541915894, "num_tokens": 220028222.0, "step": 5794 }, { "epoch": 0.737183564432006, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.071117401123047, "learning_rate": 5e-05, "loss": 0.7765, "mean_token_accuracy": 0.8602367639541626, "num_tokens": 220069005.0, "step": 5795 }, { "epoch": 0.7373107747105966, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.75494384765625, "learning_rate": 5e-05, "loss": 0.8545, "mean_token_accuracy": 0.8540622591972351, "num_tokens": 220103557.0, "step": 5796 }, { "epoch": 0.7374379849891871, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.244226455688477, "learning_rate": 5e-05, "loss": 0.779, "mean_token_accuracy": 0.8576481342315674, "num_tokens": 220141405.0, "step": 5797 }, { "epoch": 0.7375651952677776, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.536947727203369, "learning_rate": 5e-05, "loss": 0.8069, "mean_token_accuracy": 0.8572930693626404, "num_tokens": 220184913.0, "step": 5798 }, { "epoch": 0.7376924055463682, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.307680130004883, "learning_rate": 5e-05, "loss": 0.7857, "mean_token_accuracy": 0.8555032014846802, "num_tokens": 220221886.0, "step": 5799 }, { "epoch": 0.7378196158249587, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.492580413818359, "learning_rate": 5e-05, "loss": 0.762, "mean_token_accuracy": 0.8669763803482056, "num_tokens": 220260945.0, "step": 5800 }, { "epoch": 0.7379468261035491, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.251901149749756, "learning_rate": 5e-05, "loss": 0.7447, "mean_token_accuracy": 0.8624217510223389, "num_tokens": 220299580.0, "step": 5801 }, { "epoch": 0.7380740363821396, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.536657810211182, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8691762089729309, "num_tokens": 220338063.0, "step": 5802 }, { "epoch": 0.7382012466607302, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.116990566253662, "learning_rate": 5e-05, "loss": 0.7193, "mean_token_accuracy": 0.8669401407241821, "num_tokens": 220381212.0, "step": 5803 }, { "epoch": 0.7383284569393207, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.698718070983887, "learning_rate": 5e-05, "loss": 0.8088, "mean_token_accuracy": 0.8570513725280762, "num_tokens": 220422665.0, "step": 5804 }, { "epoch": 0.7384556672179112, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.867160797119141, "learning_rate": 5e-05, "loss": 0.708, "mean_token_accuracy": 0.8676077723503113, "num_tokens": 220462044.0, "step": 5805 }, { "epoch": 0.7385828774965018, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.997640132904053, "learning_rate": 5e-05, "loss": 0.8763, "mean_token_accuracy": 0.8390156030654907, "num_tokens": 220497373.0, "step": 5806 }, { "epoch": 0.7387100877750922, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.54645299911499, "learning_rate": 5e-05, "loss": 0.676, "mean_token_accuracy": 0.8722937107086182, "num_tokens": 220537612.0, "step": 5807 }, { "epoch": 0.7388372980536827, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.436986923217773, "learning_rate": 5e-05, "loss": 0.8659, "mean_token_accuracy": 0.8541019558906555, "num_tokens": 220574874.0, "step": 5808 }, { "epoch": 0.7389645083322732, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.198285102844238, "learning_rate": 5e-05, "loss": 0.6717, "mean_token_accuracy": 0.8735191226005554, "num_tokens": 220615699.0, "step": 5809 }, { "epoch": 0.7390917186108638, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.923742294311523, "learning_rate": 5e-05, "loss": 0.9443, "mean_token_accuracy": 0.8531203269958496, "num_tokens": 220654517.0, "step": 5810 }, { "epoch": 0.7392189288894543, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.15200138092041, "learning_rate": 5e-05, "loss": 0.821, "mean_token_accuracy": 0.8336042165756226, "num_tokens": 220695915.0, "step": 5811 }, { "epoch": 0.7393461391680448, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.062867164611816, "learning_rate": 5e-05, "loss": 0.9743, "mean_token_accuracy": 0.8579690456390381, "num_tokens": 220734880.0, "step": 5812 }, { "epoch": 0.7394733494466353, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.801400661468506, "learning_rate": 5e-05, "loss": 0.8373, "mean_token_accuracy": 0.8565121293067932, "num_tokens": 220767800.0, "step": 5813 }, { "epoch": 0.7396005597252258, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.231701850891113, "learning_rate": 5e-05, "loss": 0.9549, "mean_token_accuracy": 0.8486363887786865, "num_tokens": 220812253.0, "step": 5814 }, { "epoch": 0.7397277700038163, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.9140777587890625, "learning_rate": 5e-05, "loss": 0.8699, "mean_token_accuracy": 0.8733932375907898, "num_tokens": 220847777.0, "step": 5815 }, { "epoch": 0.7398549802824068, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.662442207336426, "learning_rate": 5e-05, "loss": 0.7435, "mean_token_accuracy": 0.8593788146972656, "num_tokens": 220884974.0, "step": 5816 }, { "epoch": 0.7399821905609973, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.232751846313477, "learning_rate": 5e-05, "loss": 0.9833, "mean_token_accuracy": 0.8574477434158325, "num_tokens": 220926236.0, "step": 5817 }, { "epoch": 0.7401094008395879, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.608704566955566, "learning_rate": 5e-05, "loss": 0.7496, "mean_token_accuracy": 0.8429890871047974, "num_tokens": 220967985.0, "step": 5818 }, { "epoch": 0.7402366111181784, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.675049781799316, "learning_rate": 5e-05, "loss": 1.1334, "mean_token_accuracy": 0.8581331372261047, "num_tokens": 221004308.0, "step": 5819 }, { "epoch": 0.7403638213967688, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 6.610214710235596, "learning_rate": 5e-05, "loss": 0.8309, "mean_token_accuracy": 0.877489447593689, "num_tokens": 221043455.0, "step": 5820 }, { "epoch": 0.7404910316753593, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 11.093131065368652, "learning_rate": 5e-05, "loss": 1.2853, "mean_token_accuracy": 0.8495211601257324, "num_tokens": 221084634.0, "step": 5821 }, { "epoch": 0.7406182419539499, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000827789306640625, "grad_norm": 9.07580852508545, "learning_rate": 5e-05, "loss": 1.281, "mean_token_accuracy": 0.8612518310546875, "num_tokens": 221117350.0, "step": 5822 }, { "epoch": 0.7407454522325404, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.294478416442871, "learning_rate": 5e-05, "loss": 0.9922, "mean_token_accuracy": 0.8580486178398132, "num_tokens": 221150864.0, "step": 5823 }, { "epoch": 0.7408726625111309, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000766754150390625, "grad_norm": 9.536299705505371, "learning_rate": 5e-05, "loss": 1.1766, "mean_token_accuracy": 0.8720067143440247, "num_tokens": 221186501.0, "step": 5824 }, { "epoch": 0.7409998727897215, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.782040119171143, "learning_rate": 5e-05, "loss": 0.9975, "mean_token_accuracy": 0.8402087092399597, "num_tokens": 221227734.0, "step": 5825 }, { "epoch": 0.7411270830683119, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.197013854980469, "learning_rate": 5e-05, "loss": 1.1187, "mean_token_accuracy": 0.8600153923034668, "num_tokens": 221260615.0, "step": 5826 }, { "epoch": 0.7412542933469024, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.804341793060303, "learning_rate": 5e-05, "loss": 1.0412, "mean_token_accuracy": 0.8511540293693542, "num_tokens": 221298553.0, "step": 5827 }, { "epoch": 0.7413815036254929, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.961423873901367, "learning_rate": 5e-05, "loss": 1.0461, "mean_token_accuracy": 0.8617504835128784, "num_tokens": 221338803.0, "step": 5828 }, { "epoch": 0.7415087139040835, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.187506675720215, "learning_rate": 5e-05, "loss": 1.0386, "mean_token_accuracy": 0.8619393110275269, "num_tokens": 221379557.0, "step": 5829 }, { "epoch": 0.741635924182674, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.020345687866211, "learning_rate": 5e-05, "loss": 0.9778, "mean_token_accuracy": 0.8612900376319885, "num_tokens": 221412242.0, "step": 5830 }, { "epoch": 0.7417631344612645, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.228281021118164, "learning_rate": 5e-05, "loss": 0.9413, "mean_token_accuracy": 0.8747751712799072, "num_tokens": 221449382.0, "step": 5831 }, { "epoch": 0.7418903447398549, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.101738929748535, "learning_rate": 5e-05, "loss": 0.9387, "mean_token_accuracy": 0.850674033164978, "num_tokens": 221489480.0, "step": 5832 }, { "epoch": 0.7420175550184455, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.224361419677734, "learning_rate": 5e-05, "loss": 0.9488, "mean_token_accuracy": 0.8559994697570801, "num_tokens": 221525264.0, "step": 5833 }, { "epoch": 0.742144765297036, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.522512435913086, "learning_rate": 5e-05, "loss": 0.8335, "mean_token_accuracy": 0.865381121635437, "num_tokens": 221568447.0, "step": 5834 }, { "epoch": 0.7422719755756265, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.173818588256836, "learning_rate": 5e-05, "loss": 0.9691, "mean_token_accuracy": 0.8354388475418091, "num_tokens": 221612943.0, "step": 5835 }, { "epoch": 0.742399185854217, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.286005973815918, "learning_rate": 5e-05, "loss": 0.9245, "mean_token_accuracy": 0.828803300857544, "num_tokens": 221645449.0, "step": 5836 }, { "epoch": 0.7425263961328076, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.032909393310547, "learning_rate": 5e-05, "loss": 0.8157, "mean_token_accuracy": 0.8722805976867676, "num_tokens": 221684041.0, "step": 5837 }, { "epoch": 0.742653606411398, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.194231033325195, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8686432242393494, "num_tokens": 221720200.0, "step": 5838 }, { "epoch": 0.7427808166899885, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.904126167297363, "learning_rate": 5e-05, "loss": 0.8343, "mean_token_accuracy": 0.85789954662323, "num_tokens": 221758997.0, "step": 5839 }, { "epoch": 0.742908026968579, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.171475887298584, "learning_rate": 5e-05, "loss": 0.7827, "mean_token_accuracy": 0.861190915107727, "num_tokens": 221800713.0, "step": 5840 }, { "epoch": 0.7430352372471696, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.736883640289307, "learning_rate": 5e-05, "loss": 0.8562, "mean_token_accuracy": 0.8479050993919373, "num_tokens": 221840336.0, "step": 5841 }, { "epoch": 0.7431624475257601, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.295557498931885, "learning_rate": 5e-05, "loss": 0.7138, "mean_token_accuracy": 0.8783600926399231, "num_tokens": 221876286.0, "step": 5842 }, { "epoch": 0.7432896578043506, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.446173191070557, "learning_rate": 5e-05, "loss": 0.7949, "mean_token_accuracy": 0.8528687953948975, "num_tokens": 221917596.0, "step": 5843 }, { "epoch": 0.743416868082941, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.548191070556641, "learning_rate": 5e-05, "loss": 0.7629, "mean_token_accuracy": 0.866862416267395, "num_tokens": 221954045.0, "step": 5844 }, { "epoch": 0.7435440783615316, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.074573516845703, "learning_rate": 5e-05, "loss": 0.7318, "mean_token_accuracy": 0.8635240793228149, "num_tokens": 221996493.0, "step": 5845 }, { "epoch": 0.7436712886401221, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.928210258483887, "learning_rate": 5e-05, "loss": 0.8122, "mean_token_accuracy": 0.8575596213340759, "num_tokens": 222036608.0, "step": 5846 }, { "epoch": 0.7437984989187126, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.636934757232666, "learning_rate": 5e-05, "loss": 0.7095, "mean_token_accuracy": 0.8637300133705139, "num_tokens": 222080647.0, "step": 5847 }, { "epoch": 0.7439257091973032, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.48523235321045, "learning_rate": 5e-05, "loss": 0.8891, "mean_token_accuracy": 0.8503394722938538, "num_tokens": 222114039.0, "step": 5848 }, { "epoch": 0.7440529194758937, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.1441330909729, "learning_rate": 5e-05, "loss": 0.7356, "mean_token_accuracy": 0.8518467545509338, "num_tokens": 222149233.0, "step": 5849 }, { "epoch": 0.7441801297544841, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.232845306396484, "learning_rate": 5e-05, "loss": 0.8988, "mean_token_accuracy": 0.8731661438941956, "num_tokens": 222185842.0, "step": 5850 }, { "epoch": 0.7443073400330746, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 5.895451545715332, "learning_rate": 5e-05, "loss": 0.7034, "mean_token_accuracy": 0.8650197386741638, "num_tokens": 222233576.0, "step": 5851 }, { "epoch": 0.7444345503116652, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.913237571716309, "learning_rate": 5e-05, "loss": 1.0088, "mean_token_accuracy": 0.8630015254020691, "num_tokens": 222267822.0, "step": 5852 }, { "epoch": 0.7445617605902557, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 6.700259208679199, "learning_rate": 5e-05, "loss": 0.8361, "mean_token_accuracy": 0.8709051609039307, "num_tokens": 222304694.0, "step": 5853 }, { "epoch": 0.7446889708688462, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.668559074401855, "learning_rate": 5e-05, "loss": 1.077, "mean_token_accuracy": 0.8509005904197693, "num_tokens": 222340793.0, "step": 5854 }, { "epoch": 0.7448161811474368, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.06214714050293, "learning_rate": 5e-05, "loss": 1.0504, "mean_token_accuracy": 0.8559832572937012, "num_tokens": 222380316.0, "step": 5855 }, { "epoch": 0.7449433914260272, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.139763832092285, "learning_rate": 5e-05, "loss": 0.8776, "mean_token_accuracy": 0.8584079146385193, "num_tokens": 222418867.0, "step": 5856 }, { "epoch": 0.7450706017046177, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.883293151855469, "learning_rate": 5e-05, "loss": 1.0109, "mean_token_accuracy": 0.8655728697776794, "num_tokens": 222456585.0, "step": 5857 }, { "epoch": 0.7451978119832082, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.831241607666016, "learning_rate": 5e-05, "loss": 0.7641, "mean_token_accuracy": 0.8689307570457458, "num_tokens": 222492894.0, "step": 5858 }, { "epoch": 0.7453250222617988, "ewc_loss": 0.625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.567410469055176, "learning_rate": 5e-05, "loss": 1.036, "mean_token_accuracy": 0.8618201613426208, "num_tokens": 222538214.0, "step": 5859 }, { "epoch": 0.7454522325403893, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.0075531005859375, "learning_rate": 5e-05, "loss": 0.895, "mean_token_accuracy": 0.8591009378433228, "num_tokens": 222577643.0, "step": 5860 }, { "epoch": 0.7455794428189798, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.335553169250488, "learning_rate": 5e-05, "loss": 1.0323, "mean_token_accuracy": 0.8583604693412781, "num_tokens": 222615080.0, "step": 5861 }, { "epoch": 0.7457066530975703, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 7.838758945465088, "learning_rate": 5e-05, "loss": 0.9656, "mean_token_accuracy": 0.8698889017105103, "num_tokens": 222650358.0, "step": 5862 }, { "epoch": 0.7458338633761608, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.295720100402832, "learning_rate": 5e-05, "loss": 0.9496, "mean_token_accuracy": 0.8468507528305054, "num_tokens": 222691558.0, "step": 5863 }, { "epoch": 0.7459610736547513, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.157794952392578, "learning_rate": 5e-05, "loss": 0.9878, "mean_token_accuracy": 0.8593797087669373, "num_tokens": 222734632.0, "step": 5864 }, { "epoch": 0.7460882839333418, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.58550500869751, "learning_rate": 5e-05, "loss": 0.8394, "mean_token_accuracy": 0.855957567691803, "num_tokens": 222769422.0, "step": 5865 }, { "epoch": 0.7462154942119323, "ewc_loss": 0.5, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.072507858276367, "learning_rate": 5e-05, "loss": 0.8571, "mean_token_accuracy": 0.8808207511901855, "num_tokens": 222808667.0, "step": 5866 }, { "epoch": 0.7463427044905229, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.437101364135742, "learning_rate": 5e-05, "loss": 0.8069, "mean_token_accuracy": 0.8653580546379089, "num_tokens": 222844291.0, "step": 5867 }, { "epoch": 0.7464699147691134, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.800372123718262, "learning_rate": 5e-05, "loss": 0.9082, "mean_token_accuracy": 0.844974160194397, "num_tokens": 222885812.0, "step": 5868 }, { "epoch": 0.7465971250477038, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.45645809173584, "learning_rate": 5e-05, "loss": 0.8169, "mean_token_accuracy": 0.8614274263381958, "num_tokens": 222920768.0, "step": 5869 }, { "epoch": 0.7467243353262943, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.5486063957214355, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8635488748550415, "num_tokens": 222961973.0, "step": 5870 }, { "epoch": 0.7468515456048849, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.5322723388671875, "learning_rate": 5e-05, "loss": 0.7881, "mean_token_accuracy": 0.8661595582962036, "num_tokens": 222997356.0, "step": 5871 }, { "epoch": 0.7469787558834754, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.287569522857666, "learning_rate": 5e-05, "loss": 0.7503, "mean_token_accuracy": 0.8663277626037598, "num_tokens": 223035933.0, "step": 5872 }, { "epoch": 0.7471059661620659, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.723320960998535, "learning_rate": 5e-05, "loss": 0.8307, "mean_token_accuracy": 0.8525432348251343, "num_tokens": 223074597.0, "step": 5873 }, { "epoch": 0.7472331764406565, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.002688884735107, "learning_rate": 5e-05, "loss": 0.7001, "mean_token_accuracy": 0.8738582730293274, "num_tokens": 223113736.0, "step": 5874 }, { "epoch": 0.7473603867192469, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.999157905578613, "learning_rate": 5e-05, "loss": 0.8343, "mean_token_accuracy": 0.8535386323928833, "num_tokens": 223153974.0, "step": 5875 }, { "epoch": 0.7474875969978374, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.716415882110596, "learning_rate": 5e-05, "loss": 0.7113, "mean_token_accuracy": 0.8682209253311157, "num_tokens": 223187043.0, "step": 5876 }, { "epoch": 0.7476148072764279, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.289246559143066, "learning_rate": 5e-05, "loss": 0.8045, "mean_token_accuracy": 0.8740354776382446, "num_tokens": 223221265.0, "step": 5877 }, { "epoch": 0.7477420175550185, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.4912590980529785, "learning_rate": 5e-05, "loss": 0.7263, "mean_token_accuracy": 0.856877326965332, "num_tokens": 223262167.0, "step": 5878 }, { "epoch": 0.747869227833609, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.607540130615234, "learning_rate": 5e-05, "loss": 0.9236, "mean_token_accuracy": 0.85012286901474, "num_tokens": 223302333.0, "step": 5879 }, { "epoch": 0.7479964381121995, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.321811676025391, "learning_rate": 5e-05, "loss": 0.6979, "mean_token_accuracy": 0.8677440285682678, "num_tokens": 223343337.0, "step": 5880 }, { "epoch": 0.7481236483907899, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.930532455444336, "learning_rate": 5e-05, "loss": 0.8823, "mean_token_accuracy": 0.8742823600769043, "num_tokens": 223382905.0, "step": 5881 }, { "epoch": 0.7482508586693805, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.38523006439209, "learning_rate": 5e-05, "loss": 0.7245, "mean_token_accuracy": 0.8647881746292114, "num_tokens": 223422177.0, "step": 5882 }, { "epoch": 0.748378068947971, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.990610122680664, "learning_rate": 5e-05, "loss": 0.968, "mean_token_accuracy": 0.8596758842468262, "num_tokens": 223465725.0, "step": 5883 }, { "epoch": 0.7485052792265615, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 6.8204264640808105, "learning_rate": 5e-05, "loss": 0.8476, "mean_token_accuracy": 0.8541833162307739, "num_tokens": 223507246.0, "step": 5884 }, { "epoch": 0.748632489505152, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.58676528930664, "learning_rate": 5e-05, "loss": 0.9092, "mean_token_accuracy": 0.8676837682723999, "num_tokens": 223550736.0, "step": 5885 }, { "epoch": 0.7487596997837426, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.373729228973389, "learning_rate": 5e-05, "loss": 0.9434, "mean_token_accuracy": 0.8467739224433899, "num_tokens": 223588096.0, "step": 5886 }, { "epoch": 0.748886910062333, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.900430202484131, "learning_rate": 5e-05, "loss": 0.8751, "mean_token_accuracy": 0.8594447374343872, "num_tokens": 223629414.0, "step": 5887 }, { "epoch": 0.7490141203409235, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.64866304397583, "learning_rate": 5e-05, "loss": 0.8881, "mean_token_accuracy": 0.8587903380393982, "num_tokens": 223664468.0, "step": 5888 }, { "epoch": 0.749141330619514, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.510940074920654, "learning_rate": 5e-05, "loss": 0.7774, "mean_token_accuracy": 0.8696843385696411, "num_tokens": 223699832.0, "step": 5889 }, { "epoch": 0.7492685408981046, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.555606365203857, "learning_rate": 5e-05, "loss": 0.8432, "mean_token_accuracy": 0.8589819073677063, "num_tokens": 223736774.0, "step": 5890 }, { "epoch": 0.7493957511766951, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.554107666015625, "learning_rate": 5e-05, "loss": 0.7996, "mean_token_accuracy": 0.8632758855819702, "num_tokens": 223774622.0, "step": 5891 }, { "epoch": 0.7495229614552856, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.254273891448975, "learning_rate": 5e-05, "loss": 0.7962, "mean_token_accuracy": 0.8564354181289673, "num_tokens": 223816845.0, "step": 5892 }, { "epoch": 0.749650171733876, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.780381679534912, "learning_rate": 5e-05, "loss": 0.7858, "mean_token_accuracy": 0.8761496543884277, "num_tokens": 223849964.0, "step": 5893 }, { "epoch": 0.7497773820124666, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.978352069854736, "learning_rate": 5e-05, "loss": 0.7814, "mean_token_accuracy": 0.8499510884284973, "num_tokens": 223887821.0, "step": 5894 }, { "epoch": 0.7499045922910571, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.950606346130371, "learning_rate": 5e-05, "loss": 0.7736, "mean_token_accuracy": 0.8752843141555786, "num_tokens": 223929420.0, "step": 5895 }, { "epoch": 0.7500318025696476, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.824394702911377, "learning_rate": 5e-05, "loss": 0.7169, "mean_token_accuracy": 0.8667088747024536, "num_tokens": 223964910.0, "step": 5896 }, { "epoch": 0.7501590128482382, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.090211868286133, "learning_rate": 5e-05, "loss": 0.8871, "mean_token_accuracy": 0.8427579402923584, "num_tokens": 224001131.0, "step": 5897 }, { "epoch": 0.7502862231268287, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.728716850280762, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8442980647087097, "num_tokens": 224038969.0, "step": 5898 }, { "epoch": 0.7504134334054191, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.182991981506348, "learning_rate": 5e-05, "loss": 0.8836, "mean_token_accuracy": 0.8498574495315552, "num_tokens": 224081044.0, "step": 5899 }, { "epoch": 0.7505406436840096, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.734689712524414, "learning_rate": 5e-05, "loss": 0.8312, "mean_token_accuracy": 0.8331202268600464, "num_tokens": 224117967.0, "step": 5900 }, { "epoch": 0.7506678539626002, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.166519165039062, "learning_rate": 5e-05, "loss": 0.8981, "mean_token_accuracy": 0.8445816040039062, "num_tokens": 224148890.0, "step": 5901 }, { "epoch": 0.7507950642411907, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003681182861328125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.864018440246582, "learning_rate": 5e-05, "loss": 0.8082, "mean_token_accuracy": 0.8427685499191284, "num_tokens": 224183592.0, "step": 5902 }, { "epoch": 0.7509222745197812, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.950864315032959, "learning_rate": 5e-05, "loss": 0.8142, "mean_token_accuracy": 0.8653577566146851, "num_tokens": 224225983.0, "step": 5903 }, { "epoch": 0.7510494847983717, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.10963773727417, "learning_rate": 5e-05, "loss": 0.8234, "mean_token_accuracy": 0.8473984003067017, "num_tokens": 224267068.0, "step": 5904 }, { "epoch": 0.7511766950769622, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.60202693939209, "learning_rate": 5e-05, "loss": 0.7388, "mean_token_accuracy": 0.8783726096153259, "num_tokens": 224313060.0, "step": 5905 }, { "epoch": 0.7513039053555527, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.41132926940918, "learning_rate": 5e-05, "loss": 0.7956, "mean_token_accuracy": 0.854576587677002, "num_tokens": 224355066.0, "step": 5906 }, { "epoch": 0.7514311156341432, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.245502948760986, "learning_rate": 5e-05, "loss": 0.732, "mean_token_accuracy": 0.871518611907959, "num_tokens": 224383930.0, "step": 5907 }, { "epoch": 0.7515583259127337, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.674248695373535, "learning_rate": 5e-05, "loss": 0.8362, "mean_token_accuracy": 0.8438169956207275, "num_tokens": 224418462.0, "step": 5908 }, { "epoch": 0.7516855361913243, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.931127071380615, "learning_rate": 5e-05, "loss": 0.7218, "mean_token_accuracy": 0.8643763065338135, "num_tokens": 224453672.0, "step": 5909 }, { "epoch": 0.7518127464699148, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.9411725997924805, "learning_rate": 5e-05, "loss": 0.7809, "mean_token_accuracy": 0.8637403249740601, "num_tokens": 224494093.0, "step": 5910 }, { "epoch": 0.7519399567485053, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.640678882598877, "learning_rate": 5e-05, "loss": 0.7153, "mean_token_accuracy": 0.8629217147827148, "num_tokens": 224531829.0, "step": 5911 }, { "epoch": 0.7520671670270958, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.291647911071777, "learning_rate": 5e-05, "loss": 0.8574, "mean_token_accuracy": 0.8511176705360413, "num_tokens": 224569403.0, "step": 5912 }, { "epoch": 0.7521943773056863, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.335650444030762, "learning_rate": 5e-05, "loss": 0.7327, "mean_token_accuracy": 0.8543701171875, "num_tokens": 224605189.0, "step": 5913 }, { "epoch": 0.7523215875842768, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.722152709960938, "learning_rate": 5e-05, "loss": 0.8993, "mean_token_accuracy": 0.8572721481323242, "num_tokens": 224638729.0, "step": 5914 }, { "epoch": 0.7524487978628673, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.093367099761963, "learning_rate": 5e-05, "loss": 0.6885, "mean_token_accuracy": 0.8621869087219238, "num_tokens": 224673442.0, "step": 5915 }, { "epoch": 0.7525760081414579, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.130669593811035, "learning_rate": 5e-05, "loss": 1.0065, "mean_token_accuracy": 0.8477756381034851, "num_tokens": 224705945.0, "step": 5916 }, { "epoch": 0.7527032184200484, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.297566890716553, "learning_rate": 5e-05, "loss": 0.7449, "mean_token_accuracy": 0.8600059747695923, "num_tokens": 224740035.0, "step": 5917 }, { "epoch": 0.7528304286986388, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.032602310180664, "learning_rate": 5e-05, "loss": 0.9566, "mean_token_accuracy": 0.8728141784667969, "num_tokens": 224782047.0, "step": 5918 }, { "epoch": 0.7529576389772293, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.141137599945068, "learning_rate": 5e-05, "loss": 0.8874, "mean_token_accuracy": 0.8543261289596558, "num_tokens": 224821755.0, "step": 5919 }, { "epoch": 0.7530848492558199, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.027766227722168, "learning_rate": 5e-05, "loss": 0.8667, "mean_token_accuracy": 0.8691150546073914, "num_tokens": 224854822.0, "step": 5920 }, { "epoch": 0.7532120595344104, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.033991813659668, "learning_rate": 5e-05, "loss": 0.9214, "mean_token_accuracy": 0.8655901551246643, "num_tokens": 224899570.0, "step": 5921 }, { "epoch": 0.7533392698130009, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.8734130859375, "learning_rate": 5e-05, "loss": 0.7595, "mean_token_accuracy": 0.8635039329528809, "num_tokens": 224934897.0, "step": 5922 }, { "epoch": 0.7534664800915915, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.6085786819458, "learning_rate": 5e-05, "loss": 0.8843, "mean_token_accuracy": 0.8716861009597778, "num_tokens": 224972866.0, "step": 5923 }, { "epoch": 0.7535936903701819, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.536154747009277, "learning_rate": 5e-05, "loss": 0.8023, "mean_token_accuracy": 0.8456999063491821, "num_tokens": 225015075.0, "step": 5924 }, { "epoch": 0.7537209006487724, "ewc_loss": 0.5, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.521960258483887, "learning_rate": 5e-05, "loss": 0.9069, "mean_token_accuracy": 0.8622381091117859, "num_tokens": 225046729.0, "step": 5925 }, { "epoch": 0.7538481109273629, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.943320274353027, "learning_rate": 5e-05, "loss": 0.8252, "mean_token_accuracy": 0.8543292284011841, "num_tokens": 225080117.0, "step": 5926 }, { "epoch": 0.7539753212059535, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.886209964752197, "learning_rate": 5e-05, "loss": 0.8619, "mean_token_accuracy": 0.8603636622428894, "num_tokens": 225116854.0, "step": 5927 }, { "epoch": 0.754102531484544, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.44915771484375, "learning_rate": 5e-05, "loss": 0.8061, "mean_token_accuracy": 0.8685626983642578, "num_tokens": 225154044.0, "step": 5928 }, { "epoch": 0.7542297417631345, "ewc_loss": 0.375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.33152437210083, "learning_rate": 5e-05, "loss": 0.7372, "mean_token_accuracy": 0.8800872564315796, "num_tokens": 225191393.0, "step": 5929 }, { "epoch": 0.7543569520417249, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.681745529174805, "learning_rate": 5e-05, "loss": 0.8597, "mean_token_accuracy": 0.850062370300293, "num_tokens": 225232577.0, "step": 5930 }, { "epoch": 0.7544841623203155, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.173961639404297, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.8468868732452393, "num_tokens": 225270458.0, "step": 5931 }, { "epoch": 0.754611372598906, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.616127014160156, "learning_rate": 5e-05, "loss": 0.8117, "mean_token_accuracy": 0.8598587512969971, "num_tokens": 225306777.0, "step": 5932 }, { "epoch": 0.7547385828774965, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.2329301834106445, "learning_rate": 5e-05, "loss": 0.742, "mean_token_accuracy": 0.8687955141067505, "num_tokens": 225345602.0, "step": 5933 }, { "epoch": 0.754865793156087, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.460416793823242, "learning_rate": 5e-05, "loss": 0.7489, "mean_token_accuracy": 0.8678590059280396, "num_tokens": 225378885.0, "step": 5934 }, { "epoch": 0.7549930034346776, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.328667163848877, "learning_rate": 5e-05, "loss": 0.7546, "mean_token_accuracy": 0.8615001440048218, "num_tokens": 225408421.0, "step": 5935 }, { "epoch": 0.755120213713268, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.3552703857421875, "learning_rate": 5e-05, "loss": 0.7464, "mean_token_accuracy": 0.8613622188568115, "num_tokens": 225441428.0, "step": 5936 }, { "epoch": 0.7552474239918585, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.334495544433594, "learning_rate": 5e-05, "loss": 0.7873, "mean_token_accuracy": 0.8505626916885376, "num_tokens": 225474912.0, "step": 5937 }, { "epoch": 0.755374634270449, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.341892719268799, "learning_rate": 5e-05, "loss": 0.774, "mean_token_accuracy": 0.8568196296691895, "num_tokens": 225511876.0, "step": 5938 }, { "epoch": 0.7555018445490396, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.296938896179199, "learning_rate": 5e-05, "loss": 0.7597, "mean_token_accuracy": 0.8589615225791931, "num_tokens": 225548754.0, "step": 5939 }, { "epoch": 0.7556290548276301, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.364332675933838, "learning_rate": 5e-05, "loss": 0.7754, "mean_token_accuracy": 0.8506213426589966, "num_tokens": 225585554.0, "step": 5940 }, { "epoch": 0.7557562651062206, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.243186950683594, "learning_rate": 5e-05, "loss": 0.7212, "mean_token_accuracy": 0.8666737079620361, "num_tokens": 225622727.0, "step": 5941 }, { "epoch": 0.755883475384811, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.403649806976318, "learning_rate": 5e-05, "loss": 0.7102, "mean_token_accuracy": 0.870672345161438, "num_tokens": 225662163.0, "step": 5942 }, { "epoch": 0.7560106856634016, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.175779342651367, "learning_rate": 5e-05, "loss": 0.7654, "mean_token_accuracy": 0.8488957285881042, "num_tokens": 225698932.0, "step": 5943 }, { "epoch": 0.7561378959419921, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.46586275100708, "learning_rate": 5e-05, "loss": 0.7193, "mean_token_accuracy": 0.8725699186325073, "num_tokens": 225735587.0, "step": 5944 }, { "epoch": 0.7562651062205826, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.065707206726074, "learning_rate": 5e-05, "loss": 0.7147, "mean_token_accuracy": 0.8730741739273071, "num_tokens": 225777013.0, "step": 5945 }, { "epoch": 0.7563923164991732, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.608911991119385, "learning_rate": 5e-05, "loss": 0.7372, "mean_token_accuracy": 0.8638594150543213, "num_tokens": 225808496.0, "step": 5946 }, { "epoch": 0.7565195267777637, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.863893985748291, "learning_rate": 5e-05, "loss": 0.6938, "mean_token_accuracy": 0.8683778643608093, "num_tokens": 225848629.0, "step": 5947 }, { "epoch": 0.7566467370563541, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.905163764953613, "learning_rate": 5e-05, "loss": 0.7452, "mean_token_accuracy": 0.869810163974762, "num_tokens": 225882284.0, "step": 5948 }, { "epoch": 0.7567739473349446, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00023651123046875, "grad_norm": 6.460954666137695, "learning_rate": 5e-05, "loss": 0.7162, "mean_token_accuracy": 0.8513170480728149, "num_tokens": 225919294.0, "step": 5949 }, { "epoch": 0.7569011576135352, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.529070854187012, "learning_rate": 5e-05, "loss": 0.8722, "mean_token_accuracy": 0.8470929265022278, "num_tokens": 225954742.0, "step": 5950 }, { "epoch": 0.7570283678921257, "ewc_loss": 0.244140625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00020694732666015625, "grad_norm": 5.754400730133057, "learning_rate": 5e-05, "loss": 0.5824, "mean_token_accuracy": 0.8846327066421509, "num_tokens": 225992532.0, "step": 5951 }, { "epoch": 0.7571555781707162, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.760594367980957, "learning_rate": 5e-05, "loss": 0.9868, "mean_token_accuracy": 0.8620895147323608, "num_tokens": 226034272.0, "step": 5952 }, { "epoch": 0.7572827884493067, "ewc_loss": 0.25390625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000217437744140625, "grad_norm": 5.094202995300293, "learning_rate": 5e-05, "loss": 0.6765, "mean_token_accuracy": 0.8559649586677551, "num_tokens": 226066607.0, "step": 5953 }, { "epoch": 0.7574099987278972, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 11.639101028442383, "learning_rate": 5e-05, "loss": 1.2994, "mean_token_accuracy": 0.85317063331604, "num_tokens": 226103077.0, "step": 5954 }, { "epoch": 0.7575372090064877, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 7.171214580535889, "learning_rate": 5e-05, "loss": 0.9326, "mean_token_accuracy": 0.8787664175033569, "num_tokens": 226140488.0, "step": 5955 }, { "epoch": 0.7576644192850782, "ewc_loss": 0.953125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 12.377671241760254, "learning_rate": 5e-05, "loss": 1.3078, "mean_token_accuracy": 0.8778194785118103, "num_tokens": 226177139.0, "step": 5956 }, { "epoch": 0.7577916295636687, "ewc_loss": 1.09375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0010528564453125, "grad_norm": 10.093289375305176, "learning_rate": 5e-05, "loss": 1.4931, "mean_token_accuracy": 0.8648989200592041, "num_tokens": 226216340.0, "step": 5957 }, { "epoch": 0.7579188398422593, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.966451644897461, "learning_rate": 5e-05, "loss": 1.037, "mean_token_accuracy": 0.879605770111084, "num_tokens": 226253507.0, "step": 5958 }, { "epoch": 0.7580460501208498, "ewc_loss": 0.88671875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00084686279296875, "grad_norm": 9.544153213500977, "learning_rate": 5e-05, "loss": 1.3141, "mean_token_accuracy": 0.8562561273574829, "num_tokens": 226296617.0, "step": 5959 }, { "epoch": 0.7581732603994403, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.102184295654297, "learning_rate": 5e-05, "loss": 1.0963, "mean_token_accuracy": 0.8565740585327148, "num_tokens": 226335483.0, "step": 5960 }, { "epoch": 0.7583004706780307, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 8.958978652954102, "learning_rate": 5e-05, "loss": 1.0984, "mean_token_accuracy": 0.8490354418754578, "num_tokens": 226373773.0, "step": 5961 }, { "epoch": 0.7584276809566213, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.915053367614746, "learning_rate": 5e-05, "loss": 1.0007, "mean_token_accuracy": 0.8638734817504883, "num_tokens": 226410333.0, "step": 5962 }, { "epoch": 0.7585548912352118, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.751751899719238, "learning_rate": 5e-05, "loss": 0.994, "mean_token_accuracy": 0.86240553855896, "num_tokens": 226451058.0, "step": 5963 }, { "epoch": 0.7586821015138023, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.032171249389648, "learning_rate": 5e-05, "loss": 0.8957, "mean_token_accuracy": 0.8664319515228271, "num_tokens": 226487099.0, "step": 5964 }, { "epoch": 0.7588093117923929, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.707329750061035, "learning_rate": 5e-05, "loss": 0.9441, "mean_token_accuracy": 0.8665710091590881, "num_tokens": 226521462.0, "step": 5965 }, { "epoch": 0.7589365220709834, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.417612552642822, "learning_rate": 5e-05, "loss": 0.8526, "mean_token_accuracy": 0.8612449765205383, "num_tokens": 226560573.0, "step": 5966 }, { "epoch": 0.7590637323495738, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.730046272277832, "learning_rate": 5e-05, "loss": 0.9656, "mean_token_accuracy": 0.85631263256073, "num_tokens": 226603717.0, "step": 5967 }, { "epoch": 0.7591909426281643, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.088685035705566, "learning_rate": 5e-05, "loss": 0.812, "mean_token_accuracy": 0.8664937615394592, "num_tokens": 226647049.0, "step": 5968 }, { "epoch": 0.7593181529067549, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.911590576171875, "learning_rate": 5e-05, "loss": 0.9314, "mean_token_accuracy": 0.8710771799087524, "num_tokens": 226682459.0, "step": 5969 }, { "epoch": 0.7594453631853454, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 6.9372453689575195, "learning_rate": 5e-05, "loss": 0.8376, "mean_token_accuracy": 0.8568387627601624, "num_tokens": 226727381.0, "step": 5970 }, { "epoch": 0.7595725734639359, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.098405838012695, "learning_rate": 5e-05, "loss": 1.0539, "mean_token_accuracy": 0.8437318205833435, "num_tokens": 226766732.0, "step": 5971 }, { "epoch": 0.7596997837425264, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.107060432434082, "learning_rate": 5e-05, "loss": 0.8994, "mean_token_accuracy": 0.856248140335083, "num_tokens": 226808165.0, "step": 5972 }, { "epoch": 0.7598269940211169, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.985828399658203, "learning_rate": 5e-05, "loss": 1.0195, "mean_token_accuracy": 0.8557336330413818, "num_tokens": 226848113.0, "step": 5973 }, { "epoch": 0.7599542042997074, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.508686542510986, "learning_rate": 5e-05, "loss": 0.996, "mean_token_accuracy": 0.8406668305397034, "num_tokens": 226889390.0, "step": 5974 }, { "epoch": 0.7600814145782979, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.413961410522461, "learning_rate": 5e-05, "loss": 0.9719, "mean_token_accuracy": 0.8502634763717651, "num_tokens": 226929111.0, "step": 5975 }, { "epoch": 0.7602086248568884, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.864811420440674, "learning_rate": 5e-05, "loss": 0.8822, "mean_token_accuracy": 0.8752235174179077, "num_tokens": 226966677.0, "step": 5976 }, { "epoch": 0.760335835135479, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.548737525939941, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.857845664024353, "num_tokens": 227004083.0, "step": 5977 }, { "epoch": 0.7604630454140695, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.169260025024414, "learning_rate": 5e-05, "loss": 0.9481, "mean_token_accuracy": 0.859153151512146, "num_tokens": 227049954.0, "step": 5978 }, { "epoch": 0.7605902556926599, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.864808559417725, "learning_rate": 5e-05, "loss": 0.7505, "mean_token_accuracy": 0.8617357611656189, "num_tokens": 227088532.0, "step": 5979 }, { "epoch": 0.7607174659712505, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000370025634765625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.521063804626465, "learning_rate": 5e-05, "loss": 0.8794, "mean_token_accuracy": 0.8777287006378174, "num_tokens": 227129471.0, "step": 5980 }, { "epoch": 0.760844676249841, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.4853901863098145, "learning_rate": 5e-05, "loss": 0.7388, "mean_token_accuracy": 0.8592329025268555, "num_tokens": 227167443.0, "step": 5981 }, { "epoch": 0.7609718865284315, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.06403636932373, "learning_rate": 5e-05, "loss": 0.9463, "mean_token_accuracy": 0.8703680038452148, "num_tokens": 227207577.0, "step": 5982 }, { "epoch": 0.761099096807022, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.378910541534424, "learning_rate": 5e-05, "loss": 0.7545, "mean_token_accuracy": 0.8616065382957458, "num_tokens": 227247424.0, "step": 5983 }, { "epoch": 0.7612263070856126, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.718659400939941, "learning_rate": 5e-05, "loss": 1.056, "mean_token_accuracy": 0.8639034032821655, "num_tokens": 227286361.0, "step": 5984 }, { "epoch": 0.761353517364203, "ewc_loss": 0.5, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.179243564605713, "learning_rate": 5e-05, "loss": 0.887, "mean_token_accuracy": 0.8669856786727905, "num_tokens": 227326786.0, "step": 5985 }, { "epoch": 0.7614807276427935, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.487075805664062, "learning_rate": 5e-05, "loss": 1.0869, "mean_token_accuracy": 0.851203441619873, "num_tokens": 227370237.0, "step": 5986 }, { "epoch": 0.761607937921384, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.333882331848145, "learning_rate": 5e-05, "loss": 0.9768, "mean_token_accuracy": 0.8804051876068115, "num_tokens": 227399711.0, "step": 5987 }, { "epoch": 0.7617351481999746, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.9923601150512695, "learning_rate": 5e-05, "loss": 0.8769, "mean_token_accuracy": 0.8685263395309448, "num_tokens": 227436468.0, "step": 5988 }, { "epoch": 0.7618623584785651, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.815428733825684, "learning_rate": 5e-05, "loss": 1.0287, "mean_token_accuracy": 0.8590618371963501, "num_tokens": 227476128.0, "step": 5989 }, { "epoch": 0.7619895687571556, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.910251140594482, "learning_rate": 5e-05, "loss": 0.7891, "mean_token_accuracy": 0.8624840378761292, "num_tokens": 227510499.0, "step": 5990 }, { "epoch": 0.762116779035746, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.975349426269531, "learning_rate": 5e-05, "loss": 0.992, "mean_token_accuracy": 0.859140932559967, "num_tokens": 227542886.0, "step": 5991 }, { "epoch": 0.7622439893143366, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.086147785186768, "learning_rate": 5e-05, "loss": 0.8003, "mean_token_accuracy": 0.8712266683578491, "num_tokens": 227583244.0, "step": 5992 }, { "epoch": 0.7623711995929271, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.386319160461426, "learning_rate": 5e-05, "loss": 0.9099, "mean_token_accuracy": 0.8697249889373779, "num_tokens": 227622690.0, "step": 5993 }, { "epoch": 0.7624984098715176, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.627045631408691, "learning_rate": 5e-05, "loss": 0.9005, "mean_token_accuracy": 0.8596017360687256, "num_tokens": 227663537.0, "step": 5994 }, { "epoch": 0.7626256201501082, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.549283027648926, "learning_rate": 5e-05, "loss": 0.8166, "mean_token_accuracy": 0.8624539375305176, "num_tokens": 227700679.0, "step": 5995 }, { "epoch": 0.7627528304286987, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.924222946166992, "learning_rate": 5e-05, "loss": 0.9242, "mean_token_accuracy": 0.8519536256790161, "num_tokens": 227738410.0, "step": 5996 }, { "epoch": 0.7628800407072891, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.108064651489258, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8662450909614563, "num_tokens": 227780849.0, "step": 5997 }, { "epoch": 0.7630072509858796, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.011187553405762, "learning_rate": 5e-05, "loss": 0.8409, "mean_token_accuracy": 0.8699631690979004, "num_tokens": 227815488.0, "step": 5998 }, { "epoch": 0.7631344612644702, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.957798480987549, "learning_rate": 5e-05, "loss": 0.7309, "mean_token_accuracy": 0.8690539002418518, "num_tokens": 227848627.0, "step": 5999 }, { "epoch": 0.7632616715430607, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.129261016845703, "learning_rate": 5e-05, "loss": 0.8374, "mean_token_accuracy": 0.8639440536499023, "num_tokens": 227888394.0, "step": 6000 }, { "epoch": 0.7633888818216512, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.788755416870117, "learning_rate": 5e-05, "loss": 0.7227, "mean_token_accuracy": 0.8711566925048828, "num_tokens": 227923785.0, "step": 6001 }, { "epoch": 0.7635160921002417, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.362792015075684, "learning_rate": 5e-05, "loss": 0.904, "mean_token_accuracy": 0.8538376688957214, "num_tokens": 227958714.0, "step": 6002 }, { "epoch": 0.7636433023788322, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.6423187255859375, "learning_rate": 5e-05, "loss": 0.8269, "mean_token_accuracy": 0.8349301815032959, "num_tokens": 228004414.0, "step": 6003 }, { "epoch": 0.7637705126574227, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.56064224243164, "learning_rate": 5e-05, "loss": 0.9564, "mean_token_accuracy": 0.8479118347167969, "num_tokens": 228044291.0, "step": 6004 }, { "epoch": 0.7638977229360132, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.647667407989502, "learning_rate": 5e-05, "loss": 0.7416, "mean_token_accuracy": 0.8675987124443054, "num_tokens": 228080656.0, "step": 6005 }, { "epoch": 0.7640249332146037, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.613997459411621, "learning_rate": 5e-05, "loss": 0.9211, "mean_token_accuracy": 0.8627710938453674, "num_tokens": 228117787.0, "step": 6006 }, { "epoch": 0.7641521434931943, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.813819885253906, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.8553426861763, "num_tokens": 228150620.0, "step": 6007 }, { "epoch": 0.7642793537717848, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.49145793914795, "learning_rate": 5e-05, "loss": 0.8986, "mean_token_accuracy": 0.8709594011306763, "num_tokens": 228188330.0, "step": 6008 }, { "epoch": 0.7644065640503753, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 6.999757289886475, "learning_rate": 5e-05, "loss": 0.7969, "mean_token_accuracy": 0.8690783977508545, "num_tokens": 228231369.0, "step": 6009 }, { "epoch": 0.7645337743289657, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.322169303894043, "learning_rate": 5e-05, "loss": 0.891, "mean_token_accuracy": 0.8708046674728394, "num_tokens": 228269321.0, "step": 6010 }, { "epoch": 0.7646609846075563, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.08466911315918, "learning_rate": 5e-05, "loss": 0.8014, "mean_token_accuracy": 0.8697752952575684, "num_tokens": 228305240.0, "step": 6011 }, { "epoch": 0.7647881948861468, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.218693733215332, "learning_rate": 5e-05, "loss": 0.8996, "mean_token_accuracy": 0.8627707958221436, "num_tokens": 228345100.0, "step": 6012 }, { "epoch": 0.7649154051647373, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.073539733886719, "learning_rate": 5e-05, "loss": 0.7876, "mean_token_accuracy": 0.8672174215316772, "num_tokens": 228379215.0, "step": 6013 }, { "epoch": 0.7650426154433279, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.19011116027832, "learning_rate": 5e-05, "loss": 0.9022, "mean_token_accuracy": 0.8608341217041016, "num_tokens": 228415954.0, "step": 6014 }, { "epoch": 0.7651698257219184, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.01605224609375, "learning_rate": 5e-05, "loss": 0.8364, "mean_token_accuracy": 0.849517822265625, "num_tokens": 228453569.0, "step": 6015 }, { "epoch": 0.7652970360005088, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.209423065185547, "learning_rate": 5e-05, "loss": 0.9558, "mean_token_accuracy": 0.8477358222007751, "num_tokens": 228491384.0, "step": 6016 }, { "epoch": 0.7654242462790993, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.994931221008301, "learning_rate": 5e-05, "loss": 0.8351, "mean_token_accuracy": 0.8453856706619263, "num_tokens": 228528118.0, "step": 6017 }, { "epoch": 0.7655514565576899, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.158342361450195, "learning_rate": 5e-05, "loss": 0.8986, "mean_token_accuracy": 0.8564622402191162, "num_tokens": 228568207.0, "step": 6018 }, { "epoch": 0.7656786668362804, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.079376220703125, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.8608555793762207, "num_tokens": 228604313.0, "step": 6019 }, { "epoch": 0.7658058771148709, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.955454349517822, "learning_rate": 5e-05, "loss": 0.8117, "mean_token_accuracy": 0.8773728013038635, "num_tokens": 228645615.0, "step": 6020 }, { "epoch": 0.7659330873934614, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.2427825927734375, "learning_rate": 5e-05, "loss": 0.8037, "mean_token_accuracy": 0.8577959537506104, "num_tokens": 228679137.0, "step": 6021 }, { "epoch": 0.7660602976720519, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.667104721069336, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8680605292320251, "num_tokens": 228717383.0, "step": 6022 }, { "epoch": 0.7661875079506424, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.3671112060546875, "learning_rate": 5e-05, "loss": 0.7545, "mean_token_accuracy": 0.8727664947509766, "num_tokens": 228747260.0, "step": 6023 }, { "epoch": 0.7663147182292329, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.5073466300964355, "learning_rate": 5e-05, "loss": 0.8055, "mean_token_accuracy": 0.8559778928756714, "num_tokens": 228780852.0, "step": 6024 }, { "epoch": 0.7664419285078234, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.395442485809326, "learning_rate": 5e-05, "loss": 0.7716, "mean_token_accuracy": 0.865287184715271, "num_tokens": 228820220.0, "step": 6025 }, { "epoch": 0.766569138786414, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.411565780639648, "learning_rate": 5e-05, "loss": 0.756, "mean_token_accuracy": 0.8686447143554688, "num_tokens": 228858903.0, "step": 6026 }, { "epoch": 0.7666963490650045, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.3666090965271, "learning_rate": 5e-05, "loss": 0.7359, "mean_token_accuracy": 0.8731695413589478, "num_tokens": 228894998.0, "step": 6027 }, { "epoch": 0.7668235593435949, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.384817123413086, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8667874336242676, "num_tokens": 228933333.0, "step": 6028 }, { "epoch": 0.7669507696221854, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.353126049041748, "learning_rate": 5e-05, "loss": 0.7501, "mean_token_accuracy": 0.8639019131660461, "num_tokens": 228972960.0, "step": 6029 }, { "epoch": 0.767077979900776, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.368860244750977, "learning_rate": 5e-05, "loss": 0.7867, "mean_token_accuracy": 0.8514453768730164, "num_tokens": 229010967.0, "step": 6030 }, { "epoch": 0.7672051901793665, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.30895471572876, "learning_rate": 5e-05, "loss": 0.7326, "mean_token_accuracy": 0.8667930364608765, "num_tokens": 229052424.0, "step": 6031 }, { "epoch": 0.767332400457957, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.377227783203125, "learning_rate": 5e-05, "loss": 0.7603, "mean_token_accuracy": 0.8586682081222534, "num_tokens": 229088106.0, "step": 6032 }, { "epoch": 0.7674596107365476, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.270554065704346, "learning_rate": 5e-05, "loss": 0.7409, "mean_token_accuracy": 0.8650210499763489, "num_tokens": 229122340.0, "step": 6033 }, { "epoch": 0.767586821015138, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.403538703918457, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8522897362709045, "num_tokens": 229158299.0, "step": 6034 }, { "epoch": 0.7677140312937285, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.2148847579956055, "learning_rate": 5e-05, "loss": 0.7283, "mean_token_accuracy": 0.8622273206710815, "num_tokens": 229187870.0, "step": 6035 }, { "epoch": 0.767841241572319, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.4445109367370605, "learning_rate": 5e-05, "loss": 0.7759, "mean_token_accuracy": 0.8496600389480591, "num_tokens": 229226459.0, "step": 6036 }, { "epoch": 0.7679684518509096, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.130504608154297, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.8622962236404419, "num_tokens": 229257973.0, "step": 6037 }, { "epoch": 0.7680956621295001, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.5485758781433105, "learning_rate": 5e-05, "loss": 0.7343, "mean_token_accuracy": 0.8675296306610107, "num_tokens": 229297870.0, "step": 6038 }, { "epoch": 0.7682228724080906, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.985222816467285, "learning_rate": 5e-05, "loss": 0.7491, "mean_token_accuracy": 0.8543797135353088, "num_tokens": 229331295.0, "step": 6039 }, { "epoch": 0.768350082686681, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.725894927978516, "learning_rate": 5e-05, "loss": 0.7708, "mean_token_accuracy": 0.8607525825500488, "num_tokens": 229369491.0, "step": 6040 }, { "epoch": 0.7684772929652716, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.745143890380859, "learning_rate": 5e-05, "loss": 0.7098, "mean_token_accuracy": 0.8557023406028748, "num_tokens": 229407836.0, "step": 6041 }, { "epoch": 0.7686045032438621, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.0629243850708, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8668999671936035, "num_tokens": 229444806.0, "step": 6042 }, { "epoch": 0.7687317135224526, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0002384185791015625, "grad_norm": 6.362232208251953, "learning_rate": 5e-05, "loss": 0.6712, "mean_token_accuracy": 0.8674657344818115, "num_tokens": 229485588.0, "step": 6043 }, { "epoch": 0.7688589238010431, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.657537460327148, "learning_rate": 5e-05, "loss": 0.8322, "mean_token_accuracy": 0.8679475784301758, "num_tokens": 229521924.0, "step": 6044 }, { "epoch": 0.7689861340796337, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000217437744140625, "grad_norm": 5.802401542663574, "learning_rate": 5e-05, "loss": 0.6441, "mean_token_accuracy": 0.8666768074035645, "num_tokens": 229561511.0, "step": 6045 }, { "epoch": 0.7691133443582241, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.583756446838379, "learning_rate": 5e-05, "loss": 0.969, "mean_token_accuracy": 0.8620573282241821, "num_tokens": 229607193.0, "step": 6046 }, { "epoch": 0.7692405546368146, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.595195293426514, "learning_rate": 5e-05, "loss": 0.6681, "mean_token_accuracy": 0.8721652030944824, "num_tokens": 229642775.0, "step": 6047 }, { "epoch": 0.7693677649154052, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.379400253295898, "learning_rate": 5e-05, "loss": 1.1324, "mean_token_accuracy": 0.846429705619812, "num_tokens": 229680960.0, "step": 6048 }, { "epoch": 0.7694949751939957, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.085839748382568, "learning_rate": 5e-05, "loss": 0.8816, "mean_token_accuracy": 0.8761245012283325, "num_tokens": 229713473.0, "step": 6049 }, { "epoch": 0.7696221854725862, "ewc_loss": 0.625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.583008766174316, "learning_rate": 5e-05, "loss": 0.9998, "mean_token_accuracy": 0.8727630972862244, "num_tokens": 229753919.0, "step": 6050 }, { "epoch": 0.7697493957511767, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 8.815061569213867, "learning_rate": 5e-05, "loss": 1.1807, "mean_token_accuracy": 0.8464732766151428, "num_tokens": 229788635.0, "step": 6051 }, { "epoch": 0.7698766060297672, "ewc_loss": 0.375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.128878593444824, "learning_rate": 5e-05, "loss": 0.7745, "mean_token_accuracy": 0.8678650259971619, "num_tokens": 229831221.0, "step": 6052 }, { "epoch": 0.7700038163083577, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 10.946525573730469, "learning_rate": 5e-05, "loss": 1.2615, "mean_token_accuracy": 0.8484953045845032, "num_tokens": 229866762.0, "step": 6053 }, { "epoch": 0.7701310265869482, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 6.953601837158203, "learning_rate": 5e-05, "loss": 0.8485, "mean_token_accuracy": 0.8669467568397522, "num_tokens": 229904715.0, "step": 6054 }, { "epoch": 0.7702582368655387, "ewc_loss": 0.98828125, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000949859619140625, "grad_norm": 12.829029083251953, "learning_rate": 5e-05, "loss": 1.3799, "mean_token_accuracy": 0.8634434938430786, "num_tokens": 229942938.0, "step": 6055 }, { "epoch": 0.7703854471441293, "ewc_loss": 1.046875, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.001007080078125, "grad_norm": 10.487826347351074, "learning_rate": 5e-05, "loss": 1.434, "mean_token_accuracy": 0.8661940097808838, "num_tokens": 229982468.0, "step": 6056 }, { "epoch": 0.7705126574227198, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 11.405474662780762, "learning_rate": 5e-05, "loss": 1.1429, "mean_token_accuracy": 0.8542076945304871, "num_tokens": 230022386.0, "step": 6057 }, { "epoch": 0.7706398677013102, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 9.443352699279785, "learning_rate": 5e-05, "loss": 1.2215, "mean_token_accuracy": 0.8490585684776306, "num_tokens": 230054871.0, "step": 6058 }, { "epoch": 0.7707670779799007, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 11.768963813781738, "learning_rate": 5e-05, "loss": 1.2044, "mean_token_accuracy": 0.8682956695556641, "num_tokens": 230090996.0, "step": 6059 }, { "epoch": 0.7708942882584913, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.64585018157959, "learning_rate": 5e-05, "loss": 0.953, "mean_token_accuracy": 0.8693745136260986, "num_tokens": 230124575.0, "step": 6060 }, { "epoch": 0.7710214985370818, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.0008697509765625, "grad_norm": 12.313353538513184, "learning_rate": 5e-05, "loss": 1.3863, "mean_token_accuracy": 0.8410913944244385, "num_tokens": 230168225.0, "step": 6061 }, { "epoch": 0.7711487088156723, "ewc_loss": 0.75, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 10.056107521057129, "learning_rate": 5e-05, "loss": 1.1658, "mean_token_accuracy": 0.8606923222541809, "num_tokens": 230208004.0, "step": 6062 }, { "epoch": 0.7712759190942629, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 11.449612617492676, "learning_rate": 5e-05, "loss": 1.1865, "mean_token_accuracy": 0.8545076847076416, "num_tokens": 230243247.0, "step": 6063 }, { "epoch": 0.7714031293728534, "ewc_loss": 0.75, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 10.155653953552246, "learning_rate": 5e-05, "loss": 1.1277, "mean_token_accuracy": 0.8686053156852722, "num_tokens": 230286646.0, "step": 6064 }, { "epoch": 0.7715303396514438, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.76257610321045, "learning_rate": 5e-05, "loss": 1.1339, "mean_token_accuracy": 0.8423739075660706, "num_tokens": 230317705.0, "step": 6065 }, { "epoch": 0.7716575499300343, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.37851333618164, "learning_rate": 5e-05, "loss": 1.051, "mean_token_accuracy": 0.8544592261314392, "num_tokens": 230362523.0, "step": 6066 }, { "epoch": 0.7717847602086249, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.530863761901855, "learning_rate": 5e-05, "loss": 1.0959, "mean_token_accuracy": 0.8543686866760254, "num_tokens": 230402678.0, "step": 6067 }, { "epoch": 0.7719119704872154, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.837447166442871, "learning_rate": 5e-05, "loss": 0.9629, "mean_token_accuracy": 0.8619813919067383, "num_tokens": 230440557.0, "step": 6068 }, { "epoch": 0.7720391807658059, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.53066349029541, "learning_rate": 5e-05, "loss": 1.0623, "mean_token_accuracy": 0.8627529740333557, "num_tokens": 230477602.0, "step": 6069 }, { "epoch": 0.7721663910443964, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.61739444732666, "learning_rate": 5e-05, "loss": 0.8763, "mean_token_accuracy": 0.8804931640625, "num_tokens": 230518381.0, "step": 6070 }, { "epoch": 0.7722936013229869, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.759459495544434, "learning_rate": 5e-05, "loss": 1.1233, "mean_token_accuracy": 0.8547991514205933, "num_tokens": 230559615.0, "step": 6071 }, { "epoch": 0.7724208116015774, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.797733306884766, "learning_rate": 5e-05, "loss": 0.964, "mean_token_accuracy": 0.8592735528945923, "num_tokens": 230595270.0, "step": 6072 }, { "epoch": 0.7725480218801679, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.791958808898926, "learning_rate": 5e-05, "loss": 1.1331, "mean_token_accuracy": 0.849107563495636, "num_tokens": 230630405.0, "step": 6073 }, { "epoch": 0.7726752321587584, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.130724906921387, "learning_rate": 5e-05, "loss": 1.0526, "mean_token_accuracy": 0.8425009250640869, "num_tokens": 230670836.0, "step": 6074 }, { "epoch": 0.772802442437349, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.394685745239258, "learning_rate": 5e-05, "loss": 1.1098, "mean_token_accuracy": 0.8544440269470215, "num_tokens": 230709150.0, "step": 6075 }, { "epoch": 0.7729296527159395, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.147665977478027, "learning_rate": 5e-05, "loss": 0.978, "mean_token_accuracy": 0.8564642667770386, "num_tokens": 230740306.0, "step": 6076 }, { "epoch": 0.7730568629945299, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.797856330871582, "learning_rate": 5e-05, "loss": 0.9998, "mean_token_accuracy": 0.863495945930481, "num_tokens": 230776499.0, "step": 6077 }, { "epoch": 0.7731840732731204, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.74914264678955, "learning_rate": 5e-05, "loss": 0.8984, "mean_token_accuracy": 0.86360102891922, "num_tokens": 230816614.0, "step": 6078 }, { "epoch": 0.773311283551711, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.457087516784668, "learning_rate": 5e-05, "loss": 0.992, "mean_token_accuracy": 0.8539732694625854, "num_tokens": 230865271.0, "step": 6079 }, { "epoch": 0.7734384938303015, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.203961372375488, "learning_rate": 5e-05, "loss": 0.8672, "mean_token_accuracy": 0.8583804368972778, "num_tokens": 230903092.0, "step": 6080 }, { "epoch": 0.773565704108892, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.547296524047852, "learning_rate": 5e-05, "loss": 0.9902, "mean_token_accuracy": 0.863257646560669, "num_tokens": 230942726.0, "step": 6081 }, { "epoch": 0.7736929143874826, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.878185749053955, "learning_rate": 5e-05, "loss": 0.8235, "mean_token_accuracy": 0.8585221171379089, "num_tokens": 230978694.0, "step": 6082 }, { "epoch": 0.773820124666073, "ewc_loss": 0.625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.89289379119873, "learning_rate": 5e-05, "loss": 1.0618, "mean_token_accuracy": 0.8583098649978638, "num_tokens": 231015466.0, "step": 6083 }, { "epoch": 0.7739473349446635, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.132771492004395, "learning_rate": 5e-05, "loss": 0.842, "mean_token_accuracy": 0.8635209798812866, "num_tokens": 231054823.0, "step": 6084 }, { "epoch": 0.774074545223254, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.834094047546387, "learning_rate": 5e-05, "loss": 0.9944, "mean_token_accuracy": 0.8699653744697571, "num_tokens": 231087893.0, "step": 6085 }, { "epoch": 0.7742017555018446, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.56100082397461, "learning_rate": 5e-05, "loss": 0.8993, "mean_token_accuracy": 0.8606134653091431, "num_tokens": 231123446.0, "step": 6086 }, { "epoch": 0.7743289657804351, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.154258728027344, "learning_rate": 5e-05, "loss": 0.9557, "mean_token_accuracy": 0.8647989630699158, "num_tokens": 231151558.0, "step": 6087 }, { "epoch": 0.7744561760590256, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.42164421081543, "learning_rate": 5e-05, "loss": 0.8526, "mean_token_accuracy": 0.86561518907547, "num_tokens": 231190715.0, "step": 6088 }, { "epoch": 0.774583386337616, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.558215141296387, "learning_rate": 5e-05, "loss": 0.8596, "mean_token_accuracy": 0.8701173663139343, "num_tokens": 231224634.0, "step": 6089 }, { "epoch": 0.7747105966162066, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.918268203735352, "learning_rate": 5e-05, "loss": 0.8513, "mean_token_accuracy": 0.8516777753829956, "num_tokens": 231265497.0, "step": 6090 }, { "epoch": 0.7748378068947971, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.443230628967285, "learning_rate": 5e-05, "loss": 0.8587, "mean_token_accuracy": 0.8638519048690796, "num_tokens": 231301886.0, "step": 6091 }, { "epoch": 0.7749650171733876, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.425858020782471, "learning_rate": 5e-05, "loss": 0.8089, "mean_token_accuracy": 0.85010826587677, "num_tokens": 231339580.0, "step": 6092 }, { "epoch": 0.7750922274519781, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.60041332244873, "learning_rate": 5e-05, "loss": 0.8793, "mean_token_accuracy": 0.8626269102096558, "num_tokens": 231376176.0, "step": 6093 }, { "epoch": 0.7752194377305687, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.185981750488281, "learning_rate": 5e-05, "loss": 0.7467, "mean_token_accuracy": 0.8619570732116699, "num_tokens": 231409975.0, "step": 6094 }, { "epoch": 0.7753466480091591, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.777493476867676, "learning_rate": 5e-05, "loss": 0.9244, "mean_token_accuracy": 0.8567002415657043, "num_tokens": 231447998.0, "step": 6095 }, { "epoch": 0.7754738582877496, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.216678619384766, "learning_rate": 5e-05, "loss": 0.736, "mean_token_accuracy": 0.8689437508583069, "num_tokens": 231481810.0, "step": 6096 }, { "epoch": 0.7756010685663401, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.790103912353516, "learning_rate": 5e-05, "loss": 0.877, "mean_token_accuracy": 0.8715293407440186, "num_tokens": 231521866.0, "step": 6097 }, { "epoch": 0.7757282788449307, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.48446798324585, "learning_rate": 5e-05, "loss": 0.829, "mean_token_accuracy": 0.850612223148346, "num_tokens": 231558247.0, "step": 6098 }, { "epoch": 0.7758554891235212, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.428004264831543, "learning_rate": 5e-05, "loss": 0.8718, "mean_token_accuracy": 0.8646285533905029, "num_tokens": 231594183.0, "step": 6099 }, { "epoch": 0.7759826994021117, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.709017753601074, "learning_rate": 5e-05, "loss": 0.7964, "mean_token_accuracy": 0.862118661403656, "num_tokens": 231628991.0, "step": 6100 }, { "epoch": 0.7761099096807021, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.830296516418457, "learning_rate": 5e-05, "loss": 0.8639, "mean_token_accuracy": 0.8477061986923218, "num_tokens": 231669413.0, "step": 6101 }, { "epoch": 0.7762371199592927, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.749785900115967, "learning_rate": 5e-05, "loss": 0.7782, "mean_token_accuracy": 0.8634591102600098, "num_tokens": 231708765.0, "step": 6102 }, { "epoch": 0.7763643302378832, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.427786827087402, "learning_rate": 5e-05, "loss": 0.7826, "mean_token_accuracy": 0.857361912727356, "num_tokens": 231747577.0, "step": 6103 }, { "epoch": 0.7764915405164737, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.767435550689697, "learning_rate": 5e-05, "loss": 0.777, "mean_token_accuracy": 0.857192873954773, "num_tokens": 231788256.0, "step": 6104 }, { "epoch": 0.7766187507950643, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.2041144371032715, "learning_rate": 5e-05, "loss": 0.7515, "mean_token_accuracy": 0.8598649501800537, "num_tokens": 231825709.0, "step": 6105 }, { "epoch": 0.7767459610736548, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.887513160705566, "learning_rate": 5e-05, "loss": 0.7607, "mean_token_accuracy": 0.8671689629554749, "num_tokens": 231862240.0, "step": 6106 }, { "epoch": 0.7768731713522452, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.981091022491455, "learning_rate": 5e-05, "loss": 0.6926, "mean_token_accuracy": 0.8717015981674194, "num_tokens": 231897711.0, "step": 6107 }, { "epoch": 0.7770003816308357, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.119355201721191, "learning_rate": 5e-05, "loss": 0.8234, "mean_token_accuracy": 0.8537799119949341, "num_tokens": 231932383.0, "step": 6108 }, { "epoch": 0.7771275919094263, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.78549337387085, "learning_rate": 5e-05, "loss": 0.7208, "mean_token_accuracy": 0.8576622605323792, "num_tokens": 231960919.0, "step": 6109 }, { "epoch": 0.7772548021880168, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.257965087890625, "learning_rate": 5e-05, "loss": 0.7617, "mean_token_accuracy": 0.880530834197998, "num_tokens": 231996435.0, "step": 6110 }, { "epoch": 0.7773820124666073, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.848987102508545, "learning_rate": 5e-05, "loss": 0.7405, "mean_token_accuracy": 0.8516477346420288, "num_tokens": 232036146.0, "step": 6111 }, { "epoch": 0.7775092227451978, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.11214542388916, "learning_rate": 5e-05, "loss": 0.8295, "mean_token_accuracy": 0.8535047173500061, "num_tokens": 232077656.0, "step": 6112 }, { "epoch": 0.7776364330237884, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.098855972290039, "learning_rate": 5e-05, "loss": 0.7176, "mean_token_accuracy": 0.8672485947608948, "num_tokens": 232113370.0, "step": 6113 }, { "epoch": 0.7777636433023788, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003719329833984375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.798819541931152, "learning_rate": 5e-05, "loss": 0.7592, "mean_token_accuracy": 0.8676020503044128, "num_tokens": 232155450.0, "step": 6114 }, { "epoch": 0.7778908535809693, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.341222763061523, "learning_rate": 5e-05, "loss": 0.7094, "mean_token_accuracy": 0.8678054213523865, "num_tokens": 232198076.0, "step": 6115 }, { "epoch": 0.7780180638595598, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.518886089324951, "learning_rate": 5e-05, "loss": 0.7828, "mean_token_accuracy": 0.8558309078216553, "num_tokens": 232246472.0, "step": 6116 }, { "epoch": 0.7781452741381504, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.494640350341797, "learning_rate": 5e-05, "loss": 0.7372, "mean_token_accuracy": 0.8664764761924744, "num_tokens": 232280576.0, "step": 6117 }, { "epoch": 0.7782724844167409, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.32786226272583, "learning_rate": 5e-05, "loss": 0.7248, "mean_token_accuracy": 0.8679757714271545, "num_tokens": 232322172.0, "step": 6118 }, { "epoch": 0.7783996946953314, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.548856735229492, "learning_rate": 5e-05, "loss": 0.7466, "mean_token_accuracy": 0.8637323379516602, "num_tokens": 232359317.0, "step": 6119 }, { "epoch": 0.7785269049739219, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.3109869956970215, "learning_rate": 5e-05, "loss": 0.7889, "mean_token_accuracy": 0.844787061214447, "num_tokens": 232395209.0, "step": 6120 }, { "epoch": 0.7786541152525124, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.399329662322998, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8532058596611023, "num_tokens": 232433303.0, "step": 6121 }, { "epoch": 0.7787813255311029, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.52061128616333, "learning_rate": 5e-05, "loss": 0.6927, "mean_token_accuracy": 0.8734574913978577, "num_tokens": 232464465.0, "step": 6122 }, { "epoch": 0.7789085358096934, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.0902910232543945, "learning_rate": 5e-05, "loss": 0.6708, "mean_token_accuracy": 0.877152144908905, "num_tokens": 232499846.0, "step": 6123 }, { "epoch": 0.779035746088284, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.8952507972717285, "learning_rate": 5e-05, "loss": 0.7259, "mean_token_accuracy": 0.8735111355781555, "num_tokens": 232528714.0, "step": 6124 }, { "epoch": 0.7791629563668745, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.645483493804932, "learning_rate": 5e-05, "loss": 0.7282, "mean_token_accuracy": 0.8515443801879883, "num_tokens": 232561155.0, "step": 6125 }, { "epoch": 0.7792901666454649, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00037384033203125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.460179328918457, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8749171495437622, "num_tokens": 232595685.0, "step": 6126 }, { "epoch": 0.7794173769240554, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.00022220611572265625, "grad_norm": 6.109969139099121, "learning_rate": 5e-05, "loss": 0.7353, "mean_token_accuracy": 0.8484945893287659, "num_tokens": 232638683.0, "step": 6127 }, { "epoch": 0.779544587202646, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 9.222126007080078, "learning_rate": 5e-05, "loss": 0.9044, "mean_token_accuracy": 0.8618737459182739, "num_tokens": 232677492.0, "step": 6128 }, { "epoch": 0.7796717974812365, "ewc_loss": 0.263671875, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.00022602081298828125, "grad_norm": 5.8323869705200195, "learning_rate": 5e-05, "loss": 0.7057, "mean_token_accuracy": 0.8500128984451294, "num_tokens": 232717997.0, "step": 6129 }, { "epoch": 0.779799007759827, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.7409029006958, "learning_rate": 5e-05, "loss": 1.0017, "mean_token_accuracy": 0.8555888533592224, "num_tokens": 232756784.0, "step": 6130 }, { "epoch": 0.7799262180384176, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.686067581176758, "learning_rate": 5e-05, "loss": 0.7621, "mean_token_accuracy": 0.861228883266449, "num_tokens": 232795194.0, "step": 6131 }, { "epoch": 0.780053428317008, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.015603065490723, "learning_rate": 5e-05, "loss": 0.965, "mean_token_accuracy": 0.8652653098106384, "num_tokens": 232826505.0, "step": 6132 }, { "epoch": 0.7801806385955985, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.161759376525879, "learning_rate": 5e-05, "loss": 0.8874, "mean_token_accuracy": 0.8704981803894043, "num_tokens": 232860111.0, "step": 6133 }, { "epoch": 0.780307848874189, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.048833847045898, "learning_rate": 5e-05, "loss": 0.7987, "mean_token_accuracy": 0.8536454439163208, "num_tokens": 232895558.0, "step": 6134 }, { "epoch": 0.7804350591527796, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.525524139404297, "learning_rate": 5e-05, "loss": 0.9615, "mean_token_accuracy": 0.8683524131774902, "num_tokens": 232935737.0, "step": 6135 }, { "epoch": 0.7805622694313701, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.020040988922119, "learning_rate": 5e-05, "loss": 0.7779, "mean_token_accuracy": 0.8436917662620544, "num_tokens": 232977917.0, "step": 6136 }, { "epoch": 0.7806894797099606, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.576437950134277, "learning_rate": 5e-05, "loss": 1.0491, "mean_token_accuracy": 0.8753657341003418, "num_tokens": 233016748.0, "step": 6137 }, { "epoch": 0.780816689988551, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.359064102172852, "learning_rate": 5e-05, "loss": 0.9038, "mean_token_accuracy": 0.8627674579620361, "num_tokens": 233052814.0, "step": 6138 }, { "epoch": 0.7809439002671416, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.019850730895996, "learning_rate": 5e-05, "loss": 1.0592, "mean_token_accuracy": 0.8610463738441467, "num_tokens": 233088727.0, "step": 6139 }, { "epoch": 0.7810711105457321, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.80095386505127, "learning_rate": 5e-05, "loss": 1.0434, "mean_token_accuracy": 0.8590619564056396, "num_tokens": 233126402.0, "step": 6140 }, { "epoch": 0.7811983208243226, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.325779914855957, "learning_rate": 5e-05, "loss": 0.946, "mean_token_accuracy": 0.8478298187255859, "num_tokens": 233161204.0, "step": 6141 }, { "epoch": 0.7813255311029131, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.154263496398926, "learning_rate": 5e-05, "loss": 1.0185, "mean_token_accuracy": 0.8601912260055542, "num_tokens": 233196527.0, "step": 6142 }, { "epoch": 0.7814527413815037, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.624417781829834, "learning_rate": 5e-05, "loss": 0.8825, "mean_token_accuracy": 0.8533815741539001, "num_tokens": 233236164.0, "step": 6143 }, { "epoch": 0.7815799516600941, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.74643325805664, "learning_rate": 5e-05, "loss": 0.949, "mean_token_accuracy": 0.8541630506515503, "num_tokens": 233268762.0, "step": 6144 }, { "epoch": 0.7817071619386846, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.948014736175537, "learning_rate": 5e-05, "loss": 0.9132, "mean_token_accuracy": 0.8547105193138123, "num_tokens": 233307388.0, "step": 6145 }, { "epoch": 0.7818343722172751, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.608266353607178, "learning_rate": 5e-05, "loss": 0.787, "mean_token_accuracy": 0.8664429187774658, "num_tokens": 233347902.0, "step": 6146 }, { "epoch": 0.7819615824958657, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.496411323547363, "learning_rate": 5e-05, "loss": 0.9499, "mean_token_accuracy": 0.8464246988296509, "num_tokens": 233378959.0, "step": 6147 }, { "epoch": 0.7820887927744562, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.912644386291504, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.8546266555786133, "num_tokens": 233420580.0, "step": 6148 }, { "epoch": 0.7822160030530467, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.857770919799805, "learning_rate": 5e-05, "loss": 0.907, "mean_token_accuracy": 0.8682180643081665, "num_tokens": 233451628.0, "step": 6149 }, { "epoch": 0.7823432133316371, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.800963401794434, "learning_rate": 5e-05, "loss": 0.7835, "mean_token_accuracy": 0.8531183004379272, "num_tokens": 233493994.0, "step": 6150 }, { "epoch": 0.7824704236102277, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.968840599060059, "learning_rate": 5e-05, "loss": 0.9206, "mean_token_accuracy": 0.8669532537460327, "num_tokens": 233534156.0, "step": 6151 }, { "epoch": 0.7825976338888182, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.960315704345703, "learning_rate": 5e-05, "loss": 0.795, "mean_token_accuracy": 0.8606343269348145, "num_tokens": 233573179.0, "step": 6152 }, { "epoch": 0.7827248441674087, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.863320350646973, "learning_rate": 5e-05, "loss": 0.9199, "mean_token_accuracy": 0.8704811930656433, "num_tokens": 233606424.0, "step": 6153 }, { "epoch": 0.7828520544459993, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.121616363525391, "learning_rate": 5e-05, "loss": 0.8222, "mean_token_accuracy": 0.8564872145652771, "num_tokens": 233649155.0, "step": 6154 }, { "epoch": 0.7829792647245898, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.618203163146973, "learning_rate": 5e-05, "loss": 0.958, "mean_token_accuracy": 0.8471782207489014, "num_tokens": 233677163.0, "step": 6155 }, { "epoch": 0.7831064750031802, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.18204927444458, "learning_rate": 5e-05, "loss": 0.8385, "mean_token_accuracy": 0.8494260311126709, "num_tokens": 233713120.0, "step": 6156 }, { "epoch": 0.7832336852817707, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.497734069824219, "learning_rate": 5e-05, "loss": 0.9259, "mean_token_accuracy": 0.8585034608840942, "num_tokens": 233751160.0, "step": 6157 }, { "epoch": 0.7833608955603613, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.0094804763793945, "learning_rate": 5e-05, "loss": 0.7817, "mean_token_accuracy": 0.8617323637008667, "num_tokens": 233786474.0, "step": 6158 }, { "epoch": 0.7834881058389518, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.676345825195312, "learning_rate": 5e-05, "loss": 0.9182, "mean_token_accuracy": 0.8677854537963867, "num_tokens": 233818256.0, "step": 6159 }, { "epoch": 0.7836153161175423, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.897286415100098, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.8520199060440063, "num_tokens": 233856519.0, "step": 6160 }, { "epoch": 0.7837425263961328, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.754809379577637, "learning_rate": 5e-05, "loss": 0.944, "mean_token_accuracy": 0.861883282661438, "num_tokens": 233896132.0, "step": 6161 }, { "epoch": 0.7838697366747234, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.082464218139648, "learning_rate": 5e-05, "loss": 0.8174, "mean_token_accuracy": 0.8597204685211182, "num_tokens": 233931668.0, "step": 6162 }, { "epoch": 0.7839969469533138, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.516258239746094, "learning_rate": 5e-05, "loss": 0.9002, "mean_token_accuracy": 0.8719983100891113, "num_tokens": 233965821.0, "step": 6163 }, { "epoch": 0.7841241572319043, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.378032207489014, "learning_rate": 5e-05, "loss": 0.825, "mean_token_accuracy": 0.8629766702651978, "num_tokens": 234003488.0, "step": 6164 }, { "epoch": 0.7842513675104948, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.110553741455078, "learning_rate": 5e-05, "loss": 0.9129, "mean_token_accuracy": 0.8547775745391846, "num_tokens": 234044270.0, "step": 6165 }, { "epoch": 0.7843785777890854, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.536663055419922, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.8792060017585754, "num_tokens": 234081754.0, "step": 6166 }, { "epoch": 0.7845057880676759, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.753859043121338, "learning_rate": 5e-05, "loss": 0.8063, "mean_token_accuracy": 0.8769088387489319, "num_tokens": 234118961.0, "step": 6167 }, { "epoch": 0.7846329983462664, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.565762042999268, "learning_rate": 5e-05, "loss": 0.8676, "mean_token_accuracy": 0.8465676307678223, "num_tokens": 234159084.0, "step": 6168 }, { "epoch": 0.7847602086248568, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.606780529022217, "learning_rate": 5e-05, "loss": 0.797, "mean_token_accuracy": 0.8652194142341614, "num_tokens": 234201296.0, "step": 6169 }, { "epoch": 0.7848874189034474, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003757476806640625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.452178955078125, "learning_rate": 5e-05, "loss": 0.7718, "mean_token_accuracy": 0.8665654063224792, "num_tokens": 234240661.0, "step": 6170 }, { "epoch": 0.7850146291820379, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.646075248718262, "learning_rate": 5e-05, "loss": 0.8363, "mean_token_accuracy": 0.8517253994941711, "num_tokens": 234277192.0, "step": 6171 }, { "epoch": 0.7851418394606284, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.305614471435547, "learning_rate": 5e-05, "loss": 0.7798, "mean_token_accuracy": 0.8591709136962891, "num_tokens": 234321527.0, "step": 6172 }, { "epoch": 0.785269049739219, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.715975761413574, "learning_rate": 5e-05, "loss": 0.8369, "mean_token_accuracy": 0.8472140431404114, "num_tokens": 234355578.0, "step": 6173 }, { "epoch": 0.7853962600178095, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.211340427398682, "learning_rate": 5e-05, "loss": 0.7347, "mean_token_accuracy": 0.8702123761177063, "num_tokens": 234395240.0, "step": 6174 }, { "epoch": 0.7855234702963999, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.739022254943848, "learning_rate": 5e-05, "loss": 0.854, "mean_token_accuracy": 0.8472384214401245, "num_tokens": 234436914.0, "step": 6175 }, { "epoch": 0.7856506805749904, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.1350555419921875, "learning_rate": 5e-05, "loss": 0.7549, "mean_token_accuracy": 0.8608118295669556, "num_tokens": 234477811.0, "step": 6176 }, { "epoch": 0.785777890853581, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.792155742645264, "learning_rate": 5e-05, "loss": 0.7734, "mean_token_accuracy": 0.8652423620223999, "num_tokens": 234515489.0, "step": 6177 }, { "epoch": 0.7859051011321715, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.061850070953369, "learning_rate": 5e-05, "loss": 0.7821, "mean_token_accuracy": 0.8568652272224426, "num_tokens": 234544846.0, "step": 6178 }, { "epoch": 0.786032311410762, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.844021797180176, "learning_rate": 5e-05, "loss": 0.7615, "mean_token_accuracy": 0.8719630241394043, "num_tokens": 234581497.0, "step": 6179 }, { "epoch": 0.7861595216893525, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.003994464874268, "learning_rate": 5e-05, "loss": 0.7624, "mean_token_accuracy": 0.8550963401794434, "num_tokens": 234621837.0, "step": 6180 }, { "epoch": 0.786286731967943, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.9005937576293945, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.864647388458252, "num_tokens": 234655049.0, "step": 6181 }, { "epoch": 0.7864139422465335, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.933618545532227, "learning_rate": 5e-05, "loss": 0.7571, "mean_token_accuracy": 0.8524097204208374, "num_tokens": 234689177.0, "step": 6182 }, { "epoch": 0.786541152525124, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.997966766357422, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8731154203414917, "num_tokens": 234722744.0, "step": 6183 }, { "epoch": 0.7866683628037145, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.845335483551025, "learning_rate": 5e-05, "loss": 0.706, "mean_token_accuracy": 0.8662447929382324, "num_tokens": 234762998.0, "step": 6184 }, { "epoch": 0.7867955730823051, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.126408576965332, "learning_rate": 5e-05, "loss": 0.8656, "mean_token_accuracy": 0.845816969871521, "num_tokens": 234799283.0, "step": 6185 }, { "epoch": 0.7869227833608956, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.758005142211914, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.8515207171440125, "num_tokens": 234838019.0, "step": 6186 }, { "epoch": 0.787049993639486, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.203165054321289, "learning_rate": 5e-05, "loss": 0.8522, "mean_token_accuracy": 0.8536965250968933, "num_tokens": 234874449.0, "step": 6187 }, { "epoch": 0.7871772039180766, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.769786834716797, "learning_rate": 5e-05, "loss": 0.7353, "mean_token_accuracy": 0.8635030388832092, "num_tokens": 234913868.0, "step": 6188 }, { "epoch": 0.7873044141966671, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.120966911315918, "learning_rate": 5e-05, "loss": 0.7689, "mean_token_accuracy": 0.8793142437934875, "num_tokens": 234949060.0, "step": 6189 }, { "epoch": 0.7874316244752576, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.976826190948486, "learning_rate": 5e-05, "loss": 0.7438, "mean_token_accuracy": 0.8602375984191895, "num_tokens": 234987732.0, "step": 6190 }, { "epoch": 0.7875588347538481, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.811435699462891, "learning_rate": 5e-05, "loss": 0.7698, "mean_token_accuracy": 0.8703409433364868, "num_tokens": 235023030.0, "step": 6191 }, { "epoch": 0.7876860450324387, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.31993293762207, "learning_rate": 5e-05, "loss": 0.7565, "mean_token_accuracy": 0.8631376624107361, "num_tokens": 235061560.0, "step": 6192 }, { "epoch": 0.7878132553110291, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.341219425201416, "learning_rate": 5e-05, "loss": 0.733, "mean_token_accuracy": 0.8662811517715454, "num_tokens": 235098826.0, "step": 6193 }, { "epoch": 0.7879404655896196, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.770905017852783, "learning_rate": 5e-05, "loss": 0.7962, "mean_token_accuracy": 0.857810378074646, "num_tokens": 235142116.0, "step": 6194 }, { "epoch": 0.7880676758682101, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.795909881591797, "learning_rate": 5e-05, "loss": 0.7403, "mean_token_accuracy": 0.8567216396331787, "num_tokens": 235182920.0, "step": 6195 }, { "epoch": 0.7881948861468007, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.35268783569336, "learning_rate": 5e-05, "loss": 0.835, "mean_token_accuracy": 0.8600890636444092, "num_tokens": 235214669.0, "step": 6196 }, { "epoch": 0.7883220964253912, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.00023937225341796875, "grad_norm": 6.193018913269043, "learning_rate": 5e-05, "loss": 0.6805, "mean_token_accuracy": 0.8615574836730957, "num_tokens": 235251723.0, "step": 6197 }, { "epoch": 0.7884493067039817, "ewc_loss": 0.5, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 9.145092010498047, "learning_rate": 5e-05, "loss": 0.8716, "mean_token_accuracy": 0.8676542043685913, "num_tokens": 235285931.0, "step": 6198 }, { "epoch": 0.7885765169825721, "ewc_loss": 0.265625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.00022792816162109375, "grad_norm": 5.685704708099365, "learning_rate": 5e-05, "loss": 0.6424, "mean_token_accuracy": 0.8692244291305542, "num_tokens": 235324571.0, "step": 6199 }, { "epoch": 0.7887037272611627, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 10.027165412902832, "learning_rate": 5e-05, "loss": 0.9945, "mean_token_accuracy": 0.8732802271842957, "num_tokens": 235359348.0, "step": 6200 }, { "epoch": 0.7888309375397532, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.203426837921143, "learning_rate": 5e-05, "loss": 0.7391, "mean_token_accuracy": 0.8734867572784424, "num_tokens": 235397998.0, "step": 6201 }, { "epoch": 0.7889581478183437, "ewc_loss": 0.6875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.085095405578613, "learning_rate": 5e-05, "loss": 1.0955, "mean_token_accuracy": 0.8590739965438843, "num_tokens": 235436991.0, "step": 6202 }, { "epoch": 0.7890853580969343, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.023345947265625, "learning_rate": 5e-05, "loss": 0.9714, "mean_token_accuracy": 0.8785878419876099, "num_tokens": 235478129.0, "step": 6203 }, { "epoch": 0.7892125683755248, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.295751571655273, "learning_rate": 5e-05, "loss": 0.8994, "mean_token_accuracy": 0.8663678765296936, "num_tokens": 235514630.0, "step": 6204 }, { "epoch": 0.7893397786541152, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.393383979797363, "learning_rate": 5e-05, "loss": 1.0485, "mean_token_accuracy": 0.8715835809707642, "num_tokens": 235550212.0, "step": 6205 }, { "epoch": 0.7894669889327057, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.403102874755859, "learning_rate": 5e-05, "loss": 0.7342, "mean_token_accuracy": 0.871148943901062, "num_tokens": 235592602.0, "step": 6206 }, { "epoch": 0.7895941992112963, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.000766754150390625, "grad_norm": 11.109207153320312, "learning_rate": 5e-05, "loss": 1.1768, "mean_token_accuracy": 0.8732297420501709, "num_tokens": 235634199.0, "step": 6207 }, { "epoch": 0.7897214094898868, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 7.592936038970947, "learning_rate": 5e-05, "loss": 0.9506, "mean_token_accuracy": 0.8691524267196655, "num_tokens": 235670086.0, "step": 6208 }, { "epoch": 0.7898486197684773, "ewc_loss": 0.75, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 11.181784629821777, "learning_rate": 5e-05, "loss": 1.2243, "mean_token_accuracy": 0.8464243412017822, "num_tokens": 235709749.0, "step": 6209 }, { "epoch": 0.7899758300470678, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.143387794494629, "learning_rate": 5e-05, "loss": 1.2176, "mean_token_accuracy": 0.8651673197746277, "num_tokens": 235750574.0, "step": 6210 }, { "epoch": 0.7901030403256584, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.021149635314941, "learning_rate": 5e-05, "loss": 1.0309, "mean_token_accuracy": 0.8674769997596741, "num_tokens": 235785234.0, "step": 6211 }, { "epoch": 0.7902302506042488, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 8.7855224609375, "learning_rate": 5e-05, "loss": 1.1097, "mean_token_accuracy": 0.849876880645752, "num_tokens": 235822931.0, "step": 6212 }, { "epoch": 0.7903574608828393, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.281120300292969, "learning_rate": 5e-05, "loss": 1.0516, "mean_token_accuracy": 0.8509953022003174, "num_tokens": 235863022.0, "step": 6213 }, { "epoch": 0.7904846711614298, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.10564136505127, "learning_rate": 5e-05, "loss": 0.9156, "mean_token_accuracy": 0.8648757934570312, "num_tokens": 235893275.0, "step": 6214 }, { "epoch": 0.7906118814400204, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.12844467163086, "learning_rate": 5e-05, "loss": 0.9508, "mean_token_accuracy": 0.8825637102127075, "num_tokens": 235933908.0, "step": 6215 }, { "epoch": 0.7907390917186109, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.400095462799072, "learning_rate": 5e-05, "loss": 0.8743, "mean_token_accuracy": 0.8538141250610352, "num_tokens": 235975296.0, "step": 6216 }, { "epoch": 0.7908663019972014, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.586236953735352, "learning_rate": 5e-05, "loss": 0.997, "mean_token_accuracy": 0.872100830078125, "num_tokens": 236012872.0, "step": 6217 }, { "epoch": 0.7909935122757918, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.251281261444092, "learning_rate": 5e-05, "loss": 0.8619, "mean_token_accuracy": 0.8656690120697021, "num_tokens": 236055762.0, "step": 6218 }, { "epoch": 0.7911207225543824, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.949430465698242, "learning_rate": 5e-05, "loss": 1.0485, "mean_token_accuracy": 0.8653849363327026, "num_tokens": 236099968.0, "step": 6219 }, { "epoch": 0.7912479328329729, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.768714904785156, "learning_rate": 5e-05, "loss": 0.9701, "mean_token_accuracy": 0.8543651700019836, "num_tokens": 236134000.0, "step": 6220 }, { "epoch": 0.7913751431115634, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.74933910369873, "learning_rate": 5e-05, "loss": 1.031, "mean_token_accuracy": 0.8685121536254883, "num_tokens": 236174427.0, "step": 6221 }, { "epoch": 0.791502353390154, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.156139373779297, "learning_rate": 5e-05, "loss": 1.0106, "mean_token_accuracy": 0.8544683456420898, "num_tokens": 236205643.0, "step": 6222 }, { "epoch": 0.7916295636687445, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.205552101135254, "learning_rate": 5e-05, "loss": 1.0716, "mean_token_accuracy": 0.8453407287597656, "num_tokens": 236252232.0, "step": 6223 }, { "epoch": 0.7917567739473349, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.187726974487305, "learning_rate": 5e-05, "loss": 0.9654, "mean_token_accuracy": 0.8578660488128662, "num_tokens": 236293694.0, "step": 6224 }, { "epoch": 0.7918839842259254, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.605697631835938, "learning_rate": 5e-05, "loss": 0.9506, "mean_token_accuracy": 0.8676679134368896, "num_tokens": 236335129.0, "step": 6225 }, { "epoch": 0.792011194504516, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.93335485458374, "learning_rate": 5e-05, "loss": 0.8572, "mean_token_accuracy": 0.8734590411186218, "num_tokens": 236370294.0, "step": 6226 }, { "epoch": 0.7921384047831065, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.44869613647461, "learning_rate": 5e-05, "loss": 0.9157, "mean_token_accuracy": 0.8717762231826782, "num_tokens": 236410728.0, "step": 6227 }, { "epoch": 0.792265615061697, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.374945640563965, "learning_rate": 5e-05, "loss": 0.8217, "mean_token_accuracy": 0.8631782531738281, "num_tokens": 236453053.0, "step": 6228 }, { "epoch": 0.7923928253402875, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.892560958862305, "learning_rate": 5e-05, "loss": 0.9478, "mean_token_accuracy": 0.8707574009895325, "num_tokens": 236491746.0, "step": 6229 }, { "epoch": 0.792520035618878, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.066323280334473, "learning_rate": 5e-05, "loss": 0.8395, "mean_token_accuracy": 0.8506428003311157, "num_tokens": 236530938.0, "step": 6230 }, { "epoch": 0.7926472458974685, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.296354293823242, "learning_rate": 5e-05, "loss": 0.9498, "mean_token_accuracy": 0.8876787424087524, "num_tokens": 236567765.0, "step": 6231 }, { "epoch": 0.792774456176059, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.431410789489746, "learning_rate": 5e-05, "loss": 0.8659, "mean_token_accuracy": 0.8618043065071106, "num_tokens": 236596522.0, "step": 6232 }, { "epoch": 0.7929016664546495, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.051702499389648, "learning_rate": 5e-05, "loss": 1.0221, "mean_token_accuracy": 0.8541989326477051, "num_tokens": 236637493.0, "step": 6233 }, { "epoch": 0.7930288767332401, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.974891185760498, "learning_rate": 5e-05, "loss": 0.9316, "mean_token_accuracy": 0.8589653968811035, "num_tokens": 236668405.0, "step": 6234 }, { "epoch": 0.7931560870118306, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.237784385681152, "learning_rate": 5e-05, "loss": 0.9078, "mean_token_accuracy": 0.8672257661819458, "num_tokens": 236703105.0, "step": 6235 }, { "epoch": 0.793283297290421, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.078542709350586, "learning_rate": 5e-05, "loss": 0.8988, "mean_token_accuracy": 0.864952802658081, "num_tokens": 236747017.0, "step": 6236 }, { "epoch": 0.7934105075690115, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.625315189361572, "learning_rate": 5e-05, "loss": 0.8239, "mean_token_accuracy": 0.8718551397323608, "num_tokens": 236785659.0, "step": 6237 }, { "epoch": 0.7935377178476021, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.8336896896362305, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8548146486282349, "num_tokens": 236819603.0, "step": 6238 }, { "epoch": 0.7936649281261926, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.673641681671143, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.865959644317627, "num_tokens": 236859440.0, "step": 6239 }, { "epoch": 0.7937921384047831, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.313834190368652, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8628339767456055, "num_tokens": 236892138.0, "step": 6240 }, { "epoch": 0.7939193486833737, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.125009536743164, "learning_rate": 5e-05, "loss": 0.9182, "mean_token_accuracy": 0.8446657061576843, "num_tokens": 236930559.0, "step": 6241 }, { "epoch": 0.7940465589619641, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.884444713592529, "learning_rate": 5e-05, "loss": 0.7531, "mean_token_accuracy": 0.8591809272766113, "num_tokens": 236969275.0, "step": 6242 }, { "epoch": 0.7941737692405546, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.420384407043457, "learning_rate": 5e-05, "loss": 0.8497, "mean_token_accuracy": 0.8697343468666077, "num_tokens": 237007662.0, "step": 6243 }, { "epoch": 0.7943009795191451, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.791983127593994, "learning_rate": 5e-05, "loss": 0.7884, "mean_token_accuracy": 0.8463823795318604, "num_tokens": 237046553.0, "step": 6244 }, { "epoch": 0.7944281897977357, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.53028392791748, "learning_rate": 5e-05, "loss": 0.9127, "mean_token_accuracy": 0.8597716689109802, "num_tokens": 237082309.0, "step": 6245 }, { "epoch": 0.7945554000763262, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.876784801483154, "learning_rate": 5e-05, "loss": 0.7318, "mean_token_accuracy": 0.870866596698761, "num_tokens": 237117997.0, "step": 6246 }, { "epoch": 0.7946826103549167, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.517086029052734, "learning_rate": 5e-05, "loss": 0.9637, "mean_token_accuracy": 0.8487119078636169, "num_tokens": 237159169.0, "step": 6247 }, { "epoch": 0.7948098206335071, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.1207427978515625, "learning_rate": 5e-05, "loss": 0.8034, "mean_token_accuracy": 0.8597812652587891, "num_tokens": 237200863.0, "step": 6248 }, { "epoch": 0.7949370309120977, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.169888496398926, "learning_rate": 5e-05, "loss": 0.9343, "mean_token_accuracy": 0.8491755723953247, "num_tokens": 237237318.0, "step": 6249 }, { "epoch": 0.7950642411906882, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.481250286102295, "learning_rate": 5e-05, "loss": 0.8601, "mean_token_accuracy": 0.8461904525756836, "num_tokens": 237278724.0, "step": 6250 }, { "epoch": 0.7951914514692787, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.641135215759277, "learning_rate": 5e-05, "loss": 0.8093, "mean_token_accuracy": 0.8673893809318542, "num_tokens": 237314190.0, "step": 6251 }, { "epoch": 0.7953186617478692, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.729147911071777, "learning_rate": 5e-05, "loss": 0.769, "mean_token_accuracy": 0.8742362856864929, "num_tokens": 237348560.0, "step": 6252 }, { "epoch": 0.7954458720264598, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.285046577453613, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8629623651504517, "num_tokens": 237388421.0, "step": 6253 }, { "epoch": 0.7955730823050502, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.806212425231934, "learning_rate": 5e-05, "loss": 0.8502, "mean_token_accuracy": 0.8460046052932739, "num_tokens": 237424517.0, "step": 6254 }, { "epoch": 0.7957002925836407, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.136480808258057, "learning_rate": 5e-05, "loss": 0.7413, "mean_token_accuracy": 0.8658702373504639, "num_tokens": 237466195.0, "step": 6255 }, { "epoch": 0.7958275028622313, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000377655029296875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.880440711975098, "learning_rate": 5e-05, "loss": 0.81, "mean_token_accuracy": 0.8633708953857422, "num_tokens": 237500189.0, "step": 6256 }, { "epoch": 0.7959547131408218, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.9982075691223145, "learning_rate": 5e-05, "loss": 0.7661, "mean_token_accuracy": 0.8541747331619263, "num_tokens": 237534186.0, "step": 6257 }, { "epoch": 0.7960819234194123, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.978610515594482, "learning_rate": 5e-05, "loss": 0.8237, "mean_token_accuracy": 0.8596516251564026, "num_tokens": 237572382.0, "step": 6258 }, { "epoch": 0.7962091336980028, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.9038472175598145, "learning_rate": 5e-05, "loss": 0.7271, "mean_token_accuracy": 0.8613116145133972, "num_tokens": 237614300.0, "step": 6259 }, { "epoch": 0.7963363439765934, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003795623779296875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.089998245239258, "learning_rate": 5e-05, "loss": 0.8018, "mean_token_accuracy": 0.8649528622627258, "num_tokens": 237650952.0, "step": 6260 }, { "epoch": 0.7964635542551838, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003814697265625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.807999610900879, "learning_rate": 5e-05, "loss": 0.753, "mean_token_accuracy": 0.8535232543945312, "num_tokens": 237689901.0, "step": 6261 }, { "epoch": 0.7965907645337743, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.201726913452148, "learning_rate": 5e-05, "loss": 0.8757, "mean_token_accuracy": 0.8474739789962769, "num_tokens": 237724889.0, "step": 6262 }, { "epoch": 0.7967179748123648, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.745920658111572, "learning_rate": 5e-05, "loss": 0.703, "mean_token_accuracy": 0.8674964904785156, "num_tokens": 237758853.0, "step": 6263 }, { "epoch": 0.7968451850909554, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.274184226989746, "learning_rate": 5e-05, "loss": 0.8501, "mean_token_accuracy": 0.8601479530334473, "num_tokens": 237795646.0, "step": 6264 }, { "epoch": 0.7969723953695459, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.800702095031738, "learning_rate": 5e-05, "loss": 0.7787, "mean_token_accuracy": 0.8495674133300781, "num_tokens": 237832229.0, "step": 6265 }, { "epoch": 0.7970996056481364, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.173050880432129, "learning_rate": 5e-05, "loss": 0.8673, "mean_token_accuracy": 0.8481495380401611, "num_tokens": 237870459.0, "step": 6266 }, { "epoch": 0.7972268159267268, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.974853992462158, "learning_rate": 5e-05, "loss": 0.7618, "mean_token_accuracy": 0.8568581342697144, "num_tokens": 237912367.0, "step": 6267 }, { "epoch": 0.7973540262053174, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.948810577392578, "learning_rate": 5e-05, "loss": 0.8303, "mean_token_accuracy": 0.8603291511535645, "num_tokens": 237950847.0, "step": 6268 }, { "epoch": 0.7974812364839079, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.17749547958374, "learning_rate": 5e-05, "loss": 0.7413, "mean_token_accuracy": 0.8683580160140991, "num_tokens": 237985325.0, "step": 6269 }, { "epoch": 0.7976084467624984, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.710511207580566, "learning_rate": 5e-05, "loss": 0.8242, "mean_token_accuracy": 0.8521064519882202, "num_tokens": 238021710.0, "step": 6270 }, { "epoch": 0.797735657041089, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.315585613250732, "learning_rate": 5e-05, "loss": 0.7426, "mean_token_accuracy": 0.8650568723678589, "num_tokens": 238053024.0, "step": 6271 }, { "epoch": 0.7978628673196795, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.557636260986328, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8672357797622681, "num_tokens": 238089732.0, "step": 6272 }, { "epoch": 0.7979900775982699, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.345189571380615, "learning_rate": 5e-05, "loss": 0.7423, "mean_token_accuracy": 0.863503098487854, "num_tokens": 238127583.0, "step": 6273 }, { "epoch": 0.7981172878768604, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.594559669494629, "learning_rate": 5e-05, "loss": 0.7841, "mean_token_accuracy": 0.8596958518028259, "num_tokens": 238166328.0, "step": 6274 }, { "epoch": 0.798244498155451, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.151463985443115, "learning_rate": 5e-05, "loss": 0.767, "mean_token_accuracy": 0.8519219160079956, "num_tokens": 238206111.0, "step": 6275 }, { "epoch": 0.7983717084340415, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.843053817749023, "learning_rate": 5e-05, "loss": 0.7704, "mean_token_accuracy": 0.8640909194946289, "num_tokens": 238243793.0, "step": 6276 }, { "epoch": 0.798498918712632, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.819320201873779, "learning_rate": 5e-05, "loss": 0.6914, "mean_token_accuracy": 0.8766143321990967, "num_tokens": 238284005.0, "step": 6277 }, { "epoch": 0.7986261289912225, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.254868507385254, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.8564975261688232, "num_tokens": 238321180.0, "step": 6278 }, { "epoch": 0.798753339269813, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.40348482131958, "learning_rate": 5e-05, "loss": 0.6822, "mean_token_accuracy": 0.8680471181869507, "num_tokens": 238364383.0, "step": 6279 }, { "epoch": 0.7988805495484035, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.82971477508545, "learning_rate": 5e-05, "loss": 0.9506, "mean_token_accuracy": 0.8479139804840088, "num_tokens": 238405547.0, "step": 6280 }, { "epoch": 0.799007759826994, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.075944423675537, "learning_rate": 5e-05, "loss": 0.6837, "mean_token_accuracy": 0.8636839985847473, "num_tokens": 238436810.0, "step": 6281 }, { "epoch": 0.7991349701055845, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.326190948486328, "learning_rate": 5e-05, "loss": 0.9897, "mean_token_accuracy": 0.8549985885620117, "num_tokens": 238470062.0, "step": 6282 }, { "epoch": 0.7992621803841751, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.354964733123779, "learning_rate": 5e-05, "loss": 0.6948, "mean_token_accuracy": 0.8784958124160767, "num_tokens": 238502680.0, "step": 6283 }, { "epoch": 0.7993893906627656, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.028246879577637, "learning_rate": 5e-05, "loss": 1.0126, "mean_token_accuracy": 0.8526555895805359, "num_tokens": 238534355.0, "step": 6284 }, { "epoch": 0.799516600941356, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.489654064178467, "learning_rate": 5e-05, "loss": 0.9107, "mean_token_accuracy": 0.8595438599586487, "num_tokens": 238572225.0, "step": 6285 }, { "epoch": 0.7996438112199465, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.528491497039795, "learning_rate": 5e-05, "loss": 0.8006, "mean_token_accuracy": 0.8669639825820923, "num_tokens": 238608313.0, "step": 6286 }, { "epoch": 0.7997710214985371, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.80370807647705, "learning_rate": 5e-05, "loss": 1.0152, "mean_token_accuracy": 0.8460135459899902, "num_tokens": 238652973.0, "step": 6287 }, { "epoch": 0.7998982317771276, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0003833770751953125, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 5.936129570007324, "learning_rate": 5e-05, "loss": 0.7196, "mean_token_accuracy": 0.856748104095459, "num_tokens": 238688490.0, "step": 6288 }, { "epoch": 0.8000254420557181, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.413180351257324, "learning_rate": 5e-05, "loss": 1.0444, "mean_token_accuracy": 0.8690561056137085, "num_tokens": 238728641.0, "step": 6289 }, { "epoch": 0.8001526523343087, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 5.9913811683654785, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.846585750579834, "num_tokens": 238764134.0, "step": 6290 }, { "epoch": 0.8002798626128991, "ewc_loss": 0.890625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 11.522069931030273, "learning_rate": 5e-05, "loss": 1.2783, "mean_token_accuracy": 0.86611008644104, "num_tokens": 238803705.0, "step": 6291 }, { "epoch": 0.8004070728914896, "ewc_loss": 0.82421875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 8.834827423095703, "learning_rate": 5e-05, "loss": 1.2543, "mean_token_accuracy": 0.854099690914154, "num_tokens": 238836816.0, "step": 6292 }, { "epoch": 0.8005342831700801, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.078645706176758, "learning_rate": 5e-05, "loss": 1.0286, "mean_token_accuracy": 0.8691244721412659, "num_tokens": 238870006.0, "step": 6293 }, { "epoch": 0.8006614934486707, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.486360549926758, "learning_rate": 5e-05, "loss": 1.2263, "mean_token_accuracy": 0.8599013090133667, "num_tokens": 238907018.0, "step": 6294 }, { "epoch": 0.8007887037272612, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.479143142700195, "learning_rate": 5e-05, "loss": 1.0135, "mean_token_accuracy": 0.8477470874786377, "num_tokens": 238943319.0, "step": 6295 }, { "epoch": 0.8009159140058517, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 9.768351554870605, "learning_rate": 5e-05, "loss": 1.1271, "mean_token_accuracy": 0.843370795249939, "num_tokens": 238978546.0, "step": 6296 }, { "epoch": 0.8010431242844421, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.861936569213867, "learning_rate": 5e-05, "loss": 0.9998, "mean_token_accuracy": 0.8571293354034424, "num_tokens": 239014785.0, "step": 6297 }, { "epoch": 0.8011703345630327, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.529729843139648, "learning_rate": 5e-05, "loss": 1.0823, "mean_token_accuracy": 0.8386778831481934, "num_tokens": 239048212.0, "step": 6298 }, { "epoch": 0.8012975448416232, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.046589851379395, "learning_rate": 5e-05, "loss": 1.0252, "mean_token_accuracy": 0.8501607775688171, "num_tokens": 239086946.0, "step": 6299 }, { "epoch": 0.8014247551202137, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.507648468017578, "learning_rate": 5e-05, "loss": 0.8666, "mean_token_accuracy": 0.877626895904541, "num_tokens": 239125100.0, "step": 6300 }, { "epoch": 0.8015519653988042, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.287850379943848, "learning_rate": 5e-05, "loss": 0.9805, "mean_token_accuracy": 0.853317141532898, "num_tokens": 239163118.0, "step": 6301 }, { "epoch": 0.8016791756773948, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.6549272537231445, "learning_rate": 5e-05, "loss": 0.8232, "mean_token_accuracy": 0.865691065788269, "num_tokens": 239196914.0, "step": 6302 }, { "epoch": 0.8018063859559852, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.377084732055664, "learning_rate": 5e-05, "loss": 0.9165, "mean_token_accuracy": 0.8618125915527344, "num_tokens": 239229609.0, "step": 6303 }, { "epoch": 0.8019335962345757, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.302847385406494, "learning_rate": 5e-05, "loss": 0.7589, "mean_token_accuracy": 0.8801053762435913, "num_tokens": 239269829.0, "step": 6304 }, { "epoch": 0.8020608065131662, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.259037971496582, "learning_rate": 5e-05, "loss": 0.8271, "mean_token_accuracy": 0.8757779002189636, "num_tokens": 239308563.0, "step": 6305 }, { "epoch": 0.8021880167917568, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.295402526855469, "learning_rate": 5e-05, "loss": 0.8134, "mean_token_accuracy": 0.8638803958892822, "num_tokens": 239348620.0, "step": 6306 }, { "epoch": 0.8023152270703473, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.955118179321289, "learning_rate": 5e-05, "loss": 0.8306, "mean_token_accuracy": 0.8604949712753296, "num_tokens": 239380706.0, "step": 6307 }, { "epoch": 0.8024424373489378, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.4724507331848145, "learning_rate": 5e-05, "loss": 0.8478, "mean_token_accuracy": 0.8501049280166626, "num_tokens": 239417365.0, "step": 6308 }, { "epoch": 0.8025696476275284, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.468959331512451, "learning_rate": 5e-05, "loss": 0.7619, "mean_token_accuracy": 0.860344409942627, "num_tokens": 239449926.0, "step": 6309 }, { "epoch": 0.8026968579061188, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.893052101135254, "learning_rate": 5e-05, "loss": 0.8351, "mean_token_accuracy": 0.8589317798614502, "num_tokens": 239485134.0, "step": 6310 }, { "epoch": 0.8028240681847093, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.901061058044434, "learning_rate": 5e-05, "loss": 0.7114, "mean_token_accuracy": 0.8704130053520203, "num_tokens": 239521353.0, "step": 6311 }, { "epoch": 0.8029512784632998, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.51113510131836, "learning_rate": 5e-05, "loss": 0.8883, "mean_token_accuracy": 0.8567191958427429, "num_tokens": 239554902.0, "step": 6312 }, { "epoch": 0.8030784887418904, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.371011734008789, "learning_rate": 5e-05, "loss": 0.7577, "mean_token_accuracy": 0.8509181141853333, "num_tokens": 239597436.0, "step": 6313 }, { "epoch": 0.8032056990204809, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.147284507751465, "learning_rate": 5e-05, "loss": 0.9293, "mean_token_accuracy": 0.8649239540100098, "num_tokens": 239635543.0, "step": 6314 }, { "epoch": 0.8033329092990714, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.246537685394287, "learning_rate": 5e-05, "loss": 0.698, "mean_token_accuracy": 0.8727308511734009, "num_tokens": 239673956.0, "step": 6315 }, { "epoch": 0.8034601195776618, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.498205184936523, "learning_rate": 5e-05, "loss": 1.0178, "mean_token_accuracy": 0.857377827167511, "num_tokens": 239712556.0, "step": 6316 }, { "epoch": 0.8035873298562524, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 6.793824672698975, "learning_rate": 5e-05, "loss": 0.7807, "mean_token_accuracy": 0.8754152059555054, "num_tokens": 239747034.0, "step": 6317 }, { "epoch": 0.8037145401348429, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.197432518005371, "learning_rate": 5e-05, "loss": 0.9917, "mean_token_accuracy": 0.8640048503875732, "num_tokens": 239787054.0, "step": 6318 }, { "epoch": 0.8038417504134334, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.571916103363037, "learning_rate": 5e-05, "loss": 0.9895, "mean_token_accuracy": 0.8459600210189819, "num_tokens": 239824796.0, "step": 6319 }, { "epoch": 0.803968960692024, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.334705352783203, "learning_rate": 5e-05, "loss": 0.94, "mean_token_accuracy": 0.8542112112045288, "num_tokens": 239862526.0, "step": 6320 }, { "epoch": 0.8040961709706145, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.8976616859436035, "learning_rate": 5e-05, "loss": 0.9529, "mean_token_accuracy": 0.8586368560791016, "num_tokens": 239897409.0, "step": 6321 }, { "epoch": 0.8042233812492049, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.785144805908203, "learning_rate": 5e-05, "loss": 0.8405, "mean_token_accuracy": 0.8677623271942139, "num_tokens": 239935202.0, "step": 6322 }, { "epoch": 0.8043505915277954, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.644257068634033, "learning_rate": 5e-05, "loss": 0.8758, "mean_token_accuracy": 0.8645018339157104, "num_tokens": 239967603.0, "step": 6323 }, { "epoch": 0.804477801806386, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.976867198944092, "learning_rate": 5e-05, "loss": 0.893, "mean_token_accuracy": 0.8534187078475952, "num_tokens": 240000757.0, "step": 6324 }, { "epoch": 0.8046050120849765, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.993885517120361, "learning_rate": 5e-05, "loss": 0.7889, "mean_token_accuracy": 0.8534163236618042, "num_tokens": 240039747.0, "step": 6325 }, { "epoch": 0.804732222363567, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.59729290008545, "learning_rate": 5e-05, "loss": 0.933, "mean_token_accuracy": 0.8567585945129395, "num_tokens": 240078951.0, "step": 6326 }, { "epoch": 0.8048594326421575, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.484111309051514, "learning_rate": 5e-05, "loss": 0.7983, "mean_token_accuracy": 0.8463702201843262, "num_tokens": 240120097.0, "step": 6327 }, { "epoch": 0.804986642920748, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.175786018371582, "learning_rate": 5e-05, "loss": 1.0596, "mean_token_accuracy": 0.8420741558074951, "num_tokens": 240166510.0, "step": 6328 }, { "epoch": 0.8051138531993385, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.626885414123535, "learning_rate": 5e-05, "loss": 0.7852, "mean_token_accuracy": 0.8614082932472229, "num_tokens": 240207961.0, "step": 6329 }, { "epoch": 0.805241063477929, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.171263694763184, "learning_rate": 5e-05, "loss": 0.9903, "mean_token_accuracy": 0.8628357648849487, "num_tokens": 240243257.0, "step": 6330 }, { "epoch": 0.8053682737565195, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.44729471206665, "learning_rate": 5e-05, "loss": 0.8874, "mean_token_accuracy": 0.8660584688186646, "num_tokens": 240279844.0, "step": 6331 }, { "epoch": 0.8054954840351101, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.27027702331543, "learning_rate": 5e-05, "loss": 0.9023, "mean_token_accuracy": 0.865023136138916, "num_tokens": 240316836.0, "step": 6332 }, { "epoch": 0.8056226943137006, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.156821250915527, "learning_rate": 5e-05, "loss": 1.0045, "mean_token_accuracy": 0.8449974656105042, "num_tokens": 240357794.0, "step": 6333 }, { "epoch": 0.805749904592291, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.19511079788208, "learning_rate": 5e-05, "loss": 0.8461, "mean_token_accuracy": 0.8422398567199707, "num_tokens": 240397929.0, "step": 6334 }, { "epoch": 0.8058771148708815, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.515058517456055, "learning_rate": 5e-05, "loss": 0.8944, "mean_token_accuracy": 0.8737555146217346, "num_tokens": 240434319.0, "step": 6335 }, { "epoch": 0.8060043251494721, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.77871036529541, "learning_rate": 5e-05, "loss": 0.7619, "mean_token_accuracy": 0.8584465980529785, "num_tokens": 240476755.0, "step": 6336 }, { "epoch": 0.8061315354280626, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.618165969848633, "learning_rate": 5e-05, "loss": 0.9006, "mean_token_accuracy": 0.8676483631134033, "num_tokens": 240513042.0, "step": 6337 }, { "epoch": 0.8062587457066531, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.777511119842529, "learning_rate": 5e-05, "loss": 0.7626, "mean_token_accuracy": 0.8651531934738159, "num_tokens": 240554805.0, "step": 6338 }, { "epoch": 0.8063859559852437, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.635534286499023, "learning_rate": 5e-05, "loss": 0.8871, "mean_token_accuracy": 0.874620795249939, "num_tokens": 240593196.0, "step": 6339 }, { "epoch": 0.8065131662638341, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.849898815155029, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8631577491760254, "num_tokens": 240632640.0, "step": 6340 }, { "epoch": 0.8066403765424246, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.635249137878418, "learning_rate": 5e-05, "loss": 0.9603, "mean_token_accuracy": 0.8651041984558105, "num_tokens": 240672868.0, "step": 6341 }, { "epoch": 0.8067675868210151, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.977619171142578, "learning_rate": 5e-05, "loss": 0.8555, "mean_token_accuracy": 0.8469672203063965, "num_tokens": 240713499.0, "step": 6342 }, { "epoch": 0.8068947970996057, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.490974426269531, "learning_rate": 5e-05, "loss": 0.8904, "mean_token_accuracy": 0.8730106353759766, "num_tokens": 240759684.0, "step": 6343 }, { "epoch": 0.8070220073781962, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.152575492858887, "learning_rate": 5e-05, "loss": 0.8707, "mean_token_accuracy": 0.8549536466598511, "num_tokens": 240799439.0, "step": 6344 }, { "epoch": 0.8071492176567867, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.289210319519043, "learning_rate": 5e-05, "loss": 0.9284, "mean_token_accuracy": 0.8568609356880188, "num_tokens": 240845569.0, "step": 6345 }, { "epoch": 0.8072764279353771, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.223720073699951, "learning_rate": 5e-05, "loss": 0.7777, "mean_token_accuracy": 0.8759728670120239, "num_tokens": 240879721.0, "step": 6346 }, { "epoch": 0.8074036382139677, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.140690803527832, "learning_rate": 5e-05, "loss": 0.8923, "mean_token_accuracy": 0.8627007007598877, "num_tokens": 240918433.0, "step": 6347 }, { "epoch": 0.8075308484925582, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.222184658050537, "learning_rate": 5e-05, "loss": 0.824, "mean_token_accuracy": 0.8559693694114685, "num_tokens": 240949650.0, "step": 6348 }, { "epoch": 0.8076580587711487, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.05881118774414, "learning_rate": 5e-05, "loss": 0.8494, "mean_token_accuracy": 0.8689534664154053, "num_tokens": 240984803.0, "step": 6349 }, { "epoch": 0.8077852690497392, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.194337368011475, "learning_rate": 5e-05, "loss": 0.805, "mean_token_accuracy": 0.8581270575523376, "num_tokens": 241023173.0, "step": 6350 }, { "epoch": 0.8079124793283298, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.9764275550842285, "learning_rate": 5e-05, "loss": 0.8518, "mean_token_accuracy": 0.8646052479743958, "num_tokens": 241065235.0, "step": 6351 }, { "epoch": 0.8080396896069202, "ewc_loss": 0.375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.1573710441589355, "learning_rate": 5e-05, "loss": 0.7786, "mean_token_accuracy": 0.8613609671592712, "num_tokens": 241106283.0, "step": 6352 }, { "epoch": 0.8081668998855107, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.007650375366211, "learning_rate": 5e-05, "loss": 0.886, "mean_token_accuracy": 0.8508090972900391, "num_tokens": 241145544.0, "step": 6353 }, { "epoch": 0.8082941101641012, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.070850372314453, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.8540524840354919, "num_tokens": 241183888.0, "step": 6354 }, { "epoch": 0.8084213204426918, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.041523933410645, "learning_rate": 5e-05, "loss": 0.8057, "mean_token_accuracy": 0.8764994144439697, "num_tokens": 241224650.0, "step": 6355 }, { "epoch": 0.8085485307212823, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.01002311706543, "learning_rate": 5e-05, "loss": 0.7689, "mean_token_accuracy": 0.8607155084609985, "num_tokens": 241263907.0, "step": 6356 }, { "epoch": 0.8086757409998728, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.084806442260742, "learning_rate": 5e-05, "loss": 0.9081, "mean_token_accuracy": 0.8388227224349976, "num_tokens": 241305009.0, "step": 6357 }, { "epoch": 0.8088029512784632, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.981773376464844, "learning_rate": 5e-05, "loss": 0.7437, "mean_token_accuracy": 0.8665053844451904, "num_tokens": 241344283.0, "step": 6358 }, { "epoch": 0.8089301615570538, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.096298217773438, "learning_rate": 5e-05, "loss": 0.8426, "mean_token_accuracy": 0.8627060055732727, "num_tokens": 241380243.0, "step": 6359 }, { "epoch": 0.8090573718356443, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.005451679229736, "learning_rate": 5e-05, "loss": 0.756, "mean_token_accuracy": 0.8600271344184875, "num_tokens": 241414373.0, "step": 6360 }, { "epoch": 0.8091845821142348, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.998253345489502, "learning_rate": 5e-05, "loss": 0.8331, "mean_token_accuracy": 0.8670672178268433, "num_tokens": 241456393.0, "step": 6361 }, { "epoch": 0.8093117923928254, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.128278732299805, "learning_rate": 5e-05, "loss": 0.8071, "mean_token_accuracy": 0.8510797619819641, "num_tokens": 241496086.0, "step": 6362 }, { "epoch": 0.8094390026714159, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.805966377258301, "learning_rate": 5e-05, "loss": 0.8009, "mean_token_accuracy": 0.8662468791007996, "num_tokens": 241537814.0, "step": 6363 }, { "epoch": 0.8095662129500064, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.310285568237305, "learning_rate": 5e-05, "loss": 0.7989, "mean_token_accuracy": 0.8558347225189209, "num_tokens": 241576762.0, "step": 6364 }, { "epoch": 0.8096934232285968, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.544692516326904, "learning_rate": 5e-05, "loss": 0.7506, "mean_token_accuracy": 0.8727765083312988, "num_tokens": 241614667.0, "step": 6365 }, { "epoch": 0.8098206335071874, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.489222526550293, "learning_rate": 5e-05, "loss": 0.7236, "mean_token_accuracy": 0.8756104707717896, "num_tokens": 241647216.0, "step": 6366 }, { "epoch": 0.8099478437857779, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.324253559112549, "learning_rate": 5e-05, "loss": 0.7291, "mean_token_accuracy": 0.8687703609466553, "num_tokens": 241684278.0, "step": 6367 }, { "epoch": 0.8100750540643684, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.622857093811035, "learning_rate": 5e-05, "loss": 0.7308, "mean_token_accuracy": 0.8777080774307251, "num_tokens": 241723326.0, "step": 6368 }, { "epoch": 0.8102022643429589, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.156801223754883, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.8668200969696045, "num_tokens": 241765251.0, "step": 6369 }, { "epoch": 0.8103294746215495, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.739184856414795, "learning_rate": 5e-05, "loss": 0.747, "mean_token_accuracy": 0.8714039921760559, "num_tokens": 241801862.0, "step": 6370 }, { "epoch": 0.8104566849001399, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.002967834472656, "learning_rate": 5e-05, "loss": 0.7272, "mean_token_accuracy": 0.8631315231323242, "num_tokens": 241841448.0, "step": 6371 }, { "epoch": 0.8105838951787304, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000385284423828125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.910089015960693, "learning_rate": 5e-05, "loss": 0.7935, "mean_token_accuracy": 0.8581149578094482, "num_tokens": 241883151.0, "step": 6372 }, { "epoch": 0.810711105457321, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000385284423828125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.791060447692871, "learning_rate": 5e-05, "loss": 0.7475, "mean_token_accuracy": 0.8534181118011475, "num_tokens": 241924643.0, "step": 6373 }, { "epoch": 0.8108383157359115, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000385284423828125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.162907600402832, "learning_rate": 5e-05, "loss": 0.8232, "mean_token_accuracy": 0.8555901646614075, "num_tokens": 241960361.0, "step": 6374 }, { "epoch": 0.810965526014502, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000385284423828125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.54053258895874, "learning_rate": 5e-05, "loss": 0.7563, "mean_token_accuracy": 0.8460643291473389, "num_tokens": 241999616.0, "step": 6375 }, { "epoch": 0.8110927362930925, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000385284423828125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.523990631103516, "learning_rate": 5e-05, "loss": 0.8623, "mean_token_accuracy": 0.8578188419342041, "num_tokens": 242035824.0, "step": 6376 }, { "epoch": 0.811219946571683, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.000385284423828125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.26472282409668, "learning_rate": 5e-05, "loss": 0.6913, "mean_token_accuracy": 0.8657784461975098, "num_tokens": 242072680.0, "step": 6377 }, { "epoch": 0.8113471568502735, "ewc_loss": 0.5, "ewc_loss_diag": 0.000385284423828125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.933768272399902, "learning_rate": 5e-05, "loss": 0.8932, "mean_token_accuracy": 0.8692444562911987, "num_tokens": 242111135.0, "step": 6378 }, { "epoch": 0.811474367128864, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.1815643310546875, "learning_rate": 5e-05, "loss": 0.6971, "mean_token_accuracy": 0.8698701858520508, "num_tokens": 242148993.0, "step": 6379 }, { "epoch": 0.8116015774074545, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.117412567138672, "learning_rate": 5e-05, "loss": 0.9997, "mean_token_accuracy": 0.8564786314964294, "num_tokens": 242186230.0, "step": 6380 }, { "epoch": 0.8117287876860451, "ewc_loss": 0.375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.696847438812256, "learning_rate": 5e-05, "loss": 0.7798, "mean_token_accuracy": 0.8643444776535034, "num_tokens": 242222341.0, "step": 6381 }, { "epoch": 0.8118559979646356, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.505094528198242, "learning_rate": 5e-05, "loss": 0.9503, "mean_token_accuracy": 0.8524494767189026, "num_tokens": 242255408.0, "step": 6382 }, { "epoch": 0.811983208243226, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.687178134918213, "learning_rate": 5e-05, "loss": 0.9666, "mean_token_accuracy": 0.8417102098464966, "num_tokens": 242294664.0, "step": 6383 }, { "epoch": 0.8121104185218165, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.128511905670166, "learning_rate": 5e-05, "loss": 0.7141, "mean_token_accuracy": 0.8801241517066956, "num_tokens": 242328982.0, "step": 6384 }, { "epoch": 0.8122376288004071, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.729233741760254, "learning_rate": 5e-05, "loss": 0.9855, "mean_token_accuracy": 0.8507713079452515, "num_tokens": 242369542.0, "step": 6385 }, { "epoch": 0.8123648390789976, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.128451824188232, "learning_rate": 5e-05, "loss": 0.7353, "mean_token_accuracy": 0.8557260036468506, "num_tokens": 242406699.0, "step": 6386 }, { "epoch": 0.8124920493575881, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.548683166503906, "learning_rate": 5e-05, "loss": 1.0174, "mean_token_accuracy": 0.8543100953102112, "num_tokens": 242448057.0, "step": 6387 }, { "epoch": 0.8126192596361786, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.167932510375977, "learning_rate": 5e-05, "loss": 0.7876, "mean_token_accuracy": 0.8502581119537354, "num_tokens": 242489792.0, "step": 6388 }, { "epoch": 0.8127464699147691, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.658927917480469, "learning_rate": 5e-05, "loss": 1.0616, "mean_token_accuracy": 0.854928731918335, "num_tokens": 242535458.0, "step": 6389 }, { "epoch": 0.8128736801933596, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 7.347264289855957, "learning_rate": 5e-05, "loss": 0.9511, "mean_token_accuracy": 0.8613777756690979, "num_tokens": 242577380.0, "step": 6390 }, { "epoch": 0.8130008904719501, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.570801734924316, "learning_rate": 5e-05, "loss": 0.9792, "mean_token_accuracy": 0.857239842414856, "num_tokens": 242614353.0, "step": 6391 }, { "epoch": 0.8131281007505406, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.359004974365234, "learning_rate": 5e-05, "loss": 0.9811, "mean_token_accuracy": 0.8690022230148315, "num_tokens": 242648444.0, "step": 6392 }, { "epoch": 0.8132553110291312, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.210260391235352, "learning_rate": 5e-05, "loss": 0.8306, "mean_token_accuracy": 0.8591862320899963, "num_tokens": 242687658.0, "step": 6393 }, { "epoch": 0.8133825213077217, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.797809600830078, "learning_rate": 5e-05, "loss": 0.9594, "mean_token_accuracy": 0.8607809543609619, "num_tokens": 242726097.0, "step": 6394 }, { "epoch": 0.8135097315863121, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.034977436065674, "learning_rate": 5e-05, "loss": 0.8523, "mean_token_accuracy": 0.8536572456359863, "num_tokens": 242769229.0, "step": 6395 }, { "epoch": 0.8136369418649027, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.308611869812012, "learning_rate": 5e-05, "loss": 0.9201, "mean_token_accuracy": 0.857550859451294, "num_tokens": 242805815.0, "step": 6396 }, { "epoch": 0.8137641521434932, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.597812175750732, "learning_rate": 5e-05, "loss": 0.9236, "mean_token_accuracy": 0.8452430963516235, "num_tokens": 242835929.0, "step": 6397 }, { "epoch": 0.8138913624220837, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.415709495544434, "learning_rate": 5e-05, "loss": 0.7807, "mean_token_accuracy": 0.8703258037567139, "num_tokens": 242875295.0, "step": 6398 }, { "epoch": 0.8140185727006742, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.079899787902832, "learning_rate": 5e-05, "loss": 0.8964, "mean_token_accuracy": 0.8577710390090942, "num_tokens": 242911825.0, "step": 6399 }, { "epoch": 0.8141457829792648, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.980085849761963, "learning_rate": 5e-05, "loss": 0.7483, "mean_token_accuracy": 0.8640910387039185, "num_tokens": 242950568.0, "step": 6400 }, { "epoch": 0.8142729932578552, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.230966567993164, "learning_rate": 5e-05, "loss": 0.8535, "mean_token_accuracy": 0.86893230676651, "num_tokens": 242981223.0, "step": 6401 }, { "epoch": 0.8144002035364457, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.861135959625244, "learning_rate": 5e-05, "loss": 0.6866, "mean_token_accuracy": 0.8806396722793579, "num_tokens": 243016261.0, "step": 6402 }, { "epoch": 0.8145274138150362, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.385872840881348, "learning_rate": 5e-05, "loss": 0.9409, "mean_token_accuracy": 0.847295343875885, "num_tokens": 243056213.0, "step": 6403 }, { "epoch": 0.8146546240936268, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.729776859283447, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8635689616203308, "num_tokens": 243097184.0, "step": 6404 }, { "epoch": 0.8147818343722173, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.582605361938477, "learning_rate": 5e-05, "loss": 0.9146, "mean_token_accuracy": 0.8601188659667969, "num_tokens": 243136995.0, "step": 6405 }, { "epoch": 0.8149090446508078, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.644657611846924, "learning_rate": 5e-05, "loss": 0.7313, "mean_token_accuracy": 0.8752551674842834, "num_tokens": 243173495.0, "step": 6406 }, { "epoch": 0.8150362549293982, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.722285270690918, "learning_rate": 5e-05, "loss": 0.9325, "mean_token_accuracy": 0.8585993647575378, "num_tokens": 243209198.0, "step": 6407 }, { "epoch": 0.8151634652079888, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.663894176483154, "learning_rate": 5e-05, "loss": 0.7525, "mean_token_accuracy": 0.8674337863922119, "num_tokens": 243249777.0, "step": 6408 }, { "epoch": 0.8152906754865793, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.800908088684082, "learning_rate": 5e-05, "loss": 1.0018, "mean_token_accuracy": 0.8512230515480042, "num_tokens": 243293094.0, "step": 6409 }, { "epoch": 0.8154178857651698, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 6.791377544403076, "learning_rate": 5e-05, "loss": 0.8768, "mean_token_accuracy": 0.8440309762954712, "num_tokens": 243330881.0, "step": 6410 }, { "epoch": 0.8155450960437604, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.772048950195312, "learning_rate": 5e-05, "loss": 0.9976, "mean_token_accuracy": 0.8538666367530823, "num_tokens": 243371292.0, "step": 6411 }, { "epoch": 0.8156723063223509, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.000405788421631, "learning_rate": 5e-05, "loss": 0.8773, "mean_token_accuracy": 0.8453353643417358, "num_tokens": 243404925.0, "step": 6412 }, { "epoch": 0.8157995166009414, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.5921630859375, "learning_rate": 5e-05, "loss": 0.9353, "mean_token_accuracy": 0.8693162202835083, "num_tokens": 243442377.0, "step": 6413 }, { "epoch": 0.8159267268795318, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.242100238800049, "learning_rate": 5e-05, "loss": 0.8409, "mean_token_accuracy": 0.8678265810012817, "num_tokens": 243476857.0, "step": 6414 }, { "epoch": 0.8160539371581224, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.294331550598145, "learning_rate": 5e-05, "loss": 0.874, "mean_token_accuracy": 0.8799017071723938, "num_tokens": 243506641.0, "step": 6415 }, { "epoch": 0.8161811474367129, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.360981464385986, "learning_rate": 5e-05, "loss": 0.8896, "mean_token_accuracy": 0.8566598892211914, "num_tokens": 243547555.0, "step": 6416 }, { "epoch": 0.8163083577153034, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.055421829223633, "learning_rate": 5e-05, "loss": 0.9582, "mean_token_accuracy": 0.8503707051277161, "num_tokens": 243588210.0, "step": 6417 }, { "epoch": 0.8164355679938939, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.35167932510376, "learning_rate": 5e-05, "loss": 0.8451, "mean_token_accuracy": 0.8595129251480103, "num_tokens": 243626913.0, "step": 6418 }, { "epoch": 0.8165627782724845, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.943683624267578, "learning_rate": 5e-05, "loss": 0.8554, "mean_token_accuracy": 0.8735554218292236, "num_tokens": 243662645.0, "step": 6419 }, { "epoch": 0.8166899885510749, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.289752006530762, "learning_rate": 5e-05, "loss": 0.8121, "mean_token_accuracy": 0.8648237586021423, "num_tokens": 243704603.0, "step": 6420 }, { "epoch": 0.8168171988296654, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.872696876525879, "learning_rate": 5e-05, "loss": 0.8727, "mean_token_accuracy": 0.8565068244934082, "num_tokens": 243739571.0, "step": 6421 }, { "epoch": 0.8169444091082559, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.221884250640869, "learning_rate": 5e-05, "loss": 0.7577, "mean_token_accuracy": 0.8759143352508545, "num_tokens": 243782047.0, "step": 6422 }, { "epoch": 0.8170716193868465, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.86116361618042, "learning_rate": 5e-05, "loss": 0.8889, "mean_token_accuracy": 0.843431830406189, "num_tokens": 243825159.0, "step": 6423 }, { "epoch": 0.817198829665437, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.165092468261719, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.8588608503341675, "num_tokens": 243865627.0, "step": 6424 }, { "epoch": 0.8173260399440275, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.833539962768555, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.8574051856994629, "num_tokens": 243905750.0, "step": 6425 }, { "epoch": 0.8174532502226179, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0003871917724609375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.168821334838867, "learning_rate": 5e-05, "loss": 0.7861, "mean_token_accuracy": 0.8537368774414062, "num_tokens": 243936176.0, "step": 6426 }, { "epoch": 0.8175804605012085, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.7546162605285645, "learning_rate": 5e-05, "loss": 0.781, "mean_token_accuracy": 0.8722290396690369, "num_tokens": 243979191.0, "step": 6427 }, { "epoch": 0.817707670779799, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.198265075683594, "learning_rate": 5e-05, "loss": 0.8039, "mean_token_accuracy": 0.8515652418136597, "num_tokens": 244025074.0, "step": 6428 }, { "epoch": 0.8178348810583895, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.6982645988464355, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8648576736450195, "num_tokens": 244062033.0, "step": 6429 }, { "epoch": 0.8179620913369801, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.206608295440674, "learning_rate": 5e-05, "loss": 0.7045, "mean_token_accuracy": 0.8774389028549194, "num_tokens": 244100025.0, "step": 6430 }, { "epoch": 0.8180893016155706, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.6492486000061035, "learning_rate": 5e-05, "loss": 0.7325, "mean_token_accuracy": 0.8804072141647339, "num_tokens": 244136253.0, "step": 6431 }, { "epoch": 0.818216511894161, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.2105278968811035, "learning_rate": 5e-05, "loss": 0.7509, "mean_token_accuracy": 0.8639837503433228, "num_tokens": 244174179.0, "step": 6432 }, { "epoch": 0.8183437221727515, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.620682239532471, "learning_rate": 5e-05, "loss": 0.7493, "mean_token_accuracy": 0.8734666705131531, "num_tokens": 244209096.0, "step": 6433 }, { "epoch": 0.8184709324513421, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.18243408203125, "learning_rate": 5e-05, "loss": 0.7227, "mean_token_accuracy": 0.8690606355667114, "num_tokens": 244248920.0, "step": 6434 }, { "epoch": 0.8185981427299326, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.64607572555542, "learning_rate": 5e-05, "loss": 0.7848, "mean_token_accuracy": 0.8593034744262695, "num_tokens": 244289338.0, "step": 6435 }, { "epoch": 0.8187253530085231, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.137289524078369, "learning_rate": 5e-05, "loss": 0.7123, "mean_token_accuracy": 0.8728855848312378, "num_tokens": 244324844.0, "step": 6436 }, { "epoch": 0.8188525632871136, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.687349796295166, "learning_rate": 5e-05, "loss": 0.8335, "mean_token_accuracy": 0.844085156917572, "num_tokens": 244364815.0, "step": 6437 }, { "epoch": 0.8189797735657041, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.05214786529541, "learning_rate": 5e-05, "loss": 0.7263, "mean_token_accuracy": 0.8632885813713074, "num_tokens": 244407509.0, "step": 6438 }, { "epoch": 0.8191069838442946, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.767326831817627, "learning_rate": 5e-05, "loss": 0.7444, "mean_token_accuracy": 0.8752949237823486, "num_tokens": 244446663.0, "step": 6439 }, { "epoch": 0.8192341941228851, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.953480243682861, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8573828339576721, "num_tokens": 244492292.0, "step": 6440 }, { "epoch": 0.8193614044014756, "ewc_loss": 0.375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.883210182189941, "learning_rate": 5e-05, "loss": 0.7742, "mean_token_accuracy": 0.8611790537834167, "num_tokens": 244531390.0, "step": 6441 }, { "epoch": 0.8194886146800662, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.856073379516602, "learning_rate": 5e-05, "loss": 0.7052, "mean_token_accuracy": 0.866203248500824, "num_tokens": 244567377.0, "step": 6442 }, { "epoch": 0.8196158249586567, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.000574111938477, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8765803575515747, "num_tokens": 244605690.0, "step": 6443 }, { "epoch": 0.8197430352372471, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.7365827560424805, "learning_rate": 5e-05, "loss": 0.7104, "mean_token_accuracy": 0.8629401326179504, "num_tokens": 244643512.0, "step": 6444 }, { "epoch": 0.8198702455158376, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.158393859863281, "learning_rate": 5e-05, "loss": 0.8545, "mean_token_accuracy": 0.8501508235931396, "num_tokens": 244691048.0, "step": 6445 }, { "epoch": 0.8199974557944282, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.6013078689575195, "learning_rate": 5e-05, "loss": 0.7852, "mean_token_accuracy": 0.8409078121185303, "num_tokens": 244734205.0, "step": 6446 }, { "epoch": 0.8201246660730187, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.328575134277344, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.859606921672821, "num_tokens": 244768302.0, "step": 6447 }, { "epoch": 0.8202518763516092, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.5592474937438965, "learning_rate": 5e-05, "loss": 0.7207, "mean_token_accuracy": 0.8624476194381714, "num_tokens": 244808880.0, "step": 6448 }, { "epoch": 0.8203790866301998, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.356225967407227, "learning_rate": 5e-05, "loss": 0.8277, "mean_token_accuracy": 0.8716028928756714, "num_tokens": 244845507.0, "step": 6449 }, { "epoch": 0.8205062969087902, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.715476036071777, "learning_rate": 5e-05, "loss": 0.7762, "mean_token_accuracy": 0.8474952578544617, "num_tokens": 244888142.0, "step": 6450 }, { "epoch": 0.8206335071873807, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.071459770202637, "learning_rate": 5e-05, "loss": 0.7854, "mean_token_accuracy": 0.8780149221420288, "num_tokens": 244928612.0, "step": 6451 }, { "epoch": 0.8207607174659712, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.154661655426025, "learning_rate": 5e-05, "loss": 0.7653, "mean_token_accuracy": 0.8640514612197876, "num_tokens": 244969879.0, "step": 6452 }, { "epoch": 0.8208879277445618, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.417046070098877, "learning_rate": 5e-05, "loss": 0.7813, "mean_token_accuracy": 0.8609482049942017, "num_tokens": 245004597.0, "step": 6453 }, { "epoch": 0.8210151380231523, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.828465938568115, "learning_rate": 5e-05, "loss": 0.8443, "mean_token_accuracy": 0.8538846969604492, "num_tokens": 245050087.0, "step": 6454 }, { "epoch": 0.8211423483017428, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.585997581481934, "learning_rate": 5e-05, "loss": 0.749, "mean_token_accuracy": 0.8530969619750977, "num_tokens": 245091626.0, "step": 6455 }, { "epoch": 0.8212695585803332, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.7006196975708, "learning_rate": 5e-05, "loss": 0.8732, "mean_token_accuracy": 0.8583725094795227, "num_tokens": 245126780.0, "step": 6456 }, { "epoch": 0.8213967688589238, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00022029876708984375, "grad_norm": 5.786067485809326, "learning_rate": 5e-05, "loss": 0.7229, "mean_token_accuracy": 0.8447204828262329, "num_tokens": 245169226.0, "step": 6457 }, { "epoch": 0.8215239791375143, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.848526954650879, "learning_rate": 5e-05, "loss": 1.0487, "mean_token_accuracy": 0.8489063382148743, "num_tokens": 245204983.0, "step": 6458 }, { "epoch": 0.8216511894161048, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.534234523773193, "learning_rate": 5e-05, "loss": 0.6922, "mean_token_accuracy": 0.868148922920227, "num_tokens": 245249162.0, "step": 6459 }, { "epoch": 0.8217783996946953, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 10.908516883850098, "learning_rate": 5e-05, "loss": 1.2493, "mean_token_accuracy": 0.848018229007721, "num_tokens": 245288883.0, "step": 6460 }, { "epoch": 0.8219056099732859, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 7.514387130737305, "learning_rate": 5e-05, "loss": 1.0211, "mean_token_accuracy": 0.8589158058166504, "num_tokens": 245328578.0, "step": 6461 }, { "epoch": 0.8220328202518764, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.169919967651367, "learning_rate": 5e-05, "loss": 1.1124, "mean_token_accuracy": 0.8572068810462952, "num_tokens": 245368086.0, "step": 6462 }, { "epoch": 0.8221600305304668, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 9.149799346923828, "learning_rate": 5e-05, "loss": 1.2551, "mean_token_accuracy": 0.8634437918663025, "num_tokens": 245405234.0, "step": 6463 }, { "epoch": 0.8222872408090574, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.968906402587891, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8698206543922424, "num_tokens": 245439759.0, "step": 6464 }, { "epoch": 0.8224144510876479, "ewc_loss": 0.82421875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 10.474021911621094, "learning_rate": 5e-05, "loss": 1.2393, "mean_token_accuracy": 0.8573374152183533, "num_tokens": 245474689.0, "step": 6465 }, { "epoch": 0.8225416613662384, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.300319194793701, "learning_rate": 5e-05, "loss": 0.9197, "mean_token_accuracy": 0.8628960847854614, "num_tokens": 245509494.0, "step": 6466 }, { "epoch": 0.8226688716448289, "ewc_loss": 0.8515625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 11.675313949584961, "learning_rate": 5e-05, "loss": 1.2479, "mean_token_accuracy": 0.8645327091217041, "num_tokens": 245550798.0, "step": 6467 }, { "epoch": 0.8227960819234195, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.118905067443848, "learning_rate": 5e-05, "loss": 1.2283, "mean_token_accuracy": 0.86268150806427, "num_tokens": 245588929.0, "step": 6468 }, { "epoch": 0.8229232922020099, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.976188659667969, "learning_rate": 5e-05, "loss": 1.0769, "mean_token_accuracy": 0.864995002746582, "num_tokens": 245625952.0, "step": 6469 }, { "epoch": 0.8230505024806004, "ewc_loss": 0.75, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 8.918888092041016, "learning_rate": 5e-05, "loss": 1.1812, "mean_token_accuracy": 0.8514132499694824, "num_tokens": 245668156.0, "step": 6470 }, { "epoch": 0.8231777127591909, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 10.664709091186523, "learning_rate": 5e-05, "loss": 1.1264, "mean_token_accuracy": 0.8625032901763916, "num_tokens": 245707117.0, "step": 6471 }, { "epoch": 0.8233049230377815, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.203629493713379, "learning_rate": 5e-05, "loss": 1.0229, "mean_token_accuracy": 0.8555896878242493, "num_tokens": 245747248.0, "step": 6472 }, { "epoch": 0.823432133316372, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 10.618242263793945, "learning_rate": 5e-05, "loss": 1.1397, "mean_token_accuracy": 0.8569037914276123, "num_tokens": 245795760.0, "step": 6473 }, { "epoch": 0.8235593435949625, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.016836166381836, "learning_rate": 5e-05, "loss": 0.9527, "mean_token_accuracy": 0.866629421710968, "num_tokens": 245827778.0, "step": 6474 }, { "epoch": 0.8236865538735529, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.864584922790527, "learning_rate": 5e-05, "loss": 1.1326, "mean_token_accuracy": 0.8660982847213745, "num_tokens": 245866869.0, "step": 6475 }, { "epoch": 0.8238137641521435, "ewc_loss": 0.625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.484619140625, "learning_rate": 5e-05, "loss": 0.9848, "mean_token_accuracy": 0.8752622604370117, "num_tokens": 245904686.0, "step": 6476 }, { "epoch": 0.823940974430734, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.80774211883545, "learning_rate": 5e-05, "loss": 1.1583, "mean_token_accuracy": 0.8581333756446838, "num_tokens": 245942169.0, "step": 6477 }, { "epoch": 0.8240681847093245, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 8.863670349121094, "learning_rate": 5e-05, "loss": 1.0693, "mean_token_accuracy": 0.8575236797332764, "num_tokens": 245979957.0, "step": 6478 }, { "epoch": 0.824195394987915, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.390873908996582, "learning_rate": 5e-05, "loss": 1.0736, "mean_token_accuracy": 0.8697848320007324, "num_tokens": 246019433.0, "step": 6479 }, { "epoch": 0.8243226052665056, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.8472900390625, "learning_rate": 5e-05, "loss": 1.0266, "mean_token_accuracy": 0.8600589036941528, "num_tokens": 246057595.0, "step": 6480 }, { "epoch": 0.824449815545096, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.90978717803955, "learning_rate": 5e-05, "loss": 1.0204, "mean_token_accuracy": 0.8728737831115723, "num_tokens": 246100141.0, "step": 6481 }, { "epoch": 0.8245770258236865, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.59776496887207, "learning_rate": 5e-05, "loss": 1.0445, "mean_token_accuracy": 0.8421716690063477, "num_tokens": 246141642.0, "step": 6482 }, { "epoch": 0.824704236102277, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.530173301696777, "learning_rate": 5e-05, "loss": 1.0394, "mean_token_accuracy": 0.8601747751235962, "num_tokens": 246177844.0, "step": 6483 }, { "epoch": 0.8248314463808676, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.220396041870117, "learning_rate": 5e-05, "loss": 0.9382, "mean_token_accuracy": 0.8593568801879883, "num_tokens": 246217470.0, "step": 6484 }, { "epoch": 0.8249586566594581, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.46565055847168, "learning_rate": 5e-05, "loss": 0.9993, "mean_token_accuracy": 0.8691504001617432, "num_tokens": 246256637.0, "step": 6485 }, { "epoch": 0.8250858669380486, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.889925956726074, "learning_rate": 5e-05, "loss": 0.8603, "mean_token_accuracy": 0.8704791069030762, "num_tokens": 246298894.0, "step": 6486 }, { "epoch": 0.8252130772166391, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.635002136230469, "learning_rate": 5e-05, "loss": 1.0543, "mean_token_accuracy": 0.8549913167953491, "num_tokens": 246340986.0, "step": 6487 }, { "epoch": 0.8253402874952296, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.915535926818848, "learning_rate": 5e-05, "loss": 0.9722, "mean_token_accuracy": 0.8339318633079529, "num_tokens": 246379201.0, "step": 6488 }, { "epoch": 0.8254674977738201, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.65275764465332, "learning_rate": 5e-05, "loss": 1.0149, "mean_token_accuracy": 0.870044469833374, "num_tokens": 246411605.0, "step": 6489 }, { "epoch": 0.8255947080524106, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.196170806884766, "learning_rate": 5e-05, "loss": 0.9167, "mean_token_accuracy": 0.8640092015266418, "num_tokens": 246450908.0, "step": 6490 }, { "epoch": 0.8257219183310012, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.24341869354248, "learning_rate": 5e-05, "loss": 1.0176, "mean_token_accuracy": 0.8576855659484863, "num_tokens": 246493132.0, "step": 6491 }, { "epoch": 0.8258491286095917, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.242422103881836, "learning_rate": 5e-05, "loss": 0.9482, "mean_token_accuracy": 0.8504604697227478, "num_tokens": 246528532.0, "step": 6492 }, { "epoch": 0.8259763388881821, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.757882118225098, "learning_rate": 5e-05, "loss": 0.9608, "mean_token_accuracy": 0.8647574782371521, "num_tokens": 246567328.0, "step": 6493 }, { "epoch": 0.8261035491667726, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.936581611633301, "learning_rate": 5e-05, "loss": 0.9545, "mean_token_accuracy": 0.8397074937820435, "num_tokens": 246613798.0, "step": 6494 }, { "epoch": 0.8262307594453632, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.60086441040039, "learning_rate": 5e-05, "loss": 0.9212, "mean_token_accuracy": 0.8693075776100159, "num_tokens": 246649117.0, "step": 6495 }, { "epoch": 0.8263579697239537, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.533661365509033, "learning_rate": 5e-05, "loss": 0.7962, "mean_token_accuracy": 0.8657636046409607, "num_tokens": 246690304.0, "step": 6496 }, { "epoch": 0.8264851800025442, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.722214698791504, "learning_rate": 5e-05, "loss": 0.916, "mean_token_accuracy": 0.8677690625190735, "num_tokens": 246728103.0, "step": 6497 }, { "epoch": 0.8266123902811348, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.332273483276367, "learning_rate": 5e-05, "loss": 0.7724, "mean_token_accuracy": 0.8722113966941833, "num_tokens": 246763902.0, "step": 6498 }, { "epoch": 0.8267396005597252, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.832777976989746, "learning_rate": 5e-05, "loss": 1.0164, "mean_token_accuracy": 0.8449959754943848, "num_tokens": 246803899.0, "step": 6499 }, { "epoch": 0.8268668108383157, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.390688419342041, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.8543386459350586, "num_tokens": 246840464.0, "step": 6500 }, { "epoch": 0.8269940211169062, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.733813285827637, "learning_rate": 5e-05, "loss": 0.9489, "mean_token_accuracy": 0.8632718324661255, "num_tokens": 246886743.0, "step": 6501 }, { "epoch": 0.8271212313954968, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.543242454528809, "learning_rate": 5e-05, "loss": 0.8024, "mean_token_accuracy": 0.8631702661514282, "num_tokens": 246925260.0, "step": 6502 }, { "epoch": 0.8272484416740873, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.450931549072266, "learning_rate": 5e-05, "loss": 0.9479, "mean_token_accuracy": 0.8549425601959229, "num_tokens": 246961523.0, "step": 6503 }, { "epoch": 0.8273756519526778, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.628812789916992, "learning_rate": 5e-05, "loss": 0.812, "mean_token_accuracy": 0.8663370013237, "num_tokens": 246996881.0, "step": 6504 }, { "epoch": 0.8275028622312682, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.08767032623291, "learning_rate": 5e-05, "loss": 0.9015, "mean_token_accuracy": 0.8576765060424805, "num_tokens": 247033492.0, "step": 6505 }, { "epoch": 0.8276300725098588, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.55942964553833, "learning_rate": 5e-05, "loss": 0.7887, "mean_token_accuracy": 0.8662514090538025, "num_tokens": 247065078.0, "step": 6506 }, { "epoch": 0.8277572827884493, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.909175395965576, "learning_rate": 5e-05, "loss": 0.8737, "mean_token_accuracy": 0.8522363901138306, "num_tokens": 247103876.0, "step": 6507 }, { "epoch": 0.8278844930670398, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.40570068359375, "learning_rate": 5e-05, "loss": 0.8279, "mean_token_accuracy": 0.853930652141571, "num_tokens": 247142132.0, "step": 6508 }, { "epoch": 0.8280117033456303, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.899188995361328, "learning_rate": 5e-05, "loss": 0.8222, "mean_token_accuracy": 0.8628577589988708, "num_tokens": 247182141.0, "step": 6509 }, { "epoch": 0.8281389136242209, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.265575408935547, "learning_rate": 5e-05, "loss": 0.8342, "mean_token_accuracy": 0.842671811580658, "num_tokens": 247218240.0, "step": 6510 }, { "epoch": 0.8282661239028114, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.864169120788574, "learning_rate": 5e-05, "loss": 0.8196, "mean_token_accuracy": 0.8592400550842285, "num_tokens": 247257370.0, "step": 6511 }, { "epoch": 0.8283933341814018, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.232472896575928, "learning_rate": 5e-05, "loss": 0.7521, "mean_token_accuracy": 0.8643163442611694, "num_tokens": 247294205.0, "step": 6512 }, { "epoch": 0.8285205444599923, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.84486198425293, "learning_rate": 5e-05, "loss": 0.8476, "mean_token_accuracy": 0.847962498664856, "num_tokens": 247329974.0, "step": 6513 }, { "epoch": 0.8286477547385829, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.164356708526611, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8596153855323792, "num_tokens": 247365279.0, "step": 6514 }, { "epoch": 0.8287749650171734, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.886312484741211, "learning_rate": 5e-05, "loss": 0.7538, "mean_token_accuracy": 0.8768474459648132, "num_tokens": 247405319.0, "step": 6515 }, { "epoch": 0.8289021752957639, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.081635475158691, "learning_rate": 5e-05, "loss": 0.6886, "mean_token_accuracy": 0.8742766380310059, "num_tokens": 247444399.0, "step": 6516 }, { "epoch": 0.8290293855743545, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.912724494934082, "learning_rate": 5e-05, "loss": 0.7918, "mean_token_accuracy": 0.8673866987228394, "num_tokens": 247475637.0, "step": 6517 }, { "epoch": 0.8291565958529449, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.084828853607178, "learning_rate": 5e-05, "loss": 0.7213, "mean_token_accuracy": 0.8648051023483276, "num_tokens": 247511009.0, "step": 6518 }, { "epoch": 0.8292838061315354, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.8216633796691895, "learning_rate": 5e-05, "loss": 0.793, "mean_token_accuracy": 0.8646368980407715, "num_tokens": 247549188.0, "step": 6519 }, { "epoch": 0.8294110164101259, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.199319839477539, "learning_rate": 5e-05, "loss": 0.7537, "mean_token_accuracy": 0.8565312027931213, "num_tokens": 247591914.0, "step": 6520 }, { "epoch": 0.8295382266887165, "ewc_loss": 0.375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.665345668792725, "learning_rate": 5e-05, "loss": 0.7708, "mean_token_accuracy": 0.8630561232566833, "num_tokens": 247626324.0, "step": 6521 }, { "epoch": 0.829665436967307, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.321505069732666, "learning_rate": 5e-05, "loss": 0.7577, "mean_token_accuracy": 0.8608235716819763, "num_tokens": 247662846.0, "step": 6522 }, { "epoch": 0.8297926472458975, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.471042633056641, "learning_rate": 5e-05, "loss": 0.7338, "mean_token_accuracy": 0.8673840761184692, "num_tokens": 247700831.0, "step": 6523 }, { "epoch": 0.8299198575244879, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.488420486450195, "learning_rate": 5e-05, "loss": 0.7662, "mean_token_accuracy": 0.8611379265785217, "num_tokens": 247740535.0, "step": 6524 }, { "epoch": 0.8300470678030785, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.256099700927734, "learning_rate": 5e-05, "loss": 0.7502, "mean_token_accuracy": 0.8603428602218628, "num_tokens": 247779490.0, "step": 6525 }, { "epoch": 0.830174278081669, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.683077812194824, "learning_rate": 5e-05, "loss": 0.7901, "mean_token_accuracy": 0.8524794578552246, "num_tokens": 247815867.0, "step": 6526 }, { "epoch": 0.8303014883602595, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.002810955047607, "learning_rate": 5e-05, "loss": 0.8285, "mean_token_accuracy": 0.8324871063232422, "num_tokens": 247854871.0, "step": 6527 }, { "epoch": 0.83042869863885, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.945992946624756, "learning_rate": 5e-05, "loss": 0.812, "mean_token_accuracy": 0.852487325668335, "num_tokens": 247897711.0, "step": 6528 }, { "epoch": 0.8305559089174406, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.7075514793396, "learning_rate": 5e-05, "loss": 0.6999, "mean_token_accuracy": 0.8602216839790344, "num_tokens": 247933792.0, "step": 6529 }, { "epoch": 0.830683119196031, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.296921730041504, "learning_rate": 5e-05, "loss": 0.7966, "mean_token_accuracy": 0.8651286959648132, "num_tokens": 247964183.0, "step": 6530 }, { "epoch": 0.8308103294746215, "ewc_loss": 0.28125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.408167839050293, "learning_rate": 5e-05, "loss": 0.7143, "mean_token_accuracy": 0.8514600992202759, "num_tokens": 248003064.0, "step": 6531 }, { "epoch": 0.830937539753212, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.709452629089355, "learning_rate": 5e-05, "loss": 0.8468, "mean_token_accuracy": 0.8670425415039062, "num_tokens": 248041025.0, "step": 6532 }, { "epoch": 0.8310647500318026, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.217095851898193, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8537169098854065, "num_tokens": 248079503.0, "step": 6533 }, { "epoch": 0.8311919603103931, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.915626525878906, "learning_rate": 5e-05, "loss": 0.9048, "mean_token_accuracy": 0.8603652715682983, "num_tokens": 248119541.0, "step": 6534 }, { "epoch": 0.8313191705889836, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.468676567077637, "learning_rate": 5e-05, "loss": 0.758, "mean_token_accuracy": 0.8533132076263428, "num_tokens": 248163359.0, "step": 6535 }, { "epoch": 0.831446380867574, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.567902565002441, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.8723477125167847, "num_tokens": 248204384.0, "step": 6536 }, { "epoch": 0.8315735911461646, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.197587966918945, "learning_rate": 5e-05, "loss": 0.7648, "mean_token_accuracy": 0.8696300983428955, "num_tokens": 248238812.0, "step": 6537 }, { "epoch": 0.8317008014247551, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.573731422424316, "learning_rate": 5e-05, "loss": 0.8073, "mean_token_accuracy": 0.8626031875610352, "num_tokens": 248277725.0, "step": 6538 }, { "epoch": 0.8318280117033456, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.090802192687988, "learning_rate": 5e-05, "loss": 0.8916, "mean_token_accuracy": 0.8560073971748352, "num_tokens": 248317591.0, "step": 6539 }, { "epoch": 0.8319552219819362, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.573614597320557, "learning_rate": 5e-05, "loss": 0.6956, "mean_token_accuracy": 0.87016761302948, "num_tokens": 248356330.0, "step": 6540 }, { "epoch": 0.8320824322605267, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.87696647644043, "learning_rate": 5e-05, "loss": 0.8803, "mean_token_accuracy": 0.8640766739845276, "num_tokens": 248390854.0, "step": 6541 }, { "epoch": 0.8322096425391171, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.110958576202393, "learning_rate": 5e-05, "loss": 0.6665, "mean_token_accuracy": 0.8730480074882507, "num_tokens": 248431581.0, "step": 6542 }, { "epoch": 0.8323368528177076, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.311532020568848, "learning_rate": 5e-05, "loss": 0.9434, "mean_token_accuracy": 0.8704325556755066, "num_tokens": 248474009.0, "step": 6543 }, { "epoch": 0.8324640630962982, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.506115436553955, "learning_rate": 5e-05, "loss": 0.7172, "mean_token_accuracy": 0.8722608685493469, "num_tokens": 248512552.0, "step": 6544 }, { "epoch": 0.8325912733748887, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.891778945922852, "learning_rate": 5e-05, "loss": 0.9591, "mean_token_accuracy": 0.8673503398895264, "num_tokens": 248549729.0, "step": 6545 }, { "epoch": 0.8327184836534792, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.503456115722656, "learning_rate": 5e-05, "loss": 0.8703, "mean_token_accuracy": 0.8623537421226501, "num_tokens": 248589481.0, "step": 6546 }, { "epoch": 0.8328456939320698, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.6987457275390625, "learning_rate": 5e-05, "loss": 0.8087, "mean_token_accuracy": 0.8720528483390808, "num_tokens": 248621871.0, "step": 6547 }, { "epoch": 0.8329729042106602, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.290925979614258, "learning_rate": 5e-05, "loss": 0.9141, "mean_token_accuracy": 0.8605157136917114, "num_tokens": 248659416.0, "step": 6548 }, { "epoch": 0.8331001144892507, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.9269914627075195, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.861966073513031, "num_tokens": 248700052.0, "step": 6549 }, { "epoch": 0.8332273247678412, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.376477241516113, "learning_rate": 5e-05, "loss": 0.9026, "mean_token_accuracy": 0.8526243567466736, "num_tokens": 248731965.0, "step": 6550 }, { "epoch": 0.8333545350464318, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.169874668121338, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.855363667011261, "num_tokens": 248770030.0, "step": 6551 }, { "epoch": 0.8334817453250223, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.630927562713623, "learning_rate": 5e-05, "loss": 0.8358, "mean_token_accuracy": 0.852675199508667, "num_tokens": 248811990.0, "step": 6552 }, { "epoch": 0.8336089556036128, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.880486011505127, "learning_rate": 5e-05, "loss": 0.8708, "mean_token_accuracy": 0.8471668362617493, "num_tokens": 248842728.0, "step": 6553 }, { "epoch": 0.8337361658822032, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.81210470199585, "learning_rate": 5e-05, "loss": 0.7737, "mean_token_accuracy": 0.8465877175331116, "num_tokens": 248885084.0, "step": 6554 }, { "epoch": 0.8338633761607938, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.53480052947998, "learning_rate": 5e-05, "loss": 0.8397, "mean_token_accuracy": 0.870578408241272, "num_tokens": 248923701.0, "step": 6555 }, { "epoch": 0.8339905864393843, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.365669250488281, "learning_rate": 5e-05, "loss": 0.6812, "mean_token_accuracy": 0.8725626468658447, "num_tokens": 248960796.0, "step": 6556 }, { "epoch": 0.8341177967179748, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.026094436645508, "learning_rate": 5e-05, "loss": 0.9305, "mean_token_accuracy": 0.864862859249115, "num_tokens": 249000628.0, "step": 6557 }, { "epoch": 0.8342450069965653, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.3730645179748535, "learning_rate": 5e-05, "loss": 0.7147, "mean_token_accuracy": 0.8664156198501587, "num_tokens": 249034409.0, "step": 6558 }, { "epoch": 0.8343722172751559, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.175047874450684, "learning_rate": 5e-05, "loss": 0.953, "mean_token_accuracy": 0.869595468044281, "num_tokens": 249071227.0, "step": 6559 }, { "epoch": 0.8344994275537464, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 6.839752674102783, "learning_rate": 5e-05, "loss": 0.7968, "mean_token_accuracy": 0.8642619848251343, "num_tokens": 249105859.0, "step": 6560 }, { "epoch": 0.8346266378323368, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.880191802978516, "learning_rate": 5e-05, "loss": 0.9672, "mean_token_accuracy": 0.8660669326782227, "num_tokens": 249148237.0, "step": 6561 }, { "epoch": 0.8347538481109273, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.2985405921936035, "learning_rate": 5e-05, "loss": 0.8896, "mean_token_accuracy": 0.8511433601379395, "num_tokens": 249181745.0, "step": 6562 }, { "epoch": 0.8348810583895179, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.572844505310059, "learning_rate": 5e-05, "loss": 0.9807, "mean_token_accuracy": 0.8573524951934814, "num_tokens": 249221839.0, "step": 6563 }, { "epoch": 0.8350082686681084, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.22370719909668, "learning_rate": 5e-05, "loss": 0.8754, "mean_token_accuracy": 0.8569314479827881, "num_tokens": 249264338.0, "step": 6564 }, { "epoch": 0.8351354789466989, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.755642890930176, "learning_rate": 5e-05, "loss": 0.9933, "mean_token_accuracy": 0.8589204549789429, "num_tokens": 249304624.0, "step": 6565 }, { "epoch": 0.8352626892252895, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 6.9394378662109375, "learning_rate": 5e-05, "loss": 0.8078, "mean_token_accuracy": 0.8622419238090515, "num_tokens": 249342996.0, "step": 6566 }, { "epoch": 0.8353898995038799, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.140878677368164, "learning_rate": 5e-05, "loss": 1.0256, "mean_token_accuracy": 0.859217643737793, "num_tokens": 249384795.0, "step": 6567 }, { "epoch": 0.8355171097824704, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.187941074371338, "learning_rate": 5e-05, "loss": 0.8606, "mean_token_accuracy": 0.8699173927307129, "num_tokens": 249418492.0, "step": 6568 }, { "epoch": 0.8356443200610609, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.837477684020996, "learning_rate": 5e-05, "loss": 0.9987, "mean_token_accuracy": 0.8628514409065247, "num_tokens": 249455714.0, "step": 6569 }, { "epoch": 0.8357715303396515, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.846010208129883, "learning_rate": 5e-05, "loss": 1.0145, "mean_token_accuracy": 0.8452810049057007, "num_tokens": 249493162.0, "step": 6570 }, { "epoch": 0.835898740618242, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.825666427612305, "learning_rate": 5e-05, "loss": 0.932, "mean_token_accuracy": 0.8516275882720947, "num_tokens": 249538154.0, "step": 6571 }, { "epoch": 0.8360259508968325, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.145578384399414, "learning_rate": 5e-05, "loss": 0.9077, "mean_token_accuracy": 0.8695196509361267, "num_tokens": 249576457.0, "step": 6572 }, { "epoch": 0.8361531611754229, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.286531448364258, "learning_rate": 5e-05, "loss": 0.8285, "mean_token_accuracy": 0.8582973480224609, "num_tokens": 249611408.0, "step": 6573 }, { "epoch": 0.8362803714540135, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.896617889404297, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8579393029212952, "num_tokens": 249651371.0, "step": 6574 }, { "epoch": 0.836407581732604, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.460010528564453, "learning_rate": 5e-05, "loss": 0.8221, "mean_token_accuracy": 0.8606781959533691, "num_tokens": 249697669.0, "step": 6575 }, { "epoch": 0.8365347920111945, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.442141056060791, "learning_rate": 5e-05, "loss": 0.8314, "mean_token_accuracy": 0.8537720441818237, "num_tokens": 249739640.0, "step": 6576 }, { "epoch": 0.836662002289785, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.689728736877441, "learning_rate": 5e-05, "loss": 0.8388, "mean_token_accuracy": 0.851679265499115, "num_tokens": 249778430.0, "step": 6577 }, { "epoch": 0.8367892125683756, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.282800674438477, "learning_rate": 5e-05, "loss": 0.7434, "mean_token_accuracy": 0.8718925714492798, "num_tokens": 249817875.0, "step": 6578 }, { "epoch": 0.836916422846966, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.579334259033203, "learning_rate": 5e-05, "loss": 0.7269, "mean_token_accuracy": 0.8820188045501709, "num_tokens": 249861987.0, "step": 6579 }, { "epoch": 0.8370436331255565, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.431695461273193, "learning_rate": 5e-05, "loss": 0.7643, "mean_token_accuracy": 0.8657258749008179, "num_tokens": 249903137.0, "step": 6580 }, { "epoch": 0.837170843404147, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.3605804443359375, "learning_rate": 5e-05, "loss": 0.7756, "mean_token_accuracy": 0.8547554016113281, "num_tokens": 249937760.0, "step": 6581 }, { "epoch": 0.8372980536827376, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.540923118591309, "learning_rate": 5e-05, "loss": 0.7797, "mean_token_accuracy": 0.8572056889533997, "num_tokens": 249977280.0, "step": 6582 }, { "epoch": 0.8374252639613281, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.290661811828613, "learning_rate": 5e-05, "loss": 0.7918, "mean_token_accuracy": 0.8532075881958008, "num_tokens": 250022160.0, "step": 6583 }, { "epoch": 0.8375524742399186, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.526078224182129, "learning_rate": 5e-05, "loss": 0.7512, "mean_token_accuracy": 0.8677223920822144, "num_tokens": 250063328.0, "step": 6584 }, { "epoch": 0.837679684518509, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.313462734222412, "learning_rate": 5e-05, "loss": 0.7111, "mean_token_accuracy": 0.8727572560310364, "num_tokens": 250096994.0, "step": 6585 }, { "epoch": 0.8378068947970996, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.473509788513184, "learning_rate": 5e-05, "loss": 0.7445, "mean_token_accuracy": 0.8650816082954407, "num_tokens": 250126058.0, "step": 6586 }, { "epoch": 0.8379341050756901, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.305160999298096, "learning_rate": 5e-05, "loss": 0.7901, "mean_token_accuracy": 0.8491005897521973, "num_tokens": 250171965.0, "step": 6587 }, { "epoch": 0.8380613153542806, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.497602939605713, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8562692403793335, "num_tokens": 250209862.0, "step": 6588 }, { "epoch": 0.8381885256328712, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.239004135131836, "learning_rate": 5e-05, "loss": 0.7431, "mean_token_accuracy": 0.8589763641357422, "num_tokens": 250244895.0, "step": 6589 }, { "epoch": 0.8383157359114617, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.585638523101807, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.8631976246833801, "num_tokens": 250283340.0, "step": 6590 }, { "epoch": 0.8384429461900521, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.100645065307617, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8624462485313416, "num_tokens": 250318671.0, "step": 6591 }, { "epoch": 0.8385701564686426, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.763364791870117, "learning_rate": 5e-05, "loss": 0.7399, "mean_token_accuracy": 0.868565022945404, "num_tokens": 250357718.0, "step": 6592 }, { "epoch": 0.8386973667472332, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.8569655418396, "learning_rate": 5e-05, "loss": 0.7146, "mean_token_accuracy": 0.8621243238449097, "num_tokens": 250391583.0, "step": 6593 }, { "epoch": 0.8388245770258237, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.09719181060791, "learning_rate": 5e-05, "loss": 0.8147, "mean_token_accuracy": 0.8606355786323547, "num_tokens": 250420938.0, "step": 6594 }, { "epoch": 0.8389517873044142, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.422948837280273, "learning_rate": 5e-05, "loss": 0.6614, "mean_token_accuracy": 0.8703486919403076, "num_tokens": 250455608.0, "step": 6595 }, { "epoch": 0.8390789975830047, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.792668342590332, "learning_rate": 5e-05, "loss": 0.8829, "mean_token_accuracy": 0.8574121594429016, "num_tokens": 250489745.0, "step": 6596 }, { "epoch": 0.8392062078615952, "ewc_loss": 0.25, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00021076202392578125, "grad_norm": 5.760110378265381, "learning_rate": 5e-05, "loss": 0.6624, "mean_token_accuracy": 0.8613271713256836, "num_tokens": 250523711.0, "step": 6597 }, { "epoch": 0.8393334181401857, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.920035362243652, "learning_rate": 5e-05, "loss": 1.0508, "mean_token_accuracy": 0.8493930697441101, "num_tokens": 250562044.0, "step": 6598 }, { "epoch": 0.8394606284187762, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.502681255340576, "learning_rate": 5e-05, "loss": 0.7168, "mean_token_accuracy": 0.8573138117790222, "num_tokens": 250605930.0, "step": 6599 }, { "epoch": 0.8395878386973668, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0007781982421875, "grad_norm": 11.111913681030273, "learning_rate": 5e-05, "loss": 1.2206, "mean_token_accuracy": 0.8605334162712097, "num_tokens": 250640621.0, "step": 6600 }, { "epoch": 0.8397150489759573, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 7.577522277832031, "learning_rate": 5e-05, "loss": 1.0212, "mean_token_accuracy": 0.8617458939552307, "num_tokens": 250678056.0, "step": 6601 }, { "epoch": 0.8398422592545478, "ewc_loss": 0.6875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.475454330444336, "learning_rate": 5e-05, "loss": 1.0868, "mean_token_accuracy": 0.8673667907714844, "num_tokens": 250719642.0, "step": 6602 }, { "epoch": 0.8399694695331382, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 9.35481071472168, "learning_rate": 5e-05, "loss": 1.3754, "mean_token_accuracy": 0.8395146131515503, "num_tokens": 250757643.0, "step": 6603 }, { "epoch": 0.8400966798117288, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.173445701599121, "learning_rate": 5e-05, "loss": 0.8956, "mean_token_accuracy": 0.8537589311599731, "num_tokens": 250795277.0, "step": 6604 }, { "epoch": 0.8402238900903193, "ewc_loss": 0.87890625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 10.815994262695312, "learning_rate": 5e-05, "loss": 1.2951, "mean_token_accuracy": 0.8591288924217224, "num_tokens": 250834161.0, "step": 6605 }, { "epoch": 0.8403511003689098, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 7.578578472137451, "learning_rate": 5e-05, "loss": 0.963, "mean_token_accuracy": 0.8560308218002319, "num_tokens": 250872518.0, "step": 6606 }, { "epoch": 0.8404783106475003, "ewc_loss": 0.94140625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000904083251953125, "grad_norm": 12.29399585723877, "learning_rate": 5e-05, "loss": 1.3447, "mean_token_accuracy": 0.8670510053634644, "num_tokens": 250908516.0, "step": 6607 }, { "epoch": 0.8406055209260909, "ewc_loss": 0.98046875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000942230224609375, "grad_norm": 10.025036811828613, "learning_rate": 5e-05, "loss": 1.3921, "mean_token_accuracy": 0.8564298748970032, "num_tokens": 250949351.0, "step": 6608 }, { "epoch": 0.8407327312046813, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 11.18805980682373, "learning_rate": 5e-05, "loss": 1.0812, "mean_token_accuracy": 0.8622869253158569, "num_tokens": 250987145.0, "step": 6609 }, { "epoch": 0.8408599414832718, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 9.302446365356445, "learning_rate": 5e-05, "loss": 1.1747, "mean_token_accuracy": 0.8690288066864014, "num_tokens": 251024073.0, "step": 6610 }, { "epoch": 0.8409871517618623, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 11.163670539855957, "learning_rate": 5e-05, "loss": 1.1955, "mean_token_accuracy": 0.8509273529052734, "num_tokens": 251069537.0, "step": 6611 }, { "epoch": 0.8411143620404529, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.18215560913086, "learning_rate": 5e-05, "loss": 0.997, "mean_token_accuracy": 0.854525625705719, "num_tokens": 251103343.0, "step": 6612 }, { "epoch": 0.8412415723190434, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 11.583134651184082, "learning_rate": 5e-05, "loss": 1.2279, "mean_token_accuracy": 0.8550763130187988, "num_tokens": 251140437.0, "step": 6613 }, { "epoch": 0.8413687825976339, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.317697525024414, "learning_rate": 5e-05, "loss": 0.9962, "mean_token_accuracy": 0.8618373870849609, "num_tokens": 251178002.0, "step": 6614 }, { "epoch": 0.8414959928762245, "ewc_loss": 0.890625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 12.408595085144043, "learning_rate": 5e-05, "loss": 1.2625, "mean_token_accuracy": 0.8727852702140808, "num_tokens": 251219694.0, "step": 6615 }, { "epoch": 0.8416232031548149, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 10.077035903930664, "learning_rate": 5e-05, "loss": 1.165, "mean_token_accuracy": 0.8654744625091553, "num_tokens": 251255576.0, "step": 6616 }, { "epoch": 0.8417504134334054, "ewc_loss": 0.765625, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 11.584413528442383, "learning_rate": 5e-05, "loss": 1.1912, "mean_token_accuracy": 0.8565226793289185, "num_tokens": 251287245.0, "step": 6617 }, { "epoch": 0.8418776237119959, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.174822807312012, "learning_rate": 5e-05, "loss": 1.1695, "mean_token_accuracy": 0.8617957830429077, "num_tokens": 251320435.0, "step": 6618 }, { "epoch": 0.8420048339905865, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.807341575622559, "learning_rate": 5e-05, "loss": 1.061, "mean_token_accuracy": 0.8702688217163086, "num_tokens": 251358893.0, "step": 6619 }, { "epoch": 0.842132044269177, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.566385269165039, "learning_rate": 5e-05, "loss": 0.9898, "mean_token_accuracy": 0.8741477727890015, "num_tokens": 251392805.0, "step": 6620 }, { "epoch": 0.8422592545477675, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.23515510559082, "learning_rate": 5e-05, "loss": 1.0313, "mean_token_accuracy": 0.8705006837844849, "num_tokens": 251436848.0, "step": 6621 }, { "epoch": 0.8423864648263579, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.967789649963379, "learning_rate": 5e-05, "loss": 0.9162, "mean_token_accuracy": 0.872146487236023, "num_tokens": 251471590.0, "step": 6622 }, { "epoch": 0.8425136751049485, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0003910064697265625, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.982366561889648, "learning_rate": 5e-05, "loss": 1.0353, "mean_token_accuracy": 0.8543598055839539, "num_tokens": 251509251.0, "step": 6623 }, { "epoch": 0.842640885383539, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.347443580627441, "learning_rate": 5e-05, "loss": 0.9336, "mean_token_accuracy": 0.8489915728569031, "num_tokens": 251549337.0, "step": 6624 }, { "epoch": 0.8427680956621295, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.263614654541016, "learning_rate": 5e-05, "loss": 1.1134, "mean_token_accuracy": 0.8408805131912231, "num_tokens": 251593953.0, "step": 6625 }, { "epoch": 0.84289530594072, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.143049240112305, "learning_rate": 5e-05, "loss": 0.8874, "mean_token_accuracy": 0.8588268756866455, "num_tokens": 251631797.0, "step": 6626 }, { "epoch": 0.8430225162193106, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.717073440551758, "learning_rate": 5e-05, "loss": 1.1609, "mean_token_accuracy": 0.8469254970550537, "num_tokens": 251668378.0, "step": 6627 }, { "epoch": 0.843149726497901, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.715616226196289, "learning_rate": 5e-05, "loss": 0.9774, "mean_token_accuracy": 0.8492118716239929, "num_tokens": 251712248.0, "step": 6628 }, { "epoch": 0.8432769367764915, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.47155475616455, "learning_rate": 5e-05, "loss": 1.0922, "mean_token_accuracy": 0.8605988025665283, "num_tokens": 251748770.0, "step": 6629 }, { "epoch": 0.843404147055082, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.182692527770996, "learning_rate": 5e-05, "loss": 0.9631, "mean_token_accuracy": 0.8651183843612671, "num_tokens": 251787262.0, "step": 6630 }, { "epoch": 0.8435313573336726, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.574067115783691, "learning_rate": 5e-05, "loss": 1.0329, "mean_token_accuracy": 0.8485631942749023, "num_tokens": 251820584.0, "step": 6631 }, { "epoch": 0.8436585676122631, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.863446235656738, "learning_rate": 5e-05, "loss": 0.9565, "mean_token_accuracy": 0.8528675436973572, "num_tokens": 251861394.0, "step": 6632 }, { "epoch": 0.8437857778908536, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.91104507446289, "learning_rate": 5e-05, "loss": 0.9246, "mean_token_accuracy": 0.8617441654205322, "num_tokens": 251902144.0, "step": 6633 }, { "epoch": 0.843912988169444, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.177129745483398, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8832757472991943, "num_tokens": 251936162.0, "step": 6634 }, { "epoch": 0.8440401984480346, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.858786582946777, "learning_rate": 5e-05, "loss": 0.922, "mean_token_accuracy": 0.8537882566452026, "num_tokens": 251977223.0, "step": 6635 }, { "epoch": 0.8441674087266251, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.507755279541016, "learning_rate": 5e-05, "loss": 0.7686, "mean_token_accuracy": 0.8682047724723816, "num_tokens": 252013413.0, "step": 6636 }, { "epoch": 0.8442946190052156, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.200401306152344, "learning_rate": 5e-05, "loss": 0.9643, "mean_token_accuracy": 0.8547693490982056, "num_tokens": 252051395.0, "step": 6637 }, { "epoch": 0.8444218292838062, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.281064510345459, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.8568698167800903, "num_tokens": 252092982.0, "step": 6638 }, { "epoch": 0.8445490395623967, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.60081958770752, "learning_rate": 5e-05, "loss": 0.9417, "mean_token_accuracy": 0.8820837140083313, "num_tokens": 252128953.0, "step": 6639 }, { "epoch": 0.8446762498409871, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.679506301879883, "learning_rate": 5e-05, "loss": 0.8403, "mean_token_accuracy": 0.8505157828330994, "num_tokens": 252171438.0, "step": 6640 }, { "epoch": 0.8448034601195776, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.499645233154297, "learning_rate": 5e-05, "loss": 0.9905, "mean_token_accuracy": 0.8607008457183838, "num_tokens": 252210310.0, "step": 6641 }, { "epoch": 0.8449306703981682, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.29189682006836, "learning_rate": 5e-05, "loss": 0.8245, "mean_token_accuracy": 0.8754246830940247, "num_tokens": 252242690.0, "step": 6642 }, { "epoch": 0.8450578806767587, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.610320091247559, "learning_rate": 5e-05, "loss": 0.899, "mean_token_accuracy": 0.8621735572814941, "num_tokens": 252281570.0, "step": 6643 }, { "epoch": 0.8451850909553492, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.353620529174805, "learning_rate": 5e-05, "loss": 0.89, "mean_token_accuracy": 0.847711443901062, "num_tokens": 252323478.0, "step": 6644 }, { "epoch": 0.8453123012339397, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.751498222351074, "learning_rate": 5e-05, "loss": 0.8693, "mean_token_accuracy": 0.8445901870727539, "num_tokens": 252362620.0, "step": 6645 }, { "epoch": 0.8454395115125302, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.120509147644043, "learning_rate": 5e-05, "loss": 0.7832, "mean_token_accuracy": 0.8707772493362427, "num_tokens": 252397549.0, "step": 6646 }, { "epoch": 0.8455667217911207, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.4534687995910645, "learning_rate": 5e-05, "loss": 0.7956, "mean_token_accuracy": 0.8556663990020752, "num_tokens": 252433021.0, "step": 6647 }, { "epoch": 0.8456939320697112, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.953462600708008, "learning_rate": 5e-05, "loss": 0.7956, "mean_token_accuracy": 0.861640989780426, "num_tokens": 252466455.0, "step": 6648 }, { "epoch": 0.8458211423483017, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.256842613220215, "learning_rate": 5e-05, "loss": 0.7658, "mean_token_accuracy": 0.8553293943405151, "num_tokens": 252503547.0, "step": 6649 }, { "epoch": 0.8459483526268923, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.125718116760254, "learning_rate": 5e-05, "loss": 0.8195, "mean_token_accuracy": 0.8557440638542175, "num_tokens": 252535621.0, "step": 6650 }, { "epoch": 0.8460755629054828, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.97552490234375, "learning_rate": 5e-05, "loss": 0.7525, "mean_token_accuracy": 0.853341817855835, "num_tokens": 252573699.0, "step": 6651 }, { "epoch": 0.8462027731840732, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.426817893981934, "learning_rate": 5e-05, "loss": 0.8132, "mean_token_accuracy": 0.8687983155250549, "num_tokens": 252619816.0, "step": 6652 }, { "epoch": 0.8463299834626637, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.830393314361572, "learning_rate": 5e-05, "loss": 0.6881, "mean_token_accuracy": 0.8722492456436157, "num_tokens": 252652631.0, "step": 6653 }, { "epoch": 0.8464571937412543, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.582661628723145, "learning_rate": 5e-05, "loss": 0.8484, "mean_token_accuracy": 0.8672181963920593, "num_tokens": 252690534.0, "step": 6654 }, { "epoch": 0.8465844040198448, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.952542781829834, "learning_rate": 5e-05, "loss": 0.7689, "mean_token_accuracy": 0.8508193492889404, "num_tokens": 252723090.0, "step": 6655 }, { "epoch": 0.8467116142984353, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 10.574499130249023, "learning_rate": 5e-05, "loss": 0.8715, "mean_token_accuracy": 0.8633806705474854, "num_tokens": 252755879.0, "step": 6656 }, { "epoch": 0.8468388245770259, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00038909912109375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.431612491607666, "learning_rate": 5e-05, "loss": 0.8672, "mean_token_accuracy": 0.8418437242507935, "num_tokens": 252799137.0, "step": 6657 }, { "epoch": 0.8469660348556163, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0003948211669921875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.390846252441406, "learning_rate": 5e-05, "loss": 0.9598, "mean_token_accuracy": 0.8598276376724243, "num_tokens": 252840871.0, "step": 6658 }, { "epoch": 0.8470932451342068, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000400543212890625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.682851791381836, "learning_rate": 5e-05, "loss": 0.9444, "mean_token_accuracy": 0.8559446334838867, "num_tokens": 252879708.0, "step": 6659 }, { "epoch": 0.8472204554127973, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.08026647567749, "learning_rate": 5e-05, "loss": 0.7468, "mean_token_accuracy": 0.8652164936065674, "num_tokens": 252914223.0, "step": 6660 }, { "epoch": 0.8473476656913879, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 10.062552452087402, "learning_rate": 5e-05, "loss": 1.0104, "mean_token_accuracy": 0.859927237033844, "num_tokens": 252956666.0, "step": 6661 }, { "epoch": 0.8474748759699784, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.709724426269531, "learning_rate": 5e-05, "loss": 0.7468, "mean_token_accuracy": 0.8600701093673706, "num_tokens": 252997566.0, "step": 6662 }, { "epoch": 0.8476020862485689, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 10.351798057556152, "learning_rate": 5e-05, "loss": 1.0245, "mean_token_accuracy": 0.8673219680786133, "num_tokens": 253038545.0, "step": 6663 }, { "epoch": 0.8477292965271594, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.109260559082031, "learning_rate": 5e-05, "loss": 0.8885, "mean_token_accuracy": 0.8629129528999329, "num_tokens": 253078123.0, "step": 6664 }, { "epoch": 0.8478565068057499, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.867596626281738, "learning_rate": 5e-05, "loss": 0.8795, "mean_token_accuracy": 0.8729931712150574, "num_tokens": 253118273.0, "step": 6665 }, { "epoch": 0.8479837170843404, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.93558120727539, "learning_rate": 5e-05, "loss": 1.0025, "mean_token_accuracy": 0.8465450406074524, "num_tokens": 253152994.0, "step": 6666 }, { "epoch": 0.8481109273629309, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000453948974609375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.5559234619140625, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8710728883743286, "num_tokens": 253189736.0, "step": 6667 }, { "epoch": 0.8482381376415215, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000457763671875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.249041557312012, "learning_rate": 5e-05, "loss": 0.9947, "mean_token_accuracy": 0.8489362597465515, "num_tokens": 253225872.0, "step": 6668 }, { "epoch": 0.848365347920112, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000457763671875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.595188140869141, "learning_rate": 5e-05, "loss": 0.7658, "mean_token_accuracy": 0.8762385249137878, "num_tokens": 253263902.0, "step": 6669 }, { "epoch": 0.8484925581987025, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0004596710205078125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.703929901123047, "learning_rate": 5e-05, "loss": 0.8553, "mean_token_accuracy": 0.8701245188713074, "num_tokens": 253297536.0, "step": 6670 }, { "epoch": 0.8486197684772929, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0004596710205078125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.923077583312988, "learning_rate": 5e-05, "loss": 0.7933, "mean_token_accuracy": 0.8721119165420532, "num_tokens": 253334624.0, "step": 6671 }, { "epoch": 0.8487469787558835, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0004596710205078125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.806379318237305, "learning_rate": 5e-05, "loss": 0.8345, "mean_token_accuracy": 0.8562805652618408, "num_tokens": 253379203.0, "step": 6672 }, { "epoch": 0.848874189034474, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.210601806640625, "learning_rate": 5e-05, "loss": 0.8086, "mean_token_accuracy": 0.8709206581115723, "num_tokens": 253417209.0, "step": 6673 }, { "epoch": 0.8490013993130645, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.4100022315979, "learning_rate": 5e-05, "loss": 0.8002, "mean_token_accuracy": 0.8493014574050903, "num_tokens": 253453909.0, "step": 6674 }, { "epoch": 0.849128609591655, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00046539306640625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.099517822265625, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.828484058380127, "num_tokens": 253489448.0, "step": 6675 }, { "epoch": 0.8492558198702456, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00046539306640625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.519674301147461, "learning_rate": 5e-05, "loss": 0.764, "mean_token_accuracy": 0.8606212735176086, "num_tokens": 253517544.0, "step": 6676 }, { "epoch": 0.849383030148836, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00046539306640625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.679542541503906, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8672857880592346, "num_tokens": 253558829.0, "step": 6677 }, { "epoch": 0.8495102404274265, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.740302085876465, "learning_rate": 5e-05, "loss": 0.7852, "mean_token_accuracy": 0.8646180629730225, "num_tokens": 253593944.0, "step": 6678 }, { "epoch": 0.849637450706017, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.407703399658203, "learning_rate": 5e-05, "loss": 0.7691, "mean_token_accuracy": 0.8566663861274719, "num_tokens": 253633579.0, "step": 6679 }, { "epoch": 0.8497646609846076, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.818118572235107, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8559006452560425, "num_tokens": 253680247.0, "step": 6680 }, { "epoch": 0.8498918712631981, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.324800968170166, "learning_rate": 5e-05, "loss": 0.7436, "mean_token_accuracy": 0.8656079173088074, "num_tokens": 253714766.0, "step": 6681 }, { "epoch": 0.8500190815417886, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.760519504547119, "learning_rate": 5e-05, "loss": 0.7902, "mean_token_accuracy": 0.8524149656295776, "num_tokens": 253747643.0, "step": 6682 }, { "epoch": 0.850146291820379, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.382037162780762, "learning_rate": 5e-05, "loss": 0.8082, "mean_token_accuracy": 0.8405463099479675, "num_tokens": 253791252.0, "step": 6683 }, { "epoch": 0.8502735020989696, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.606509208679199, "learning_rate": 5e-05, "loss": 0.7382, "mean_token_accuracy": 0.8675878047943115, "num_tokens": 253822244.0, "step": 6684 }, { "epoch": 0.8504007123775601, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.476184844970703, "learning_rate": 5e-05, "loss": 0.766, "mean_token_accuracy": 0.8552006483078003, "num_tokens": 253858680.0, "step": 6685 }, { "epoch": 0.8505279226561506, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.519720554351807, "learning_rate": 5e-05, "loss": 0.7205, "mean_token_accuracy": 0.8714669942855835, "num_tokens": 253896266.0, "step": 6686 }, { "epoch": 0.8506551329347412, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004596710205078125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.401968002319336, "learning_rate": 5e-05, "loss": 0.7467, "mean_token_accuracy": 0.8621680736541748, "num_tokens": 253934059.0, "step": 6687 }, { "epoch": 0.8507823432133317, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004596710205078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.686371803283691, "learning_rate": 5e-05, "loss": 0.7478, "mean_token_accuracy": 0.8620407581329346, "num_tokens": 253967862.0, "step": 6688 }, { "epoch": 0.8509095534919221, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000457763671875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 7.1559038162231445, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.8541719913482666, "num_tokens": 254008277.0, "step": 6689 }, { "epoch": 0.8510367637705126, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000453948974609375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.93905782699585, "learning_rate": 5e-05, "loss": 0.7771, "mean_token_accuracy": 0.8627702593803406, "num_tokens": 254039939.0, "step": 6690 }, { "epoch": 0.8511639740491032, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000453948974609375, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.867007255554199, "learning_rate": 5e-05, "loss": 0.6991, "mean_token_accuracy": 0.8667611479759216, "num_tokens": 254080497.0, "step": 6691 }, { "epoch": 0.8512911843276937, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000453948974609375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.306384086608887, "learning_rate": 5e-05, "loss": 0.7974, "mean_token_accuracy": 0.86181640625, "num_tokens": 254111917.0, "step": 6692 }, { "epoch": 0.8514183946062842, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.000453948974609375, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 6.464566230773926, "learning_rate": 5e-05, "loss": 0.6604, "mean_token_accuracy": 0.870579719543457, "num_tokens": 254150245.0, "step": 6693 }, { "epoch": 0.8515456048848747, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000453948974609375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.880911827087402, "learning_rate": 5e-05, "loss": 0.8748, "mean_token_accuracy": 0.8597381114959717, "num_tokens": 254187106.0, "step": 6694 }, { "epoch": 0.8516728151634652, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.000453948974609375, "ewc_loss_parallel": 0.0002117156982421875, "grad_norm": 5.892702579498291, "learning_rate": 5e-05, "loss": 0.6842, "mean_token_accuracy": 0.8526574373245239, "num_tokens": 254219877.0, "step": 6695 }, { "epoch": 0.8518000254420557, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0004520416259765625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.819594383239746, "learning_rate": 5e-05, "loss": 0.9981, "mean_token_accuracy": 0.8541863560676575, "num_tokens": 254264234.0, "step": 6696 }, { "epoch": 0.8519272357206462, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.00022983551025390625, "grad_norm": 5.540815353393555, "learning_rate": 5e-05, "loss": 0.683, "mean_token_accuracy": 0.8676742911338806, "num_tokens": 254299762.0, "step": 6697 }, { "epoch": 0.8520544459992367, "ewc_loss": 0.71875, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.878419876098633, "learning_rate": 5e-05, "loss": 1.1651, "mean_token_accuracy": 0.8464422821998596, "num_tokens": 254336403.0, "step": 6698 }, { "epoch": 0.8521816562778273, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 6.848299026489258, "learning_rate": 5e-05, "loss": 0.8767, "mean_token_accuracy": 0.8676701784133911, "num_tokens": 254375971.0, "step": 6699 }, { "epoch": 0.8523088665564178, "ewc_loss": 0.78125, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.000732421875, "grad_norm": 10.98915958404541, "learning_rate": 5e-05, "loss": 1.2092, "mean_token_accuracy": 0.852853000164032, "num_tokens": 254418430.0, "step": 6700 }, { "epoch": 0.8524360768350082, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 9.00069808959961, "learning_rate": 5e-05, "loss": 1.2065, "mean_token_accuracy": 0.8614682555198669, "num_tokens": 254457275.0, "step": 6701 }, { "epoch": 0.8525632871135987, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.127853393554688, "learning_rate": 5e-05, "loss": 0.9887, "mean_token_accuracy": 0.8566383719444275, "num_tokens": 254499280.0, "step": 6702 }, { "epoch": 0.8526904973921893, "ewc_loss": 0.765625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 9.72684383392334, "learning_rate": 5e-05, "loss": 1.1938, "mean_token_accuracy": 0.8557513952255249, "num_tokens": 254536571.0, "step": 6703 }, { "epoch": 0.8528177076707798, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0004405975341796875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.46028995513916, "learning_rate": 5e-05, "loss": 0.8694, "mean_token_accuracy": 0.8510503768920898, "num_tokens": 254576900.0, "step": 6704 }, { "epoch": 0.8529449179493703, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.0004405975341796875, "ewc_loss_parallel": 0.000732421875, "grad_norm": 10.689471244812012, "learning_rate": 5e-05, "loss": 1.2671, "mean_token_accuracy": 0.8370778560638428, "num_tokens": 254617791.0, "step": 6705 }, { "epoch": 0.8530721282279609, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0004405975341796875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 7.899566650390625, "learning_rate": 5e-05, "loss": 0.9876, "mean_token_accuracy": 0.8686329126358032, "num_tokens": 254658300.0, "step": 6706 }, { "epoch": 0.8531993385065513, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0004405975341796875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.502205848693848, "learning_rate": 5e-05, "loss": 1.1119, "mean_token_accuracy": 0.8585309982299805, "num_tokens": 254698133.0, "step": 6707 }, { "epoch": 0.8533265487851418, "ewc_loss": 0.703125, "ewc_loss_diag": 0.0004405975341796875, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 8.798246383666992, "learning_rate": 5e-05, "loss": 1.0911, "mean_token_accuracy": 0.8661799430847168, "num_tokens": 254736099.0, "step": 6708 }, { "epoch": 0.8534537590637323, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.280797004699707, "learning_rate": 5e-05, "loss": 0.9573, "mean_token_accuracy": 0.8718692064285278, "num_tokens": 254773460.0, "step": 6709 }, { "epoch": 0.8535809693423229, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 8.863970756530762, "learning_rate": 5e-05, "loss": 1.0285, "mean_token_accuracy": 0.8674064874649048, "num_tokens": 254812722.0, "step": 6710 }, { "epoch": 0.8537081796209134, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0004367828369140625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.241796493530273, "learning_rate": 5e-05, "loss": 0.8948, "mean_token_accuracy": 0.8653807044029236, "num_tokens": 254851994.0, "step": 6711 }, { "epoch": 0.8538353898995039, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0004367828369140625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.806679725646973, "learning_rate": 5e-05, "loss": 1.0225, "mean_token_accuracy": 0.8544190526008606, "num_tokens": 254889111.0, "step": 6712 }, { "epoch": 0.8539626001780944, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.704815864562988, "learning_rate": 5e-05, "loss": 0.8898, "mean_token_accuracy": 0.8569223880767822, "num_tokens": 254931779.0, "step": 6713 }, { "epoch": 0.8540898104566849, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.705431938171387, "learning_rate": 5e-05, "loss": 0.9411, "mean_token_accuracy": 0.8651683330535889, "num_tokens": 254968709.0, "step": 6714 }, { "epoch": 0.8542170207352754, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.506324768066406, "learning_rate": 5e-05, "loss": 0.8514, "mean_token_accuracy": 0.8649319410324097, "num_tokens": 255000235.0, "step": 6715 }, { "epoch": 0.8543442310138659, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.477303504943848, "learning_rate": 5e-05, "loss": 0.9083, "mean_token_accuracy": 0.8572996854782104, "num_tokens": 255041157.0, "step": 6716 }, { "epoch": 0.8544714412924564, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.450584411621094, "learning_rate": 5e-05, "loss": 0.8597, "mean_token_accuracy": 0.8593516945838928, "num_tokens": 255077001.0, "step": 6717 }, { "epoch": 0.854598651571047, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.195913314819336, "learning_rate": 5e-05, "loss": 0.8302, "mean_token_accuracy": 0.869077205657959, "num_tokens": 255113581.0, "step": 6718 }, { "epoch": 0.8547258618496375, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.481405258178711, "learning_rate": 5e-05, "loss": 0.799, "mean_token_accuracy": 0.8769009113311768, "num_tokens": 255153183.0, "step": 6719 }, { "epoch": 0.8548530721282279, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.782712459564209, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.868973970413208, "num_tokens": 255192216.0, "step": 6720 }, { "epoch": 0.8549802824068184, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.780813217163086, "learning_rate": 5e-05, "loss": 0.8194, "mean_token_accuracy": 0.8675378561019897, "num_tokens": 255229969.0, "step": 6721 }, { "epoch": 0.855107492685409, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.202929496765137, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8516988158226013, "num_tokens": 255267232.0, "step": 6722 }, { "epoch": 0.8552347029639995, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.391469955444336, "learning_rate": 5e-05, "loss": 0.871, "mean_token_accuracy": 0.8633989691734314, "num_tokens": 255304923.0, "step": 6723 }, { "epoch": 0.85536191324259, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.523019790649414, "learning_rate": 5e-05, "loss": 0.7813, "mean_token_accuracy": 0.8419253826141357, "num_tokens": 255342422.0, "step": 6724 }, { "epoch": 0.8554891235211806, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.275800704956055, "learning_rate": 5e-05, "loss": 0.9317, "mean_token_accuracy": 0.8704095482826233, "num_tokens": 255380218.0, "step": 6725 }, { "epoch": 0.855616333799771, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.146777153015137, "learning_rate": 5e-05, "loss": 0.7237, "mean_token_accuracy": 0.8625956177711487, "num_tokens": 255413964.0, "step": 6726 }, { "epoch": 0.8557435440783615, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.105462074279785, "learning_rate": 5e-05, "loss": 1.0223, "mean_token_accuracy": 0.869444727897644, "num_tokens": 255450421.0, "step": 6727 }, { "epoch": 0.855870754356952, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 6.963870525360107, "learning_rate": 5e-05, "loss": 0.89, "mean_token_accuracy": 0.8614094257354736, "num_tokens": 255487239.0, "step": 6728 }, { "epoch": 0.8559979646355426, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.013397216796875, "learning_rate": 5e-05, "loss": 1.0735, "mean_token_accuracy": 0.8594474792480469, "num_tokens": 255529363.0, "step": 6729 }, { "epoch": 0.8561251749141331, "ewc_loss": 0.625, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.349166870117188, "learning_rate": 5e-05, "loss": 0.9483, "mean_token_accuracy": 0.885619044303894, "num_tokens": 255570790.0, "step": 6730 }, { "epoch": 0.8562523851927236, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.600461959838867, "learning_rate": 5e-05, "loss": 0.9346, "mean_token_accuracy": 0.8580388426780701, "num_tokens": 255608188.0, "step": 6731 }, { "epoch": 0.856379595471314, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.82562255859375, "learning_rate": 5e-05, "loss": 1.0232, "mean_token_accuracy": 0.8622878789901733, "num_tokens": 255640929.0, "step": 6732 }, { "epoch": 0.8565068057499046, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.501476764678955, "learning_rate": 5e-05, "loss": 0.8096, "mean_token_accuracy": 0.8647913336753845, "num_tokens": 255684579.0, "step": 6733 }, { "epoch": 0.8566340160284951, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.840603828430176, "learning_rate": 5e-05, "loss": 0.988, "mean_token_accuracy": 0.8617094159126282, "num_tokens": 255721195.0, "step": 6734 }, { "epoch": 0.8567612263070856, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.464190483093262, "learning_rate": 5e-05, "loss": 0.8654, "mean_token_accuracy": 0.8615986704826355, "num_tokens": 255761697.0, "step": 6735 }, { "epoch": 0.8568884365856761, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.280566215515137, "learning_rate": 5e-05, "loss": 0.8953, "mean_token_accuracy": 0.8616098761558533, "num_tokens": 255800492.0, "step": 6736 }, { "epoch": 0.8570156468642667, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.758193016052246, "learning_rate": 5e-05, "loss": 0.8684, "mean_token_accuracy": 0.8640699982643127, "num_tokens": 255837074.0, "step": 6737 }, { "epoch": 0.8571428571428571, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.628168106079102, "learning_rate": 5e-05, "loss": 0.8012, "mean_token_accuracy": 0.8614764213562012, "num_tokens": 255876238.0, "step": 6738 }, { "epoch": 0.8572700674214476, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.036392211914062, "learning_rate": 5e-05, "loss": 0.9308, "mean_token_accuracy": 0.8470803499221802, "num_tokens": 255914622.0, "step": 6739 }, { "epoch": 0.8573972777000382, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.137747287750244, "learning_rate": 5e-05, "loss": 0.7533, "mean_token_accuracy": 0.8643976449966431, "num_tokens": 255954027.0, "step": 6740 }, { "epoch": 0.8575244879786287, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.404519081115723, "learning_rate": 5e-05, "loss": 0.9191, "mean_token_accuracy": 0.8515939116477966, "num_tokens": 255997416.0, "step": 6741 }, { "epoch": 0.8576516982572192, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.777375221252441, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8618050813674927, "num_tokens": 256030447.0, "step": 6742 }, { "epoch": 0.8577789085358097, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.834604263305664, "learning_rate": 5e-05, "loss": 0.8972, "mean_token_accuracy": 0.8675637245178223, "num_tokens": 256068533.0, "step": 6743 }, { "epoch": 0.8579061188144002, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.662476062774658, "learning_rate": 5e-05, "loss": 0.7035, "mean_token_accuracy": 0.8761872053146362, "num_tokens": 256104783.0, "step": 6744 }, { "epoch": 0.8580333290929907, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.94548511505127, "learning_rate": 5e-05, "loss": 0.9325, "mean_token_accuracy": 0.8653343319892883, "num_tokens": 256138272.0, "step": 6745 }, { "epoch": 0.8581605393715812, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 6.982743740081787, "learning_rate": 5e-05, "loss": 0.8429, "mean_token_accuracy": 0.8526839017868042, "num_tokens": 256171102.0, "step": 6746 }, { "epoch": 0.8582877496501717, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.59503173828125, "learning_rate": 5e-05, "loss": 0.9052, "mean_token_accuracy": 0.8638439178466797, "num_tokens": 256210593.0, "step": 6747 }, { "epoch": 0.8584149599287623, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.342221736907959, "learning_rate": 5e-05, "loss": 0.8286, "mean_token_accuracy": 0.8717080354690552, "num_tokens": 256248693.0, "step": 6748 }, { "epoch": 0.8585421702073528, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.203495025634766, "learning_rate": 5e-05, "loss": 0.8385, "mean_token_accuracy": 0.8685827255249023, "num_tokens": 256284875.0, "step": 6749 }, { "epoch": 0.8586693804859432, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.400575160980225, "learning_rate": 5e-05, "loss": 0.8184, "mean_token_accuracy": 0.86737459897995, "num_tokens": 256319121.0, "step": 6750 }, { "epoch": 0.8587965907645337, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.06149673461914, "learning_rate": 5e-05, "loss": 0.8081, "mean_token_accuracy": 0.8709604144096375, "num_tokens": 256354112.0, "step": 6751 }, { "epoch": 0.8589238010431243, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.237909317016602, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8723322153091431, "num_tokens": 256387612.0, "step": 6752 }, { "epoch": 0.8590510113217148, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.148945808410645, "learning_rate": 5e-05, "loss": 0.8441, "mean_token_accuracy": 0.8613399863243103, "num_tokens": 256419720.0, "step": 6753 }, { "epoch": 0.8591782216003053, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.021320819854736, "learning_rate": 5e-05, "loss": 0.7849, "mean_token_accuracy": 0.8528942465782166, "num_tokens": 256456717.0, "step": 6754 }, { "epoch": 0.8593054318788959, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.315834045410156, "learning_rate": 5e-05, "loss": 0.9195, "mean_token_accuracy": 0.8415839076042175, "num_tokens": 256496480.0, "step": 6755 }, { "epoch": 0.8594326421574863, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.86125373840332, "learning_rate": 5e-05, "loss": 0.7354, "mean_token_accuracy": 0.8642442226409912, "num_tokens": 256531974.0, "step": 6756 }, { "epoch": 0.8595598524360768, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.439901351928711, "learning_rate": 5e-05, "loss": 0.9062, "mean_token_accuracy": 0.8563156127929688, "num_tokens": 256565983.0, "step": 6757 }, { "epoch": 0.8596870627146673, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.878400802612305, "learning_rate": 5e-05, "loss": 0.7227, "mean_token_accuracy": 0.867829442024231, "num_tokens": 256607215.0, "step": 6758 }, { "epoch": 0.8598142729932579, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.35617446899414, "learning_rate": 5e-05, "loss": 0.8192, "mean_token_accuracy": 0.8750280141830444, "num_tokens": 256646213.0, "step": 6759 }, { "epoch": 0.8599414832718484, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.061161994934082, "learning_rate": 5e-05, "loss": 0.748, "mean_token_accuracy": 0.8703211545944214, "num_tokens": 256687638.0, "step": 6760 }, { "epoch": 0.8600686935504389, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.115943908691406, "learning_rate": 5e-05, "loss": 0.855, "mean_token_accuracy": 0.8584939241409302, "num_tokens": 256716852.0, "step": 6761 }, { "epoch": 0.8601959038290294, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.267910957336426, "learning_rate": 5e-05, "loss": 0.7721, "mean_token_accuracy": 0.8610116243362427, "num_tokens": 256748260.0, "step": 6762 }, { "epoch": 0.8603231141076199, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.84138822555542, "learning_rate": 5e-05, "loss": 0.8837, "mean_token_accuracy": 0.8456039428710938, "num_tokens": 256786400.0, "step": 6763 }, { "epoch": 0.8604503243862104, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.480406761169434, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8713145852088928, "num_tokens": 256820186.0, "step": 6764 }, { "epoch": 0.8605775346648009, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.533388137817383, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8627817034721375, "num_tokens": 256855735.0, "step": 6765 }, { "epoch": 0.8607047449433914, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.657285690307617, "learning_rate": 5e-05, "loss": 0.7521, "mean_token_accuracy": 0.8714607954025269, "num_tokens": 256887809.0, "step": 6766 }, { "epoch": 0.860831955221982, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.324883937835693, "learning_rate": 5e-05, "loss": 0.7412, "mean_token_accuracy": 0.865777850151062, "num_tokens": 256921801.0, "step": 6767 }, { "epoch": 0.8609591655005725, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.769212245941162, "learning_rate": 5e-05, "loss": 0.7411, "mean_token_accuracy": 0.8756980895996094, "num_tokens": 256960399.0, "step": 6768 }, { "epoch": 0.8610863757791629, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.199367523193359, "learning_rate": 5e-05, "loss": 0.726, "mean_token_accuracy": 0.8658415079116821, "num_tokens": 256995094.0, "step": 6769 }, { "epoch": 0.8612135860577534, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.819729328155518, "learning_rate": 5e-05, "loss": 0.8597, "mean_token_accuracy": 0.8420051336288452, "num_tokens": 257037792.0, "step": 6770 }, { "epoch": 0.861340796336344, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.130443572998047, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8655455112457275, "num_tokens": 257075376.0, "step": 6771 }, { "epoch": 0.8614680066149345, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.856428146362305, "learning_rate": 5e-05, "loss": 0.814, "mean_token_accuracy": 0.8548978567123413, "num_tokens": 257115878.0, "step": 6772 }, { "epoch": 0.861595216893525, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.0778656005859375, "learning_rate": 5e-05, "loss": 0.7304, "mean_token_accuracy": 0.8617252111434937, "num_tokens": 257150780.0, "step": 6773 }, { "epoch": 0.8617224271721156, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.893784999847412, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8641605377197266, "num_tokens": 257194704.0, "step": 6774 }, { "epoch": 0.861849637450706, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.006112098693848, "learning_rate": 5e-05, "loss": 0.7675, "mean_token_accuracy": 0.8497557640075684, "num_tokens": 257235337.0, "step": 6775 }, { "epoch": 0.8619768477292965, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.995864391326904, "learning_rate": 5e-05, "loss": 0.8204, "mean_token_accuracy": 0.8499312400817871, "num_tokens": 257268976.0, "step": 6776 }, { "epoch": 0.862104058007887, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.844306468963623, "learning_rate": 5e-05, "loss": 0.7014, "mean_token_accuracy": 0.8638412952423096, "num_tokens": 257304638.0, "step": 6777 }, { "epoch": 0.8622312682864776, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.224867820739746, "learning_rate": 5e-05, "loss": 0.8353, "mean_token_accuracy": 0.8553272485733032, "num_tokens": 257341773.0, "step": 6778 }, { "epoch": 0.8623584785650681, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.621675491333008, "learning_rate": 5e-05, "loss": 0.708, "mean_token_accuracy": 0.8657825589179993, "num_tokens": 257375145.0, "step": 6779 }, { "epoch": 0.8624856888436586, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.540751457214355, "learning_rate": 5e-05, "loss": 0.8287, "mean_token_accuracy": 0.8681566715240479, "num_tokens": 257411070.0, "step": 6780 }, { "epoch": 0.862612899122249, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.351297855377197, "learning_rate": 5e-05, "loss": 0.7292, "mean_token_accuracy": 0.8504693508148193, "num_tokens": 257447283.0, "step": 6781 }, { "epoch": 0.8627401094008396, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.943344116210938, "learning_rate": 5e-05, "loss": 0.8921, "mean_token_accuracy": 0.86326664686203, "num_tokens": 257487370.0, "step": 6782 }, { "epoch": 0.8628673196794301, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.227581024169922, "learning_rate": 5e-05, "loss": 0.7583, "mean_token_accuracy": 0.8486887216567993, "num_tokens": 257525028.0, "step": 6783 }, { "epoch": 0.8629945299580206, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.168272972106934, "learning_rate": 5e-05, "loss": 0.9696, "mean_token_accuracy": 0.8590256571769714, "num_tokens": 257563884.0, "step": 6784 }, { "epoch": 0.8631217402366111, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.637557506561279, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8658711314201355, "num_tokens": 257599431.0, "step": 6785 }, { "epoch": 0.8632489505152017, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.674079895019531, "learning_rate": 5e-05, "loss": 0.9423, "mean_token_accuracy": 0.8622297644615173, "num_tokens": 257636060.0, "step": 6786 }, { "epoch": 0.8633761607937921, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.544657230377197, "learning_rate": 5e-05, "loss": 0.8586, "mean_token_accuracy": 0.8698296546936035, "num_tokens": 257672374.0, "step": 6787 }, { "epoch": 0.8635033710723826, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.459316253662109, "learning_rate": 5e-05, "loss": 0.8432, "mean_token_accuracy": 0.8550323247909546, "num_tokens": 257710875.0, "step": 6788 }, { "epoch": 0.8636305813509731, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.488580703735352, "learning_rate": 5e-05, "loss": 0.968, "mean_token_accuracy": 0.8502017259597778, "num_tokens": 257756697.0, "step": 6789 }, { "epoch": 0.8637577916295637, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.412621021270752, "learning_rate": 5e-05, "loss": 0.7349, "mean_token_accuracy": 0.8545979261398315, "num_tokens": 257795281.0, "step": 6790 }, { "epoch": 0.8638850019081542, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.258069038391113, "learning_rate": 5e-05, "loss": 0.9433, "mean_token_accuracy": 0.8721651434898376, "num_tokens": 257835743.0, "step": 6791 }, { "epoch": 0.8640122121867447, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.1439208984375, "learning_rate": 5e-05, "loss": 0.6847, "mean_token_accuracy": 0.8754228353500366, "num_tokens": 257878546.0, "step": 6792 }, { "epoch": 0.8641394224653351, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.61629581451416, "learning_rate": 5e-05, "loss": 1.0035, "mean_token_accuracy": 0.8659759163856506, "num_tokens": 257915480.0, "step": 6793 }, { "epoch": 0.8642666327439257, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 6.869853496551514, "learning_rate": 5e-05, "loss": 0.8611, "mean_token_accuracy": 0.860084056854248, "num_tokens": 257951215.0, "step": 6794 }, { "epoch": 0.8643938430225162, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.023873329162598, "learning_rate": 5e-05, "loss": 1.0186, "mean_token_accuracy": 0.8576740026473999, "num_tokens": 257987623.0, "step": 6795 }, { "epoch": 0.8645210533011067, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.932648181915283, "learning_rate": 5e-05, "loss": 1.01, "mean_token_accuracy": 0.8494714498519897, "num_tokens": 258030078.0, "step": 6796 }, { "epoch": 0.8646482635796973, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.756474494934082, "learning_rate": 5e-05, "loss": 0.835, "mean_token_accuracy": 0.8724368810653687, "num_tokens": 258064598.0, "step": 6797 }, { "epoch": 0.8647754738582878, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.473021507263184, "learning_rate": 5e-05, "loss": 0.9439, "mean_token_accuracy": 0.8655718564987183, "num_tokens": 258100121.0, "step": 6798 }, { "epoch": 0.8649026841368782, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.17657995223999, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.8598823547363281, "num_tokens": 258138398.0, "step": 6799 }, { "epoch": 0.8650298944154687, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.200196266174316, "learning_rate": 5e-05, "loss": 0.911, "mean_token_accuracy": 0.8605042695999146, "num_tokens": 258178730.0, "step": 6800 }, { "epoch": 0.8651571046940593, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.579260349273682, "learning_rate": 5e-05, "loss": 0.8436, "mean_token_accuracy": 0.8649904727935791, "num_tokens": 258211122.0, "step": 6801 }, { "epoch": 0.8652843149726498, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.370438575744629, "learning_rate": 5e-05, "loss": 0.7806, "mean_token_accuracy": 0.868091881275177, "num_tokens": 258249753.0, "step": 6802 }, { "epoch": 0.8654115252512403, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.1348295211792, "learning_rate": 5e-05, "loss": 0.8746, "mean_token_accuracy": 0.8594678044319153, "num_tokens": 258287951.0, "step": 6803 }, { "epoch": 0.8655387355298308, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.852487087249756, "learning_rate": 5e-05, "loss": 0.7797, "mean_token_accuracy": 0.8536652326583862, "num_tokens": 258333216.0, "step": 6804 }, { "epoch": 0.8656659458084213, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.45638370513916, "learning_rate": 5e-05, "loss": 0.8866, "mean_token_accuracy": 0.8648599982261658, "num_tokens": 258370380.0, "step": 6805 }, { "epoch": 0.8657931560870118, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.647890090942383, "learning_rate": 5e-05, "loss": 0.7901, "mean_token_accuracy": 0.8487348556518555, "num_tokens": 258413730.0, "step": 6806 }, { "epoch": 0.8659203663656023, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.779346466064453, "learning_rate": 5e-05, "loss": 0.9216, "mean_token_accuracy": 0.8662146329879761, "num_tokens": 258451644.0, "step": 6807 }, { "epoch": 0.8660475766441929, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.5132060050964355, "learning_rate": 5e-05, "loss": 0.7092, "mean_token_accuracy": 0.8754907846450806, "num_tokens": 258490084.0, "step": 6808 }, { "epoch": 0.8661747869227834, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.093279838562012, "learning_rate": 5e-05, "loss": 1.0443, "mean_token_accuracy": 0.8443394303321838, "num_tokens": 258530769.0, "step": 6809 }, { "epoch": 0.8663019972013739, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.6606597900390625, "learning_rate": 5e-05, "loss": 0.8201, "mean_token_accuracy": 0.8549513816833496, "num_tokens": 258567578.0, "step": 6810 }, { "epoch": 0.8664292074799644, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.091015815734863, "learning_rate": 5e-05, "loss": 0.9759, "mean_token_accuracy": 0.8614137768745422, "num_tokens": 258605498.0, "step": 6811 }, { "epoch": 0.8665564177585549, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.131685733795166, "learning_rate": 5e-05, "loss": 0.8275, "mean_token_accuracy": 0.8786877989768982, "num_tokens": 258646931.0, "step": 6812 }, { "epoch": 0.8666836280371454, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.711581230163574, "learning_rate": 5e-05, "loss": 0.9513, "mean_token_accuracy": 0.8686947226524353, "num_tokens": 258683853.0, "step": 6813 }, { "epoch": 0.8668108383157359, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.460813999176025, "learning_rate": 5e-05, "loss": 0.8929, "mean_token_accuracy": 0.8643523454666138, "num_tokens": 258723706.0, "step": 6814 }, { "epoch": 0.8669380485943264, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.364219665527344, "learning_rate": 5e-05, "loss": 0.945, "mean_token_accuracy": 0.8628634214401245, "num_tokens": 258762404.0, "step": 6815 }, { "epoch": 0.867065258872917, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.431126117706299, "learning_rate": 5e-05, "loss": 0.9222, "mean_token_accuracy": 0.8507142066955566, "num_tokens": 258803209.0, "step": 6816 }, { "epoch": 0.8671924691515075, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.29178237915039, "learning_rate": 5e-05, "loss": 0.9383, "mean_token_accuracy": 0.8624938726425171, "num_tokens": 258840001.0, "step": 6817 }, { "epoch": 0.8673196794300979, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.263192176818848, "learning_rate": 5e-05, "loss": 0.7897, "mean_token_accuracy": 0.8810036778450012, "num_tokens": 258871401.0, "step": 6818 }, { "epoch": 0.8674468897086884, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.312328338623047, "learning_rate": 5e-05, "loss": 0.9766, "mean_token_accuracy": 0.8461100459098816, "num_tokens": 258902898.0, "step": 6819 }, { "epoch": 0.867574099987279, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.206853866577148, "learning_rate": 5e-05, "loss": 0.8266, "mean_token_accuracy": 0.867053747177124, "num_tokens": 258942136.0, "step": 6820 }, { "epoch": 0.8677013102658695, "ewc_loss": 0.5, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.188138008117676, "learning_rate": 5e-05, "loss": 0.9023, "mean_token_accuracy": 0.863129198551178, "num_tokens": 258976599.0, "step": 6821 }, { "epoch": 0.86782852054446, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.311441898345947, "learning_rate": 5e-05, "loss": 0.8398, "mean_token_accuracy": 0.8595133423805237, "num_tokens": 259010457.0, "step": 6822 }, { "epoch": 0.8679557308230506, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.9041032791137695, "learning_rate": 5e-05, "loss": 0.8605, "mean_token_accuracy": 0.8626033663749695, "num_tokens": 259046368.0, "step": 6823 }, { "epoch": 0.868082941101641, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.440300464630127, "learning_rate": 5e-05, "loss": 0.804, "mean_token_accuracy": 0.8687654733657837, "num_tokens": 259087240.0, "step": 6824 }, { "epoch": 0.8682101513802315, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.699934005737305, "learning_rate": 5e-05, "loss": 0.8376, "mean_token_accuracy": 0.8638709783554077, "num_tokens": 259128297.0, "step": 6825 }, { "epoch": 0.868337361658822, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.457923412322998, "learning_rate": 5e-05, "loss": 0.8201, "mean_token_accuracy": 0.8603090047836304, "num_tokens": 259170242.0, "step": 6826 }, { "epoch": 0.8684645719374126, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.634744644165039, "learning_rate": 5e-05, "loss": 0.8051, "mean_token_accuracy": 0.8618474006652832, "num_tokens": 259208580.0, "step": 6827 }, { "epoch": 0.8685917822160031, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.403954029083252, "learning_rate": 5e-05, "loss": 0.8671, "mean_token_accuracy": 0.8426609635353088, "num_tokens": 259249721.0, "step": 6828 }, { "epoch": 0.8687189924945936, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.613714218139648, "learning_rate": 5e-05, "loss": 0.7773, "mean_token_accuracy": 0.8697314262390137, "num_tokens": 259292056.0, "step": 6829 }, { "epoch": 0.868846202773184, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.377493381500244, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8693746328353882, "num_tokens": 259323390.0, "step": 6830 }, { "epoch": 0.8689734130517746, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.571651935577393, "learning_rate": 5e-05, "loss": 0.7498, "mean_token_accuracy": 0.8732923269271851, "num_tokens": 259360728.0, "step": 6831 }, { "epoch": 0.8691006233303651, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.380500316619873, "learning_rate": 5e-05, "loss": 0.7632, "mean_token_accuracy": 0.8618152141571045, "num_tokens": 259397619.0, "step": 6832 }, { "epoch": 0.8692278336089556, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.5192155838012695, "learning_rate": 5e-05, "loss": 0.8165, "mean_token_accuracy": 0.8479071855545044, "num_tokens": 259440049.0, "step": 6833 }, { "epoch": 0.8693550438875461, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.384920597076416, "learning_rate": 5e-05, "loss": 0.7475, "mean_token_accuracy": 0.8666641712188721, "num_tokens": 259475657.0, "step": 6834 }, { "epoch": 0.8694822541661367, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.505377769470215, "learning_rate": 5e-05, "loss": 0.8327, "mean_token_accuracy": 0.8405799865722656, "num_tokens": 259511289.0, "step": 6835 }, { "epoch": 0.8696094644447271, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.329379081726074, "learning_rate": 5e-05, "loss": 0.7192, "mean_token_accuracy": 0.8748233914375305, "num_tokens": 259546704.0, "step": 6836 }, { "epoch": 0.8697366747233176, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.5412163734436035, "learning_rate": 5e-05, "loss": 0.7373, "mean_token_accuracy": 0.8742718696594238, "num_tokens": 259588490.0, "step": 6837 }, { "epoch": 0.8698638850019081, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.275598049163818, "learning_rate": 5e-05, "loss": 0.7542, "mean_token_accuracy": 0.8618303537368774, "num_tokens": 259619838.0, "step": 6838 }, { "epoch": 0.8699910952804987, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.610101222991943, "learning_rate": 5e-05, "loss": 0.7966, "mean_token_accuracy": 0.8522868752479553, "num_tokens": 259662232.0, "step": 6839 }, { "epoch": 0.8701183055590892, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.175008773803711, "learning_rate": 5e-05, "loss": 0.8051, "mean_token_accuracy": 0.8392272591590881, "num_tokens": 259699824.0, "step": 6840 }, { "epoch": 0.8702455158376797, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.718215465545654, "learning_rate": 5e-05, "loss": 0.7715, "mean_token_accuracy": 0.8605777621269226, "num_tokens": 259728217.0, "step": 6841 }, { "epoch": 0.8703727261162701, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.016533374786377, "learning_rate": 5e-05, "loss": 0.6756, "mean_token_accuracy": 0.876233696937561, "num_tokens": 259763522.0, "step": 6842 }, { "epoch": 0.8704999363948607, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.960378646850586, "learning_rate": 5e-05, "loss": 0.8161, "mean_token_accuracy": 0.8512899279594421, "num_tokens": 259797650.0, "step": 6843 }, { "epoch": 0.8706271466734512, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.70815372467041, "learning_rate": 5e-05, "loss": 0.7506, "mean_token_accuracy": 0.8523268103599548, "num_tokens": 259837265.0, "step": 6844 }, { "epoch": 0.8707543569520417, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.392305374145508, "learning_rate": 5e-05, "loss": 0.8169, "mean_token_accuracy": 0.864956259727478, "num_tokens": 259873040.0, "step": 6845 }, { "epoch": 0.8708815672306323, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00023555755615234375, "grad_norm": 6.260813236236572, "learning_rate": 5e-05, "loss": 0.7397, "mean_token_accuracy": 0.8446069955825806, "num_tokens": 259907627.0, "step": 6846 }, { "epoch": 0.8710087775092228, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 9.070387840270996, "learning_rate": 5e-05, "loss": 0.9616, "mean_token_accuracy": 0.8437118530273438, "num_tokens": 259948213.0, "step": 6847 }, { "epoch": 0.8711359877878132, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 5.792283058166504, "learning_rate": 5e-05, "loss": 0.6933, "mean_token_accuracy": 0.8560380935668945, "num_tokens": 259985239.0, "step": 6848 }, { "epoch": 0.8712631980664037, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.919000625610352, "learning_rate": 5e-05, "loss": 1.1196, "mean_token_accuracy": 0.8331258296966553, "num_tokens": 260023046.0, "step": 6849 }, { "epoch": 0.8713904083449943, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.116856098175049, "learning_rate": 5e-05, "loss": 0.7238, "mean_token_accuracy": 0.8718578815460205, "num_tokens": 260060716.0, "step": 6850 }, { "epoch": 0.8715176186235848, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.047194480895996, "learning_rate": 5e-05, "loss": 1.037, "mean_token_accuracy": 0.8809267282485962, "num_tokens": 260093939.0, "step": 6851 }, { "epoch": 0.8716448289021753, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 7.84667444229126, "learning_rate": 5e-05, "loss": 1.0111, "mean_token_accuracy": 0.8623813390731812, "num_tokens": 260129282.0, "step": 6852 }, { "epoch": 0.8717720391807658, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.301823616027832, "learning_rate": 5e-05, "loss": 0.8767, "mean_token_accuracy": 0.8762044906616211, "num_tokens": 260172167.0, "step": 6853 }, { "epoch": 0.8718992494593563, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 9.425969123840332, "learning_rate": 5e-05, "loss": 1.0712, "mean_token_accuracy": 0.8682819604873657, "num_tokens": 260212613.0, "step": 6854 }, { "epoch": 0.8720264597379468, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.128102779388428, "learning_rate": 5e-05, "loss": 0.7142, "mean_token_accuracy": 0.8677650094032288, "num_tokens": 260248774.0, "step": 6855 }, { "epoch": 0.8721536700165373, "ewc_loss": 0.91796875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000873565673828125, "grad_norm": 11.882163047790527, "learning_rate": 5e-05, "loss": 1.3744, "mean_token_accuracy": 0.8437729477882385, "num_tokens": 260291105.0, "step": 6856 }, { "epoch": 0.8722808802951278, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 7.923644065856934, "learning_rate": 5e-05, "loss": 1.0098, "mean_token_accuracy": 0.8679900169372559, "num_tokens": 260332316.0, "step": 6857 }, { "epoch": 0.8724080905737184, "ewc_loss": 0.90234375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 12.481206893920898, "learning_rate": 5e-05, "loss": 1.2703, "mean_token_accuracy": 0.8715403079986572, "num_tokens": 260365016.0, "step": 6858 }, { "epoch": 0.8725353008523089, "ewc_loss": 1.0546875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00101470947265625, "grad_norm": 10.376077651977539, "learning_rate": 5e-05, "loss": 1.3966, "mean_token_accuracy": 0.880984902381897, "num_tokens": 260396131.0, "step": 6859 }, { "epoch": 0.8726625111308994, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 11.337189674377441, "learning_rate": 5e-05, "loss": 1.2095, "mean_token_accuracy": 0.847374677658081, "num_tokens": 260438053.0, "step": 6860 }, { "epoch": 0.8727897214094898, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 9.357353210449219, "learning_rate": 5e-05, "loss": 1.1474, "mean_token_accuracy": 0.8577359914779663, "num_tokens": 260479523.0, "step": 6861 }, { "epoch": 0.8729169316880804, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 11.149811744689941, "learning_rate": 5e-05, "loss": 1.1899, "mean_token_accuracy": 0.8676673173904419, "num_tokens": 260516623.0, "step": 6862 }, { "epoch": 0.8730441419666709, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.726404190063477, "learning_rate": 5e-05, "loss": 0.9882, "mean_token_accuracy": 0.8672620058059692, "num_tokens": 260552499.0, "step": 6863 }, { "epoch": 0.8731713522452614, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 10.941827774047852, "learning_rate": 5e-05, "loss": 1.1555, "mean_token_accuracy": 0.8641577959060669, "num_tokens": 260595636.0, "step": 6864 }, { "epoch": 0.873298562523852, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.900580406188965, "learning_rate": 5e-05, "loss": 1.0196, "mean_token_accuracy": 0.8686490058898926, "num_tokens": 260639787.0, "step": 6865 }, { "epoch": 0.8734257728024425, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.622868537902832, "learning_rate": 5e-05, "loss": 1.0774, "mean_token_accuracy": 0.8763903379440308, "num_tokens": 260674117.0, "step": 6866 }, { "epoch": 0.8735529830810329, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.002152442932129, "learning_rate": 5e-05, "loss": 1.0461, "mean_token_accuracy": 0.8618425130844116, "num_tokens": 260712757.0, "step": 6867 }, { "epoch": 0.8736801933596234, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.135152816772461, "learning_rate": 5e-05, "loss": 1.1166, "mean_token_accuracy": 0.8442223072052002, "num_tokens": 260750756.0, "step": 6868 }, { "epoch": 0.873807403638214, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.684103012084961, "learning_rate": 5e-05, "loss": 1.0009, "mean_token_accuracy": 0.8592473268508911, "num_tokens": 260782929.0, "step": 6869 }, { "epoch": 0.8739346139168045, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.869660377502441, "learning_rate": 5e-05, "loss": 1.0373, "mean_token_accuracy": 0.8620879054069519, "num_tokens": 260819762.0, "step": 6870 }, { "epoch": 0.874061824195395, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.22082233428955, "learning_rate": 5e-05, "loss": 0.9469, "mean_token_accuracy": 0.8563005328178406, "num_tokens": 260862543.0, "step": 6871 }, { "epoch": 0.8741890344739855, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.030989646911621, "learning_rate": 5e-05, "loss": 1.123, "mean_token_accuracy": 0.8493748903274536, "num_tokens": 260899835.0, "step": 6872 }, { "epoch": 0.874316244752576, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.111018180847168, "learning_rate": 5e-05, "loss": 0.9653, "mean_token_accuracy": 0.8491755723953247, "num_tokens": 260940593.0, "step": 6873 }, { "epoch": 0.8744434550311665, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.244604110717773, "learning_rate": 5e-05, "loss": 1.1012, "mean_token_accuracy": 0.8624186515808105, "num_tokens": 260977454.0, "step": 6874 }, { "epoch": 0.874570665309757, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.477928161621094, "learning_rate": 5e-05, "loss": 0.9725, "mean_token_accuracy": 0.8613405227661133, "num_tokens": 261019095.0, "step": 6875 }, { "epoch": 0.8746978755883476, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.982422828674316, "learning_rate": 5e-05, "loss": 1.1039, "mean_token_accuracy": 0.8535877466201782, "num_tokens": 261056428.0, "step": 6876 }, { "epoch": 0.8748250858669381, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.740486145019531, "learning_rate": 5e-05, "loss": 0.9391, "mean_token_accuracy": 0.8739484548568726, "num_tokens": 261090679.0, "step": 6877 }, { "epoch": 0.8749522961455286, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.331188201904297, "learning_rate": 5e-05, "loss": 1.0144, "mean_token_accuracy": 0.8607511520385742, "num_tokens": 261130640.0, "step": 6878 }, { "epoch": 0.875079506424119, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.48819637298584, "learning_rate": 5e-05, "loss": 0.8993, "mean_token_accuracy": 0.875970184803009, "num_tokens": 261169064.0, "step": 6879 }, { "epoch": 0.8752067167027096, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.916464805603027, "learning_rate": 5e-05, "loss": 0.9587, "mean_token_accuracy": 0.8637458682060242, "num_tokens": 261205625.0, "step": 6880 }, { "epoch": 0.8753339269813001, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.939992427825928, "learning_rate": 5e-05, "loss": 0.8245, "mean_token_accuracy": 0.8712983727455139, "num_tokens": 261251859.0, "step": 6881 }, { "epoch": 0.8754611372598906, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.055805206298828, "learning_rate": 5e-05, "loss": 0.9674, "mean_token_accuracy": 0.862655520439148, "num_tokens": 261289604.0, "step": 6882 }, { "epoch": 0.8755883475384811, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.487751483917236, "learning_rate": 5e-05, "loss": 0.7429, "mean_token_accuracy": 0.8836625814437866, "num_tokens": 261330053.0, "step": 6883 }, { "epoch": 0.8757155578170717, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.49411392211914, "learning_rate": 5e-05, "loss": 1.0405, "mean_token_accuracy": 0.8558171987533569, "num_tokens": 261371042.0, "step": 6884 }, { "epoch": 0.8758427680956621, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.642839431762695, "learning_rate": 5e-05, "loss": 0.8747, "mean_token_accuracy": 0.8543882369995117, "num_tokens": 261406911.0, "step": 6885 }, { "epoch": 0.8759699783742526, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.580371856689453, "learning_rate": 5e-05, "loss": 1.0285, "mean_token_accuracy": 0.8653366565704346, "num_tokens": 261443856.0, "step": 6886 }, { "epoch": 0.8760971886528431, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.213155746459961, "learning_rate": 5e-05, "loss": 0.9442, "mean_token_accuracy": 0.8570191860198975, "num_tokens": 261485987.0, "step": 6887 }, { "epoch": 0.8762243989314337, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.90307903289795, "learning_rate": 5e-05, "loss": 0.9974, "mean_token_accuracy": 0.8575316667556763, "num_tokens": 261519888.0, "step": 6888 }, { "epoch": 0.8763516092100242, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.328288078308105, "learning_rate": 5e-05, "loss": 0.9198, "mean_token_accuracy": 0.8633891344070435, "num_tokens": 261557435.0, "step": 6889 }, { "epoch": 0.8764788194886147, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.19091796875, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.8604919910430908, "num_tokens": 261598209.0, "step": 6890 }, { "epoch": 0.8766060297672051, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.914546489715576, "learning_rate": 5e-05, "loss": 0.8765, "mean_token_accuracy": 0.8488236665725708, "num_tokens": 261637972.0, "step": 6891 }, { "epoch": 0.8767332400457957, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.138044357299805, "learning_rate": 5e-05, "loss": 0.9233, "mean_token_accuracy": 0.8499833345413208, "num_tokens": 261676717.0, "step": 6892 }, { "epoch": 0.8768604503243862, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.400658130645752, "learning_rate": 5e-05, "loss": 0.8482, "mean_token_accuracy": 0.84291672706604, "num_tokens": 261711434.0, "step": 6893 }, { "epoch": 0.8769876606029767, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.382542610168457, "learning_rate": 5e-05, "loss": 0.8732, "mean_token_accuracy": 0.8641119003295898, "num_tokens": 261749527.0, "step": 6894 }, { "epoch": 0.8771148708815673, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.086524486541748, "learning_rate": 5e-05, "loss": 0.7372, "mean_token_accuracy": 0.8716509342193604, "num_tokens": 261792031.0, "step": 6895 }, { "epoch": 0.8772420811601578, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.6229887008667, "learning_rate": 5e-05, "loss": 0.9726, "mean_token_accuracy": 0.8452037572860718, "num_tokens": 261831862.0, "step": 6896 }, { "epoch": 0.8773692914387482, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.008965969085693, "learning_rate": 5e-05, "loss": 0.7418, "mean_token_accuracy": 0.8674806356430054, "num_tokens": 261869949.0, "step": 6897 }, { "epoch": 0.8774965017173387, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.76872444152832, "learning_rate": 5e-05, "loss": 0.918, "mean_token_accuracy": 0.8633503913879395, "num_tokens": 261904080.0, "step": 6898 }, { "epoch": 0.8776237119959293, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.211948871612549, "learning_rate": 5e-05, "loss": 0.8107, "mean_token_accuracy": 0.8492647409439087, "num_tokens": 261932809.0, "step": 6899 }, { "epoch": 0.8777509222745198, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.586893081665039, "learning_rate": 5e-05, "loss": 0.915, "mean_token_accuracy": 0.8605145215988159, "num_tokens": 261974313.0, "step": 6900 }, { "epoch": 0.8778781325531103, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.549050331115723, "learning_rate": 5e-05, "loss": 0.7693, "mean_token_accuracy": 0.8765868544578552, "num_tokens": 262011505.0, "step": 6901 }, { "epoch": 0.8780053428317008, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.064190864562988, "learning_rate": 5e-05, "loss": 0.865, "mean_token_accuracy": 0.8648147583007812, "num_tokens": 262042571.0, "step": 6902 }, { "epoch": 0.8781325531102913, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.762099742889404, "learning_rate": 5e-05, "loss": 0.851, "mean_token_accuracy": 0.8546534776687622, "num_tokens": 262088038.0, "step": 6903 }, { "epoch": 0.8782597633888818, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.549023628234863, "learning_rate": 5e-05, "loss": 0.7711, "mean_token_accuracy": 0.8726382255554199, "num_tokens": 262126662.0, "step": 6904 }, { "epoch": 0.8783869736674723, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.850553512573242, "learning_rate": 5e-05, "loss": 0.8108, "mean_token_accuracy": 0.8656457662582397, "num_tokens": 262170525.0, "step": 6905 }, { "epoch": 0.8785141839460628, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.321649074554443, "learning_rate": 5e-05, "loss": 0.8225, "mean_token_accuracy": 0.84868323802948, "num_tokens": 262210524.0, "step": 6906 }, { "epoch": 0.8786413942246534, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.866237640380859, "learning_rate": 5e-05, "loss": 0.8733, "mean_token_accuracy": 0.8442649245262146, "num_tokens": 262248636.0, "step": 6907 }, { "epoch": 0.8787686045032439, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.196749687194824, "learning_rate": 5e-05, "loss": 0.7847, "mean_token_accuracy": 0.8521646857261658, "num_tokens": 262291795.0, "step": 6908 }, { "epoch": 0.8788958147818343, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.93075704574585, "learning_rate": 5e-05, "loss": 0.7935, "mean_token_accuracy": 0.8661741018295288, "num_tokens": 262328484.0, "step": 6909 }, { "epoch": 0.8790230250604248, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.08118200302124, "learning_rate": 5e-05, "loss": 0.7229, "mean_token_accuracy": 0.8678644299507141, "num_tokens": 262362262.0, "step": 6910 }, { "epoch": 0.8791502353390154, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.004042625427246, "learning_rate": 5e-05, "loss": 0.8419, "mean_token_accuracy": 0.8539531230926514, "num_tokens": 262399082.0, "step": 6911 }, { "epoch": 0.8792774456176059, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.040284633636475, "learning_rate": 5e-05, "loss": 0.7303, "mean_token_accuracy": 0.8640420436859131, "num_tokens": 262434968.0, "step": 6912 }, { "epoch": 0.8794046558961964, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.990469932556152, "learning_rate": 5e-05, "loss": 0.7868, "mean_token_accuracy": 0.8665574789047241, "num_tokens": 262472989.0, "step": 6913 }, { "epoch": 0.879531866174787, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.119815349578857, "learning_rate": 5e-05, "loss": 0.7356, "mean_token_accuracy": 0.8623637557029724, "num_tokens": 262512526.0, "step": 6914 }, { "epoch": 0.8796590764533775, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.835806846618652, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8688112497329712, "num_tokens": 262548082.0, "step": 6915 }, { "epoch": 0.8797862867319679, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.2665486335754395, "learning_rate": 5e-05, "loss": 0.7925, "mean_token_accuracy": 0.8477435111999512, "num_tokens": 262584786.0, "step": 6916 }, { "epoch": 0.8799134970105584, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.644948482513428, "learning_rate": 5e-05, "loss": 0.7683, "mean_token_accuracy": 0.8674691319465637, "num_tokens": 262625898.0, "step": 6917 }, { "epoch": 0.880040707289149, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.401279926300049, "learning_rate": 5e-05, "loss": 0.7408, "mean_token_accuracy": 0.8647757768630981, "num_tokens": 262664041.0, "step": 6918 }, { "epoch": 0.8801679175677395, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.487602710723877, "learning_rate": 5e-05, "loss": 0.8007, "mean_token_accuracy": 0.850123405456543, "num_tokens": 262702597.0, "step": 6919 }, { "epoch": 0.88029512784633, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.498105525970459, "learning_rate": 5e-05, "loss": 0.8153, "mean_token_accuracy": 0.8439078330993652, "num_tokens": 262734951.0, "step": 6920 }, { "epoch": 0.8804223381249205, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.402912139892578, "learning_rate": 5e-05, "loss": 0.728, "mean_token_accuracy": 0.8697965145111084, "num_tokens": 262771513.0, "step": 6921 }, { "epoch": 0.880549548403511, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.5117292404174805, "learning_rate": 5e-05, "loss": 0.7615, "mean_token_accuracy": 0.85517418384552, "num_tokens": 262813045.0, "step": 6922 }, { "epoch": 0.8806767586821015, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.405292987823486, "learning_rate": 5e-05, "loss": 0.701, "mean_token_accuracy": 0.8734561204910278, "num_tokens": 262846663.0, "step": 6923 }, { "epoch": 0.880803968960692, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.470313549041748, "learning_rate": 5e-05, "loss": 0.739, "mean_token_accuracy": 0.8642817735671997, "num_tokens": 262880056.0, "step": 6924 }, { "epoch": 0.8809311792392825, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.437323570251465, "learning_rate": 5e-05, "loss": 0.7154, "mean_token_accuracy": 0.8709765672683716, "num_tokens": 262914819.0, "step": 6925 }, { "epoch": 0.8810583895178731, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.407699108123779, "learning_rate": 5e-05, "loss": 0.7496, "mean_token_accuracy": 0.8573665022850037, "num_tokens": 262953167.0, "step": 6926 }, { "epoch": 0.8811855997964636, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.520580291748047, "learning_rate": 5e-05, "loss": 0.7658, "mean_token_accuracy": 0.8561645746231079, "num_tokens": 262991628.0, "step": 6927 }, { "epoch": 0.881312810075054, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.2536163330078125, "learning_rate": 5e-05, "loss": 0.705, "mean_token_accuracy": 0.8671410083770752, "num_tokens": 263029287.0, "step": 6928 }, { "epoch": 0.8814400203536445, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.695040702819824, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8603187799453735, "num_tokens": 263068982.0, "step": 6929 }, { "epoch": 0.8815672306322351, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 7.0082478523254395, "learning_rate": 5e-05, "loss": 0.8018, "mean_token_accuracy": 0.8385601043701172, "num_tokens": 263111260.0, "step": 6930 }, { "epoch": 0.8816944409108256, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.015702247619629, "learning_rate": 5e-05, "loss": 0.7667, "mean_token_accuracy": 0.8656647205352783, "num_tokens": 263149310.0, "step": 6931 }, { "epoch": 0.8818216511894161, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00023651123046875, "grad_norm": 6.630132675170898, "learning_rate": 5e-05, "loss": 0.6499, "mean_token_accuracy": 0.8743974566459656, "num_tokens": 263188440.0, "step": 6932 }, { "epoch": 0.8819488614680067, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.570465087890625, "learning_rate": 5e-05, "loss": 0.8611, "mean_token_accuracy": 0.8524523377418518, "num_tokens": 263229079.0, "step": 6933 }, { "epoch": 0.8820760717465971, "ewc_loss": 0.25, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000209808349609375, "grad_norm": 6.015937805175781, "learning_rate": 5e-05, "loss": 0.6825, "mean_token_accuracy": 0.8543329834938049, "num_tokens": 263260591.0, "step": 6934 }, { "epoch": 0.8822032820251876, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.567778587341309, "learning_rate": 5e-05, "loss": 0.9719, "mean_token_accuracy": 0.8506943583488464, "num_tokens": 263301256.0, "step": 6935 }, { "epoch": 0.8823304923037781, "ewc_loss": 0.2421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00020122528076171875, "grad_norm": 5.3513994216918945, "learning_rate": 5e-05, "loss": 0.626, "mean_token_accuracy": 0.8677116632461548, "num_tokens": 263337600.0, "step": 6936 }, { "epoch": 0.8824577025823687, "ewc_loss": 0.71875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.963313102722168, "learning_rate": 5e-05, "loss": 1.1414, "mean_token_accuracy": 0.8580079674720764, "num_tokens": 263383124.0, "step": 6937 }, { "epoch": 0.8825849128609592, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 6.423493385314941, "learning_rate": 5e-05, "loss": 0.7997, "mean_token_accuracy": 0.8682374954223633, "num_tokens": 263416047.0, "step": 6938 }, { "epoch": 0.8827121231395497, "ewc_loss": 0.8828125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 11.486123085021973, "learning_rate": 5e-05, "loss": 1.2461, "mean_token_accuracy": 0.874570369720459, "num_tokens": 263458515.0, "step": 6939 }, { "epoch": 0.8828393334181401, "ewc_loss": 0.859375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00081634521484375, "grad_norm": 9.205920219421387, "learning_rate": 5e-05, "loss": 1.2992, "mean_token_accuracy": 0.8525868058204651, "num_tokens": 263499833.0, "step": 6940 }, { "epoch": 0.8829665436967307, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.449654579162598, "learning_rate": 5e-05, "loss": 0.9786, "mean_token_accuracy": 0.8586032390594482, "num_tokens": 263531680.0, "step": 6941 }, { "epoch": 0.8830937539753212, "ewc_loss": 0.84375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 10.117497444152832, "learning_rate": 5e-05, "loss": 1.2895, "mean_token_accuracy": 0.8484832048416138, "num_tokens": 263565889.0, "step": 6942 }, { "epoch": 0.8832209642539117, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.664785861968994, "learning_rate": 5e-05, "loss": 0.8897, "mean_token_accuracy": 0.855044960975647, "num_tokens": 263604974.0, "step": 6943 }, { "epoch": 0.8833481745325023, "ewc_loss": 0.921875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00087738037109375, "grad_norm": 11.836788177490234, "learning_rate": 5e-05, "loss": 1.3289, "mean_token_accuracy": 0.8618921637535095, "num_tokens": 263646960.0, "step": 6944 }, { "epoch": 0.8834753848110928, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 9.11538314819336, "learning_rate": 5e-05, "loss": 1.1622, "mean_token_accuracy": 0.8591632843017578, "num_tokens": 263686427.0, "step": 6945 }, { "epoch": 0.8836025950896832, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 11.521474838256836, "learning_rate": 5e-05, "loss": 1.0925, "mean_token_accuracy": 0.8715478181838989, "num_tokens": 263717520.0, "step": 6946 }, { "epoch": 0.8837298053682737, "ewc_loss": 0.796875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 9.478304862976074, "learning_rate": 5e-05, "loss": 1.2195, "mean_token_accuracy": 0.8579591512680054, "num_tokens": 263756226.0, "step": 6947 }, { "epoch": 0.8838570156468643, "ewc_loss": 0.6875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 11.06391716003418, "learning_rate": 5e-05, "loss": 1.0629, "mean_token_accuracy": 0.8746497631072998, "num_tokens": 263793961.0, "step": 6948 }, { "epoch": 0.8839842259254548, "ewc_loss": 0.625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.610187530517578, "learning_rate": 5e-05, "loss": 1.0409, "mean_token_accuracy": 0.8584915399551392, "num_tokens": 263827325.0, "step": 6949 }, { "epoch": 0.8841114362040453, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 10.9205322265625, "learning_rate": 5e-05, "loss": 1.1809, "mean_token_accuracy": 0.8405921459197998, "num_tokens": 263862854.0, "step": 6950 }, { "epoch": 0.8842386464826358, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.030985832214355, "learning_rate": 5e-05, "loss": 1.0399, "mean_token_accuracy": 0.8362864255905151, "num_tokens": 263907035.0, "step": 6951 }, { "epoch": 0.8843658567612263, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 11.247751235961914, "learning_rate": 5e-05, "loss": 1.1716, "mean_token_accuracy": 0.8508137464523315, "num_tokens": 263949550.0, "step": 6952 }, { "epoch": 0.8844930670398168, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.318743705749512, "learning_rate": 5e-05, "loss": 0.986, "mean_token_accuracy": 0.8623531460762024, "num_tokens": 263988966.0, "step": 6953 }, { "epoch": 0.8846202773184073, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000732421875, "grad_norm": 11.639720916748047, "learning_rate": 5e-05, "loss": 1.1378, "mean_token_accuracy": 0.8756442070007324, "num_tokens": 264023442.0, "step": 6954 }, { "epoch": 0.8847474875969978, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 9.23528003692627, "learning_rate": 5e-05, "loss": 1.0597, "mean_token_accuracy": 0.8702212572097778, "num_tokens": 264064185.0, "step": 6955 }, { "epoch": 0.8848746978755884, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 11.240945816040039, "learning_rate": 5e-05, "loss": 1.1607, "mean_token_accuracy": 0.8529358506202698, "num_tokens": 264107455.0, "step": 6956 }, { "epoch": 0.8850019081541789, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.503716468811035, "learning_rate": 5e-05, "loss": 1.0825, "mean_token_accuracy": 0.8557217121124268, "num_tokens": 264143090.0, "step": 6957 }, { "epoch": 0.8851291184327693, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.564746856689453, "learning_rate": 5e-05, "loss": 1.085, "mean_token_accuracy": 0.8588166832923889, "num_tokens": 264180002.0, "step": 6958 }, { "epoch": 0.8852563287113598, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.240660667419434, "learning_rate": 5e-05, "loss": 0.9869, "mean_token_accuracy": 0.8713166117668152, "num_tokens": 264215444.0, "step": 6959 }, { "epoch": 0.8853835389899504, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.90333080291748, "learning_rate": 5e-05, "loss": 1.0138, "mean_token_accuracy": 0.8642602562904358, "num_tokens": 264250758.0, "step": 6960 }, { "epoch": 0.8855107492685409, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.732237815856934, "learning_rate": 5e-05, "loss": 0.9944, "mean_token_accuracy": 0.8466671705245972, "num_tokens": 264290181.0, "step": 6961 }, { "epoch": 0.8856379595471314, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.585451126098633, "learning_rate": 5e-05, "loss": 0.9963, "mean_token_accuracy": 0.862805962562561, "num_tokens": 264327480.0, "step": 6962 }, { "epoch": 0.885765169825722, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.100319862365723, "learning_rate": 5e-05, "loss": 0.8678, "mean_token_accuracy": 0.8635596036911011, "num_tokens": 264368762.0, "step": 6963 }, { "epoch": 0.8858923801043125, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.77455997467041, "learning_rate": 5e-05, "loss": 1.0131, "mean_token_accuracy": 0.862507164478302, "num_tokens": 264410261.0, "step": 6964 }, { "epoch": 0.8860195903829029, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.811376571655273, "learning_rate": 5e-05, "loss": 0.8159, "mean_token_accuracy": 0.867411196231842, "num_tokens": 264449111.0, "step": 6965 }, { "epoch": 0.8861468006614934, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.094624519348145, "learning_rate": 5e-05, "loss": 1.0657, "mean_token_accuracy": 0.8563542366027832, "num_tokens": 264487045.0, "step": 6966 }, { "epoch": 0.886274010940084, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.135764122009277, "learning_rate": 5e-05, "loss": 0.8736, "mean_token_accuracy": 0.865233302116394, "num_tokens": 264524025.0, "step": 6967 }, { "epoch": 0.8864012212186745, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.995270729064941, "learning_rate": 5e-05, "loss": 1.0657, "mean_token_accuracy": 0.8564320802688599, "num_tokens": 264561361.0, "step": 6968 }, { "epoch": 0.886528431497265, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.601323127746582, "learning_rate": 5e-05, "loss": 1.0106, "mean_token_accuracy": 0.8362069129943848, "num_tokens": 264603557.0, "step": 6969 }, { "epoch": 0.8866556417758555, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.279579162597656, "learning_rate": 5e-05, "loss": 0.941, "mean_token_accuracy": 0.8752967119216919, "num_tokens": 264638539.0, "step": 6970 }, { "epoch": 0.886782852054446, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.521575927734375, "learning_rate": 5e-05, "loss": 0.8996, "mean_token_accuracy": 0.8630005121231079, "num_tokens": 264676592.0, "step": 6971 }, { "epoch": 0.8869100623330365, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.638100624084473, "learning_rate": 5e-05, "loss": 0.8751, "mean_token_accuracy": 0.8734168410301208, "num_tokens": 264708339.0, "step": 6972 }, { "epoch": 0.887037272611627, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.02155876159668, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.8617795705795288, "num_tokens": 264741705.0, "step": 6973 }, { "epoch": 0.8871644828902175, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.550028800964355, "learning_rate": 5e-05, "loss": 0.8856, "mean_token_accuracy": 0.860589325428009, "num_tokens": 264773095.0, "step": 6974 }, { "epoch": 0.8872916931688081, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.448578834533691, "learning_rate": 5e-05, "loss": 0.7359, "mean_token_accuracy": 0.8788487315177917, "num_tokens": 264813499.0, "step": 6975 }, { "epoch": 0.8874189034473986, "ewc_loss": 0.5, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.768099784851074, "learning_rate": 5e-05, "loss": 0.8907, "mean_token_accuracy": 0.8694180250167847, "num_tokens": 264851314.0, "step": 6976 }, { "epoch": 0.887546113725989, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.17664909362793, "learning_rate": 5e-05, "loss": 0.7478, "mean_token_accuracy": 0.8644586205482483, "num_tokens": 264886908.0, "step": 6977 }, { "epoch": 0.8876733240045795, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.979249954223633, "learning_rate": 5e-05, "loss": 1.0204, "mean_token_accuracy": 0.8347418308258057, "num_tokens": 264927516.0, "step": 6978 }, { "epoch": 0.8878005342831701, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.224527359008789, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.8677552342414856, "num_tokens": 264966758.0, "step": 6979 }, { "epoch": 0.8879277445617606, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.016559600830078, "learning_rate": 5e-05, "loss": 0.9769, "mean_token_accuracy": 0.8526250123977661, "num_tokens": 265002267.0, "step": 6980 }, { "epoch": 0.8880549548403511, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.545701026916504, "learning_rate": 5e-05, "loss": 0.7713, "mean_token_accuracy": 0.8701666593551636, "num_tokens": 265034957.0, "step": 6981 }, { "epoch": 0.8881821651189417, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.617804527282715, "learning_rate": 5e-05, "loss": 0.9032, "mean_token_accuracy": 0.8633958101272583, "num_tokens": 265075587.0, "step": 6982 }, { "epoch": 0.8883093753975321, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.832126617431641, "learning_rate": 5e-05, "loss": 0.77, "mean_token_accuracy": 0.8756003379821777, "num_tokens": 265111385.0, "step": 6983 }, { "epoch": 0.8884365856761226, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.98257303237915, "learning_rate": 5e-05, "loss": 0.8284, "mean_token_accuracy": 0.8719466924667358, "num_tokens": 265153674.0, "step": 6984 }, { "epoch": 0.8885637959547131, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.8768391609191895, "learning_rate": 5e-05, "loss": 0.7655, "mean_token_accuracy": 0.8737516403198242, "num_tokens": 265190168.0, "step": 6985 }, { "epoch": 0.8886910062333037, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.643961429595947, "learning_rate": 5e-05, "loss": 0.759, "mean_token_accuracy": 0.8743619918823242, "num_tokens": 265230145.0, "step": 6986 }, { "epoch": 0.8888182165118942, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.770756721496582, "learning_rate": 5e-05, "loss": 0.8169, "mean_token_accuracy": 0.8503327369689941, "num_tokens": 265263229.0, "step": 6987 }, { "epoch": 0.8889454267904847, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.571889877319336, "learning_rate": 5e-05, "loss": 0.788, "mean_token_accuracy": 0.8542367815971375, "num_tokens": 265305426.0, "step": 6988 }, { "epoch": 0.8890726370690751, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.662937164306641, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.8517065644264221, "num_tokens": 265345677.0, "step": 6989 }, { "epoch": 0.8891998473476657, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.520352363586426, "learning_rate": 5e-05, "loss": 0.723, "mean_token_accuracy": 0.8699765801429749, "num_tokens": 265383987.0, "step": 6990 }, { "epoch": 0.8893270576262562, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.6191864013671875, "learning_rate": 5e-05, "loss": 0.814, "mean_token_accuracy": 0.8453444242477417, "num_tokens": 265423209.0, "step": 6991 }, { "epoch": 0.8894542679048467, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.542654037475586, "learning_rate": 5e-05, "loss": 0.7762, "mean_token_accuracy": 0.8524364233016968, "num_tokens": 265463304.0, "step": 6992 }, { "epoch": 0.8895814781834372, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.453062534332275, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8538708686828613, "num_tokens": 265506991.0, "step": 6993 }, { "epoch": 0.8897086884620278, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.740714073181152, "learning_rate": 5e-05, "loss": 0.7742, "mean_token_accuracy": 0.8623775243759155, "num_tokens": 265539050.0, "step": 6994 }, { "epoch": 0.8898358987406182, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004062652587890625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.146864891052246, "learning_rate": 5e-05, "loss": 0.7403, "mean_token_accuracy": 0.8581693172454834, "num_tokens": 265583593.0, "step": 6995 }, { "epoch": 0.8899631090192087, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 8.051199913024902, "learning_rate": 5e-05, "loss": 0.8482, "mean_token_accuracy": 0.8424108624458313, "num_tokens": 265623873.0, "step": 6996 }, { "epoch": 0.8900903192977992, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.839139461517334, "learning_rate": 5e-05, "loss": 0.722, "mean_token_accuracy": 0.8581291437149048, "num_tokens": 265666191.0, "step": 6997 }, { "epoch": 0.8902175295763898, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.39795207977295, "learning_rate": 5e-05, "loss": 0.8189, "mean_token_accuracy": 0.8593378067016602, "num_tokens": 265703929.0, "step": 6998 }, { "epoch": 0.8903447398549803, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.596648216247559, "learning_rate": 5e-05, "loss": 0.6969, "mean_token_accuracy": 0.8637682199478149, "num_tokens": 265737246.0, "step": 6999 }, { "epoch": 0.8904719501335708, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.72574234008789, "learning_rate": 5e-05, "loss": 0.9135, "mean_token_accuracy": 0.852607011795044, "num_tokens": 265774549.0, "step": 7000 }, { "epoch": 0.8905991604121613, "ewc_loss": 0.296875, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.479393482208252, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8545670509338379, "num_tokens": 265806972.0, "step": 7001 }, { "epoch": 0.8907263706907518, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.870691299438477, "learning_rate": 5e-05, "loss": 0.903, "mean_token_accuracy": 0.8543435335159302, "num_tokens": 265844505.0, "step": 7002 }, { "epoch": 0.8908535809693423, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.754090785980225, "learning_rate": 5e-05, "loss": 0.7211, "mean_token_accuracy": 0.8622995615005493, "num_tokens": 265885350.0, "step": 7003 }, { "epoch": 0.8909807912479328, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.485326766967773, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8653520345687866, "num_tokens": 265925345.0, "step": 7004 }, { "epoch": 0.8911080015265234, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.381301403045654, "learning_rate": 5e-05, "loss": 0.7601, "mean_token_accuracy": 0.867332935333252, "num_tokens": 265961329.0, "step": 7005 }, { "epoch": 0.8912352118051139, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.6814470291137695, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8354350328445435, "num_tokens": 266000836.0, "step": 7006 }, { "epoch": 0.8913624220837043, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.94356632232666, "learning_rate": 5e-05, "loss": 0.821, "mean_token_accuracy": 0.8548401594161987, "num_tokens": 266037178.0, "step": 7007 }, { "epoch": 0.8914896323622948, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.101522922515869, "learning_rate": 5e-05, "loss": 0.7149, "mean_token_accuracy": 0.869735062122345, "num_tokens": 266077125.0, "step": 7008 }, { "epoch": 0.8916168426408854, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000408172607421875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.205262184143066, "learning_rate": 5e-05, "loss": 0.8692, "mean_token_accuracy": 0.8483538627624512, "num_tokens": 266117313.0, "step": 7009 }, { "epoch": 0.8917440529194759, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.949919700622559, "learning_rate": 5e-05, "loss": 0.7248, "mean_token_accuracy": 0.8626411557197571, "num_tokens": 266157432.0, "step": 7010 }, { "epoch": 0.8918712631980664, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.219917297363281, "learning_rate": 5e-05, "loss": 0.8239, "mean_token_accuracy": 0.8604799509048462, "num_tokens": 266197170.0, "step": 7011 }, { "epoch": 0.891998473476657, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.973836898803711, "learning_rate": 5e-05, "loss": 0.7512, "mean_token_accuracy": 0.8535593152046204, "num_tokens": 266230383.0, "step": 7012 }, { "epoch": 0.8921256837552475, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.17385482788086, "learning_rate": 5e-05, "loss": 0.8279, "mean_token_accuracy": 0.8624384999275208, "num_tokens": 266268685.0, "step": 7013 }, { "epoch": 0.8922528940338379, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.995352745056152, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8570795655250549, "num_tokens": 266308875.0, "step": 7014 }, { "epoch": 0.8923801043124284, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.176992416381836, "learning_rate": 5e-05, "loss": 0.9176, "mean_token_accuracy": 0.8324311375617981, "num_tokens": 266346540.0, "step": 7015 }, { "epoch": 0.892507314591019, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.983109951019287, "learning_rate": 5e-05, "loss": 0.7334, "mean_token_accuracy": 0.8618952035903931, "num_tokens": 266383359.0, "step": 7016 }, { "epoch": 0.8926345248696095, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.207732200622559, "learning_rate": 5e-05, "loss": 0.859, "mean_token_accuracy": 0.8555048704147339, "num_tokens": 266420235.0, "step": 7017 }, { "epoch": 0.8927617351482, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.927736759185791, "learning_rate": 5e-05, "loss": 0.7437, "mean_token_accuracy": 0.857781708240509, "num_tokens": 266455097.0, "step": 7018 }, { "epoch": 0.8928889454267905, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.284285545349121, "learning_rate": 5e-05, "loss": 0.8283, "mean_token_accuracy": 0.866165041923523, "num_tokens": 266488471.0, "step": 7019 }, { "epoch": 0.893016155705381, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.858638286590576, "learning_rate": 5e-05, "loss": 0.7405, "mean_token_accuracy": 0.8574713468551636, "num_tokens": 266528107.0, "step": 7020 }, { "epoch": 0.8931433659839715, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.389216423034668, "learning_rate": 5e-05, "loss": 0.8262, "mean_token_accuracy": 0.8695287108421326, "num_tokens": 266567859.0, "step": 7021 }, { "epoch": 0.893270576262562, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.777685642242432, "learning_rate": 5e-05, "loss": 0.7098, "mean_token_accuracy": 0.8694583177566528, "num_tokens": 266609722.0, "step": 7022 }, { "epoch": 0.8933977865411525, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.498412132263184, "learning_rate": 5e-05, "loss": 0.8215, "mean_token_accuracy": 0.8733054399490356, "num_tokens": 266650546.0, "step": 7023 }, { "epoch": 0.8935249968197431, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.790472030639648, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.858121395111084, "num_tokens": 266686508.0, "step": 7024 }, { "epoch": 0.8936522070983336, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.456544876098633, "learning_rate": 5e-05, "loss": 0.8243, "mean_token_accuracy": 0.8759891986846924, "num_tokens": 266725528.0, "step": 7025 }, { "epoch": 0.893779417376924, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.954159259796143, "learning_rate": 5e-05, "loss": 0.686, "mean_token_accuracy": 0.8764092326164246, "num_tokens": 266754818.0, "step": 7026 }, { "epoch": 0.8939066276555145, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.237767219543457, "learning_rate": 5e-05, "loss": 0.8694, "mean_token_accuracy": 0.859528660774231, "num_tokens": 266790723.0, "step": 7027 }, { "epoch": 0.8940338379341051, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.200335502624512, "learning_rate": 5e-05, "loss": 0.7508, "mean_token_accuracy": 0.865469753742218, "num_tokens": 266824889.0, "step": 7028 }, { "epoch": 0.8941610482126956, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.933570384979248, "learning_rate": 5e-05, "loss": 0.8342, "mean_token_accuracy": 0.8600009083747864, "num_tokens": 266866129.0, "step": 7029 }, { "epoch": 0.8942882584912861, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.396678924560547, "learning_rate": 5e-05, "loss": 0.7923, "mean_token_accuracy": 0.8575267195701599, "num_tokens": 266910005.0, "step": 7030 }, { "epoch": 0.8944154687698767, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.688445091247559, "learning_rate": 5e-05, "loss": 0.7807, "mean_token_accuracy": 0.8674254417419434, "num_tokens": 266947366.0, "step": 7031 }, { "epoch": 0.8945426790484671, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.476723670959473, "learning_rate": 5e-05, "loss": 0.7787, "mean_token_accuracy": 0.8578122854232788, "num_tokens": 266982402.0, "step": 7032 }, { "epoch": 0.8946698893270576, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.6181135177612305, "learning_rate": 5e-05, "loss": 0.8272, "mean_token_accuracy": 0.8506811261177063, "num_tokens": 267015397.0, "step": 7033 }, { "epoch": 0.8947970996056481, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.391690731048584, "learning_rate": 5e-05, "loss": 0.7077, "mean_token_accuracy": 0.8748189210891724, "num_tokens": 267050204.0, "step": 7034 }, { "epoch": 0.8949243098842387, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.691619396209717, "learning_rate": 5e-05, "loss": 0.7383, "mean_token_accuracy": 0.8781067728996277, "num_tokens": 267085127.0, "step": 7035 }, { "epoch": 0.8950515201628292, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.236510753631592, "learning_rate": 5e-05, "loss": 0.7825, "mean_token_accuracy": 0.8531011939048767, "num_tokens": 267122210.0, "step": 7036 }, { "epoch": 0.8951787304414197, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.825319766998291, "learning_rate": 5e-05, "loss": 0.7904, "mean_token_accuracy": 0.8613693118095398, "num_tokens": 267161124.0, "step": 7037 }, { "epoch": 0.8953059407200101, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.085208415985107, "learning_rate": 5e-05, "loss": 0.7165, "mean_token_accuracy": 0.870858371257782, "num_tokens": 267195339.0, "step": 7038 }, { "epoch": 0.8954331509986007, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.962315082550049, "learning_rate": 5e-05, "loss": 0.8119, "mean_token_accuracy": 0.8571393489837646, "num_tokens": 267228712.0, "step": 7039 }, { "epoch": 0.8955603612771912, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.941223621368408, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.851814329624176, "num_tokens": 267262796.0, "step": 7040 }, { "epoch": 0.8956875715557817, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.105558395385742, "learning_rate": 5e-05, "loss": 0.8269, "mean_token_accuracy": 0.8544296026229858, "num_tokens": 267304101.0, "step": 7041 }, { "epoch": 0.8958147818343722, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.815674304962158, "learning_rate": 5e-05, "loss": 0.7868, "mean_token_accuracy": 0.8463551998138428, "num_tokens": 267348943.0, "step": 7042 }, { "epoch": 0.8959419921129628, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.25328540802002, "learning_rate": 5e-05, "loss": 0.8112, "mean_token_accuracy": 0.8669818639755249, "num_tokens": 267390436.0, "step": 7043 }, { "epoch": 0.8960692023915532, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.724347114562988, "learning_rate": 5e-05, "loss": 0.6928, "mean_token_accuracy": 0.8726309537887573, "num_tokens": 267428069.0, "step": 7044 }, { "epoch": 0.8961964126701437, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.35387897491455, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.8458603620529175, "num_tokens": 267468029.0, "step": 7045 }, { "epoch": 0.8963236229487342, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.771108150482178, "learning_rate": 5e-05, "loss": 0.7634, "mean_token_accuracy": 0.8540019392967224, "num_tokens": 267504013.0, "step": 7046 }, { "epoch": 0.8964508332273248, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.228271484375, "learning_rate": 5e-05, "loss": 0.8082, "mean_token_accuracy": 0.8699555993080139, "num_tokens": 267537696.0, "step": 7047 }, { "epoch": 0.8965780435059153, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.030568599700928, "learning_rate": 5e-05, "loss": 0.7253, "mean_token_accuracy": 0.8691904544830322, "num_tokens": 267568971.0, "step": 7048 }, { "epoch": 0.8967052537845058, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.836986541748047, "learning_rate": 5e-05, "loss": 0.8145, "mean_token_accuracy": 0.8578850030899048, "num_tokens": 267607829.0, "step": 7049 }, { "epoch": 0.8968324640630962, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.482216835021973, "learning_rate": 5e-05, "loss": 0.7601, "mean_token_accuracy": 0.864153265953064, "num_tokens": 267649338.0, "step": 7050 }, { "epoch": 0.8969596743416868, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.226129531860352, "learning_rate": 5e-05, "loss": 0.7149, "mean_token_accuracy": 0.8692975044250488, "num_tokens": 267685053.0, "step": 7051 }, { "epoch": 0.8970868846202773, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004138946533203125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.074511528015137, "learning_rate": 5e-05, "loss": 0.8005, "mean_token_accuracy": 0.8676988482475281, "num_tokens": 267722291.0, "step": 7052 }, { "epoch": 0.8972140948988678, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.548366069793701, "learning_rate": 5e-05, "loss": 0.7079, "mean_token_accuracy": 0.8624071478843689, "num_tokens": 267760533.0, "step": 7053 }, { "epoch": 0.8973413051774584, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.81910228729248, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.8636115789413452, "num_tokens": 267796655.0, "step": 7054 }, { "epoch": 0.8974685154560489, "ewc_loss": 0.265625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00022411346435546875, "grad_norm": 5.906210899353027, "learning_rate": 5e-05, "loss": 0.6664, "mean_token_accuracy": 0.8641801476478577, "num_tokens": 267835950.0, "step": 7055 }, { "epoch": 0.8975957257346393, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.811942100524902, "learning_rate": 5e-05, "loss": 0.9819, "mean_token_accuracy": 0.8641781806945801, "num_tokens": 267871796.0, "step": 7056 }, { "epoch": 0.8977229360132298, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 5.702037811279297, "learning_rate": 5e-05, "loss": 0.6865, "mean_token_accuracy": 0.8647891283035278, "num_tokens": 267910562.0, "step": 7057 }, { "epoch": 0.8978501462918204, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.610870361328125, "learning_rate": 5e-05, "loss": 1.1401, "mean_token_accuracy": 0.8628866672515869, "num_tokens": 267948962.0, "step": 7058 }, { "epoch": 0.8979773565704109, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.343568801879883, "learning_rate": 5e-05, "loss": 1.0062, "mean_token_accuracy": 0.8533802032470703, "num_tokens": 267993158.0, "step": 7059 }, { "epoch": 0.8981045668490014, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.70254898071289, "learning_rate": 5e-05, "loss": 1.0907, "mean_token_accuracy": 0.8528420329093933, "num_tokens": 268033366.0, "step": 7060 }, { "epoch": 0.898231777127592, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 9.040467262268066, "learning_rate": 5e-05, "loss": 1.1454, "mean_token_accuracy": 0.8609158396720886, "num_tokens": 268074125.0, "step": 7061 }, { "epoch": 0.8983589874061825, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.288946628570557, "learning_rate": 5e-05, "loss": 0.8383, "mean_token_accuracy": 0.8580294847488403, "num_tokens": 268112132.0, "step": 7062 }, { "epoch": 0.8984861976847729, "ewc_loss": 0.8125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00077056884765625, "grad_norm": 10.77923583984375, "learning_rate": 5e-05, "loss": 1.2902, "mean_token_accuracy": 0.8403463363647461, "num_tokens": 268148161.0, "step": 7063 }, { "epoch": 0.8986134079633634, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 6.963427543640137, "learning_rate": 5e-05, "loss": 0.8007, "mean_token_accuracy": 0.8875138163566589, "num_tokens": 268184977.0, "step": 7064 }, { "epoch": 0.898740618241954, "ewc_loss": 0.890625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 12.02120590209961, "learning_rate": 5e-05, "loss": 1.3036, "mean_token_accuracy": 0.861045241355896, "num_tokens": 268219355.0, "step": 7065 }, { "epoch": 0.8988678285205445, "ewc_loss": 0.875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000835418701171875, "grad_norm": 9.382610321044922, "learning_rate": 5e-05, "loss": 1.2686, "mean_token_accuracy": 0.8651973009109497, "num_tokens": 268260533.0, "step": 7066 }, { "epoch": 0.898995038799135, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 11.030341148376465, "learning_rate": 5e-05, "loss": 1.1188, "mean_token_accuracy": 0.8573784232139587, "num_tokens": 268294415.0, "step": 7067 }, { "epoch": 0.8991222490777255, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 9.093144416809082, "learning_rate": 5e-05, "loss": 1.1467, "mean_token_accuracy": 0.8689369559288025, "num_tokens": 268331038.0, "step": 7068 }, { "epoch": 0.899249459356316, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 10.600780487060547, "learning_rate": 5e-05, "loss": 1.1733, "mean_token_accuracy": 0.8445192575454712, "num_tokens": 268369438.0, "step": 7069 }, { "epoch": 0.8993766696349065, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.320280075073242, "learning_rate": 5e-05, "loss": 0.982, "mean_token_accuracy": 0.8631857633590698, "num_tokens": 268403808.0, "step": 7070 }, { "epoch": 0.899503879913497, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.385570526123047, "learning_rate": 5e-05, "loss": 1.0887, "mean_token_accuracy": 0.8728964328765869, "num_tokens": 268442575.0, "step": 7071 }, { "epoch": 0.8996310901920875, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.055760383605957, "learning_rate": 5e-05, "loss": 0.9644, "mean_token_accuracy": 0.8596505522727966, "num_tokens": 268475387.0, "step": 7072 }, { "epoch": 0.8997583004706781, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.532559394836426, "learning_rate": 5e-05, "loss": 1.153, "mean_token_accuracy": 0.8598175048828125, "num_tokens": 268513514.0, "step": 7073 }, { "epoch": 0.8998855107492686, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.38078784942627, "learning_rate": 5e-05, "loss": 0.9941, "mean_token_accuracy": 0.8655864000320435, "num_tokens": 268553690.0, "step": 7074 }, { "epoch": 0.900012721027859, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.407835006713867, "learning_rate": 5e-05, "loss": 1.0865, "mean_token_accuracy": 0.870038628578186, "num_tokens": 268593285.0, "step": 7075 }, { "epoch": 0.9001399313064495, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.624627113342285, "learning_rate": 5e-05, "loss": 1.0234, "mean_token_accuracy": 0.8629289269447327, "num_tokens": 268631929.0, "step": 7076 }, { "epoch": 0.9002671415850401, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.016914367675781, "learning_rate": 5e-05, "loss": 1.1619, "mean_token_accuracy": 0.8355549573898315, "num_tokens": 268669882.0, "step": 7077 }, { "epoch": 0.9003943518636306, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.54384994506836, "learning_rate": 5e-05, "loss": 1.0424, "mean_token_accuracy": 0.8474581837654114, "num_tokens": 268712709.0, "step": 7078 }, { "epoch": 0.9005215621422211, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.687257766723633, "learning_rate": 5e-05, "loss": 1.0524, "mean_token_accuracy": 0.8579497337341309, "num_tokens": 268751514.0, "step": 7079 }, { "epoch": 0.9006487724208116, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.332297325134277, "learning_rate": 5e-05, "loss": 1.0018, "mean_token_accuracy": 0.8475689888000488, "num_tokens": 268796841.0, "step": 7080 }, { "epoch": 0.9007759826994021, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.472898483276367, "learning_rate": 5e-05, "loss": 1.005, "mean_token_accuracy": 0.870879054069519, "num_tokens": 268841536.0, "step": 7081 }, { "epoch": 0.9009031929779926, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.114363670349121, "learning_rate": 5e-05, "loss": 0.9204, "mean_token_accuracy": 0.862259566783905, "num_tokens": 268884636.0, "step": 7082 }, { "epoch": 0.9010304032565831, "ewc_loss": 0.625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.431196212768555, "learning_rate": 5e-05, "loss": 1.0404, "mean_token_accuracy": 0.8588762879371643, "num_tokens": 268924114.0, "step": 7083 }, { "epoch": 0.9011576135351737, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.944991111755371, "learning_rate": 5e-05, "loss": 0.8888, "mean_token_accuracy": 0.86992347240448, "num_tokens": 268966476.0, "step": 7084 }, { "epoch": 0.9012848238137642, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.503260612487793, "learning_rate": 5e-05, "loss": 1.0217, "mean_token_accuracy": 0.8721951246261597, "num_tokens": 268998224.0, "step": 7085 }, { "epoch": 0.9014120340923547, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.984644889831543, "learning_rate": 5e-05, "loss": 0.9443, "mean_token_accuracy": 0.856381893157959, "num_tokens": 269043552.0, "step": 7086 }, { "epoch": 0.9015392443709451, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.406671524047852, "learning_rate": 5e-05, "loss": 1.0914, "mean_token_accuracy": 0.8495476245880127, "num_tokens": 269085180.0, "step": 7087 }, { "epoch": 0.9016664546495357, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.1434965133667, "learning_rate": 5e-05, "loss": 0.9719, "mean_token_accuracy": 0.8505439758300781, "num_tokens": 269121694.0, "step": 7088 }, { "epoch": 0.9017936649281262, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.028314590454102, "learning_rate": 5e-05, "loss": 1.0811, "mean_token_accuracy": 0.8446435332298279, "num_tokens": 269164896.0, "step": 7089 }, { "epoch": 0.9019208752067167, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.049649238586426, "learning_rate": 5e-05, "loss": 0.8375, "mean_token_accuracy": 0.8834295272827148, "num_tokens": 269201095.0, "step": 7090 }, { "epoch": 0.9020480854853072, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.813665390014648, "learning_rate": 5e-05, "loss": 0.9803, "mean_token_accuracy": 0.8658645153045654, "num_tokens": 269243055.0, "step": 7091 }, { "epoch": 0.9021752957638978, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.777564525604248, "learning_rate": 5e-05, "loss": 0.8472, "mean_token_accuracy": 0.8720241785049438, "num_tokens": 269279902.0, "step": 7092 }, { "epoch": 0.9023025060424882, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.818109512329102, "learning_rate": 5e-05, "loss": 0.9444, "mean_token_accuracy": 0.876625657081604, "num_tokens": 269315773.0, "step": 7093 }, { "epoch": 0.9024297163210787, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.591517925262451, "learning_rate": 5e-05, "loss": 0.9096, "mean_token_accuracy": 0.844498872756958, "num_tokens": 269353127.0, "step": 7094 }, { "epoch": 0.9025569265996692, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.869869232177734, "learning_rate": 5e-05, "loss": 0.9913, "mean_token_accuracy": 0.8640021085739136, "num_tokens": 269387023.0, "step": 7095 }, { "epoch": 0.9026841368782598, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.580545425415039, "learning_rate": 5e-05, "loss": 0.9191, "mean_token_accuracy": 0.843269407749176, "num_tokens": 269419983.0, "step": 7096 }, { "epoch": 0.9028113471568503, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.774523735046387, "learning_rate": 5e-05, "loss": 0.9738, "mean_token_accuracy": 0.8679677248001099, "num_tokens": 269460580.0, "step": 7097 }, { "epoch": 0.9029385574354408, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.650783538818359, "learning_rate": 5e-05, "loss": 0.8097, "mean_token_accuracy": 0.8780392408370972, "num_tokens": 269497711.0, "step": 7098 }, { "epoch": 0.9030657677140312, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.566529273986816, "learning_rate": 5e-05, "loss": 0.9653, "mean_token_accuracy": 0.8636635541915894, "num_tokens": 269538111.0, "step": 7099 }, { "epoch": 0.9031929779926218, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.653501510620117, "learning_rate": 5e-05, "loss": 0.7901, "mean_token_accuracy": 0.88128662109375, "num_tokens": 269570695.0, "step": 7100 }, { "epoch": 0.9033201882712123, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.357890129089355, "learning_rate": 5e-05, "loss": 0.9255, "mean_token_accuracy": 0.8644344806671143, "num_tokens": 269605849.0, "step": 7101 }, { "epoch": 0.9034473985498028, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.577649116516113, "learning_rate": 5e-05, "loss": 0.8376, "mean_token_accuracy": 0.8590700626373291, "num_tokens": 269646445.0, "step": 7102 }, { "epoch": 0.9035746088283934, "ewc_loss": 0.5, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.247220039367676, "learning_rate": 5e-05, "loss": 0.9563, "mean_token_accuracy": 0.8474600315093994, "num_tokens": 269686806.0, "step": 7103 }, { "epoch": 0.9037018191069839, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.45611572265625, "learning_rate": 5e-05, "loss": 0.8419, "mean_token_accuracy": 0.8538169860839844, "num_tokens": 269727311.0, "step": 7104 }, { "epoch": 0.9038290293855743, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.196086883544922, "learning_rate": 5e-05, "loss": 0.9696, "mean_token_accuracy": 0.8432912826538086, "num_tokens": 269765066.0, "step": 7105 }, { "epoch": 0.9039562396641648, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.371143817901611, "learning_rate": 5e-05, "loss": 0.8215, "mean_token_accuracy": 0.8544254302978516, "num_tokens": 269807065.0, "step": 7106 }, { "epoch": 0.9040834499427554, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.157865524291992, "learning_rate": 5e-05, "loss": 0.8359, "mean_token_accuracy": 0.8771320581436157, "num_tokens": 269839874.0, "step": 7107 }, { "epoch": 0.9042106602213459, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.311919689178467, "learning_rate": 5e-05, "loss": 0.7449, "mean_token_accuracy": 0.8709813356399536, "num_tokens": 269877647.0, "step": 7108 }, { "epoch": 0.9043378704999364, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.129886627197266, "learning_rate": 5e-05, "loss": 0.8623, "mean_token_accuracy": 0.8617303371429443, "num_tokens": 269912403.0, "step": 7109 }, { "epoch": 0.9044650807785269, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.309436321258545, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8705794811248779, "num_tokens": 269946086.0, "step": 7110 }, { "epoch": 0.9045922910571175, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.021307945251465, "learning_rate": 5e-05, "loss": 0.8126, "mean_token_accuracy": 0.8739931583404541, "num_tokens": 269985820.0, "step": 7111 }, { "epoch": 0.9047195013357079, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.3553032875061035, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8713462352752686, "num_tokens": 270021377.0, "step": 7112 }, { "epoch": 0.9048467116142984, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.888199806213379, "learning_rate": 5e-05, "loss": 0.8826, "mean_token_accuracy": 0.8479881882667542, "num_tokens": 270056928.0, "step": 7113 }, { "epoch": 0.9049739218928889, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.403358459472656, "learning_rate": 5e-05, "loss": 0.7367, "mean_token_accuracy": 0.8721884489059448, "num_tokens": 270093720.0, "step": 7114 }, { "epoch": 0.9051011321714795, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.7498369216918945, "learning_rate": 5e-05, "loss": 0.8342, "mean_token_accuracy": 0.8539389371871948, "num_tokens": 270140151.0, "step": 7115 }, { "epoch": 0.90522834245007, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.467957496643066, "learning_rate": 5e-05, "loss": 0.7658, "mean_token_accuracy": 0.8655073642730713, "num_tokens": 270173970.0, "step": 7116 }, { "epoch": 0.9053555527286605, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.634862899780273, "learning_rate": 5e-05, "loss": 0.7662, "mean_token_accuracy": 0.8647282123565674, "num_tokens": 270207014.0, "step": 7117 }, { "epoch": 0.905482763007251, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.51257848739624, "learning_rate": 5e-05, "loss": 0.8232, "mean_token_accuracy": 0.846221387386322, "num_tokens": 270246635.0, "step": 7118 }, { "epoch": 0.9056099732858415, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.536546230316162, "learning_rate": 5e-05, "loss": 0.7768, "mean_token_accuracy": 0.8592386245727539, "num_tokens": 270289902.0, "step": 7119 }, { "epoch": 0.905737183564432, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.554168224334717, "learning_rate": 5e-05, "loss": 0.8002, "mean_token_accuracy": 0.8487120866775513, "num_tokens": 270321173.0, "step": 7120 }, { "epoch": 0.9058643938430225, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.437610149383545, "learning_rate": 5e-05, "loss": 0.7235, "mean_token_accuracy": 0.8718621134757996, "num_tokens": 270360281.0, "step": 7121 }, { "epoch": 0.9059916041216131, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.598074436187744, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8676586151123047, "num_tokens": 270396609.0, "step": 7122 }, { "epoch": 0.9061188144002036, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.362064361572266, "learning_rate": 5e-05, "loss": 0.7911, "mean_token_accuracy": 0.85060715675354, "num_tokens": 270436365.0, "step": 7123 }, { "epoch": 0.906246024678794, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.624556541442871, "learning_rate": 5e-05, "loss": 0.7437, "mean_token_accuracy": 0.866881251335144, "num_tokens": 270471314.0, "step": 7124 }, { "epoch": 0.9063732349573845, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.329899787902832, "learning_rate": 5e-05, "loss": 0.7725, "mean_token_accuracy": 0.8525584936141968, "num_tokens": 270510169.0, "step": 7125 }, { "epoch": 0.9065004452359751, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.647222518920898, "learning_rate": 5e-05, "loss": 0.7531, "mean_token_accuracy": 0.8688243627548218, "num_tokens": 270549787.0, "step": 7126 }, { "epoch": 0.9066276555145656, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.266806125640869, "learning_rate": 5e-05, "loss": 0.7038, "mean_token_accuracy": 0.87060546875, "num_tokens": 270584059.0, "step": 7127 }, { "epoch": 0.9067548657931561, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.680056095123291, "learning_rate": 5e-05, "loss": 0.7394, "mean_token_accuracy": 0.8688617944717407, "num_tokens": 270618059.0, "step": 7128 }, { "epoch": 0.9068820760717466, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.231241703033447, "learning_rate": 5e-05, "loss": 0.721, "mean_token_accuracy": 0.8632387518882751, "num_tokens": 270656968.0, "step": 7129 }, { "epoch": 0.9070092863503371, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.724623203277588, "learning_rate": 5e-05, "loss": 0.7629, "mean_token_accuracy": 0.8617513179779053, "num_tokens": 270692088.0, "step": 7130 }, { "epoch": 0.9071364966289276, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.170334339141846, "learning_rate": 5e-05, "loss": 0.7128, "mean_token_accuracy": 0.8679138422012329, "num_tokens": 270729480.0, "step": 7131 }, { "epoch": 0.9072637069075181, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.779603958129883, "learning_rate": 5e-05, "loss": 0.7293, "mean_token_accuracy": 0.8708552122116089, "num_tokens": 270769062.0, "step": 7132 }, { "epoch": 0.9073909171861086, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 7.064017295837402, "learning_rate": 5e-05, "loss": 0.7222, "mean_token_accuracy": 0.857884407043457, "num_tokens": 270808142.0, "step": 7133 }, { "epoch": 0.9075181274646992, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.943783760070801, "learning_rate": 5e-05, "loss": 0.7876, "mean_token_accuracy": 0.8572298884391785, "num_tokens": 270842232.0, "step": 7134 }, { "epoch": 0.9076453377432897, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.861843109130859, "learning_rate": 5e-05, "loss": 0.7474, "mean_token_accuracy": 0.8508480191230774, "num_tokens": 270881588.0, "step": 7135 }, { "epoch": 0.9077725480218801, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.223543167114258, "learning_rate": 5e-05, "loss": 0.8225, "mean_token_accuracy": 0.8570501208305359, "num_tokens": 270921122.0, "step": 7136 }, { "epoch": 0.9078997583004706, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00024318695068359375, "grad_norm": 6.533204078674316, "learning_rate": 5e-05, "loss": 0.7113, "mean_token_accuracy": 0.8582513332366943, "num_tokens": 270959037.0, "step": 7137 }, { "epoch": 0.9080269685790612, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.725468635559082, "learning_rate": 5e-05, "loss": 0.8468, "mean_token_accuracy": 0.8637259602546692, "num_tokens": 271002731.0, "step": 7138 }, { "epoch": 0.9081541788576517, "ewc_loss": 0.265625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00022411346435546875, "grad_norm": 6.069458961486816, "learning_rate": 5e-05, "loss": 0.6595, "mean_token_accuracy": 0.8654141426086426, "num_tokens": 271038487.0, "step": 7139 }, { "epoch": 0.9082813891362422, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.457048416137695, "learning_rate": 5e-05, "loss": 0.9475, "mean_token_accuracy": 0.8588577508926392, "num_tokens": 271068811.0, "step": 7140 }, { "epoch": 0.9084085994148328, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 5.833693504333496, "learning_rate": 5e-05, "loss": 0.6647, "mean_token_accuracy": 0.8747340440750122, "num_tokens": 271111019.0, "step": 7141 }, { "epoch": 0.9085358096934232, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.980447769165039, "learning_rate": 5e-05, "loss": 1.033, "mean_token_accuracy": 0.8597720861434937, "num_tokens": 271149317.0, "step": 7142 }, { "epoch": 0.9086630199720137, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 6.787596225738525, "learning_rate": 5e-05, "loss": 0.7626, "mean_token_accuracy": 0.8797544836997986, "num_tokens": 271190664.0, "step": 7143 }, { "epoch": 0.9087902302506042, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.148052215576172, "learning_rate": 5e-05, "loss": 0.981, "mean_token_accuracy": 0.8673335313796997, "num_tokens": 271225071.0, "step": 7144 }, { "epoch": 0.9089174405291948, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.447093963623047, "learning_rate": 5e-05, "loss": 1.0022, "mean_token_accuracy": 0.862541675567627, "num_tokens": 271264228.0, "step": 7145 }, { "epoch": 0.9090446508077853, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.822785377502441, "learning_rate": 5e-05, "loss": 0.755, "mean_token_accuracy": 0.8630689978599548, "num_tokens": 271300888.0, "step": 7146 }, { "epoch": 0.9091718610863758, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.455448150634766, "learning_rate": 5e-05, "loss": 1.0923, "mean_token_accuracy": 0.8627967834472656, "num_tokens": 271339629.0, "step": 7147 }, { "epoch": 0.9092990713649662, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 5.865515232086182, "learning_rate": 5e-05, "loss": 0.7363, "mean_token_accuracy": 0.8635481595993042, "num_tokens": 271379116.0, "step": 7148 }, { "epoch": 0.9094262816435568, "ewc_loss": 1.015625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0009765625, "grad_norm": 12.633549690246582, "learning_rate": 5e-05, "loss": 1.4617, "mean_token_accuracy": 0.8487977981567383, "num_tokens": 271420829.0, "step": 7149 }, { "epoch": 0.9095534919221473, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 9.505148887634277, "learning_rate": 5e-05, "loss": 1.3062, "mean_token_accuracy": 0.8595341444015503, "num_tokens": 271457934.0, "step": 7150 }, { "epoch": 0.9096807022007378, "ewc_loss": 0.734375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 11.425995826721191, "learning_rate": 5e-05, "loss": 1.1408, "mean_token_accuracy": 0.8626996278762817, "num_tokens": 271497441.0, "step": 7151 }, { "epoch": 0.9098079124793284, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 9.868341445922852, "learning_rate": 5e-05, "loss": 1.3131, "mean_token_accuracy": 0.855657696723938, "num_tokens": 271532358.0, "step": 7152 }, { "epoch": 0.9099351227579189, "ewc_loss": 0.71875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.675371170043945, "learning_rate": 5e-05, "loss": 1.1096, "mean_token_accuracy": 0.8668867945671082, "num_tokens": 271570103.0, "step": 7153 }, { "epoch": 0.9100623330365093, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.995038986206055, "learning_rate": 5e-05, "loss": 1.042, "mean_token_accuracy": 0.8662897944450378, "num_tokens": 271604542.0, "step": 7154 }, { "epoch": 0.9101895433150998, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.845586776733398, "learning_rate": 5e-05, "loss": 1.123, "mean_token_accuracy": 0.8522140383720398, "num_tokens": 271646901.0, "step": 7155 }, { "epoch": 0.9103167535936904, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.494512557983398, "learning_rate": 5e-05, "loss": 1.0218, "mean_token_accuracy": 0.8446617126464844, "num_tokens": 271684922.0, "step": 7156 }, { "epoch": 0.9104439638722809, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.463770866394043, "learning_rate": 5e-05, "loss": 0.9736, "mean_token_accuracy": 0.8775036334991455, "num_tokens": 271721206.0, "step": 7157 }, { "epoch": 0.9105711741508714, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.040487289428711, "learning_rate": 5e-05, "loss": 0.8948, "mean_token_accuracy": 0.8735141754150391, "num_tokens": 271757978.0, "step": 7158 }, { "epoch": 0.9106983844294619, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.37348461151123, "learning_rate": 5e-05, "loss": 1.0895, "mean_token_accuracy": 0.8353722095489502, "num_tokens": 271792208.0, "step": 7159 }, { "epoch": 0.9108255947080524, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.864867210388184, "learning_rate": 5e-05, "loss": 0.9357, "mean_token_accuracy": 0.8543397188186646, "num_tokens": 271828853.0, "step": 7160 }, { "epoch": 0.9109528049866429, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.225775718688965, "learning_rate": 5e-05, "loss": 1.0392, "mean_token_accuracy": 0.8465126156806946, "num_tokens": 271868596.0, "step": 7161 }, { "epoch": 0.9110800152652334, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.624307632446289, "learning_rate": 5e-05, "loss": 0.8608, "mean_token_accuracy": 0.8678649067878723, "num_tokens": 271911292.0, "step": 7162 }, { "epoch": 0.9112072255438239, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.41206169128418, "learning_rate": 5e-05, "loss": 1.0021, "mean_token_accuracy": 0.8694352507591248, "num_tokens": 271954512.0, "step": 7163 }, { "epoch": 0.9113344358224145, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.319340705871582, "learning_rate": 5e-05, "loss": 0.8591, "mean_token_accuracy": 0.8579420447349548, "num_tokens": 271994229.0, "step": 7164 }, { "epoch": 0.911461646101005, "ewc_loss": 0.6875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 9.990851402282715, "learning_rate": 5e-05, "loss": 1.0735, "mean_token_accuracy": 0.8690265417098999, "num_tokens": 272031712.0, "step": 7165 }, { "epoch": 0.9115888563795955, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.670089244842529, "learning_rate": 5e-05, "loss": 0.9206, "mean_token_accuracy": 0.8561123609542847, "num_tokens": 272070269.0, "step": 7166 }, { "epoch": 0.9117160666581859, "ewc_loss": 0.71875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.09136962890625, "learning_rate": 5e-05, "loss": 1.1254, "mean_token_accuracy": 0.8621283769607544, "num_tokens": 272115407.0, "step": 7167 }, { "epoch": 0.9118432769367765, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.55506706237793, "learning_rate": 5e-05, "loss": 1.0275, "mean_token_accuracy": 0.8620274662971497, "num_tokens": 272161463.0, "step": 7168 }, { "epoch": 0.911970487215367, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.060171127319336, "learning_rate": 5e-05, "loss": 0.9952, "mean_token_accuracy": 0.8692135810852051, "num_tokens": 272196712.0, "step": 7169 }, { "epoch": 0.9120976974939575, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.741922378540039, "learning_rate": 5e-05, "loss": 1.0327, "mean_token_accuracy": 0.8565100431442261, "num_tokens": 272235617.0, "step": 7170 }, { "epoch": 0.912224907772548, "ewc_loss": 0.5, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.998439311981201, "learning_rate": 5e-05, "loss": 0.8898, "mean_token_accuracy": 0.866346001625061, "num_tokens": 272266464.0, "step": 7171 }, { "epoch": 0.9123521180511386, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.423595428466797, "learning_rate": 5e-05, "loss": 0.96, "mean_token_accuracy": 0.8490580320358276, "num_tokens": 272301069.0, "step": 7172 }, { "epoch": 0.912479328329729, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.743448734283447, "learning_rate": 5e-05, "loss": 0.9166, "mean_token_accuracy": 0.8460869789123535, "num_tokens": 272333119.0, "step": 7173 }, { "epoch": 0.9126065386083195, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.729429721832275, "learning_rate": 5e-05, "loss": 0.8571, "mean_token_accuracy": 0.8502094745635986, "num_tokens": 272365560.0, "step": 7174 }, { "epoch": 0.9127337488869101, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0004100799560546875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.174959182739258, "learning_rate": 5e-05, "loss": 0.8147, "mean_token_accuracy": 0.8801394104957581, "num_tokens": 272401290.0, "step": 7175 }, { "epoch": 0.9128609591655006, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.038476467132568, "learning_rate": 5e-05, "loss": 0.7522, "mean_token_accuracy": 0.8675683736801147, "num_tokens": 272442925.0, "step": 7176 }, { "epoch": 0.9129881694440911, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.676959991455078, "learning_rate": 5e-05, "loss": 0.9358, "mean_token_accuracy": 0.8565473556518555, "num_tokens": 272477192.0, "step": 7177 }, { "epoch": 0.9131153797226816, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.73839807510376, "learning_rate": 5e-05, "loss": 0.7944, "mean_token_accuracy": 0.8449852466583252, "num_tokens": 272506987.0, "step": 7178 }, { "epoch": 0.9132425900012721, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.099573135375977, "learning_rate": 5e-05, "loss": 0.964, "mean_token_accuracy": 0.8667104244232178, "num_tokens": 272544399.0, "step": 7179 }, { "epoch": 0.9133698002798626, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.7800750732421875, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8617385625839233, "num_tokens": 272577639.0, "step": 7180 }, { "epoch": 0.9134970105584531, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.3903169631958, "learning_rate": 5e-05, "loss": 1.0072, "mean_token_accuracy": 0.8663442134857178, "num_tokens": 272615358.0, "step": 7181 }, { "epoch": 0.9136242208370436, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0004119873046875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.329065799713135, "learning_rate": 5e-05, "loss": 0.8653, "mean_token_accuracy": 0.8572676777839661, "num_tokens": 272652927.0, "step": 7182 }, { "epoch": 0.9137514311156342, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0004138946533203125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.046713829040527, "learning_rate": 5e-05, "loss": 1.0297, "mean_token_accuracy": 0.8573070168495178, "num_tokens": 272682509.0, "step": 7183 }, { "epoch": 0.9138786413942247, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.044073104858398, "learning_rate": 5e-05, "loss": 0.9831, "mean_token_accuracy": 0.8462182283401489, "num_tokens": 272723295.0, "step": 7184 }, { "epoch": 0.9140058516728151, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.015088081359863, "learning_rate": 5e-05, "loss": 0.9351, "mean_token_accuracy": 0.848563551902771, "num_tokens": 272764985.0, "step": 7185 }, { "epoch": 0.9141330619514056, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000415802001953125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.319454193115234, "learning_rate": 5e-05, "loss": 0.9197, "mean_token_accuracy": 0.8628289699554443, "num_tokens": 272807926.0, "step": 7186 }, { "epoch": 0.9142602722299962, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0004177093505859375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.330871105194092, "learning_rate": 5e-05, "loss": 0.7839, "mean_token_accuracy": 0.8658217191696167, "num_tokens": 272844249.0, "step": 7187 }, { "epoch": 0.9143874825085867, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.22762393951416, "learning_rate": 5e-05, "loss": 0.8787, "mean_token_accuracy": 0.8622271418571472, "num_tokens": 272878029.0, "step": 7188 }, { "epoch": 0.9145146927871772, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.251616954803467, "learning_rate": 5e-05, "loss": 0.8127, "mean_token_accuracy": 0.8500123023986816, "num_tokens": 272913185.0, "step": 7189 }, { "epoch": 0.9146419030657678, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.143887519836426, "learning_rate": 5e-05, "loss": 0.8439, "mean_token_accuracy": 0.8651756048202515, "num_tokens": 272946835.0, "step": 7190 }, { "epoch": 0.9147691133443582, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.112203121185303, "learning_rate": 5e-05, "loss": 0.7789, "mean_token_accuracy": 0.8547625541687012, "num_tokens": 272983467.0, "step": 7191 }, { "epoch": 0.9148963236229487, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.324844360351562, "learning_rate": 5e-05, "loss": 0.8683, "mean_token_accuracy": 0.8627419471740723, "num_tokens": 273022464.0, "step": 7192 }, { "epoch": 0.9150235339015392, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.911717414855957, "learning_rate": 5e-05, "loss": 0.7974, "mean_token_accuracy": 0.8448387384414673, "num_tokens": 273059746.0, "step": 7193 }, { "epoch": 0.9151507441801298, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.487592697143555, "learning_rate": 5e-05, "loss": 0.8667, "mean_token_accuracy": 0.8728454113006592, "num_tokens": 273100815.0, "step": 7194 }, { "epoch": 0.9152779544587203, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.906601428985596, "learning_rate": 5e-05, "loss": 0.7815, "mean_token_accuracy": 0.8515028953552246, "num_tokens": 273139429.0, "step": 7195 }, { "epoch": 0.9154051647373108, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.425664901733398, "learning_rate": 5e-05, "loss": 0.8832, "mean_token_accuracy": 0.8685179948806763, "num_tokens": 273174159.0, "step": 7196 }, { "epoch": 0.9155323750159012, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.068970680236816, "learning_rate": 5e-05, "loss": 0.7306, "mean_token_accuracy": 0.8703025579452515, "num_tokens": 273211434.0, "step": 7197 }, { "epoch": 0.9156595852944918, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.221878051757812, "learning_rate": 5e-05, "loss": 0.8973, "mean_token_accuracy": 0.8536176681518555, "num_tokens": 273243463.0, "step": 7198 }, { "epoch": 0.9157867955730823, "ewc_loss": 0.375, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.2346038818359375, "learning_rate": 5e-05, "loss": 0.7699, "mean_token_accuracy": 0.8644855618476868, "num_tokens": 273283515.0, "step": 7199 }, { "epoch": 0.9159140058516728, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.303536415100098, "learning_rate": 5e-05, "loss": 0.9291, "mean_token_accuracy": 0.8541198968887329, "num_tokens": 273322927.0, "step": 7200 }, { "epoch": 0.9160412161302633, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.940301418304443, "learning_rate": 5e-05, "loss": 0.7464, "mean_token_accuracy": 0.8694368004798889, "num_tokens": 273359157.0, "step": 7201 }, { "epoch": 0.9161684264088539, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.470056533813477, "learning_rate": 5e-05, "loss": 0.8922, "mean_token_accuracy": 0.8640679121017456, "num_tokens": 273394385.0, "step": 7202 }, { "epoch": 0.9162956366874443, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.9794511795043945, "learning_rate": 5e-05, "loss": 0.7723, "mean_token_accuracy": 0.8536988496780396, "num_tokens": 273430286.0, "step": 7203 }, { "epoch": 0.9164228469660348, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.350919723510742, "learning_rate": 5e-05, "loss": 0.8595, "mean_token_accuracy": 0.8701356053352356, "num_tokens": 273467821.0, "step": 7204 }, { "epoch": 0.9165500572446253, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.139527320861816, "learning_rate": 5e-05, "loss": 0.8101, "mean_token_accuracy": 0.8547079563140869, "num_tokens": 273507659.0, "step": 7205 }, { "epoch": 0.9166772675232159, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0004367828369140625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.177149772644043, "learning_rate": 5e-05, "loss": 0.9186, "mean_token_accuracy": 0.8475750088691711, "num_tokens": 273546445.0, "step": 7206 }, { "epoch": 0.9168044778018064, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0004367828369140625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.197305679321289, "learning_rate": 5e-05, "loss": 0.8009, "mean_token_accuracy": 0.8548377752304077, "num_tokens": 273580354.0, "step": 7207 }, { "epoch": 0.9169316880803969, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.154870986938477, "learning_rate": 5e-05, "loss": 0.8846, "mean_token_accuracy": 0.8583102226257324, "num_tokens": 273619194.0, "step": 7208 }, { "epoch": 0.9170588983589874, "ewc_loss": 0.375, "ewc_loss_diag": 0.0004405975341796875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.162479400634766, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8675573468208313, "num_tokens": 273662888.0, "step": 7209 }, { "epoch": 0.9171861086375779, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.165946006774902, "learning_rate": 5e-05, "loss": 0.8499, "mean_token_accuracy": 0.8666415214538574, "num_tokens": 273697823.0, "step": 7210 }, { "epoch": 0.9173133189161684, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.127965927124023, "learning_rate": 5e-05, "loss": 0.818, "mean_token_accuracy": 0.8515533208847046, "num_tokens": 273737541.0, "step": 7211 }, { "epoch": 0.9174405291947589, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.160364151000977, "learning_rate": 5e-05, "loss": 0.8959, "mean_token_accuracy": 0.853806734085083, "num_tokens": 273777565.0, "step": 7212 }, { "epoch": 0.9175677394733495, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.1655592918396, "learning_rate": 5e-05, "loss": 0.8166, "mean_token_accuracy": 0.8494862914085388, "num_tokens": 273814269.0, "step": 7213 }, { "epoch": 0.91769494975194, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.060978889465332, "learning_rate": 5e-05, "loss": 0.8578, "mean_token_accuracy": 0.8634519577026367, "num_tokens": 273855828.0, "step": 7214 }, { "epoch": 0.9178221600305305, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.258503437042236, "learning_rate": 5e-05, "loss": 0.7906, "mean_token_accuracy": 0.8596778512001038, "num_tokens": 273887514.0, "step": 7215 }, { "epoch": 0.9179493703091209, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.8946990966796875, "learning_rate": 5e-05, "loss": 0.862, "mean_token_accuracy": 0.8504924774169922, "num_tokens": 273923134.0, "step": 7216 }, { "epoch": 0.9180765805877115, "ewc_loss": 0.375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.368885040283203, "learning_rate": 5e-05, "loss": 0.759, "mean_token_accuracy": 0.8660562038421631, "num_tokens": 273962089.0, "step": 7217 }, { "epoch": 0.918203790866302, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.71279239654541, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.8626642823219299, "num_tokens": 273997150.0, "step": 7218 }, { "epoch": 0.9183310011448925, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0004405975341796875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.469192028045654, "learning_rate": 5e-05, "loss": 0.7282, "mean_token_accuracy": 0.8770008087158203, "num_tokens": 274032144.0, "step": 7219 }, { "epoch": 0.918458211423483, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.589728832244873, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8766664862632751, "num_tokens": 274063344.0, "step": 7220 }, { "epoch": 0.9185854217020736, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.515343189239502, "learning_rate": 5e-05, "loss": 0.7653, "mean_token_accuracy": 0.8675292730331421, "num_tokens": 274102242.0, "step": 7221 }, { "epoch": 0.918712631980664, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.498476982116699, "learning_rate": 5e-05, "loss": 0.7721, "mean_token_accuracy": 0.8621324896812439, "num_tokens": 274146560.0, "step": 7222 }, { "epoch": 0.9188398422592545, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.5534348487854, "learning_rate": 5e-05, "loss": 0.8019, "mean_token_accuracy": 0.8513388633728027, "num_tokens": 274180793.0, "step": 7223 }, { "epoch": 0.918967052537845, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.408261299133301, "learning_rate": 5e-05, "loss": 0.7491, "mean_token_accuracy": 0.865068793296814, "num_tokens": 274221207.0, "step": 7224 }, { "epoch": 0.9190942628164356, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.5988054275512695, "learning_rate": 5e-05, "loss": 0.7953, "mean_token_accuracy": 0.855314314365387, "num_tokens": 274265413.0, "step": 7225 }, { "epoch": 0.9192214730950261, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.336186408996582, "learning_rate": 5e-05, "loss": 0.8272, "mean_token_accuracy": 0.836871862411499, "num_tokens": 274305875.0, "step": 7226 }, { "epoch": 0.9193486833736166, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.631725788116455, "learning_rate": 5e-05, "loss": 0.7641, "mean_token_accuracy": 0.8579695224761963, "num_tokens": 274344223.0, "step": 7227 }, { "epoch": 0.919475893652207, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.3004469871521, "learning_rate": 5e-05, "loss": 0.7639, "mean_token_accuracy": 0.8591689467430115, "num_tokens": 274382671.0, "step": 7228 }, { "epoch": 0.9196031039307976, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.650835037231445, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8615595102310181, "num_tokens": 274413475.0, "step": 7229 }, { "epoch": 0.9197303142093881, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0004367828369140625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.23308801651001, "learning_rate": 5e-05, "loss": 0.6824, "mean_token_accuracy": 0.8778030872344971, "num_tokens": 274453714.0, "step": 7230 }, { "epoch": 0.9198575244879786, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004367828369140625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.721582889556885, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.8562936782836914, "num_tokens": 274490388.0, "step": 7231 }, { "epoch": 0.9199847347665692, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 7.116195201873779, "learning_rate": 5e-05, "loss": 0.7068, "mean_token_accuracy": 0.8655902147293091, "num_tokens": 274529939.0, "step": 7232 }, { "epoch": 0.9201119450451597, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.8655900955200195, "learning_rate": 5e-05, "loss": 0.8005, "mean_token_accuracy": 0.8583246469497681, "num_tokens": 274570967.0, "step": 7233 }, { "epoch": 0.9202391553237501, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.927646160125732, "learning_rate": 5e-05, "loss": 0.7197, "mean_token_accuracy": 0.8617100715637207, "num_tokens": 274601977.0, "step": 7234 }, { "epoch": 0.9203663656023406, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.144732475280762, "learning_rate": 5e-05, "loss": 0.7851, "mean_token_accuracy": 0.8677569627761841, "num_tokens": 274638509.0, "step": 7235 }, { "epoch": 0.9204935758809312, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.603757858276367, "learning_rate": 5e-05, "loss": 0.6235, "mean_token_accuracy": 0.8813470602035522, "num_tokens": 274674325.0, "step": 7236 }, { "epoch": 0.9206207861595217, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.647064208984375, "learning_rate": 5e-05, "loss": 0.8436, "mean_token_accuracy": 0.8614069819450378, "num_tokens": 274709686.0, "step": 7237 }, { "epoch": 0.9207479964381122, "ewc_loss": 0.265625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00022125244140625, "grad_norm": 6.061923027038574, "learning_rate": 5e-05, "loss": 0.6988, "mean_token_accuracy": 0.8558401465415955, "num_tokens": 274748917.0, "step": 7238 }, { "epoch": 0.9208752067167028, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0004367828369140625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.519183158874512, "learning_rate": 5e-05, "loss": 0.974, "mean_token_accuracy": 0.8518388271331787, "num_tokens": 274794119.0, "step": 7239 }, { "epoch": 0.9210024169952932, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 5.587336540222168, "learning_rate": 5e-05, "loss": 0.7113, "mean_token_accuracy": 0.8503119945526123, "num_tokens": 274827907.0, "step": 7240 }, { "epoch": 0.9211296272738837, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.524148941040039, "learning_rate": 5e-05, "loss": 1.0148, "mean_token_accuracy": 0.8821982145309448, "num_tokens": 274865821.0, "step": 7241 }, { "epoch": 0.9212568375524742, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 6.559443950653076, "learning_rate": 5e-05, "loss": 0.8561, "mean_token_accuracy": 0.8652517795562744, "num_tokens": 274903550.0, "step": 7242 }, { "epoch": 0.9213840478310648, "ewc_loss": 0.71875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.350663185119629, "learning_rate": 5e-05, "loss": 1.1168, "mean_token_accuracy": 0.8631851673126221, "num_tokens": 274944212.0, "step": 7243 }, { "epoch": 0.9215112581096553, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 8.692960739135742, "learning_rate": 5e-05, "loss": 1.1311, "mean_token_accuracy": 0.8619178533554077, "num_tokens": 274980175.0, "step": 7244 }, { "epoch": 0.9216384683882458, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.959193229675293, "learning_rate": 5e-05, "loss": 0.8786, "mean_token_accuracy": 0.8548086285591125, "num_tokens": 275020018.0, "step": 7245 }, { "epoch": 0.9217656786668362, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 10.564711570739746, "learning_rate": 5e-05, "loss": 1.264, "mean_token_accuracy": 0.8568070530891418, "num_tokens": 275063538.0, "step": 7246 }, { "epoch": 0.9218928889454268, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.507026195526123, "learning_rate": 5e-05, "loss": 0.794, "mean_token_accuracy": 0.8592986464500427, "num_tokens": 275099884.0, "step": 7247 }, { "epoch": 0.9220200992240173, "ewc_loss": 1.125, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00107574462890625, "grad_norm": 13.291934967041016, "learning_rate": 5e-05, "loss": 1.5498, "mean_token_accuracy": 0.8543922305107117, "num_tokens": 275140292.0, "step": 7248 }, { "epoch": 0.9221473095026078, "ewc_loss": 1.078125, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00102996826171875, "grad_norm": 10.535911560058594, "learning_rate": 5e-05, "loss": 1.4861, "mean_token_accuracy": 0.8627610206604004, "num_tokens": 275178480.0, "step": 7249 }, { "epoch": 0.9222745197811983, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 11.842495918273926, "learning_rate": 5e-05, "loss": 1.1711, "mean_token_accuracy": 0.8556869029998779, "num_tokens": 275215855.0, "step": 7250 }, { "epoch": 0.9224017300597889, "ewc_loss": 0.87890625, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.000835418701171875, "grad_norm": 9.898946762084961, "learning_rate": 5e-05, "loss": 1.3398, "mean_token_accuracy": 0.845396876335144, "num_tokens": 275254895.0, "step": 7251 }, { "epoch": 0.9225289403383793, "ewc_loss": 0.88671875, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.000843048095703125, "grad_norm": 12.145832061767578, "learning_rate": 5e-05, "loss": 1.2975, "mean_token_accuracy": 0.8607901334762573, "num_tokens": 275292352.0, "step": 7252 }, { "epoch": 0.9226561506169698, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.848142623901367, "learning_rate": 5e-05, "loss": 1.0594, "mean_token_accuracy": 0.8544985055923462, "num_tokens": 275335867.0, "step": 7253 }, { "epoch": 0.9227833608955603, "ewc_loss": 0.97265625, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.0009307861328125, "grad_norm": 12.803797721862793, "learning_rate": 5e-05, "loss": 1.4093, "mean_token_accuracy": 0.8548600673675537, "num_tokens": 275376551.0, "step": 7254 }, { "epoch": 0.9229105711741509, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.000732421875, "grad_norm": 10.235963821411133, "learning_rate": 5e-05, "loss": 1.199, "mean_token_accuracy": 0.8544804453849792, "num_tokens": 275412568.0, "step": 7255 }, { "epoch": 0.9230377814527414, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.000774383544921875, "grad_norm": 12.069416046142578, "learning_rate": 5e-05, "loss": 1.2261, "mean_token_accuracy": 0.8623378276824951, "num_tokens": 275450534.0, "step": 7256 }, { "epoch": 0.9231649917313319, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.000766754150390625, "grad_norm": 10.672656059265137, "learning_rate": 5e-05, "loss": 1.2332, "mean_token_accuracy": 0.8568080067634583, "num_tokens": 275493165.0, "step": 7257 }, { "epoch": 0.9232922020099223, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 11.201803207397461, "learning_rate": 5e-05, "loss": 1.126, "mean_token_accuracy": 0.8625556230545044, "num_tokens": 275533935.0, "step": 7258 }, { "epoch": 0.9234194122885129, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.83557415008545, "learning_rate": 5e-05, "loss": 1.0848, "mean_token_accuracy": 0.8562077283859253, "num_tokens": 275570937.0, "step": 7259 }, { "epoch": 0.9235466225671034, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.924288749694824, "learning_rate": 5e-05, "loss": 1.0832, "mean_token_accuracy": 0.8685982823371887, "num_tokens": 275606036.0, "step": 7260 }, { "epoch": 0.9236738328456939, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.24169921875, "learning_rate": 5e-05, "loss": 0.9764, "mean_token_accuracy": 0.8682433366775513, "num_tokens": 275645849.0, "step": 7261 }, { "epoch": 0.9238010431242845, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.817550659179688, "learning_rate": 5e-05, "loss": 1.0238, "mean_token_accuracy": 0.8734461069107056, "num_tokens": 275682192.0, "step": 7262 }, { "epoch": 0.923928253402875, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.924625396728516, "learning_rate": 5e-05, "loss": 0.9973, "mean_token_accuracy": 0.8530980944633484, "num_tokens": 275722074.0, "step": 7263 }, { "epoch": 0.9240554636814655, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.972222328186035, "learning_rate": 5e-05, "loss": 1.1051, "mean_token_accuracy": 0.8557446002960205, "num_tokens": 275764966.0, "step": 7264 }, { "epoch": 0.9241826739600559, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.924118995666504, "learning_rate": 5e-05, "loss": 0.9267, "mean_token_accuracy": 0.8753535747528076, "num_tokens": 275805085.0, "step": 7265 }, { "epoch": 0.9243098842386465, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 11.151276588439941, "learning_rate": 5e-05, "loss": 1.1128, "mean_token_accuracy": 0.866457462310791, "num_tokens": 275837533.0, "step": 7266 }, { "epoch": 0.924437094517237, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.277120590209961, "learning_rate": 5e-05, "loss": 0.9861, "mean_token_accuracy": 0.8590123653411865, "num_tokens": 275877168.0, "step": 7267 }, { "epoch": 0.9245643047958275, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.907330513000488, "learning_rate": 5e-05, "loss": 1.119, "mean_token_accuracy": 0.8521124124526978, "num_tokens": 275917861.0, "step": 7268 }, { "epoch": 0.924691515074418, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.512516021728516, "learning_rate": 5e-05, "loss": 1.027, "mean_token_accuracy": 0.848751425743103, "num_tokens": 275952433.0, "step": 7269 }, { "epoch": 0.9248187253530086, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 10.226522445678711, "learning_rate": 5e-05, "loss": 1.0142, "mean_token_accuracy": 0.8627719879150391, "num_tokens": 275993124.0, "step": 7270 }, { "epoch": 0.924945935631599, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.179729461669922, "learning_rate": 5e-05, "loss": 0.9205, "mean_token_accuracy": 0.8698744773864746, "num_tokens": 276032967.0, "step": 7271 }, { "epoch": 0.9250731459101895, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.724519729614258, "learning_rate": 5e-05, "loss": 0.9701, "mean_token_accuracy": 0.8657752275466919, "num_tokens": 276068840.0, "step": 7272 }, { "epoch": 0.92520035618878, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.58594799041748, "learning_rate": 5e-05, "loss": 0.8939, "mean_token_accuracy": 0.8582607507705688, "num_tokens": 276109196.0, "step": 7273 }, { "epoch": 0.9253275664673706, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.668806076049805, "learning_rate": 5e-05, "loss": 0.9705, "mean_token_accuracy": 0.8707780838012695, "num_tokens": 276151544.0, "step": 7274 }, { "epoch": 0.9254547767459611, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.079297065734863, "learning_rate": 5e-05, "loss": 0.8299, "mean_token_accuracy": 0.8583290576934814, "num_tokens": 276192849.0, "step": 7275 }, { "epoch": 0.9255819870245516, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 10.04324722290039, "learning_rate": 5e-05, "loss": 0.9793, "mean_token_accuracy": 0.8751477003097534, "num_tokens": 276226543.0, "step": 7276 }, { "epoch": 0.925709197303142, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.144331932067871, "learning_rate": 5e-05, "loss": 0.8257, "mean_token_accuracy": 0.8684727549552917, "num_tokens": 276267199.0, "step": 7277 }, { "epoch": 0.9258364075817326, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 10.23612117767334, "learning_rate": 5e-05, "loss": 1.056, "mean_token_accuracy": 0.8581734299659729, "num_tokens": 276306786.0, "step": 7278 }, { "epoch": 0.9259636178603231, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.699126243591309, "learning_rate": 5e-05, "loss": 0.9134, "mean_token_accuracy": 0.860857367515564, "num_tokens": 276343479.0, "step": 7279 }, { "epoch": 0.9260908281389136, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.649625778198242, "learning_rate": 5e-05, "loss": 1.0001, "mean_token_accuracy": 0.8606077432632446, "num_tokens": 276378067.0, "step": 7280 }, { "epoch": 0.9262180384175042, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.796719551086426, "learning_rate": 5e-05, "loss": 0.898, "mean_token_accuracy": 0.8614009022712708, "num_tokens": 276407610.0, "step": 7281 }, { "epoch": 0.9263452486960947, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.846470832824707, "learning_rate": 5e-05, "loss": 0.9343, "mean_token_accuracy": 0.8537875413894653, "num_tokens": 276443664.0, "step": 7282 }, { "epoch": 0.9264724589746851, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.317708015441895, "learning_rate": 5e-05, "loss": 0.8245, "mean_token_accuracy": 0.8651140332221985, "num_tokens": 276480446.0, "step": 7283 }, { "epoch": 0.9265996692532756, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.538922309875488, "learning_rate": 5e-05, "loss": 0.8474, "mean_token_accuracy": 0.8684768676757812, "num_tokens": 276515012.0, "step": 7284 }, { "epoch": 0.9267268795318662, "ewc_loss": 0.375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.732173442840576, "learning_rate": 5e-05, "loss": 0.7291, "mean_token_accuracy": 0.8819841146469116, "num_tokens": 276555804.0, "step": 7285 }, { "epoch": 0.9268540898104567, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.608732223510742, "learning_rate": 5e-05, "loss": 0.9077, "mean_token_accuracy": 0.8483859896659851, "num_tokens": 276596290.0, "step": 7286 }, { "epoch": 0.9269813000890472, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.367674827575684, "learning_rate": 5e-05, "loss": 0.7823, "mean_token_accuracy": 0.8522830605506897, "num_tokens": 276641767.0, "step": 7287 }, { "epoch": 0.9271085103676378, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.760743141174316, "learning_rate": 5e-05, "loss": 0.8342, "mean_token_accuracy": 0.8759772777557373, "num_tokens": 276678375.0, "step": 7288 }, { "epoch": 0.9272357206462282, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.235387802124023, "learning_rate": 5e-05, "loss": 0.7553, "mean_token_accuracy": 0.8630462884902954, "num_tokens": 276715835.0, "step": 7289 }, { "epoch": 0.9273629309248187, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.888608932495117, "learning_rate": 5e-05, "loss": 0.9081, "mean_token_accuracy": 0.858873724937439, "num_tokens": 276750323.0, "step": 7290 }, { "epoch": 0.9274901412034092, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.322936534881592, "learning_rate": 5e-05, "loss": 0.7381, "mean_token_accuracy": 0.8629053831100464, "num_tokens": 276786867.0, "step": 7291 }, { "epoch": 0.9276173514819998, "ewc_loss": 0.5, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.848916053771973, "learning_rate": 5e-05, "loss": 0.9806, "mean_token_accuracy": 0.8397287726402283, "num_tokens": 276831263.0, "step": 7292 }, { "epoch": 0.9277445617605903, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.596797943115234, "learning_rate": 5e-05, "loss": 0.7652, "mean_token_accuracy": 0.8653896450996399, "num_tokens": 276867573.0, "step": 7293 }, { "epoch": 0.9278717720391808, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.400609970092773, "learning_rate": 5e-05, "loss": 0.7819, "mean_token_accuracy": 0.8831876516342163, "num_tokens": 276900414.0, "step": 7294 }, { "epoch": 0.9279989823177712, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.866721153259277, "learning_rate": 5e-05, "loss": 0.8032, "mean_token_accuracy": 0.8570932149887085, "num_tokens": 276938336.0, "step": 7295 }, { "epoch": 0.9281261925963618, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.714413166046143, "learning_rate": 5e-05, "loss": 0.7676, "mean_token_accuracy": 0.8707574605941772, "num_tokens": 276981277.0, "step": 7296 }, { "epoch": 0.9282534028749523, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.9978508949279785, "learning_rate": 5e-05, "loss": 0.84, "mean_token_accuracy": 0.8538957834243774, "num_tokens": 277023346.0, "step": 7297 }, { "epoch": 0.9283806131535428, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.314932823181152, "learning_rate": 5e-05, "loss": 0.7317, "mean_token_accuracy": 0.8619999289512634, "num_tokens": 277054821.0, "step": 7298 }, { "epoch": 0.9285078234321333, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 8.169021606445312, "learning_rate": 5e-05, "loss": 0.7869, "mean_token_accuracy": 0.8640881180763245, "num_tokens": 277092124.0, "step": 7299 }, { "epoch": 0.9286350337107239, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.984858512878418, "learning_rate": 5e-05, "loss": 0.7261, "mean_token_accuracy": 0.8621194958686829, "num_tokens": 277130119.0, "step": 7300 }, { "epoch": 0.9287622439893143, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.501367568969727, "learning_rate": 5e-05, "loss": 0.7686, "mean_token_accuracy": 0.8805180191993713, "num_tokens": 277167777.0, "step": 7301 }, { "epoch": 0.9288894542679048, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.770925998687744, "learning_rate": 5e-05, "loss": 0.6983, "mean_token_accuracy": 0.869326114654541, "num_tokens": 277207273.0, "step": 7302 }, { "epoch": 0.9290166645464953, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.694555282592773, "learning_rate": 5e-05, "loss": 0.8227, "mean_token_accuracy": 0.8699173927307129, "num_tokens": 277248996.0, "step": 7303 }, { "epoch": 0.9291438748250859, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.962662696838379, "learning_rate": 5e-05, "loss": 0.7788, "mean_token_accuracy": 0.8475353717803955, "num_tokens": 277286212.0, "step": 7304 }, { "epoch": 0.9292710851036764, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.348129272460938, "learning_rate": 5e-05, "loss": 0.8873, "mean_token_accuracy": 0.84464031457901, "num_tokens": 277330482.0, "step": 7305 }, { "epoch": 0.9293982953822669, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.460689544677734, "learning_rate": 5e-05, "loss": 0.7259, "mean_token_accuracy": 0.869407594203949, "num_tokens": 277364965.0, "step": 7306 }, { "epoch": 0.9295255056608573, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.750602722167969, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8625446557998657, "num_tokens": 277392227.0, "step": 7307 }, { "epoch": 0.9296527159394479, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.808200359344482, "learning_rate": 5e-05, "loss": 0.743, "mean_token_accuracy": 0.8720674514770508, "num_tokens": 277435658.0, "step": 7308 }, { "epoch": 0.9297799262180384, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.350978374481201, "learning_rate": 5e-05, "loss": 0.7603, "mean_token_accuracy": 0.8611193299293518, "num_tokens": 277472513.0, "step": 7309 }, { "epoch": 0.9299071364966289, "ewc_loss": 0.375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.9796929359436035, "learning_rate": 5e-05, "loss": 0.8145, "mean_token_accuracy": 0.8502277135848999, "num_tokens": 277517740.0, "step": 7310 }, { "epoch": 0.9300343467752195, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.20038366317749, "learning_rate": 5e-05, "loss": 0.6857, "mean_token_accuracy": 0.8760366439819336, "num_tokens": 277547304.0, "step": 7311 }, { "epoch": 0.93016155705381, "ewc_loss": 0.375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.973420143127441, "learning_rate": 5e-05, "loss": 0.8264, "mean_token_accuracy": 0.8515281677246094, "num_tokens": 277581380.0, "step": 7312 }, { "epoch": 0.9302887673324005, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.274743556976318, "learning_rate": 5e-05, "loss": 0.7052, "mean_token_accuracy": 0.8705052733421326, "num_tokens": 277621615.0, "step": 7313 }, { "epoch": 0.9304159776109909, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.714658260345459, "learning_rate": 5e-05, "loss": 0.7964, "mean_token_accuracy": 0.8516587018966675, "num_tokens": 277655157.0, "step": 7314 }, { "epoch": 0.9305431878895815, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.595170497894287, "learning_rate": 5e-05, "loss": 0.7986, "mean_token_accuracy": 0.8478904962539673, "num_tokens": 277688634.0, "step": 7315 }, { "epoch": 0.930670398168172, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.283901691436768, "learning_rate": 5e-05, "loss": 0.7909, "mean_token_accuracy": 0.848932683467865, "num_tokens": 277736123.0, "step": 7316 }, { "epoch": 0.9307976084467625, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 8.026395797729492, "learning_rate": 5e-05, "loss": 0.7825, "mean_token_accuracy": 0.8611300587654114, "num_tokens": 277776197.0, "step": 7317 }, { "epoch": 0.930924818725353, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.842434406280518, "learning_rate": 5e-05, "loss": 0.6729, "mean_token_accuracy": 0.8740639090538025, "num_tokens": 277820477.0, "step": 7318 }, { "epoch": 0.9310520290039436, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.475340843200684, "learning_rate": 5e-05, "loss": 0.8208, "mean_token_accuracy": 0.8652064800262451, "num_tokens": 277864660.0, "step": 7319 }, { "epoch": 0.931179239282534, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.487789630889893, "learning_rate": 5e-05, "loss": 0.7353, "mean_token_accuracy": 0.8479579091072083, "num_tokens": 277905702.0, "step": 7320 }, { "epoch": 0.9313064495611245, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.894187927246094, "learning_rate": 5e-05, "loss": 0.8563, "mean_token_accuracy": 0.8655809164047241, "num_tokens": 277946033.0, "step": 7321 }, { "epoch": 0.931433659839715, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.346187114715576, "learning_rate": 5e-05, "loss": 0.6911, "mean_token_accuracy": 0.8611143827438354, "num_tokens": 277981556.0, "step": 7322 }, { "epoch": 0.9315608701183056, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 9.0974760055542, "learning_rate": 5e-05, "loss": 0.8704, "mean_token_accuracy": 0.8735560178756714, "num_tokens": 278021455.0, "step": 7323 }, { "epoch": 0.9316880803968961, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.607197284698486, "learning_rate": 5e-05, "loss": 0.6937, "mean_token_accuracy": 0.8691401481628418, "num_tokens": 278057840.0, "step": 7324 }, { "epoch": 0.9318152906754866, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.821584701538086, "learning_rate": 5e-05, "loss": 0.9413, "mean_token_accuracy": 0.8552775382995605, "num_tokens": 278093186.0, "step": 7325 }, { "epoch": 0.931942500954077, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.179953098297119, "learning_rate": 5e-05, "loss": 0.7667, "mean_token_accuracy": 0.861987829208374, "num_tokens": 278127235.0, "step": 7326 }, { "epoch": 0.9320697112326676, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.216622352600098, "learning_rate": 5e-05, "loss": 0.8606, "mean_token_accuracy": 0.8645471334457397, "num_tokens": 278165236.0, "step": 7327 }, { "epoch": 0.9321969215112581, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.560075283050537, "learning_rate": 5e-05, "loss": 0.8192, "mean_token_accuracy": 0.8487123250961304, "num_tokens": 278205153.0, "step": 7328 }, { "epoch": 0.9323241317898486, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.903175354003906, "learning_rate": 5e-05, "loss": 0.7811, "mean_token_accuracy": 0.8761289119720459, "num_tokens": 278240808.0, "step": 7329 }, { "epoch": 0.9324513420684392, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.421175956726074, "learning_rate": 5e-05, "loss": 0.7875, "mean_token_accuracy": 0.8539367914199829, "num_tokens": 278281364.0, "step": 7330 }, { "epoch": 0.9325785523470297, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.148058891296387, "learning_rate": 5e-05, "loss": 0.8207, "mean_token_accuracy": 0.8670498728752136, "num_tokens": 278316021.0, "step": 7331 }, { "epoch": 0.9327057626256201, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.9898529052734375, "learning_rate": 5e-05, "loss": 0.7616, "mean_token_accuracy": 0.8556040525436401, "num_tokens": 278357744.0, "step": 7332 }, { "epoch": 0.9328329729042106, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.537137031555176, "learning_rate": 5e-05, "loss": 0.8584, "mean_token_accuracy": 0.862804114818573, "num_tokens": 278390826.0, "step": 7333 }, { "epoch": 0.9329601831828012, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.817845344543457, "learning_rate": 5e-05, "loss": 0.7872, "mean_token_accuracy": 0.8458670377731323, "num_tokens": 278424395.0, "step": 7334 }, { "epoch": 0.9330873934613917, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.566118240356445, "learning_rate": 5e-05, "loss": 0.8376, "mean_token_accuracy": 0.8751189708709717, "num_tokens": 278462180.0, "step": 7335 }, { "epoch": 0.9332146037399822, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.122791290283203, "learning_rate": 5e-05, "loss": 0.773, "mean_token_accuracy": 0.8569863438606262, "num_tokens": 278499349.0, "step": 7336 }, { "epoch": 0.9333418140185727, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.032575607299805, "learning_rate": 5e-05, "loss": 0.8739, "mean_token_accuracy": 0.851199746131897, "num_tokens": 278530782.0, "step": 7337 }, { "epoch": 0.9334690242971632, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.661064147949219, "learning_rate": 5e-05, "loss": 0.8007, "mean_token_accuracy": 0.8557885885238647, "num_tokens": 278563262.0, "step": 7338 }, { "epoch": 0.9335962345757537, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.368467330932617, "learning_rate": 5e-05, "loss": 0.77, "mean_token_accuracy": 0.8617655038833618, "num_tokens": 278598882.0, "step": 7339 }, { "epoch": 0.9337234448543442, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.060897827148438, "learning_rate": 5e-05, "loss": 0.8029, "mean_token_accuracy": 0.8630692958831787, "num_tokens": 278628535.0, "step": 7340 }, { "epoch": 0.9338506551329347, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.99885368347168, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.8568412065505981, "num_tokens": 278667732.0, "step": 7341 }, { "epoch": 0.9339778654115253, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.22790241241455, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8630295991897583, "num_tokens": 278706925.0, "step": 7342 }, { "epoch": 0.9341050756901158, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.886228084564209, "learning_rate": 5e-05, "loss": 0.6865, "mean_token_accuracy": 0.8740115761756897, "num_tokens": 278740491.0, "step": 7343 }, { "epoch": 0.9342322859687062, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.311052322387695, "learning_rate": 5e-05, "loss": 0.8627, "mean_token_accuracy": 0.8503119945526123, "num_tokens": 278776739.0, "step": 7344 }, { "epoch": 0.9343594962472968, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.807877063751221, "learning_rate": 5e-05, "loss": 0.6999, "mean_token_accuracy": 0.8710334300994873, "num_tokens": 278813019.0, "step": 7345 }, { "epoch": 0.9344867065258873, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.41810131072998, "learning_rate": 5e-05, "loss": 0.8599, "mean_token_accuracy": 0.8616676330566406, "num_tokens": 278853807.0, "step": 7346 }, { "epoch": 0.9346139168044778, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.7499213218688965, "learning_rate": 5e-05, "loss": 0.6748, "mean_token_accuracy": 0.8793882131576538, "num_tokens": 278891643.0, "step": 7347 }, { "epoch": 0.9347411270830683, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.518448829650879, "learning_rate": 5e-05, "loss": 0.972, "mean_token_accuracy": 0.8418456315994263, "num_tokens": 278929915.0, "step": 7348 }, { "epoch": 0.9348683373616589, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.782041072845459, "learning_rate": 5e-05, "loss": 0.7822, "mean_token_accuracy": 0.8436924815177917, "num_tokens": 278964400.0, "step": 7349 }, { "epoch": 0.9349955476402493, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.482979774475098, "learning_rate": 5e-05, "loss": 0.8725, "mean_token_accuracy": 0.8633064031600952, "num_tokens": 279003322.0, "step": 7350 }, { "epoch": 0.9351227579188398, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.979467868804932, "learning_rate": 5e-05, "loss": 0.7877, "mean_token_accuracy": 0.8503061532974243, "num_tokens": 279045630.0, "step": 7351 }, { "epoch": 0.9352499681974303, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.186976432800293, "learning_rate": 5e-05, "loss": 0.8748, "mean_token_accuracy": 0.8615506887435913, "num_tokens": 279085330.0, "step": 7352 }, { "epoch": 0.9353771784760209, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.313037395477295, "learning_rate": 5e-05, "loss": 0.7528, "mean_token_accuracy": 0.8678014278411865, "num_tokens": 279125499.0, "step": 7353 }, { "epoch": 0.9355043887546114, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.771085739135742, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.8656564950942993, "num_tokens": 279154772.0, "step": 7354 }, { "epoch": 0.9356315990332019, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.601256847381592, "learning_rate": 5e-05, "loss": 0.8171, "mean_token_accuracy": 0.8565092086791992, "num_tokens": 279196947.0, "step": 7355 }, { "epoch": 0.9357588093117923, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.4697136878967285, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.877435564994812, "num_tokens": 279236872.0, "step": 7356 }, { "epoch": 0.9358860195903829, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.711609840393066, "learning_rate": 5e-05, "loss": 0.7542, "mean_token_accuracy": 0.8705726861953735, "num_tokens": 279275456.0, "step": 7357 }, { "epoch": 0.9360132298689734, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.375521659851074, "learning_rate": 5e-05, "loss": 0.7696, "mean_token_accuracy": 0.857567310333252, "num_tokens": 279308406.0, "step": 7358 }, { "epoch": 0.9361404401475639, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.6533355712890625, "learning_rate": 5e-05, "loss": 0.772, "mean_token_accuracy": 0.8597384095191956, "num_tokens": 279349522.0, "step": 7359 }, { "epoch": 0.9362676504261545, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.447256565093994, "learning_rate": 5e-05, "loss": 0.7173, "mean_token_accuracy": 0.8725471496582031, "num_tokens": 279388242.0, "step": 7360 }, { "epoch": 0.936394860704745, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.500773906707764, "learning_rate": 5e-05, "loss": 0.7134, "mean_token_accuracy": 0.8745187520980835, "num_tokens": 279421239.0, "step": 7361 }, { "epoch": 0.9365220709833355, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.588052272796631, "learning_rate": 5e-05, "loss": 0.8408, "mean_token_accuracy": 0.8365737199783325, "num_tokens": 279464823.0, "step": 7362 }, { "epoch": 0.9366492812619259, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.287017345428467, "learning_rate": 5e-05, "loss": 0.7787, "mean_token_accuracy": 0.8494380712509155, "num_tokens": 279507645.0, "step": 7363 }, { "epoch": 0.9367764915405165, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.821086883544922, "learning_rate": 5e-05, "loss": 0.7796, "mean_token_accuracy": 0.8582593202590942, "num_tokens": 279539905.0, "step": 7364 }, { "epoch": 0.936903701819107, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.974236965179443, "learning_rate": 5e-05, "loss": 0.6924, "mean_token_accuracy": 0.8707265853881836, "num_tokens": 279577929.0, "step": 7365 }, { "epoch": 0.9370309120976975, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.203169822692871, "learning_rate": 5e-05, "loss": 0.8004, "mean_token_accuracy": 0.8646445274353027, "num_tokens": 279618421.0, "step": 7366 }, { "epoch": 0.937158122376288, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.548366069793701, "learning_rate": 5e-05, "loss": 0.6876, "mean_token_accuracy": 0.8670242428779602, "num_tokens": 279656563.0, "step": 7367 }, { "epoch": 0.9372853326548786, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.800528526306152, "learning_rate": 5e-05, "loss": 0.8882, "mean_token_accuracy": 0.8545446395874023, "num_tokens": 279693159.0, "step": 7368 }, { "epoch": 0.937412542933469, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0002269744873046875, "grad_norm": 6.01320743560791, "learning_rate": 5e-05, "loss": 0.6308, "mean_token_accuracy": 0.8759585022926331, "num_tokens": 279731492.0, "step": 7369 }, { "epoch": 0.9375397532120595, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.669418334960938, "learning_rate": 5e-05, "loss": 0.9708, "mean_token_accuracy": 0.8675090074539185, "num_tokens": 279774906.0, "step": 7370 }, { "epoch": 0.93766696349065, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.825656414031982, "learning_rate": 5e-05, "loss": 0.7782, "mean_token_accuracy": 0.8424901962280273, "num_tokens": 279809797.0, "step": 7371 }, { "epoch": 0.9377941737692406, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.243451118469238, "learning_rate": 5e-05, "loss": 1.0476, "mean_token_accuracy": 0.8671737909317017, "num_tokens": 279850107.0, "step": 7372 }, { "epoch": 0.9379213840478311, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.083353042602539, "learning_rate": 5e-05, "loss": 0.8258, "mean_token_accuracy": 0.8849349021911621, "num_tokens": 279885115.0, "step": 7373 }, { "epoch": 0.9380485943264216, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.366140365600586, "learning_rate": 5e-05, "loss": 1.0004, "mean_token_accuracy": 0.869685173034668, "num_tokens": 279921343.0, "step": 7374 }, { "epoch": 0.938175804605012, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.735672950744629, "learning_rate": 5e-05, "loss": 0.9976, "mean_token_accuracy": 0.8729943037033081, "num_tokens": 279953059.0, "step": 7375 }, { "epoch": 0.9383030148836026, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.030760765075684, "learning_rate": 5e-05, "loss": 0.765, "mean_token_accuracy": 0.8698910474777222, "num_tokens": 279987993.0, "step": 7376 }, { "epoch": 0.9384302251621931, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.531365394592285, "learning_rate": 5e-05, "loss": 1.1488, "mean_token_accuracy": 0.8543243408203125, "num_tokens": 280024620.0, "step": 7377 }, { "epoch": 0.9385574354407836, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.3495001792907715, "learning_rate": 5e-05, "loss": 0.8607, "mean_token_accuracy": 0.8448063135147095, "num_tokens": 280069011.0, "step": 7378 }, { "epoch": 0.9386846457193742, "ewc_loss": 0.94140625, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0009002685546875, "grad_norm": 12.112176895141602, "learning_rate": 5e-05, "loss": 1.2945, "mean_token_accuracy": 0.8811306357383728, "num_tokens": 280105877.0, "step": 7379 }, { "epoch": 0.9388118559979647, "ewc_loss": 0.859375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.00081634521484375, "grad_norm": 9.295036315917969, "learning_rate": 5e-05, "loss": 1.214, "mean_token_accuracy": 0.8733419179916382, "num_tokens": 280143155.0, "step": 7380 }, { "epoch": 0.9389390662765551, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.863176345825195, "learning_rate": 5e-05, "loss": 1.0534, "mean_token_accuracy": 0.8702813386917114, "num_tokens": 280181490.0, "step": 7381 }, { "epoch": 0.9390662765551456, "ewc_loss": 0.83203125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 9.525126457214355, "learning_rate": 5e-05, "loss": 1.2752, "mean_token_accuracy": 0.8532381057739258, "num_tokens": 280221959.0, "step": 7382 }, { "epoch": 0.9391934868337362, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.889245986938477, "learning_rate": 5e-05, "loss": 0.9873, "mean_token_accuracy": 0.8836259841918945, "num_tokens": 280265224.0, "step": 7383 }, { "epoch": 0.9393206971123267, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.919402122497559, "learning_rate": 5e-05, "loss": 1.038, "mean_token_accuracy": 0.8645329475402832, "num_tokens": 280302039.0, "step": 7384 }, { "epoch": 0.9394479073909172, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.09152889251709, "learning_rate": 5e-05, "loss": 1.0249, "mean_token_accuracy": 0.8632297515869141, "num_tokens": 280338559.0, "step": 7385 }, { "epoch": 0.9395751176695077, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.196857452392578, "learning_rate": 5e-05, "loss": 0.8685, "mean_token_accuracy": 0.8753044009208679, "num_tokens": 280376731.0, "step": 7386 }, { "epoch": 0.9397023279480982, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.082036018371582, "learning_rate": 5e-05, "loss": 1.0035, "mean_token_accuracy": 0.8608109354972839, "num_tokens": 280414201.0, "step": 7387 }, { "epoch": 0.9398295382266887, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.403036594390869, "learning_rate": 5e-05, "loss": 0.8271, "mean_token_accuracy": 0.8647938966751099, "num_tokens": 280456490.0, "step": 7388 }, { "epoch": 0.9399567485052792, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.638480186462402, "learning_rate": 5e-05, "loss": 1.03, "mean_token_accuracy": 0.8608980178833008, "num_tokens": 280499593.0, "step": 7389 }, { "epoch": 0.9400839587838697, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.222601890563965, "learning_rate": 5e-05, "loss": 0.8758, "mean_token_accuracy": 0.8547958135604858, "num_tokens": 280535722.0, "step": 7390 }, { "epoch": 0.9402111690624603, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.036154747009277, "learning_rate": 5e-05, "loss": 1.045, "mean_token_accuracy": 0.867925763130188, "num_tokens": 280569666.0, "step": 7391 }, { "epoch": 0.9403383793410508, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.727053165435791, "learning_rate": 5e-05, "loss": 0.8594, "mean_token_accuracy": 0.8891383409500122, "num_tokens": 280607359.0, "step": 7392 }, { "epoch": 0.9404655896196412, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.913203239440918, "learning_rate": 5e-05, "loss": 1.0646, "mean_token_accuracy": 0.8605790734291077, "num_tokens": 280641054.0, "step": 7393 }, { "epoch": 0.9405927998982317, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.188488960266113, "learning_rate": 5e-05, "loss": 1.0259, "mean_token_accuracy": 0.8518846035003662, "num_tokens": 280683798.0, "step": 7394 }, { "epoch": 0.9407200101768223, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.372536659240723, "learning_rate": 5e-05, "loss": 1.0677, "mean_token_accuracy": 0.8533594012260437, "num_tokens": 280729407.0, "step": 7395 }, { "epoch": 0.9408472204554128, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.287203788757324, "learning_rate": 5e-05, "loss": 0.9812, "mean_token_accuracy": 0.8648617267608643, "num_tokens": 280764327.0, "step": 7396 }, { "epoch": 0.9409744307340033, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.755658149719238, "learning_rate": 5e-05, "loss": 1.0109, "mean_token_accuracy": 0.860077440738678, "num_tokens": 280802056.0, "step": 7397 }, { "epoch": 0.9411016410125939, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.003134727478027, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.885086178779602, "num_tokens": 280837269.0, "step": 7398 }, { "epoch": 0.9412288512911843, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.64030647277832, "learning_rate": 5e-05, "loss": 1.0497, "mean_token_accuracy": 0.8409549593925476, "num_tokens": 280873914.0, "step": 7399 }, { "epoch": 0.9413560615697748, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.453047275543213, "learning_rate": 5e-05, "loss": 0.8331, "mean_token_accuracy": 0.8658576011657715, "num_tokens": 280911436.0, "step": 7400 }, { "epoch": 0.9414832718483653, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0004215240478515625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.075183868408203, "learning_rate": 5e-05, "loss": 0.9961, "mean_token_accuracy": 0.8738452196121216, "num_tokens": 280952395.0, "step": 7401 }, { "epoch": 0.9416104821269559, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.200491905212402, "learning_rate": 5e-05, "loss": 0.8778, "mean_token_accuracy": 0.8460854291915894, "num_tokens": 280990455.0, "step": 7402 }, { "epoch": 0.9417376924055464, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.463582992553711, "learning_rate": 5e-05, "loss": 1.0575, "mean_token_accuracy": 0.8653329610824585, "num_tokens": 281020584.0, "step": 7403 }, { "epoch": 0.9418649026841369, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.622591495513916, "learning_rate": 5e-05, "loss": 0.8878, "mean_token_accuracy": 0.8730038404464722, "num_tokens": 281059907.0, "step": 7404 }, { "epoch": 0.9419921129627273, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.176032066345215, "learning_rate": 5e-05, "loss": 1.0369, "mean_token_accuracy": 0.8670989274978638, "num_tokens": 281097425.0, "step": 7405 }, { "epoch": 0.9421193232413179, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.16464900970459, "learning_rate": 5e-05, "loss": 0.9996, "mean_token_accuracy": 0.8606287837028503, "num_tokens": 281146696.0, "step": 7406 }, { "epoch": 0.9422465335199084, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.30881118774414, "learning_rate": 5e-05, "loss": 0.9826, "mean_token_accuracy": 0.849605917930603, "num_tokens": 281186166.0, "step": 7407 }, { "epoch": 0.9423737437984989, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.261734008789062, "learning_rate": 5e-05, "loss": 1.042, "mean_token_accuracy": 0.8328977823257446, "num_tokens": 281230372.0, "step": 7408 }, { "epoch": 0.9425009540770894, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.642199516296387, "learning_rate": 5e-05, "loss": 0.8553, "mean_token_accuracy": 0.8648606538772583, "num_tokens": 281256035.0, "step": 7409 }, { "epoch": 0.94262816435568, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.000259399414062, "learning_rate": 5e-05, "loss": 0.8112, "mean_token_accuracy": 0.8770172595977783, "num_tokens": 281290792.0, "step": 7410 }, { "epoch": 0.9427553746342705, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.70595645904541, "learning_rate": 5e-05, "loss": 0.8394, "mean_token_accuracy": 0.8689457178115845, "num_tokens": 281329949.0, "step": 7411 }, { "epoch": 0.9428825849128609, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.497706890106201, "learning_rate": 5e-05, "loss": 0.7809, "mean_token_accuracy": 0.8716624975204468, "num_tokens": 281366383.0, "step": 7412 }, { "epoch": 0.9430097951914514, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.050704002380371, "learning_rate": 5e-05, "loss": 0.891, "mean_token_accuracy": 0.85467529296875, "num_tokens": 281408537.0, "step": 7413 }, { "epoch": 0.943137005470042, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.101021766662598, "learning_rate": 5e-05, "loss": 0.775, "mean_token_accuracy": 0.8567821979522705, "num_tokens": 281447745.0, "step": 7414 }, { "epoch": 0.9432642157486325, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.318672180175781, "learning_rate": 5e-05, "loss": 0.8577, "mean_token_accuracy": 0.8681522607803345, "num_tokens": 281484430.0, "step": 7415 }, { "epoch": 0.943391426027223, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.892742156982422, "learning_rate": 5e-05, "loss": 0.6927, "mean_token_accuracy": 0.8800787329673767, "num_tokens": 281519720.0, "step": 7416 }, { "epoch": 0.9435186363058136, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.582975387573242, "learning_rate": 5e-05, "loss": 0.9466, "mean_token_accuracy": 0.8480594754219055, "num_tokens": 281556039.0, "step": 7417 }, { "epoch": 0.943645846584404, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.7620954513549805, "learning_rate": 5e-05, "loss": 0.767, "mean_token_accuracy": 0.8540144562721252, "num_tokens": 281594781.0, "step": 7418 }, { "epoch": 0.9437730568629945, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.846800804138184, "learning_rate": 5e-05, "loss": 0.9875, "mean_token_accuracy": 0.851770281791687, "num_tokens": 281632743.0, "step": 7419 }, { "epoch": 0.943900267141585, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.884416103363037, "learning_rate": 5e-05, "loss": 0.7846, "mean_token_accuracy": 0.8630407452583313, "num_tokens": 281669706.0, "step": 7420 }, { "epoch": 0.9440274774201756, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.736863136291504, "learning_rate": 5e-05, "loss": 0.9389, "mean_token_accuracy": 0.8671533465385437, "num_tokens": 281706108.0, "step": 7421 }, { "epoch": 0.9441546876987661, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.373674392700195, "learning_rate": 5e-05, "loss": 0.8565, "mean_token_accuracy": 0.856726884841919, "num_tokens": 281743437.0, "step": 7422 }, { "epoch": 0.9442818979773566, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.130287170410156, "learning_rate": 5e-05, "loss": 0.8549, "mean_token_accuracy": 0.8768875598907471, "num_tokens": 281779652.0, "step": 7423 }, { "epoch": 0.944409108255947, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.799498558044434, "learning_rate": 5e-05, "loss": 0.8527, "mean_token_accuracy": 0.867476761341095, "num_tokens": 281815841.0, "step": 7424 }, { "epoch": 0.9445363185345376, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.521120071411133, "learning_rate": 5e-05, "loss": 0.8167, "mean_token_accuracy": 0.8627631068229675, "num_tokens": 281853991.0, "step": 7425 }, { "epoch": 0.9446635288131281, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.010705947875977, "learning_rate": 5e-05, "loss": 0.8632, "mean_token_accuracy": 0.8579628467559814, "num_tokens": 281883264.0, "step": 7426 }, { "epoch": 0.9447907390917186, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.186309337615967, "learning_rate": 5e-05, "loss": 0.7284, "mean_token_accuracy": 0.8768912553787231, "num_tokens": 281920266.0, "step": 7427 }, { "epoch": 0.9449179493703092, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.133445739746094, "learning_rate": 5e-05, "loss": 0.8382, "mean_token_accuracy": 0.8650859594345093, "num_tokens": 281959949.0, "step": 7428 }, { "epoch": 0.9450451596488997, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.988711833953857, "learning_rate": 5e-05, "loss": 0.7288, "mean_token_accuracy": 0.8703391551971436, "num_tokens": 281992255.0, "step": 7429 }, { "epoch": 0.9451723699274901, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.293437004089355, "learning_rate": 5e-05, "loss": 0.8479, "mean_token_accuracy": 0.8646852970123291, "num_tokens": 282028000.0, "step": 7430 }, { "epoch": 0.9452995802060806, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.824462890625, "learning_rate": 5e-05, "loss": 0.7304, "mean_token_accuracy": 0.8643163442611694, "num_tokens": 282064332.0, "step": 7431 }, { "epoch": 0.9454267904846712, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.525684356689453, "learning_rate": 5e-05, "loss": 0.9047, "mean_token_accuracy": 0.8516325354576111, "num_tokens": 282102185.0, "step": 7432 }, { "epoch": 0.9455540007632617, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.684138298034668, "learning_rate": 5e-05, "loss": 0.8416, "mean_token_accuracy": 0.827476978302002, "num_tokens": 282138708.0, "step": 7433 }, { "epoch": 0.9456812110418522, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00041961669921875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.732718467712402, "learning_rate": 5e-05, "loss": 0.8786, "mean_token_accuracy": 0.8738987445831299, "num_tokens": 282168223.0, "step": 7434 }, { "epoch": 0.9458084213204427, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.742887020111084, "learning_rate": 5e-05, "loss": 0.7089, "mean_token_accuracy": 0.8717553019523621, "num_tokens": 282205226.0, "step": 7435 }, { "epoch": 0.9459356315990332, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.636297225952148, "learning_rate": 5e-05, "loss": 0.9648, "mean_token_accuracy": 0.8499638438224792, "num_tokens": 282239131.0, "step": 7436 }, { "epoch": 0.9460628418776237, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.115683078765869, "learning_rate": 5e-05, "loss": 0.796, "mean_token_accuracy": 0.8638399243354797, "num_tokens": 282276175.0, "step": 7437 }, { "epoch": 0.9461900521562142, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.179649353027344, "learning_rate": 5e-05, "loss": 0.9204, "mean_token_accuracy": 0.8549953699111938, "num_tokens": 282308598.0, "step": 7438 }, { "epoch": 0.9463172624348047, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000423431396484375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.517319679260254, "learning_rate": 5e-05, "loss": 0.7655, "mean_token_accuracy": 0.8822206854820251, "num_tokens": 282342193.0, "step": 7439 }, { "epoch": 0.9464444727133953, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.704764366149902, "learning_rate": 5e-05, "loss": 0.824, "mean_token_accuracy": 0.8635510206222534, "num_tokens": 282380531.0, "step": 7440 }, { "epoch": 0.9465716829919858, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.703083515167236, "learning_rate": 5e-05, "loss": 0.8126, "mean_token_accuracy": 0.8585069179534912, "num_tokens": 282410713.0, "step": 7441 }, { "epoch": 0.9466988932705762, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.571060657501221, "learning_rate": 5e-05, "loss": 0.792, "mean_token_accuracy": 0.8659756183624268, "num_tokens": 282451881.0, "step": 7442 }, { "epoch": 0.9468261035491667, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.524251461029053, "learning_rate": 5e-05, "loss": 0.8343, "mean_token_accuracy": 0.8469914197921753, "num_tokens": 282486031.0, "step": 7443 }, { "epoch": 0.9469533138277573, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.772267818450928, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8746903538703918, "num_tokens": 282523753.0, "step": 7444 }, { "epoch": 0.9470805241063478, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.177026271820068, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8572304844856262, "num_tokens": 282563322.0, "step": 7445 }, { "epoch": 0.9472077343849383, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.125503540039062, "learning_rate": 5e-05, "loss": 0.8403, "mean_token_accuracy": 0.8600614666938782, "num_tokens": 282602574.0, "step": 7446 }, { "epoch": 0.9473349446635289, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.780440330505371, "learning_rate": 5e-05, "loss": 0.7011, "mean_token_accuracy": 0.8697452545166016, "num_tokens": 282636608.0, "step": 7447 }, { "epoch": 0.9474621549421193, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.542348861694336, "learning_rate": 5e-05, "loss": 0.8144, "mean_token_accuracy": 0.878248393535614, "num_tokens": 282675320.0, "step": 7448 }, { "epoch": 0.9475893652207098, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.48148775100708, "learning_rate": 5e-05, "loss": 0.7597, "mean_token_accuracy": 0.8568747639656067, "num_tokens": 282712537.0, "step": 7449 }, { "epoch": 0.9477165754993003, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.952144622802734, "learning_rate": 5e-05, "loss": 0.9001, "mean_token_accuracy": 0.8656967878341675, "num_tokens": 282744620.0, "step": 7450 }, { "epoch": 0.9478437857778909, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.427379608154297, "learning_rate": 5e-05, "loss": 0.7106, "mean_token_accuracy": 0.8708922266960144, "num_tokens": 282786480.0, "step": 7451 }, { "epoch": 0.9479709960564814, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.060786247253418, "learning_rate": 5e-05, "loss": 0.9305, "mean_token_accuracy": 0.87192702293396, "num_tokens": 282819143.0, "step": 7452 }, { "epoch": 0.9480982063350719, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 6.941072940826416, "learning_rate": 5e-05, "loss": 0.7985, "mean_token_accuracy": 0.8653551340103149, "num_tokens": 282859658.0, "step": 7453 }, { "epoch": 0.9482254166136623, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.394498825073242, "learning_rate": 5e-05, "loss": 0.9255, "mean_token_accuracy": 0.861065149307251, "num_tokens": 282896190.0, "step": 7454 }, { "epoch": 0.9483526268922529, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.800103187561035, "learning_rate": 5e-05, "loss": 0.9295, "mean_token_accuracy": 0.8527172803878784, "num_tokens": 282937458.0, "step": 7455 }, { "epoch": 0.9484798371708434, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.2807936668396, "learning_rate": 5e-05, "loss": 0.7885, "mean_token_accuracy": 0.8633689880371094, "num_tokens": 282970669.0, "step": 7456 }, { "epoch": 0.9486070474494339, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.503028869628906, "learning_rate": 5e-05, "loss": 0.9166, "mean_token_accuracy": 0.8595180511474609, "num_tokens": 283012372.0, "step": 7457 }, { "epoch": 0.9487342577280244, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.612832069396973, "learning_rate": 5e-05, "loss": 0.7662, "mean_token_accuracy": 0.8497376441955566, "num_tokens": 283046165.0, "step": 7458 }, { "epoch": 0.948861468006615, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.868998527526855, "learning_rate": 5e-05, "loss": 0.9365, "mean_token_accuracy": 0.8609205484390259, "num_tokens": 283086539.0, "step": 7459 }, { "epoch": 0.9489886782852054, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.47930383682251, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.859201192855835, "num_tokens": 283123643.0, "step": 7460 }, { "epoch": 0.9491158885637959, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.05309772491455, "learning_rate": 5e-05, "loss": 0.9023, "mean_token_accuracy": 0.880895733833313, "num_tokens": 283162354.0, "step": 7461 }, { "epoch": 0.9492430988423864, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.703580856323242, "learning_rate": 5e-05, "loss": 0.7447, "mean_token_accuracy": 0.871573805809021, "num_tokens": 283196814.0, "step": 7462 }, { "epoch": 0.949370309120977, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.926666259765625, "learning_rate": 5e-05, "loss": 0.9962, "mean_token_accuracy": 0.8546795845031738, "num_tokens": 283238051.0, "step": 7463 }, { "epoch": 0.9494975193995675, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.2350921630859375, "learning_rate": 5e-05, "loss": 0.8976, "mean_token_accuracy": 0.8532458543777466, "num_tokens": 283278025.0, "step": 7464 }, { "epoch": 0.949624729678158, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.373908042907715, "learning_rate": 5e-05, "loss": 0.9132, "mean_token_accuracy": 0.8680734634399414, "num_tokens": 283311279.0, "step": 7465 }, { "epoch": 0.9497519399567486, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.667142868041992, "learning_rate": 5e-05, "loss": 0.9145, "mean_token_accuracy": 0.8555781841278076, "num_tokens": 283351350.0, "step": 7466 }, { "epoch": 0.949879150235339, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.834766387939453, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8683863878250122, "num_tokens": 283387374.0, "step": 7467 }, { "epoch": 0.9500063605139295, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.678603649139404, "learning_rate": 5e-05, "loss": 0.8787, "mean_token_accuracy": 0.8551422357559204, "num_tokens": 283431305.0, "step": 7468 }, { "epoch": 0.95013357079252, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.867087364196777, "learning_rate": 5e-05, "loss": 0.8559, "mean_token_accuracy": 0.8640589118003845, "num_tokens": 283464941.0, "step": 7469 }, { "epoch": 0.9502607810711106, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.226160049438477, "learning_rate": 5e-05, "loss": 0.8688, "mean_token_accuracy": 0.8444284200668335, "num_tokens": 283507254.0, "step": 7470 }, { "epoch": 0.9503879913497011, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.280494689941406, "learning_rate": 5e-05, "loss": 0.919, "mean_token_accuracy": 0.8506795167922974, "num_tokens": 283545117.0, "step": 7471 }, { "epoch": 0.9505152016282916, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.813962936401367, "learning_rate": 5e-05, "loss": 0.7902, "mean_token_accuracy": 0.8526471257209778, "num_tokens": 283577786.0, "step": 7472 }, { "epoch": 0.950642411906882, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.616063117980957, "learning_rate": 5e-05, "loss": 0.8891, "mean_token_accuracy": 0.8678142428398132, "num_tokens": 283613486.0, "step": 7473 }, { "epoch": 0.9507696221854726, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.673671722412109, "learning_rate": 5e-05, "loss": 0.7531, "mean_token_accuracy": 0.8620727062225342, "num_tokens": 283654480.0, "step": 7474 }, { "epoch": 0.9508968324640631, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.777002334594727, "learning_rate": 5e-05, "loss": 0.9511, "mean_token_accuracy": 0.8562113046646118, "num_tokens": 283686089.0, "step": 7475 }, { "epoch": 0.9510240427426536, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.811969757080078, "learning_rate": 5e-05, "loss": 0.7749, "mean_token_accuracy": 0.8631258606910706, "num_tokens": 283726483.0, "step": 7476 }, { "epoch": 0.9511512530212441, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.656890869140625, "learning_rate": 5e-05, "loss": 0.9274, "mean_token_accuracy": 0.8684441447257996, "num_tokens": 283756170.0, "step": 7477 }, { "epoch": 0.9512784632998347, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.220637321472168, "learning_rate": 5e-05, "loss": 0.8403, "mean_token_accuracy": 0.8628112077713013, "num_tokens": 283794950.0, "step": 7478 }, { "epoch": 0.9514056735784251, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.139187812805176, "learning_rate": 5e-05, "loss": 0.9768, "mean_token_accuracy": 0.8409017324447632, "num_tokens": 283835596.0, "step": 7479 }, { "epoch": 0.9515328838570156, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.645039081573486, "learning_rate": 5e-05, "loss": 0.8952, "mean_token_accuracy": 0.8553289175033569, "num_tokens": 283875380.0, "step": 7480 }, { "epoch": 0.9516600941356061, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.585498809814453, "learning_rate": 5e-05, "loss": 0.8342, "mean_token_accuracy": 0.8568712472915649, "num_tokens": 283910026.0, "step": 7481 }, { "epoch": 0.9517873044141967, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.862875938415527, "learning_rate": 5e-05, "loss": 0.8415, "mean_token_accuracy": 0.8676122426986694, "num_tokens": 283950533.0, "step": 7482 }, { "epoch": 0.9519145146927872, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.316656112670898, "learning_rate": 5e-05, "loss": 0.7897, "mean_token_accuracy": 0.8635891079902649, "num_tokens": 283980939.0, "step": 7483 }, { "epoch": 0.9520417249713777, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.8702569007873535, "learning_rate": 5e-05, "loss": 0.8209, "mean_token_accuracy": 0.8633906841278076, "num_tokens": 284018923.0, "step": 7484 }, { "epoch": 0.9521689352499682, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.245529651641846, "learning_rate": 5e-05, "loss": 0.7167, "mean_token_accuracy": 0.8790974020957947, "num_tokens": 284059752.0, "step": 7485 }, { "epoch": 0.9522961455285587, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.880649089813232, "learning_rate": 5e-05, "loss": 0.7965, "mean_token_accuracy": 0.8711358308792114, "num_tokens": 284103783.0, "step": 7486 }, { "epoch": 0.9524233558071492, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.148447513580322, "learning_rate": 5e-05, "loss": 0.8334, "mean_token_accuracy": 0.841140866279602, "num_tokens": 284144595.0, "step": 7487 }, { "epoch": 0.9525505660857397, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.968395709991455, "learning_rate": 5e-05, "loss": 0.8323, "mean_token_accuracy": 0.858418881893158, "num_tokens": 284182373.0, "step": 7488 }, { "epoch": 0.9526777763643303, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.016297340393066, "learning_rate": 5e-05, "loss": 0.8125, "mean_token_accuracy": 0.8402348756790161, "num_tokens": 284222537.0, "step": 7489 }, { "epoch": 0.9528049866429208, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.089871406555176, "learning_rate": 5e-05, "loss": 0.8414, "mean_token_accuracy": 0.8551865220069885, "num_tokens": 284264780.0, "step": 7490 }, { "epoch": 0.9529321969215112, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.872899055480957, "learning_rate": 5e-05, "loss": 0.6988, "mean_token_accuracy": 0.8756815195083618, "num_tokens": 284298663.0, "step": 7491 }, { "epoch": 0.9530594072001017, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.263511657714844, "learning_rate": 5e-05, "loss": 0.8037, "mean_token_accuracy": 0.8688222765922546, "num_tokens": 284330516.0, "step": 7492 }, { "epoch": 0.9531866174786923, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.742433071136475, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8503494262695312, "num_tokens": 284370414.0, "step": 7493 }, { "epoch": 0.9533138277572828, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.450798988342285, "learning_rate": 5e-05, "loss": 0.7894, "mean_token_accuracy": 0.8847168684005737, "num_tokens": 284409079.0, "step": 7494 }, { "epoch": 0.9534410380358733, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.631044864654541, "learning_rate": 5e-05, "loss": 0.7044, "mean_token_accuracy": 0.8710802793502808, "num_tokens": 284451937.0, "step": 7495 }, { "epoch": 0.9535682483144639, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.617084503173828, "learning_rate": 5e-05, "loss": 0.9077, "mean_token_accuracy": 0.8575561046600342, "num_tokens": 284490174.0, "step": 7496 }, { "epoch": 0.9536954585930543, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.667769908905029, "learning_rate": 5e-05, "loss": 0.7215, "mean_token_accuracy": 0.8704813718795776, "num_tokens": 284530578.0, "step": 7497 }, { "epoch": 0.9538226688716448, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.585758209228516, "learning_rate": 5e-05, "loss": 0.9376, "mean_token_accuracy": 0.8493841290473938, "num_tokens": 284562412.0, "step": 7498 }, { "epoch": 0.9539498791502353, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.95200777053833, "learning_rate": 5e-05, "loss": 0.7822, "mean_token_accuracy": 0.8617956638336182, "num_tokens": 284601846.0, "step": 7499 }, { "epoch": 0.9540770894288259, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.202662467956543, "learning_rate": 5e-05, "loss": 0.8841, "mean_token_accuracy": 0.8564507961273193, "num_tokens": 284640918.0, "step": 7500 }, { "epoch": 0.9542042997074164, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.391607761383057, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8707441687583923, "num_tokens": 284672268.0, "step": 7501 }, { "epoch": 0.9543315099860069, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.625842571258545, "learning_rate": 5e-05, "loss": 0.8077, "mean_token_accuracy": 0.8690524101257324, "num_tokens": 284708638.0, "step": 7502 }, { "epoch": 0.9544587202645973, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.809183597564697, "learning_rate": 5e-05, "loss": 0.9004, "mean_token_accuracy": 0.8420307040214539, "num_tokens": 284747670.0, "step": 7503 }, { "epoch": 0.9545859305431879, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.1534271240234375, "learning_rate": 5e-05, "loss": 0.7818, "mean_token_accuracy": 0.858407735824585, "num_tokens": 284786561.0, "step": 7504 }, { "epoch": 0.9547131408217784, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.033039093017578, "learning_rate": 5e-05, "loss": 0.8214, "mean_token_accuracy": 0.8664494752883911, "num_tokens": 284828631.0, "step": 7505 }, { "epoch": 0.9548403511003689, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.985483646392822, "learning_rate": 5e-05, "loss": 0.7391, "mean_token_accuracy": 0.8638190031051636, "num_tokens": 284865782.0, "step": 7506 }, { "epoch": 0.9549675613789594, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.050577163696289, "learning_rate": 5e-05, "loss": 0.7994, "mean_token_accuracy": 0.8691074252128601, "num_tokens": 284904000.0, "step": 7507 }, { "epoch": 0.95509477165755, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.011687755584717, "learning_rate": 5e-05, "loss": 0.7418, "mean_token_accuracy": 0.8654621839523315, "num_tokens": 284946432.0, "step": 7508 }, { "epoch": 0.9552219819361404, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.967879295349121, "learning_rate": 5e-05, "loss": 0.8186, "mean_token_accuracy": 0.8638830780982971, "num_tokens": 284985534.0, "step": 7509 }, { "epoch": 0.9553491922147309, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.06222677230835, "learning_rate": 5e-05, "loss": 0.7093, "mean_token_accuracy": 0.875278890132904, "num_tokens": 285023710.0, "step": 7510 }, { "epoch": 0.9554764024933214, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.903910160064697, "learning_rate": 5e-05, "loss": 0.8319, "mean_token_accuracy": 0.8494466543197632, "num_tokens": 285064182.0, "step": 7511 }, { "epoch": 0.955603612771912, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.085537910461426, "learning_rate": 5e-05, "loss": 0.721, "mean_token_accuracy": 0.8683189749717712, "num_tokens": 285100896.0, "step": 7512 }, { "epoch": 0.9557308230505025, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.885270118713379, "learning_rate": 5e-05, "loss": 0.7977, "mean_token_accuracy": 0.8629922866821289, "num_tokens": 285139347.0, "step": 7513 }, { "epoch": 0.955858033329093, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.060919761657715, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8611994981765747, "num_tokens": 285183047.0, "step": 7514 }, { "epoch": 0.9559852436076836, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.925330638885498, "learning_rate": 5e-05, "loss": 0.7979, "mean_token_accuracy": 0.8616546988487244, "num_tokens": 285225407.0, "step": 7515 }, { "epoch": 0.956112453886274, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.997900009155273, "learning_rate": 5e-05, "loss": 0.7038, "mean_token_accuracy": 0.8687119483947754, "num_tokens": 285258465.0, "step": 7516 }, { "epoch": 0.9562396641648645, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.01751708984375, "learning_rate": 5e-05, "loss": 0.8051, "mean_token_accuracy": 0.8628697395324707, "num_tokens": 285297221.0, "step": 7517 }, { "epoch": 0.956366874443455, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.839174270629883, "learning_rate": 5e-05, "loss": 0.7206, "mean_token_accuracy": 0.8672561049461365, "num_tokens": 285338145.0, "step": 7518 }, { "epoch": 0.9564940847220456, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.261072158813477, "learning_rate": 5e-05, "loss": 0.8142, "mean_token_accuracy": 0.8645178079605103, "num_tokens": 285381647.0, "step": 7519 }, { "epoch": 0.9566212950006361, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.586121559143066, "learning_rate": 5e-05, "loss": 0.7746, "mean_token_accuracy": 0.8421086072921753, "num_tokens": 285424468.0, "step": 7520 }, { "epoch": 0.9567485052792266, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.619048118591309, "learning_rate": 5e-05, "loss": 0.8561, "mean_token_accuracy": 0.8675099611282349, "num_tokens": 285470415.0, "step": 7521 }, { "epoch": 0.956875715557817, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.326144218444824, "learning_rate": 5e-05, "loss": 0.7379, "mean_token_accuracy": 0.8503640294075012, "num_tokens": 285513299.0, "step": 7522 }, { "epoch": 0.9570029258364076, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 9.021161079406738, "learning_rate": 5e-05, "loss": 0.9822, "mean_token_accuracy": 0.8476314544677734, "num_tokens": 285555855.0, "step": 7523 }, { "epoch": 0.9571301361149981, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.312811851501465, "learning_rate": 5e-05, "loss": 0.7208, "mean_token_accuracy": 0.8652090430259705, "num_tokens": 285589670.0, "step": 7524 }, { "epoch": 0.9572573463935886, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.104671478271484, "learning_rate": 5e-05, "loss": 1.0249, "mean_token_accuracy": 0.8486601114273071, "num_tokens": 285627257.0, "step": 7525 }, { "epoch": 0.9573845566721791, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 6.938096523284912, "learning_rate": 5e-05, "loss": 0.8313, "mean_token_accuracy": 0.8673794269561768, "num_tokens": 285670100.0, "step": 7526 }, { "epoch": 0.9575117669507697, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.291691780090332, "learning_rate": 5e-05, "loss": 0.9062, "mean_token_accuracy": 0.8669352531433105, "num_tokens": 285706902.0, "step": 7527 }, { "epoch": 0.9576389772293601, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.953415393829346, "learning_rate": 5e-05, "loss": 0.9413, "mean_token_accuracy": 0.8599693775177002, "num_tokens": 285746952.0, "step": 7528 }, { "epoch": 0.9577661875079506, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.92736291885376, "learning_rate": 5e-05, "loss": 0.7342, "mean_token_accuracy": 0.8762050867080688, "num_tokens": 285781532.0, "step": 7529 }, { "epoch": 0.9578933977865411, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.949700355529785, "learning_rate": 5e-05, "loss": 0.9786, "mean_token_accuracy": 0.8565486669540405, "num_tokens": 285818725.0, "step": 7530 }, { "epoch": 0.9580206080651317, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.040038585662842, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8544619679450989, "num_tokens": 285855502.0, "step": 7531 }, { "epoch": 0.9581478183437222, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.821979522705078, "learning_rate": 5e-05, "loss": 1.055, "mean_token_accuracy": 0.8587226867675781, "num_tokens": 285893815.0, "step": 7532 }, { "epoch": 0.9582750286223127, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.257288455963135, "learning_rate": 5e-05, "loss": 0.8062, "mean_token_accuracy": 0.8594428300857544, "num_tokens": 285928709.0, "step": 7533 }, { "epoch": 0.9584022389009031, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 9.99289321899414, "learning_rate": 5e-05, "loss": 1.0386, "mean_token_accuracy": 0.8782407641410828, "num_tokens": 285959358.0, "step": 7534 }, { "epoch": 0.9585294491794937, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 7.734545707702637, "learning_rate": 5e-05, "loss": 1.0719, "mean_token_accuracy": 0.8457566499710083, "num_tokens": 285997643.0, "step": 7535 }, { "epoch": 0.9586566594580842, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.72933292388916, "learning_rate": 5e-05, "loss": 0.98, "mean_token_accuracy": 0.8667995929718018, "num_tokens": 286032866.0, "step": 7536 }, { "epoch": 0.9587838697366747, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 8.797788619995117, "learning_rate": 5e-05, "loss": 1.0673, "mean_token_accuracy": 0.8634171485900879, "num_tokens": 286066585.0, "step": 7537 }, { "epoch": 0.9589110800152653, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.122785568237305, "learning_rate": 5e-05, "loss": 0.878, "mean_token_accuracy": 0.8468307852745056, "num_tokens": 286106151.0, "step": 7538 }, { "epoch": 0.9590382902938558, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.443390846252441, "learning_rate": 5e-05, "loss": 1.022, "mean_token_accuracy": 0.8675633668899536, "num_tokens": 286142409.0, "step": 7539 }, { "epoch": 0.9591655005724462, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.043691635131836, "learning_rate": 5e-05, "loss": 0.929, "mean_token_accuracy": 0.8523620367050171, "num_tokens": 286179408.0, "step": 7540 }, { "epoch": 0.9592927108510367, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.974873542785645, "learning_rate": 5e-05, "loss": 1.0182, "mean_token_accuracy": 0.861602783203125, "num_tokens": 286212767.0, "step": 7541 }, { "epoch": 0.9594199211296273, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 7.87903356552124, "learning_rate": 5e-05, "loss": 0.9887, "mean_token_accuracy": 0.8611108064651489, "num_tokens": 286250176.0, "step": 7542 }, { "epoch": 0.9595471314082178, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.706284046173096, "learning_rate": 5e-05, "loss": 0.8667, "mean_token_accuracy": 0.8590244054794312, "num_tokens": 286292060.0, "step": 7543 }, { "epoch": 0.9596743416868083, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.51040267944336, "learning_rate": 5e-05, "loss": 0.9649, "mean_token_accuracy": 0.8681854605674744, "num_tokens": 286333047.0, "step": 7544 }, { "epoch": 0.9598015519653988, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.023298263549805, "learning_rate": 5e-05, "loss": 0.7845, "mean_token_accuracy": 0.8651031255722046, "num_tokens": 286373465.0, "step": 7545 }, { "epoch": 0.9599287622439893, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.711076736450195, "learning_rate": 5e-05, "loss": 1.0202, "mean_token_accuracy": 0.8501261472702026, "num_tokens": 286414431.0, "step": 7546 }, { "epoch": 0.9600559725225798, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.050113677978516, "learning_rate": 5e-05, "loss": 0.8461, "mean_token_accuracy": 0.8607994318008423, "num_tokens": 286449431.0, "step": 7547 }, { "epoch": 0.9601831828011703, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.569342613220215, "learning_rate": 5e-05, "loss": 0.9334, "mean_token_accuracy": 0.8643627166748047, "num_tokens": 286486563.0, "step": 7548 }, { "epoch": 0.9603103930797608, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.234094142913818, "learning_rate": 5e-05, "loss": 0.9164, "mean_token_accuracy": 0.8524272441864014, "num_tokens": 286528691.0, "step": 7549 }, { "epoch": 0.9604376033583514, "ewc_loss": 0.5, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.288495063781738, "learning_rate": 5e-05, "loss": 0.9193, "mean_token_accuracy": 0.8595694303512573, "num_tokens": 286568726.0, "step": 7550 }, { "epoch": 0.9605648136369419, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.312699317932129, "learning_rate": 5e-05, "loss": 0.8465, "mean_token_accuracy": 0.8702911734580994, "num_tokens": 286608159.0, "step": 7551 }, { "epoch": 0.9606920239155323, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.066143035888672, "learning_rate": 5e-05, "loss": 0.9022, "mean_token_accuracy": 0.8575516939163208, "num_tokens": 286645494.0, "step": 7552 }, { "epoch": 0.9608192341941229, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.255645751953125, "learning_rate": 5e-05, "loss": 0.7956, "mean_token_accuracy": 0.8704639673233032, "num_tokens": 286679819.0, "step": 7553 }, { "epoch": 0.9609464444727134, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.036839485168457, "learning_rate": 5e-05, "loss": 0.9047, "mean_token_accuracy": 0.8468501567840576, "num_tokens": 286714116.0, "step": 7554 }, { "epoch": 0.9610736547513039, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.047369003295898, "learning_rate": 5e-05, "loss": 0.7502, "mean_token_accuracy": 0.8752332925796509, "num_tokens": 286746905.0, "step": 7555 }, { "epoch": 0.9612008650298944, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.21549129486084, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.882762610912323, "num_tokens": 286784724.0, "step": 7556 }, { "epoch": 0.961328075308485, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.791475296020508, "learning_rate": 5e-05, "loss": 0.7383, "mean_token_accuracy": 0.8672113418579102, "num_tokens": 286825672.0, "step": 7557 }, { "epoch": 0.9614552855870754, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.484810829162598, "learning_rate": 5e-05, "loss": 0.9466, "mean_token_accuracy": 0.8487969636917114, "num_tokens": 286862072.0, "step": 7558 }, { "epoch": 0.9615824958656659, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.588362216949463, "learning_rate": 5e-05, "loss": 0.7445, "mean_token_accuracy": 0.8650029897689819, "num_tokens": 286903111.0, "step": 7559 }, { "epoch": 0.9617097061442564, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.758495330810547, "learning_rate": 5e-05, "loss": 0.9137, "mean_token_accuracy": 0.866517186164856, "num_tokens": 286947957.0, "step": 7560 }, { "epoch": 0.961836916422847, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.523377418518066, "learning_rate": 5e-05, "loss": 0.7466, "mean_token_accuracy": 0.8613287210464478, "num_tokens": 286983254.0, "step": 7561 }, { "epoch": 0.9619641267014375, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.953582763671875, "learning_rate": 5e-05, "loss": 0.8998, "mean_token_accuracy": 0.8822441697120667, "num_tokens": 287015294.0, "step": 7562 }, { "epoch": 0.962091336980028, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 6.7440338134765625, "learning_rate": 5e-05, "loss": 0.7831, "mean_token_accuracy": 0.8681954145431519, "num_tokens": 287053439.0, "step": 7563 }, { "epoch": 0.9622185472586186, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.7686185836792, "learning_rate": 5e-05, "loss": 1.0403, "mean_token_accuracy": 0.8405852317810059, "num_tokens": 287096970.0, "step": 7564 }, { "epoch": 0.962345757537209, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.264009475708008, "learning_rate": 5e-05, "loss": 0.8983, "mean_token_accuracy": 0.8566412925720215, "num_tokens": 287139179.0, "step": 7565 }, { "epoch": 0.9624729678157995, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.130331039428711, "learning_rate": 5e-05, "loss": 0.9281, "mean_token_accuracy": 0.8556185960769653, "num_tokens": 287177375.0, "step": 7566 }, { "epoch": 0.96260017809439, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.746628761291504, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.8656165599822998, "num_tokens": 287211296.0, "step": 7567 }, { "epoch": 0.9627273883729806, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.428391933441162, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.8695635795593262, "num_tokens": 287246913.0, "step": 7568 }, { "epoch": 0.9628545986515711, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.040630340576172, "learning_rate": 5e-05, "loss": 0.9109, "mean_token_accuracy": 0.8590790033340454, "num_tokens": 287283639.0, "step": 7569 }, { "epoch": 0.9629818089301616, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.0495758056640625, "learning_rate": 5e-05, "loss": 0.7628, "mean_token_accuracy": 0.8703968524932861, "num_tokens": 287322694.0, "step": 7570 }, { "epoch": 0.963109019208752, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.188505172729492, "learning_rate": 5e-05, "loss": 0.8973, "mean_token_accuracy": 0.8521541357040405, "num_tokens": 287361103.0, "step": 7571 }, { "epoch": 0.9632362294873426, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.904605865478516, "learning_rate": 5e-05, "loss": 0.8133, "mean_token_accuracy": 0.8458384275436401, "num_tokens": 287400666.0, "step": 7572 }, { "epoch": 0.9633634397659331, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.305120468139648, "learning_rate": 5e-05, "loss": 0.8254, "mean_token_accuracy": 0.8786722421646118, "num_tokens": 287434729.0, "step": 7573 }, { "epoch": 0.9634906500445236, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.768080711364746, "learning_rate": 5e-05, "loss": 0.7111, "mean_token_accuracy": 0.8740693926811218, "num_tokens": 287472194.0, "step": 7574 }, { "epoch": 0.9636178603231141, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.499096870422363, "learning_rate": 5e-05, "loss": 0.9114, "mean_token_accuracy": 0.8563769459724426, "num_tokens": 287510139.0, "step": 7575 }, { "epoch": 0.9637450706017047, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.673312664031982, "learning_rate": 5e-05, "loss": 0.7909, "mean_token_accuracy": 0.8508853912353516, "num_tokens": 287549529.0, "step": 7576 }, { "epoch": 0.9638722808802951, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.639203071594238, "learning_rate": 5e-05, "loss": 0.9603, "mean_token_accuracy": 0.8475049734115601, "num_tokens": 287596525.0, "step": 7577 }, { "epoch": 0.9639994911588856, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.75023889541626, "learning_rate": 5e-05, "loss": 0.7336, "mean_token_accuracy": 0.8767679929733276, "num_tokens": 287632284.0, "step": 7578 }, { "epoch": 0.9641267014374761, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.584432601928711, "learning_rate": 5e-05, "loss": 0.9196, "mean_token_accuracy": 0.8684052228927612, "num_tokens": 287671867.0, "step": 7579 }, { "epoch": 0.9642539117160667, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.012157440185547, "learning_rate": 5e-05, "loss": 0.8498, "mean_token_accuracy": 0.8531707525253296, "num_tokens": 287706715.0, "step": 7580 }, { "epoch": 0.9643811219946572, "ewc_loss": 0.5, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.26005744934082, "learning_rate": 5e-05, "loss": 0.8974, "mean_token_accuracy": 0.8641910552978516, "num_tokens": 287746269.0, "step": 7581 }, { "epoch": 0.9645083322732477, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.304810047149658, "learning_rate": 5e-05, "loss": 0.8857, "mean_token_accuracy": 0.8528528809547424, "num_tokens": 287786194.0, "step": 7582 }, { "epoch": 0.9646355425518381, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.934871673583984, "learning_rate": 5e-05, "loss": 0.8891, "mean_token_accuracy": 0.8546202182769775, "num_tokens": 287820873.0, "step": 7583 }, { "epoch": 0.9647627528304287, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.427224159240723, "learning_rate": 5e-05, "loss": 0.8322, "mean_token_accuracy": 0.8634167313575745, "num_tokens": 287847356.0, "step": 7584 }, { "epoch": 0.9648899631090192, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.777492046356201, "learning_rate": 5e-05, "loss": 0.8709, "mean_token_accuracy": 0.8569685220718384, "num_tokens": 287879611.0, "step": 7585 }, { "epoch": 0.9650171733876097, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.315950870513916, "learning_rate": 5e-05, "loss": 0.8621, "mean_token_accuracy": 0.8498953580856323, "num_tokens": 287920865.0, "step": 7586 }, { "epoch": 0.9651443836662003, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.88274621963501, "learning_rate": 5e-05, "loss": 0.8621, "mean_token_accuracy": 0.8553644418716431, "num_tokens": 287963997.0, "step": 7587 }, { "epoch": 0.9652715939447908, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.096686840057373, "learning_rate": 5e-05, "loss": 0.7967, "mean_token_accuracy": 0.8540177345275879, "num_tokens": 287997283.0, "step": 7588 }, { "epoch": 0.9653988042233812, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.068826675415039, "learning_rate": 5e-05, "loss": 0.9212, "mean_token_accuracy": 0.8353763818740845, "num_tokens": 288037876.0, "step": 7589 }, { "epoch": 0.9655260145019717, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.898090839385986, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.8466081023216248, "num_tokens": 288078835.0, "step": 7590 }, { "epoch": 0.9656532247805623, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.245061874389648, "learning_rate": 5e-05, "loss": 0.8887, "mean_token_accuracy": 0.856230616569519, "num_tokens": 288118342.0, "step": 7591 }, { "epoch": 0.9657804350591528, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.774261951446533, "learning_rate": 5e-05, "loss": 0.7492, "mean_token_accuracy": 0.8636739253997803, "num_tokens": 288159629.0, "step": 7592 }, { "epoch": 0.9659076453377433, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.402631759643555, "learning_rate": 5e-05, "loss": 0.9058, "mean_token_accuracy": 0.8554542064666748, "num_tokens": 288199887.0, "step": 7593 }, { "epoch": 0.9660348556163338, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.779187202453613, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8508156538009644, "num_tokens": 288233958.0, "step": 7594 }, { "epoch": 0.9661620658949243, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.354415893554688, "learning_rate": 5e-05, "loss": 0.8905, "mean_token_accuracy": 0.8604140877723694, "num_tokens": 288274735.0, "step": 7595 }, { "epoch": 0.9662892761735148, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.995731353759766, "learning_rate": 5e-05, "loss": 0.7738, "mean_token_accuracy": 0.8646667003631592, "num_tokens": 288309699.0, "step": 7596 }, { "epoch": 0.9664164864521053, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.03779125213623, "learning_rate": 5e-05, "loss": 0.8643, "mean_token_accuracy": 0.862095057964325, "num_tokens": 288344397.0, "step": 7597 }, { "epoch": 0.9665436967306958, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.360499382019043, "learning_rate": 5e-05, "loss": 0.8071, "mean_token_accuracy": 0.8680205941200256, "num_tokens": 288385612.0, "step": 7598 }, { "epoch": 0.9666709070092864, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.541773796081543, "learning_rate": 5e-05, "loss": 0.8652, "mean_token_accuracy": 0.8468772172927856, "num_tokens": 288420570.0, "step": 7599 }, { "epoch": 0.9667981172878769, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.731943607330322, "learning_rate": 5e-05, "loss": 0.8814, "mean_token_accuracy": 0.8482265472412109, "num_tokens": 288456917.0, "step": 7600 }, { "epoch": 0.9669253275664673, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.099370956420898, "learning_rate": 5e-05, "loss": 0.7985, "mean_token_accuracy": 0.8462849259376526, "num_tokens": 288498228.0, "step": 7601 }, { "epoch": 0.9670525378450578, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.04078197479248, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.8413225412368774, "num_tokens": 288534082.0, "step": 7602 }, { "epoch": 0.9671797481236484, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.753567218780518, "learning_rate": 5e-05, "loss": 0.7504, "mean_token_accuracy": 0.8560792207717896, "num_tokens": 288576142.0, "step": 7603 }, { "epoch": 0.9673069584022389, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.379610061645508, "learning_rate": 5e-05, "loss": 0.8683, "mean_token_accuracy": 0.8550310134887695, "num_tokens": 288609707.0, "step": 7604 }, { "epoch": 0.9674341686808294, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.479410648345947, "learning_rate": 5e-05, "loss": 0.7812, "mean_token_accuracy": 0.8455252647399902, "num_tokens": 288648228.0, "step": 7605 }, { "epoch": 0.96756137895942, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.71151351928711, "learning_rate": 5e-05, "loss": 0.8792, "mean_token_accuracy": 0.866389274597168, "num_tokens": 288689845.0, "step": 7606 }, { "epoch": 0.9676885892380104, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.320668697357178, "learning_rate": 5e-05, "loss": 0.68, "mean_token_accuracy": 0.8735783100128174, "num_tokens": 288729571.0, "step": 7607 }, { "epoch": 0.9678157995166009, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.99942684173584, "learning_rate": 5e-05, "loss": 0.9102, "mean_token_accuracy": 0.8748732805252075, "num_tokens": 288764070.0, "step": 7608 }, { "epoch": 0.9679430097951914, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.518977642059326, "learning_rate": 5e-05, "loss": 0.7579, "mean_token_accuracy": 0.8642475008964539, "num_tokens": 288805557.0, "step": 7609 }, { "epoch": 0.968070220073782, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.819482803344727, "learning_rate": 5e-05, "loss": 0.9407, "mean_token_accuracy": 0.8707584142684937, "num_tokens": 288840567.0, "step": 7610 }, { "epoch": 0.9681974303523725, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.188395977020264, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.8592169284820557, "num_tokens": 288881952.0, "step": 7611 }, { "epoch": 0.968324640630963, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.9863810539245605, "learning_rate": 5e-05, "loss": 0.9046, "mean_token_accuracy": 0.8596484661102295, "num_tokens": 288919208.0, "step": 7612 }, { "epoch": 0.9684518509095535, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.898327350616455, "learning_rate": 5e-05, "loss": 0.9065, "mean_token_accuracy": 0.8616875410079956, "num_tokens": 288961622.0, "step": 7613 }, { "epoch": 0.968579061188144, "ewc_loss": 0.375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.100532531738281, "learning_rate": 5e-05, "loss": 0.7507, "mean_token_accuracy": 0.8710206151008606, "num_tokens": 288999799.0, "step": 7614 }, { "epoch": 0.9687062714667345, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.323819160461426, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.8716472387313843, "num_tokens": 289040434.0, "step": 7615 }, { "epoch": 0.968833481745325, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.820723533630371, "learning_rate": 5e-05, "loss": 0.7569, "mean_token_accuracy": 0.8600767850875854, "num_tokens": 289072847.0, "step": 7616 }, { "epoch": 0.9689606920239155, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.227240562438965, "learning_rate": 5e-05, "loss": 0.9346, "mean_token_accuracy": 0.8480310440063477, "num_tokens": 289114838.0, "step": 7617 }, { "epoch": 0.9690879023025061, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.049443244934082, "learning_rate": 5e-05, "loss": 0.7986, "mean_token_accuracy": 0.8553630709648132, "num_tokens": 289151937.0, "step": 7618 }, { "epoch": 0.9692151125810966, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.833645343780518, "learning_rate": 5e-05, "loss": 0.8246, "mean_token_accuracy": 0.8626374006271362, "num_tokens": 289188630.0, "step": 7619 }, { "epoch": 0.969342322859687, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.373276710510254, "learning_rate": 5e-05, "loss": 0.7959, "mean_token_accuracy": 0.8642755746841431, "num_tokens": 289229412.0, "step": 7620 }, { "epoch": 0.9694695331382776, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.482507705688477, "learning_rate": 5e-05, "loss": 0.7504, "mean_token_accuracy": 0.8710033893585205, "num_tokens": 289265195.0, "step": 7621 }, { "epoch": 0.9695967434168681, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.534989356994629, "learning_rate": 5e-05, "loss": 0.8192, "mean_token_accuracy": 0.8530654311180115, "num_tokens": 289306203.0, "step": 7622 }, { "epoch": 0.9697239536954586, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.3385491371154785, "learning_rate": 5e-05, "loss": 0.7938, "mean_token_accuracy": 0.8555020093917847, "num_tokens": 289345196.0, "step": 7623 }, { "epoch": 0.9698511639740491, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.573897361755371, "learning_rate": 5e-05, "loss": 0.7607, "mean_token_accuracy": 0.872810423374176, "num_tokens": 289377221.0, "step": 7624 }, { "epoch": 0.9699783742526397, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.295542240142822, "learning_rate": 5e-05, "loss": 0.7848, "mean_token_accuracy": 0.8537924289703369, "num_tokens": 289413196.0, "step": 7625 }, { "epoch": 0.9701055845312301, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.563083648681641, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.8546577095985413, "num_tokens": 289452774.0, "step": 7626 }, { "epoch": 0.9702327948098206, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.262013912200928, "learning_rate": 5e-05, "loss": 0.7435, "mean_token_accuracy": 0.8619455099105835, "num_tokens": 289489806.0, "step": 7627 }, { "epoch": 0.9703600050884111, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.592100143432617, "learning_rate": 5e-05, "loss": 0.7771, "mean_token_accuracy": 0.8619312047958374, "num_tokens": 289528362.0, "step": 7628 }, { "epoch": 0.9704872153670017, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.187508583068848, "learning_rate": 5e-05, "loss": 0.7261, "mean_token_accuracy": 0.8633266091346741, "num_tokens": 289562725.0, "step": 7629 }, { "epoch": 0.9706144256455922, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.66499662399292, "learning_rate": 5e-05, "loss": 0.7946, "mean_token_accuracy": 0.8571555018424988, "num_tokens": 289605968.0, "step": 7630 }, { "epoch": 0.9707416359241827, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.0462327003479, "learning_rate": 5e-05, "loss": 0.7496, "mean_token_accuracy": 0.854356586933136, "num_tokens": 289639065.0, "step": 7631 }, { "epoch": 0.9708688462027731, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.847841739654541, "learning_rate": 5e-05, "loss": 0.8533, "mean_token_accuracy": 0.8395649194717407, "num_tokens": 289679616.0, "step": 7632 }, { "epoch": 0.9709960564813637, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.796607494354248, "learning_rate": 5e-05, "loss": 0.7195, "mean_token_accuracy": 0.861090362071991, "num_tokens": 289713189.0, "step": 7633 }, { "epoch": 0.9711232667599542, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.21924114227295, "learning_rate": 5e-05, "loss": 0.7989, "mean_token_accuracy": 0.8643246293067932, "num_tokens": 289751406.0, "step": 7634 }, { "epoch": 0.9712504770385447, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00024318695068359375, "grad_norm": 6.353580951690674, "learning_rate": 5e-05, "loss": 0.6878, "mean_token_accuracy": 0.8610724806785583, "num_tokens": 289795735.0, "step": 7635 }, { "epoch": 0.9713776873171353, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.94094181060791, "learning_rate": 5e-05, "loss": 0.9265, "mean_token_accuracy": 0.8500860929489136, "num_tokens": 289833574.0, "step": 7636 }, { "epoch": 0.9715048975957258, "ewc_loss": 0.265625, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00022220611572265625, "grad_norm": 5.695430278778076, "learning_rate": 5e-05, "loss": 0.6375, "mean_token_accuracy": 0.8725075125694275, "num_tokens": 289871922.0, "step": 7637 }, { "epoch": 0.9716321078743162, "ewc_loss": 0.625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 10.119379043579102, "learning_rate": 5e-05, "loss": 1.0668, "mean_token_accuracy": 0.8545014262199402, "num_tokens": 289913517.0, "step": 7638 }, { "epoch": 0.9717593181529067, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 5.672857761383057, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.8596108555793762, "num_tokens": 289957046.0, "step": 7639 }, { "epoch": 0.9718865284314973, "ewc_loss": 0.85546875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 11.197864532470703, "learning_rate": 5e-05, "loss": 1.2703, "mean_token_accuracy": 0.8618912696838379, "num_tokens": 289997983.0, "step": 7640 }, { "epoch": 0.9720137387100878, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 7.973360538482666, "learning_rate": 5e-05, "loss": 1.0902, "mean_token_accuracy": 0.8643377423286438, "num_tokens": 290035923.0, "step": 7641 }, { "epoch": 0.9721409489886783, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.24194622039795, "learning_rate": 5e-05, "loss": 1.0483, "mean_token_accuracy": 0.8696097135543823, "num_tokens": 290075152.0, "step": 7642 }, { "epoch": 0.9722681592672688, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000865936279296875, "grad_norm": 9.39035415649414, "learning_rate": 5e-05, "loss": 1.2706, "mean_token_accuracy": 0.8768284320831299, "num_tokens": 290108360.0, "step": 7643 }, { "epoch": 0.9723953695458593, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.065605163574219, "learning_rate": 5e-05, "loss": 0.8552, "mean_token_accuracy": 0.869812548160553, "num_tokens": 290148698.0, "step": 7644 }, { "epoch": 0.9725225798244498, "ewc_loss": 0.8828125, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 10.742286682128906, "learning_rate": 5e-05, "loss": 1.2662, "mean_token_accuracy": 0.868914008140564, "num_tokens": 290185847.0, "step": 7645 }, { "epoch": 0.9726497901030403, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.566458702087402, "learning_rate": 5e-05, "loss": 0.9564, "mean_token_accuracy": 0.8613563776016235, "num_tokens": 290228763.0, "step": 7646 }, { "epoch": 0.9727770003816308, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000865936279296875, "grad_norm": 11.991988182067871, "learning_rate": 5e-05, "loss": 1.3566, "mean_token_accuracy": 0.8564143776893616, "num_tokens": 290262040.0, "step": 7647 }, { "epoch": 0.9729042106602214, "ewc_loss": 0.9453125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000904083251953125, "grad_norm": 9.59804916381836, "learning_rate": 5e-05, "loss": 1.3312, "mean_token_accuracy": 0.8674793243408203, "num_tokens": 290302205.0, "step": 7648 }, { "epoch": 0.9730314209388119, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 11.082416534423828, "learning_rate": 5e-05, "loss": 1.0863, "mean_token_accuracy": 0.8664542436599731, "num_tokens": 290340132.0, "step": 7649 }, { "epoch": 0.9731586312174023, "ewc_loss": 0.82421875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.214170455932617, "learning_rate": 5e-05, "loss": 1.1498, "mean_token_accuracy": 0.8865567445755005, "num_tokens": 290375808.0, "step": 7650 }, { "epoch": 0.9732858414959928, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.818605422973633, "learning_rate": 5e-05, "loss": 1.1832, "mean_token_accuracy": 0.8570417165756226, "num_tokens": 290415538.0, "step": 7651 }, { "epoch": 0.9734130517745834, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.31563663482666, "learning_rate": 5e-05, "loss": 0.9932, "mean_token_accuracy": 0.8684683442115784, "num_tokens": 290453363.0, "step": 7652 }, { "epoch": 0.9735402620531739, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 10.794376373291016, "learning_rate": 5e-05, "loss": 1.1645, "mean_token_accuracy": 0.8536653518676758, "num_tokens": 290483869.0, "step": 7653 }, { "epoch": 0.9736674723317644, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.959816932678223, "learning_rate": 5e-05, "loss": 0.9386, "mean_token_accuracy": 0.8708205819129944, "num_tokens": 290521772.0, "step": 7654 }, { "epoch": 0.973794682610355, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 11.303120613098145, "learning_rate": 5e-05, "loss": 1.225, "mean_token_accuracy": 0.8462401628494263, "num_tokens": 290563226.0, "step": 7655 }, { "epoch": 0.9739218928889454, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 8.660788536071777, "learning_rate": 5e-05, "loss": 1.0533, "mean_token_accuracy": 0.8598396182060242, "num_tokens": 290602209.0, "step": 7656 }, { "epoch": 0.9740491031675359, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 11.383234977722168, "learning_rate": 5e-05, "loss": 1.2019, "mean_token_accuracy": 0.8522144556045532, "num_tokens": 290645554.0, "step": 7657 }, { "epoch": 0.9741763134461264, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 9.430334091186523, "learning_rate": 5e-05, "loss": 1.1083, "mean_token_accuracy": 0.8700587749481201, "num_tokens": 290683982.0, "step": 7658 }, { "epoch": 0.974303523724717, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.698186874389648, "learning_rate": 5e-05, "loss": 1.1405, "mean_token_accuracy": 0.8533841371536255, "num_tokens": 290717720.0, "step": 7659 }, { "epoch": 0.9744307340033075, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.33342456817627, "learning_rate": 5e-05, "loss": 1.0607, "mean_token_accuracy": 0.8613672852516174, "num_tokens": 290759711.0, "step": 7660 }, { "epoch": 0.974557944281898, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.0301513671875, "learning_rate": 5e-05, "loss": 1.0505, "mean_token_accuracy": 0.8656099438667297, "num_tokens": 290795767.0, "step": 7661 }, { "epoch": 0.9746851545604885, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.886530876159668, "learning_rate": 5e-05, "loss": 0.9648, "mean_token_accuracy": 0.8680922985076904, "num_tokens": 290839148.0, "step": 7662 }, { "epoch": 0.974812364839079, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.540095329284668, "learning_rate": 5e-05, "loss": 1.0133, "mean_token_accuracy": 0.8645776510238647, "num_tokens": 290878325.0, "step": 7663 }, { "epoch": 0.9749395751176695, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.305683135986328, "learning_rate": 5e-05, "loss": 0.9417, "mean_token_accuracy": 0.8587878942489624, "num_tokens": 290913053.0, "step": 7664 }, { "epoch": 0.97506678539626, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.550023078918457, "learning_rate": 5e-05, "loss": 0.9875, "mean_token_accuracy": 0.8689787983894348, "num_tokens": 290954984.0, "step": 7665 }, { "epoch": 0.9751939956748505, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.800687789916992, "learning_rate": 5e-05, "loss": 0.8838, "mean_token_accuracy": 0.8548648357391357, "num_tokens": 290996052.0, "step": 7666 }, { "epoch": 0.9753212059534411, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.958061218261719, "learning_rate": 5e-05, "loss": 1.0839, "mean_token_accuracy": 0.8511806726455688, "num_tokens": 291034021.0, "step": 7667 }, { "epoch": 0.9754484162320316, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.914417266845703, "learning_rate": 5e-05, "loss": 0.8377, "mean_token_accuracy": 0.8724488019943237, "num_tokens": 291064216.0, "step": 7668 }, { "epoch": 0.975575626510622, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.111796379089355, "learning_rate": 5e-05, "loss": 1.0586, "mean_token_accuracy": 0.8699573278427124, "num_tokens": 291100119.0, "step": 7669 }, { "epoch": 0.9757028367892125, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.501724243164062, "learning_rate": 5e-05, "loss": 0.9554, "mean_token_accuracy": 0.8618226051330566, "num_tokens": 291136709.0, "step": 7670 }, { "epoch": 0.9758300470678031, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.530405044555664, "learning_rate": 5e-05, "loss": 1.076, "mean_token_accuracy": 0.8440012335777283, "num_tokens": 291175038.0, "step": 7671 }, { "epoch": 0.9759572573463936, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.617164611816406, "learning_rate": 5e-05, "loss": 0.9305, "mean_token_accuracy": 0.8648074865341187, "num_tokens": 291206950.0, "step": 7672 }, { "epoch": 0.9760844676249841, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.774697303771973, "learning_rate": 5e-05, "loss": 0.9706, "mean_token_accuracy": 0.8538843989372253, "num_tokens": 291247847.0, "step": 7673 }, { "epoch": 0.9762116779035747, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.210354804992676, "learning_rate": 5e-05, "loss": 0.819, "mean_token_accuracy": 0.8773477077484131, "num_tokens": 291282262.0, "step": 7674 }, { "epoch": 0.9763388881821651, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.484869003295898, "learning_rate": 5e-05, "loss": 0.9078, "mean_token_accuracy": 0.8639690279960632, "num_tokens": 291315090.0, "step": 7675 }, { "epoch": 0.9764660984607556, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.594335079193115, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8693007826805115, "num_tokens": 291356380.0, "step": 7676 }, { "epoch": 0.9765933087393461, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.692386627197266, "learning_rate": 5e-05, "loss": 0.9242, "mean_token_accuracy": 0.8629120588302612, "num_tokens": 291401133.0, "step": 7677 }, { "epoch": 0.9767205190179367, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0004291534423828125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.169508934020996, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.8652175664901733, "num_tokens": 291439999.0, "step": 7678 }, { "epoch": 0.9768477292965272, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.034735679626465, "learning_rate": 5e-05, "loss": 0.9505, "mean_token_accuracy": 0.8670607209205627, "num_tokens": 291476784.0, "step": 7679 }, { "epoch": 0.9769749395751177, "ewc_loss": 0.375, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.165410041809082, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.8733630180358887, "num_tokens": 291517677.0, "step": 7680 }, { "epoch": 0.9771021498537081, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.190258979797363, "learning_rate": 5e-05, "loss": 0.9568, "mean_token_accuracy": 0.8686530590057373, "num_tokens": 291551838.0, "step": 7681 }, { "epoch": 0.9772293601322987, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.566061973571777, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.8721320033073425, "num_tokens": 291591192.0, "step": 7682 }, { "epoch": 0.9773565704108892, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.863694190979004, "learning_rate": 5e-05, "loss": 0.9784, "mean_token_accuracy": 0.8627389669418335, "num_tokens": 291626885.0, "step": 7683 }, { "epoch": 0.9774837806894797, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.964531421661377, "learning_rate": 5e-05, "loss": 0.8476, "mean_token_accuracy": 0.8639600872993469, "num_tokens": 291666421.0, "step": 7684 }, { "epoch": 0.9776109909680702, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.085692405700684, "learning_rate": 5e-05, "loss": 0.8906, "mean_token_accuracy": 0.8595425486564636, "num_tokens": 291704973.0, "step": 7685 }, { "epoch": 0.9777382012466608, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.994767189025879, "learning_rate": 5e-05, "loss": 0.8451, "mean_token_accuracy": 0.857883095741272, "num_tokens": 291750279.0, "step": 7686 }, { "epoch": 0.9778654115252512, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.610651016235352, "learning_rate": 5e-05, "loss": 0.837, "mean_token_accuracy": 0.8600410223007202, "num_tokens": 291784390.0, "step": 7687 }, { "epoch": 0.9779926218038417, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.797505855560303, "learning_rate": 5e-05, "loss": 0.784, "mean_token_accuracy": 0.8702396154403687, "num_tokens": 291826385.0, "step": 7688 }, { "epoch": 0.9781198320824323, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.568722724914551, "learning_rate": 5e-05, "loss": 0.8146, "mean_token_accuracy": 0.8545690178871155, "num_tokens": 291858129.0, "step": 7689 }, { "epoch": 0.9782470423610228, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.6124773025512695, "learning_rate": 5e-05, "loss": 0.8492, "mean_token_accuracy": 0.8473647832870483, "num_tokens": 291902763.0, "step": 7690 }, { "epoch": 0.9783742526396133, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.503999710083008, "learning_rate": 5e-05, "loss": 0.7283, "mean_token_accuracy": 0.8754663467407227, "num_tokens": 291940686.0, "step": 7691 }, { "epoch": 0.9785014629182038, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.598842144012451, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8648656606674194, "num_tokens": 291976849.0, "step": 7692 }, { "epoch": 0.9786286731967943, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.439701080322266, "learning_rate": 5e-05, "loss": 0.7135, "mean_token_accuracy": 0.8785526156425476, "num_tokens": 292014105.0, "step": 7693 }, { "epoch": 0.9787558834753848, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004253387451171875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.527447700500488, "learning_rate": 5e-05, "loss": 0.7582, "mean_token_accuracy": 0.8656520843505859, "num_tokens": 292055083.0, "step": 7694 }, { "epoch": 0.9788830937539753, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.519375324249268, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.8654577136039734, "num_tokens": 292089713.0, "step": 7695 }, { "epoch": 0.9790103040325658, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.360742092132568, "learning_rate": 5e-05, "loss": 0.7747, "mean_token_accuracy": 0.8520187735557556, "num_tokens": 292128139.0, "step": 7696 }, { "epoch": 0.9791375143111564, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.638530731201172, "learning_rate": 5e-05, "loss": 0.7813, "mean_token_accuracy": 0.8601951599121094, "num_tokens": 292164674.0, "step": 7697 }, { "epoch": 0.9792647245897469, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.209591865539551, "learning_rate": 5e-05, "loss": 0.7167, "mean_token_accuracy": 0.8639580011367798, "num_tokens": 292197175.0, "step": 7698 }, { "epoch": 0.9793919348683373, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.785393714904785, "learning_rate": 5e-05, "loss": 0.8172, "mean_token_accuracy": 0.8480095863342285, "num_tokens": 292237051.0, "step": 7699 }, { "epoch": 0.9795191451469278, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00042724609375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.039896011352539, "learning_rate": 5e-05, "loss": 0.6834, "mean_token_accuracy": 0.8731386661529541, "num_tokens": 292272817.0, "step": 7700 }, { "epoch": 0.9796463554255184, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 8.011894226074219, "learning_rate": 5e-05, "loss": 0.793, "mean_token_accuracy": 0.8632661700248718, "num_tokens": 292310239.0, "step": 7701 }, { "epoch": 0.9797735657041089, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.739335536956787, "learning_rate": 5e-05, "loss": 0.7145, "mean_token_accuracy": 0.8609762191772461, "num_tokens": 292351090.0, "step": 7702 }, { "epoch": 0.9799007759826994, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.44344711303711, "learning_rate": 5e-05, "loss": 0.8964, "mean_token_accuracy": 0.8424113988876343, "num_tokens": 292390500.0, "step": 7703 }, { "epoch": 0.98002798626129, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00023937225341796875, "grad_norm": 6.37889289855957, "learning_rate": 5e-05, "loss": 0.7162, "mean_token_accuracy": 0.8552629947662354, "num_tokens": 292423483.0, "step": 7704 }, { "epoch": 0.9801551965398804, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.921116828918457, "learning_rate": 5e-05, "loss": 0.8436, "mean_token_accuracy": 0.8762092590332031, "num_tokens": 292460791.0, "step": 7705 }, { "epoch": 0.9802824068184709, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.215915203094482, "learning_rate": 5e-05, "loss": 0.6461, "mean_token_accuracy": 0.877762496471405, "num_tokens": 292502214.0, "step": 7706 }, { "epoch": 0.9804096170970614, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.196930885314941, "learning_rate": 5e-05, "loss": 0.9816, "mean_token_accuracy": 0.849838376045227, "num_tokens": 292541845.0, "step": 7707 }, { "epoch": 0.980536827375652, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.635308265686035, "learning_rate": 5e-05, "loss": 0.7042, "mean_token_accuracy": 0.8746452331542969, "num_tokens": 292573932.0, "step": 7708 }, { "epoch": 0.9806640376542425, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.761695861816406, "learning_rate": 5e-05, "loss": 0.9104, "mean_token_accuracy": 0.869293749332428, "num_tokens": 292615932.0, "step": 7709 }, { "epoch": 0.980791247932833, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.544833183288574, "learning_rate": 5e-05, "loss": 0.8587, "mean_token_accuracy": 0.8508071303367615, "num_tokens": 292648680.0, "step": 7710 }, { "epoch": 0.9809184582114235, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.593466758728027, "learning_rate": 5e-05, "loss": 0.7737, "mean_token_accuracy": 0.8752332925796509, "num_tokens": 292687987.0, "step": 7711 }, { "epoch": 0.981045668490014, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.323972702026367, "learning_rate": 5e-05, "loss": 0.8768, "mean_token_accuracy": 0.858507513999939, "num_tokens": 292732243.0, "step": 7712 }, { "epoch": 0.9811728787686045, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.713008403778076, "learning_rate": 5e-05, "loss": 0.7094, "mean_token_accuracy": 0.8690915107727051, "num_tokens": 292764869.0, "step": 7713 }, { "epoch": 0.981300089047195, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.79297161102295, "learning_rate": 5e-05, "loss": 0.8743, "mean_token_accuracy": 0.8721972703933716, "num_tokens": 292806401.0, "step": 7714 }, { "epoch": 0.9814272993257855, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.560312271118164, "learning_rate": 5e-05, "loss": 0.7096, "mean_token_accuracy": 0.869951605796814, "num_tokens": 292843165.0, "step": 7715 }, { "epoch": 0.9815545096043761, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.789023399353027, "learning_rate": 5e-05, "loss": 0.9225, "mean_token_accuracy": 0.8613411784172058, "num_tokens": 292885092.0, "step": 7716 }, { "epoch": 0.9816817198829666, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.969681262969971, "learning_rate": 5e-05, "loss": 0.7865, "mean_token_accuracy": 0.8536769151687622, "num_tokens": 292920595.0, "step": 7717 }, { "epoch": 0.981808930161557, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.299363136291504, "learning_rate": 5e-05, "loss": 0.9099, "mean_token_accuracy": 0.8538916110992432, "num_tokens": 292954079.0, "step": 7718 }, { "epoch": 0.9819361404401475, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.463129043579102, "learning_rate": 5e-05, "loss": 0.7815, "mean_token_accuracy": 0.8679845333099365, "num_tokens": 292985554.0, "step": 7719 }, { "epoch": 0.9820633507187381, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.717901706695557, "learning_rate": 5e-05, "loss": 0.81, "mean_token_accuracy": 0.8654245138168335, "num_tokens": 293021832.0, "step": 7720 }, { "epoch": 0.9821905609973286, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.7003960609436035, "learning_rate": 5e-05, "loss": 0.8062, "mean_token_accuracy": 0.862249493598938, "num_tokens": 293054220.0, "step": 7721 }, { "epoch": 0.9823177712759191, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.461948394775391, "learning_rate": 5e-05, "loss": 0.8174, "mean_token_accuracy": 0.853650689125061, "num_tokens": 293099348.0, "step": 7722 }, { "epoch": 0.9824449815545097, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.6033525466918945, "learning_rate": 5e-05, "loss": 0.7869, "mean_token_accuracy": 0.8602440357208252, "num_tokens": 293134801.0, "step": 7723 }, { "epoch": 0.9825721918331001, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0004329681396484375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.583222389221191, "learning_rate": 5e-05, "loss": 0.8468, "mean_token_accuracy": 0.839322030544281, "num_tokens": 293172023.0, "step": 7724 }, { "epoch": 0.9826994021116906, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.274954319000244, "learning_rate": 5e-05, "loss": 0.7416, "mean_token_accuracy": 0.8634360432624817, "num_tokens": 293210019.0, "step": 7725 }, { "epoch": 0.9828266123902811, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0004367828369140625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.9023261070251465, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.860874354839325, "num_tokens": 293249614.0, "step": 7726 }, { "epoch": 0.9829538226688717, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.8758368492126465, "learning_rate": 5e-05, "loss": 0.6986, "mean_token_accuracy": 0.8718841075897217, "num_tokens": 293289860.0, "step": 7727 }, { "epoch": 0.9830810329474622, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0004367828369140625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.301082611083984, "learning_rate": 5e-05, "loss": 0.8873, "mean_token_accuracy": 0.8502712249755859, "num_tokens": 293330337.0, "step": 7728 }, { "epoch": 0.9832082432260527, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0004367828369140625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.487130165100098, "learning_rate": 5e-05, "loss": 0.719, "mean_token_accuracy": 0.8626520037651062, "num_tokens": 293370062.0, "step": 7729 }, { "epoch": 0.9833354535046431, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.829211235046387, "learning_rate": 5e-05, "loss": 0.888, "mean_token_accuracy": 0.8608541488647461, "num_tokens": 293409544.0, "step": 7730 }, { "epoch": 0.9834626637832337, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.181994438171387, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8481337428092957, "num_tokens": 293453924.0, "step": 7731 }, { "epoch": 0.9835898740618242, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.33033561706543, "learning_rate": 5e-05, "loss": 0.9536, "mean_token_accuracy": 0.868060290813446, "num_tokens": 293493759.0, "step": 7732 }, { "epoch": 0.9837170843404147, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.360541343688965, "learning_rate": 5e-05, "loss": 0.8115, "mean_token_accuracy": 0.8394048810005188, "num_tokens": 293530155.0, "step": 7733 }, { "epoch": 0.9838442946190052, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.306577682495117, "learning_rate": 5e-05, "loss": 1.0156, "mean_token_accuracy": 0.8599233627319336, "num_tokens": 293573083.0, "step": 7734 }, { "epoch": 0.9839715048975958, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.241426467895508, "learning_rate": 5e-05, "loss": 0.845, "mean_token_accuracy": 0.8712385892868042, "num_tokens": 293608312.0, "step": 7735 }, { "epoch": 0.9840987151761862, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.43871021270752, "learning_rate": 5e-05, "loss": 0.9436, "mean_token_accuracy": 0.859874963760376, "num_tokens": 293645036.0, "step": 7736 }, { "epoch": 0.9842259254547767, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.0092191696167, "learning_rate": 5e-05, "loss": 1.0101, "mean_token_accuracy": 0.8359907865524292, "num_tokens": 293679063.0, "step": 7737 }, { "epoch": 0.9843531357333672, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.420307636260986, "learning_rate": 5e-05, "loss": 0.8067, "mean_token_accuracy": 0.8639525175094604, "num_tokens": 293720239.0, "step": 7738 }, { "epoch": 0.9844803460119578, "ewc_loss": 0.5, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.256417274475098, "learning_rate": 5e-05, "loss": 0.8556, "mean_token_accuracy": 0.877353310585022, "num_tokens": 293757154.0, "step": 7739 }, { "epoch": 0.9846075562905483, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.167439937591553, "learning_rate": 5e-05, "loss": 0.8251, "mean_token_accuracy": 0.8491467237472534, "num_tokens": 293797181.0, "step": 7740 }, { "epoch": 0.9847347665691388, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.970200061798096, "learning_rate": 5e-05, "loss": 0.867, "mean_token_accuracy": 0.8574874401092529, "num_tokens": 293838130.0, "step": 7741 }, { "epoch": 0.9848619768477292, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.35966157913208, "learning_rate": 5e-05, "loss": 0.8028, "mean_token_accuracy": 0.8575200438499451, "num_tokens": 293877341.0, "step": 7742 }, { "epoch": 0.9849891871263198, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.704829216003418, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8670434355735779, "num_tokens": 293913764.0, "step": 7743 }, { "epoch": 0.9851163974049103, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.334169864654541, "learning_rate": 5e-05, "loss": 0.7871, "mean_token_accuracy": 0.856663703918457, "num_tokens": 293954142.0, "step": 7744 }, { "epoch": 0.9852436076835008, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.833133220672607, "learning_rate": 5e-05, "loss": 0.842, "mean_token_accuracy": 0.8548458814620972, "num_tokens": 293999016.0, "step": 7745 }, { "epoch": 0.9853708179620914, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.009769916534424, "learning_rate": 5e-05, "loss": 0.7456, "mean_token_accuracy": 0.8633689880371094, "num_tokens": 294033545.0, "step": 7746 }, { "epoch": 0.9854980282406819, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.205619812011719, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.8787938952445984, "num_tokens": 294076763.0, "step": 7747 }, { "epoch": 0.9856252385192723, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.637430191040039, "learning_rate": 5e-05, "loss": 0.6609, "mean_token_accuracy": 0.881757378578186, "num_tokens": 294114661.0, "step": 7748 }, { "epoch": 0.9857524487978628, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.611358642578125, "learning_rate": 5e-05, "loss": 0.8954, "mean_token_accuracy": 0.8587969541549683, "num_tokens": 294148819.0, "step": 7749 }, { "epoch": 0.9858796590764534, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.442397117614746, "learning_rate": 5e-05, "loss": 0.7131, "mean_token_accuracy": 0.8656473755836487, "num_tokens": 294190820.0, "step": 7750 }, { "epoch": 0.9860068693550439, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.872659683227539, "learning_rate": 5e-05, "loss": 0.9243, "mean_token_accuracy": 0.8617075681686401, "num_tokens": 294230814.0, "step": 7751 }, { "epoch": 0.9861340796336344, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.5417327880859375, "learning_rate": 5e-05, "loss": 0.7665, "mean_token_accuracy": 0.8602248430252075, "num_tokens": 294271811.0, "step": 7752 }, { "epoch": 0.986261289912225, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.8263578414917, "learning_rate": 5e-05, "loss": 0.955, "mean_token_accuracy": 0.8593952655792236, "num_tokens": 294313901.0, "step": 7753 }, { "epoch": 0.9863885001908154, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.014256954193115, "learning_rate": 5e-05, "loss": 0.841, "mean_token_accuracy": 0.8514127731323242, "num_tokens": 294353275.0, "step": 7754 }, { "epoch": 0.9865157104694059, "ewc_loss": 0.5, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.283306121826172, "learning_rate": 5e-05, "loss": 0.8446, "mean_token_accuracy": 0.8803091049194336, "num_tokens": 294390569.0, "step": 7755 }, { "epoch": 0.9866429207479964, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.567885398864746, "learning_rate": 5e-05, "loss": 0.8892, "mean_token_accuracy": 0.8509383201599121, "num_tokens": 294433050.0, "step": 7756 }, { "epoch": 0.986770131026587, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.5611090660095215, "learning_rate": 5e-05, "loss": 0.8011, "mean_token_accuracy": 0.8726890087127686, "num_tokens": 294471198.0, "step": 7757 }, { "epoch": 0.9868973413051775, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.894505023956299, "learning_rate": 5e-05, "loss": 0.8466, "mean_token_accuracy": 0.8660111427307129, "num_tokens": 294511855.0, "step": 7758 }, { "epoch": 0.987024551583768, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.2264018058776855, "learning_rate": 5e-05, "loss": 0.758, "mean_token_accuracy": 0.8690393567085266, "num_tokens": 294547056.0, "step": 7759 }, { "epoch": 0.9871517618623584, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.854266166687012, "learning_rate": 5e-05, "loss": 0.8583, "mean_token_accuracy": 0.8577760457992554, "num_tokens": 294584516.0, "step": 7760 }, { "epoch": 0.987278972140949, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.271032810211182, "learning_rate": 5e-05, "loss": 0.7778, "mean_token_accuracy": 0.8601107597351074, "num_tokens": 294622253.0, "step": 7761 }, { "epoch": 0.9874061824195395, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.624621391296387, "learning_rate": 5e-05, "loss": 0.766, "mean_token_accuracy": 0.8709559440612793, "num_tokens": 294664842.0, "step": 7762 }, { "epoch": 0.98753339269813, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.449517250061035, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8563107252120972, "num_tokens": 294707135.0, "step": 7763 }, { "epoch": 0.9876606029767205, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.371333599090576, "learning_rate": 5e-05, "loss": 0.7984, "mean_token_accuracy": 0.8496053218841553, "num_tokens": 294751356.0, "step": 7764 }, { "epoch": 0.9877878132553111, "ewc_loss": 0.375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.625503063201904, "learning_rate": 5e-05, "loss": 0.7889, "mean_token_accuracy": 0.8549832105636597, "num_tokens": 294789444.0, "step": 7765 }, { "epoch": 0.9879150235339016, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.17253303527832, "learning_rate": 5e-05, "loss": 0.7736, "mean_token_accuracy": 0.853224515914917, "num_tokens": 294823247.0, "step": 7766 }, { "epoch": 0.988042233812492, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.814326286315918, "learning_rate": 5e-05, "loss": 0.8309, "mean_token_accuracy": 0.8503783941268921, "num_tokens": 294862932.0, "step": 7767 }, { "epoch": 0.9881694440910825, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.904261589050293, "learning_rate": 5e-05, "loss": 0.7304, "mean_token_accuracy": 0.8620848655700684, "num_tokens": 294902686.0, "step": 7768 }, { "epoch": 0.9882966543696731, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.14246940612793, "learning_rate": 5e-05, "loss": 0.8071, "mean_token_accuracy": 0.8680753111839294, "num_tokens": 294942028.0, "step": 7769 }, { "epoch": 0.9884238646482636, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.543841361999512, "learning_rate": 5e-05, "loss": 0.7251, "mean_token_accuracy": 0.8616236448287964, "num_tokens": 294977699.0, "step": 7770 }, { "epoch": 0.9885510749268541, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.658467292785645, "learning_rate": 5e-05, "loss": 0.8885, "mean_token_accuracy": 0.8607975244522095, "num_tokens": 295014441.0, "step": 7771 }, { "epoch": 0.9886782852054447, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.000431060791015625, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.11455774307251, "learning_rate": 5e-05, "loss": 0.6733, "mean_token_accuracy": 0.8716331720352173, "num_tokens": 295053569.0, "step": 7772 }, { "epoch": 0.9888054954840351, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00043487548828125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.352300643920898, "learning_rate": 5e-05, "loss": 0.9671, "mean_token_accuracy": 0.8604874610900879, "num_tokens": 295083607.0, "step": 7773 }, { "epoch": 0.9889327057626256, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 5.999263286590576, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.8517577648162842, "num_tokens": 295118074.0, "step": 7774 }, { "epoch": 0.9890599160412161, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.789122581481934, "learning_rate": 5e-05, "loss": 1.024, "mean_token_accuracy": 0.8624985218048096, "num_tokens": 295157926.0, "step": 7775 }, { "epoch": 0.9891871263198067, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 6.923946380615234, "learning_rate": 5e-05, "loss": 0.8897, "mean_token_accuracy": 0.8611078858375549, "num_tokens": 295201297.0, "step": 7776 }, { "epoch": 0.9893143365983972, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.100132942199707, "learning_rate": 5e-05, "loss": 1.0478, "mean_token_accuracy": 0.8556715250015259, "num_tokens": 295238178.0, "step": 7777 }, { "epoch": 0.9894415468769877, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.249847412109375, "learning_rate": 5e-05, "loss": 0.9993, "mean_token_accuracy": 0.8645116686820984, "num_tokens": 295273843.0, "step": 7778 }, { "epoch": 0.9895687571555781, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.299797534942627, "learning_rate": 5e-05, "loss": 0.7972, "mean_token_accuracy": 0.8650966882705688, "num_tokens": 295308703.0, "step": 7779 }, { "epoch": 0.9896959674341687, "ewc_loss": 0.625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.343650817871094, "learning_rate": 5e-05, "loss": 1.0338, "mean_token_accuracy": 0.8590337038040161, "num_tokens": 295347710.0, "step": 7780 }, { "epoch": 0.9898231777127592, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.300962448120117, "learning_rate": 5e-05, "loss": 0.7298, "mean_token_accuracy": 0.8689650893211365, "num_tokens": 295382255.0, "step": 7781 }, { "epoch": 0.9899503879913497, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.067803382873535, "learning_rate": 5e-05, "loss": 1.0539, "mean_token_accuracy": 0.8666260242462158, "num_tokens": 295418185.0, "step": 7782 }, { "epoch": 0.9900775982699402, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.1381516456604, "learning_rate": 5e-05, "loss": 0.9852, "mean_token_accuracy": 0.8420581221580505, "num_tokens": 295455124.0, "step": 7783 }, { "epoch": 0.9902048085485308, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.477493286132812, "learning_rate": 5e-05, "loss": 1.0922, "mean_token_accuracy": 0.8496401309967041, "num_tokens": 295498058.0, "step": 7784 }, { "epoch": 0.9903320188271212, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.340922355651855, "learning_rate": 5e-05, "loss": 0.9981, "mean_token_accuracy": 0.8782927989959717, "num_tokens": 295534623.0, "step": 7785 }, { "epoch": 0.9904592291057117, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.980203151702881, "learning_rate": 5e-05, "loss": 0.9237, "mean_token_accuracy": 0.854108989238739, "num_tokens": 295580064.0, "step": 7786 }, { "epoch": 0.9905864393843022, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.82502555847168, "learning_rate": 5e-05, "loss": 1.0622, "mean_token_accuracy": 0.849126935005188, "num_tokens": 295620720.0, "step": 7787 }, { "epoch": 0.9907136496628928, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.285983562469482, "learning_rate": 5e-05, "loss": 0.837, "mean_token_accuracy": 0.8674142360687256, "num_tokens": 295662157.0, "step": 7788 }, { "epoch": 0.9908408599414833, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.688748359680176, "learning_rate": 5e-05, "loss": 0.9746, "mean_token_accuracy": 0.8616607785224915, "num_tokens": 295700215.0, "step": 7789 }, { "epoch": 0.9909680702200738, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.512145519256592, "learning_rate": 5e-05, "loss": 0.8235, "mean_token_accuracy": 0.8833045959472656, "num_tokens": 295738363.0, "step": 7790 }, { "epoch": 0.9910952804986642, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.978592395782471, "learning_rate": 5e-05, "loss": 0.8681, "mean_token_accuracy": 0.8742626905441284, "num_tokens": 295773911.0, "step": 7791 }, { "epoch": 0.9912224907772548, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.847832202911377, "learning_rate": 5e-05, "loss": 0.8901, "mean_token_accuracy": 0.860901951789856, "num_tokens": 295809341.0, "step": 7792 }, { "epoch": 0.9913497010558453, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.406641006469727, "learning_rate": 5e-05, "loss": 0.8258, "mean_token_accuracy": 0.8521162271499634, "num_tokens": 295855952.0, "step": 7793 }, { "epoch": 0.9914769113344358, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.011780738830566, "learning_rate": 5e-05, "loss": 0.8608, "mean_token_accuracy": 0.8649789094924927, "num_tokens": 295892047.0, "step": 7794 }, { "epoch": 0.9916041216130264, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.12129020690918, "learning_rate": 5e-05, "loss": 0.8297, "mean_token_accuracy": 0.8435453176498413, "num_tokens": 295926259.0, "step": 7795 }, { "epoch": 0.9917313318916169, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.224929809570312, "learning_rate": 5e-05, "loss": 0.8822, "mean_token_accuracy": 0.8566136360168457, "num_tokens": 295965584.0, "step": 7796 }, { "epoch": 0.9918585421702073, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.845541477203369, "learning_rate": 5e-05, "loss": 0.7293, "mean_token_accuracy": 0.8713720440864563, "num_tokens": 296002754.0, "step": 7797 }, { "epoch": 0.9919857524487978, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.5172119140625, "learning_rate": 5e-05, "loss": 0.9001, "mean_token_accuracy": 0.8616666793823242, "num_tokens": 296042161.0, "step": 7798 }, { "epoch": 0.9921129627273884, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.6826581954956055, "learning_rate": 5e-05, "loss": 0.7928, "mean_token_accuracy": 0.8523043990135193, "num_tokens": 296079153.0, "step": 7799 }, { "epoch": 0.9922401730059789, "ewc_loss": 0.5, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.634148597717285, "learning_rate": 5e-05, "loss": 0.9361, "mean_token_accuracy": 0.8531991243362427, "num_tokens": 296124064.0, "step": 7800 }, { "epoch": 0.9923673832845694, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.716277122497559, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8678469657897949, "num_tokens": 296166035.0, "step": 7801 }, { "epoch": 0.9924945935631599, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.646097183227539, "learning_rate": 5e-05, "loss": 0.8707, "mean_token_accuracy": 0.8763628005981445, "num_tokens": 296206566.0, "step": 7802 }, { "epoch": 0.9926218038417504, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.773626804351807, "learning_rate": 5e-05, "loss": 0.784, "mean_token_accuracy": 0.8677385449409485, "num_tokens": 296243587.0, "step": 7803 }, { "epoch": 0.9927490141203409, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.677078247070312, "learning_rate": 5e-05, "loss": 0.9583, "mean_token_accuracy": 0.8588371276855469, "num_tokens": 296282704.0, "step": 7804 }, { "epoch": 0.9928762243989314, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.844055652618408, "learning_rate": 5e-05, "loss": 0.8439, "mean_token_accuracy": 0.8476592898368835, "num_tokens": 296317964.0, "step": 7805 }, { "epoch": 0.993003434677522, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.612485885620117, "learning_rate": 5e-05, "loss": 0.9632, "mean_token_accuracy": 0.8542968034744263, "num_tokens": 296352861.0, "step": 7806 }, { "epoch": 0.9931306449561125, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 6.994343280792236, "learning_rate": 5e-05, "loss": 0.8222, "mean_token_accuracy": 0.8601857423782349, "num_tokens": 296391302.0, "step": 7807 }, { "epoch": 0.993257855234703, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.447880744934082, "learning_rate": 5e-05, "loss": 0.9309, "mean_token_accuracy": 0.8549035787582397, "num_tokens": 296428026.0, "step": 7808 }, { "epoch": 0.9933850655132934, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.138005256652832, "learning_rate": 5e-05, "loss": 0.8408, "mean_token_accuracy": 0.8611284494400024, "num_tokens": 296468866.0, "step": 7809 }, { "epoch": 0.993512275791884, "ewc_loss": 0.5, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.24759292602539, "learning_rate": 5e-05, "loss": 0.9122, "mean_token_accuracy": 0.8629592061042786, "num_tokens": 296514806.0, "step": 7810 }, { "epoch": 0.9936394860704745, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.237693786621094, "learning_rate": 5e-05, "loss": 0.8095, "mean_token_accuracy": 0.8663887977600098, "num_tokens": 296552328.0, "step": 7811 }, { "epoch": 0.993766696349065, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.061436653137207, "learning_rate": 5e-05, "loss": 0.9047, "mean_token_accuracy": 0.8542001247406006, "num_tokens": 296589908.0, "step": 7812 }, { "epoch": 0.9938939066276555, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.28767728805542, "learning_rate": 5e-05, "loss": 0.8554, "mean_token_accuracy": 0.8530364036560059, "num_tokens": 296631147.0, "step": 7813 }, { "epoch": 0.9940211169062461, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.9129557609558105, "learning_rate": 5e-05, "loss": 0.8922, "mean_token_accuracy": 0.8535245656967163, "num_tokens": 296673633.0, "step": 7814 }, { "epoch": 0.9941483271848366, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.296640396118164, "learning_rate": 5e-05, "loss": 0.797, "mean_token_accuracy": 0.8617002964019775, "num_tokens": 296714782.0, "step": 7815 }, { "epoch": 0.994275537463427, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.826862812042236, "learning_rate": 5e-05, "loss": 0.8556, "mean_token_accuracy": 0.8570442795753479, "num_tokens": 296751556.0, "step": 7816 }, { "epoch": 0.9944027477420175, "ewc_loss": 0.375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.262361526489258, "learning_rate": 5e-05, "loss": 0.8418, "mean_token_accuracy": 0.8413211107254028, "num_tokens": 296788319.0, "step": 7817 }, { "epoch": 0.9945299580206081, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.809594631195068, "learning_rate": 5e-05, "loss": 0.851, "mean_token_accuracy": 0.8545901775360107, "num_tokens": 296823513.0, "step": 7818 }, { "epoch": 0.9946571682991986, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.204199314117432, "learning_rate": 5e-05, "loss": 0.7942, "mean_token_accuracy": 0.8541591167449951, "num_tokens": 296862059.0, "step": 7819 }, { "epoch": 0.9947843785777891, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.807497501373291, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.8661636114120483, "num_tokens": 296898051.0, "step": 7820 }, { "epoch": 0.9949115888563796, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.158215522766113, "learning_rate": 5e-05, "loss": 0.7485, "mean_token_accuracy": 0.8672974705696106, "num_tokens": 296935266.0, "step": 7821 }, { "epoch": 0.9950387991349701, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.819690704345703, "learning_rate": 5e-05, "loss": 0.8617, "mean_token_accuracy": 0.8471554517745972, "num_tokens": 296971653.0, "step": 7822 }, { "epoch": 0.9951660094135606, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.0948100090026855, "learning_rate": 5e-05, "loss": 0.8431, "mean_token_accuracy": 0.8358319401741028, "num_tokens": 297015825.0, "step": 7823 }, { "epoch": 0.9952932196921511, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.87239933013916, "learning_rate": 5e-05, "loss": 0.863, "mean_token_accuracy": 0.8484789133071899, "num_tokens": 297059115.0, "step": 7824 }, { "epoch": 0.9954204299707416, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.038642406463623, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8564345836639404, "num_tokens": 297092634.0, "step": 7825 }, { "epoch": 0.9955476402493322, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.8856706619262695, "learning_rate": 5e-05, "loss": 0.8229, "mean_token_accuracy": 0.8618788719177246, "num_tokens": 297137154.0, "step": 7826 }, { "epoch": 0.9956748505279227, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.043116569519043, "learning_rate": 5e-05, "loss": 0.7919, "mean_token_accuracy": 0.8501375913619995, "num_tokens": 297175029.0, "step": 7827 }, { "epoch": 0.9958020608065131, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.890761375427246, "learning_rate": 5e-05, "loss": 0.8014, "mean_token_accuracy": 0.8632158041000366, "num_tokens": 297207397.0, "step": 7828 }, { "epoch": 0.9959292710851037, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.027570724487305, "learning_rate": 5e-05, "loss": 0.7734, "mean_token_accuracy": 0.8570009469985962, "num_tokens": 297250114.0, "step": 7829 }, { "epoch": 0.9960564813636942, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.85499382019043, "learning_rate": 5e-05, "loss": 0.8095, "mean_token_accuracy": 0.8640726804733276, "num_tokens": 297289765.0, "step": 7830 }, { "epoch": 0.9961836916422847, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.0566887855529785, "learning_rate": 5e-05, "loss": 0.6827, "mean_token_accuracy": 0.8818122148513794, "num_tokens": 297327359.0, "step": 7831 }, { "epoch": 0.9963109019208752, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.824721336364746, "learning_rate": 5e-05, "loss": 0.7897, "mean_token_accuracy": 0.8649532794952393, "num_tokens": 297360851.0, "step": 7832 }, { "epoch": 0.9964381121994658, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.088438034057617, "learning_rate": 5e-05, "loss": 0.7523, "mean_token_accuracy": 0.8584904670715332, "num_tokens": 297396957.0, "step": 7833 }, { "epoch": 0.9965653224780562, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.774250507354736, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.8618199229240417, "num_tokens": 297427900.0, "step": 7834 }, { "epoch": 0.9966925327566467, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000438690185546875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.121119499206543, "learning_rate": 5e-05, "loss": 0.7816, "mean_token_accuracy": 0.8501375913619995, "num_tokens": 297464910.0, "step": 7835 }, { "epoch": 0.9968197430352372, "ewc_loss": 0.375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.719232082366943, "learning_rate": 5e-05, "loss": 0.7373, "mean_token_accuracy": 0.873440146446228, "num_tokens": 297497585.0, "step": 7836 }, { "epoch": 0.9969469533138278, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.170351505279541, "learning_rate": 5e-05, "loss": 0.768, "mean_token_accuracy": 0.8508382439613342, "num_tokens": 297537074.0, "step": 7837 }, { "epoch": 0.9970741635924183, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.623692989349365, "learning_rate": 5e-05, "loss": 0.7381, "mean_token_accuracy": 0.8713144063949585, "num_tokens": 297570848.0, "step": 7838 }, { "epoch": 0.9972013738710088, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.237407207489014, "learning_rate": 5e-05, "loss": 0.7741, "mean_token_accuracy": 0.8497037291526794, "num_tokens": 297607520.0, "step": 7839 }, { "epoch": 0.9973285841495992, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.512536525726318, "learning_rate": 5e-05, "loss": 0.761, "mean_token_accuracy": 0.8675467371940613, "num_tokens": 297648442.0, "step": 7840 }, { "epoch": 0.9974557944281898, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.336756229400635, "learning_rate": 5e-05, "loss": 0.8081, "mean_token_accuracy": 0.8431638479232788, "num_tokens": 297686163.0, "step": 7841 }, { "epoch": 0.9975830047067803, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.3886823654174805, "learning_rate": 5e-05, "loss": 0.7102, "mean_token_accuracy": 0.8711351156234741, "num_tokens": 297724191.0, "step": 7842 }, { "epoch": 0.9977102149853708, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.446021556854248, "learning_rate": 5e-05, "loss": 0.7768, "mean_token_accuracy": 0.8538005948066711, "num_tokens": 297766205.0, "step": 7843 }, { "epoch": 0.9978374252639614, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.237572193145752, "learning_rate": 5e-05, "loss": 0.7166, "mean_token_accuracy": 0.8694769144058228, "num_tokens": 297802153.0, "step": 7844 }, { "epoch": 0.9979646355425519, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.597257137298584, "learning_rate": 5e-05, "loss": 0.741, "mean_token_accuracy": 0.8687189221382141, "num_tokens": 297841252.0, "step": 7845 }, { "epoch": 0.9980918458211423, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 7.028803825378418, "learning_rate": 5e-05, "loss": 0.6271, "mean_token_accuracy": 0.8891618251800537, "num_tokens": 297875460.0, "step": 7846 }, { "epoch": 0.9982190560997328, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.862880706787109, "learning_rate": 5e-05, "loss": 0.7853, "mean_token_accuracy": 0.8602272272109985, "num_tokens": 297908534.0, "step": 7847 }, { "epoch": 0.9983462663783234, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.699542045593262, "learning_rate": 5e-05, "loss": 0.7617, "mean_token_accuracy": 0.8385807871818542, "num_tokens": 297950798.0, "step": 7848 }, { "epoch": 0.9984734766569139, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.322537422180176, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8683943748474121, "num_tokens": 297989696.0, "step": 7849 }, { "epoch": 0.9986006869355044, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00022220611572265625, "grad_norm": 6.141613483428955, "learning_rate": 5e-05, "loss": 0.6847, "mean_token_accuracy": 0.8581609129905701, "num_tokens": 298033702.0, "step": 7850 }, { "epoch": 0.9987278972140949, "ewc_loss": 0.5, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 9.218459129333496, "learning_rate": 5e-05, "loss": 0.8644, "mean_token_accuracy": 0.8770341873168945, "num_tokens": 298071867.0, "step": 7851 }, { "epoch": 0.9988551074926854, "ewc_loss": 0.2451171875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00020122528076171875, "grad_norm": 5.390256404876709, "learning_rate": 5e-05, "loss": 0.695, "mean_token_accuracy": 0.8492944240570068, "num_tokens": 298108726.0, "step": 7852 }, { "epoch": 0.9989823177712759, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.674460411071777, "learning_rate": 5e-05, "loss": 1.1068, "mean_token_accuracy": 0.8550872206687927, "num_tokens": 298145174.0, "step": 7853 }, { "epoch": 0.9991095280498664, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 5.834519386291504, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8733060956001282, "num_tokens": 298181300.0, "step": 7854 }, { "epoch": 0.9992367383284569, "ewc_loss": 0.95703125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 11.908584594726562, "learning_rate": 5e-05, "loss": 1.3668, "mean_token_accuracy": 0.8615376353263855, "num_tokens": 298220592.0, "step": 7855 }, { "epoch": 0.9993639486070475, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 9.00251293182373, "learning_rate": 5e-05, "loss": 1.3376, "mean_token_accuracy": 0.8560106754302979, "num_tokens": 298260796.0, "step": 7856 }, { "epoch": 0.999491158885638, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.450138092041016, "learning_rate": 5e-05, "loss": 1.1076, "mean_token_accuracy": 0.8435484170913696, "num_tokens": 298300640.0, "step": 7857 }, { "epoch": 0.9996183691642284, "ewc_loss": 0.9375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000896453857421875, "grad_norm": 9.642032623291016, "learning_rate": 5e-05, "loss": 1.3024, "mean_token_accuracy": 0.87507164478302, "num_tokens": 298336433.0, "step": 7858 }, { "epoch": 0.9997455794428189, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.82094955444336, "learning_rate": 5e-05, "loss": 0.9375, "mean_token_accuracy": 0.8723732233047485, "num_tokens": 298370355.0, "step": 7859 }, { "epoch": 0.9998727897214095, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 10.212105751037598, "learning_rate": 5e-05, "loss": 1.1999, "mean_token_accuracy": 0.8603301644325256, "num_tokens": 298406262.0, "step": 7860 }, { "epoch": 1.0, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.08582592010498, "learning_rate": 5e-05, "loss": 1.0867, "mean_token_accuracy": 0.8415590524673462, "num_tokens": 298443310.0, "step": 7861 }, { "epoch": 1.0001272102785905, "ewc_loss": 0.75, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 10.671899795532227, "learning_rate": 5e-05, "loss": 1.1304, "mean_token_accuracy": 0.869056224822998, "num_tokens": 298482309.0, "step": 7862 }, { "epoch": 1.000254420557181, "ewc_loss": 0.71875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 8.600030899047852, "learning_rate": 5e-05, "loss": 1.1051, "mean_token_accuracy": 0.8654623627662659, "num_tokens": 298522555.0, "step": 7863 }, { "epoch": 1.0003816308357716, "ewc_loss": 0.625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.936549186706543, "learning_rate": 5e-05, "loss": 0.9571, "mean_token_accuracy": 0.8834208250045776, "num_tokens": 298559559.0, "step": 7864 }, { "epoch": 1.0005088411143621, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 8.641286849975586, "learning_rate": 5e-05, "loss": 1.0706, "mean_token_accuracy": 0.86993408203125, "num_tokens": 298596604.0, "step": 7865 }, { "epoch": 1.0006360513929526, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.836601257324219, "learning_rate": 5e-05, "loss": 0.8666, "mean_token_accuracy": 0.8879650235176086, "num_tokens": 298634590.0, "step": 7866 }, { "epoch": 1.0007632616715432, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.756409645080566, "learning_rate": 5e-05, "loss": 0.9518, "mean_token_accuracy": 0.8816428184509277, "num_tokens": 298670714.0, "step": 7867 }, { "epoch": 1.0008904719501335, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.655220031738281, "learning_rate": 5e-05, "loss": 0.8196, "mean_token_accuracy": 0.8744176626205444, "num_tokens": 298713924.0, "step": 7868 }, { "epoch": 1.001017682228724, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.468812942504883, "learning_rate": 5e-05, "loss": 0.9796, "mean_token_accuracy": 0.8728035688400269, "num_tokens": 298749537.0, "step": 7869 }, { "epoch": 1.0011448925073145, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.044069766998291, "learning_rate": 5e-05, "loss": 0.8248, "mean_token_accuracy": 0.8665071129798889, "num_tokens": 298789887.0, "step": 7870 }, { "epoch": 1.001272102785905, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.241042137145996, "learning_rate": 5e-05, "loss": 1.0393, "mean_token_accuracy": 0.8737033605575562, "num_tokens": 298826790.0, "step": 7871 }, { "epoch": 1.0013993130644956, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 7.64628791809082, "learning_rate": 5e-05, "loss": 0.9814, "mean_token_accuracy": 0.8640759587287903, "num_tokens": 298860427.0, "step": 7872 }, { "epoch": 1.0015265233430861, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.12732219696045, "learning_rate": 5e-05, "loss": 1.023, "mean_token_accuracy": 0.8760286569595337, "num_tokens": 298892627.0, "step": 7873 }, { "epoch": 1.0016537336216766, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 8.312007904052734, "learning_rate": 5e-05, "loss": 1.0745, "mean_token_accuracy": 0.8628070950508118, "num_tokens": 298930513.0, "step": 7874 }, { "epoch": 1.0017809439002672, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.331459045410156, "learning_rate": 5e-05, "loss": 0.8718, "mean_token_accuracy": 0.8942500352859497, "num_tokens": 298971234.0, "step": 7875 }, { "epoch": 1.0019081541788577, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.275903701782227, "learning_rate": 5e-05, "loss": 0.9769, "mean_token_accuracy": 0.8761864304542542, "num_tokens": 299013224.0, "step": 7876 }, { "epoch": 1.0020353644574482, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.552696228027344, "learning_rate": 5e-05, "loss": 0.9349, "mean_token_accuracy": 0.8513614535331726, "num_tokens": 299055093.0, "step": 7877 }, { "epoch": 1.0021625747360388, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.043987274169922, "learning_rate": 5e-05, "loss": 0.8663, "mean_token_accuracy": 0.8874155282974243, "num_tokens": 299085368.0, "step": 7878 }, { "epoch": 1.0022897850146293, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.02090835571289, "learning_rate": 5e-05, "loss": 0.8656, "mean_token_accuracy": 0.8661496043205261, "num_tokens": 299128279.0, "step": 7879 }, { "epoch": 1.0024169952932196, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.857168197631836, "learning_rate": 5e-05, "loss": 0.845, "mean_token_accuracy": 0.874299168586731, "num_tokens": 299161714.0, "step": 7880 }, { "epoch": 1.0025442055718101, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.656728744506836, "learning_rate": 5e-05, "loss": 0.8175, "mean_token_accuracy": 0.8643778562545776, "num_tokens": 299200565.0, "step": 7881 }, { "epoch": 1.0026714158504006, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.85504150390625, "learning_rate": 5e-05, "loss": 0.8221, "mean_token_accuracy": 0.8722219467163086, "num_tokens": 299242242.0, "step": 7882 }, { "epoch": 1.0027986261289912, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.303341388702393, "learning_rate": 5e-05, "loss": 0.7086, "mean_token_accuracy": 0.8843542337417603, "num_tokens": 299278141.0, "step": 7883 }, { "epoch": 1.0029258364075817, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.030268669128418, "learning_rate": 5e-05, "loss": 0.7875, "mean_token_accuracy": 0.8777107000350952, "num_tokens": 299312413.0, "step": 7884 }, { "epoch": 1.0030530466861722, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.040644645690918, "learning_rate": 5e-05, "loss": 0.6897, "mean_token_accuracy": 0.8821472525596619, "num_tokens": 299348158.0, "step": 7885 }, { "epoch": 1.0031802569647628, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.182175636291504, "learning_rate": 5e-05, "loss": 0.7955, "mean_token_accuracy": 0.8798243999481201, "num_tokens": 299386963.0, "step": 7886 }, { "epoch": 1.0033074672433533, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.8750386238098145, "learning_rate": 5e-05, "loss": 0.6836, "mean_token_accuracy": 0.8805804252624512, "num_tokens": 299424929.0, "step": 7887 }, { "epoch": 1.0034346775219438, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.36739444732666, "learning_rate": 5e-05, "loss": 0.8499, "mean_token_accuracy": 0.8645420670509338, "num_tokens": 299461588.0, "step": 7888 }, { "epoch": 1.0035618878005343, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.724523544311523, "learning_rate": 5e-05, "loss": 0.6754, "mean_token_accuracy": 0.8809462785720825, "num_tokens": 299498226.0, "step": 7889 }, { "epoch": 1.0036890980791249, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.573623657226562, "learning_rate": 5e-05, "loss": 0.8389, "mean_token_accuracy": 0.8748584985733032, "num_tokens": 299536806.0, "step": 7890 }, { "epoch": 1.0038163083577154, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.7160749435424805, "learning_rate": 5e-05, "loss": 0.7321, "mean_token_accuracy": 0.8683898448944092, "num_tokens": 299583658.0, "step": 7891 }, { "epoch": 1.0039435186363057, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.599944114685059, "learning_rate": 5e-05, "loss": 0.8723, "mean_token_accuracy": 0.8663409948348999, "num_tokens": 299621381.0, "step": 7892 }, { "epoch": 1.0040707289148962, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 6.9791388511657715, "learning_rate": 5e-05, "loss": 0.7764, "mean_token_accuracy": 0.8711963891983032, "num_tokens": 299661055.0, "step": 7893 }, { "epoch": 1.0041979391934868, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.210373878479004, "learning_rate": 5e-05, "loss": 0.8148, "mean_token_accuracy": 0.878747820854187, "num_tokens": 299702500.0, "step": 7894 }, { "epoch": 1.0043251494720773, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.4611897468566895, "learning_rate": 5e-05, "loss": 0.8153, "mean_token_accuracy": 0.8708382844924927, "num_tokens": 299742590.0, "step": 7895 }, { "epoch": 1.0044523597506678, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.502476215362549, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8702741861343384, "num_tokens": 299777716.0, "step": 7896 }, { "epoch": 1.0045795700292584, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.003458976745605, "learning_rate": 5e-05, "loss": 0.81, "mean_token_accuracy": 0.8830174803733826, "num_tokens": 299808829.0, "step": 7897 }, { "epoch": 1.0047067803078489, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.756802082061768, "learning_rate": 5e-05, "loss": 0.6956, "mean_token_accuracy": 0.8740237951278687, "num_tokens": 299845582.0, "step": 7898 }, { "epoch": 1.0048339905864394, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.707245826721191, "learning_rate": 5e-05, "loss": 0.8588, "mean_token_accuracy": 0.875585675239563, "num_tokens": 299886373.0, "step": 7899 }, { "epoch": 1.00496120086503, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.2076616287231445, "learning_rate": 5e-05, "loss": 0.647, "mean_token_accuracy": 0.8778907656669617, "num_tokens": 299917464.0, "step": 7900 }, { "epoch": 1.0050884111436205, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.471959114074707, "learning_rate": 5e-05, "loss": 0.9668, "mean_token_accuracy": 0.8761184215545654, "num_tokens": 299953455.0, "step": 7901 }, { "epoch": 1.005215621422211, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.329793453216553, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8647723197937012, "num_tokens": 299993707.0, "step": 7902 }, { "epoch": 1.0053428317008015, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0004425048828125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.673138618469238, "learning_rate": 5e-05, "loss": 1.011, "mean_token_accuracy": 0.8740739822387695, "num_tokens": 300033021.0, "step": 7903 }, { "epoch": 1.0054700419793918, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.5395307540893555, "learning_rate": 5e-05, "loss": 0.911, "mean_token_accuracy": 0.873056948184967, "num_tokens": 300072497.0, "step": 7904 }, { "epoch": 1.0055972522579824, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004444122314453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.590153694152832, "learning_rate": 5e-05, "loss": 0.8924, "mean_token_accuracy": 0.8803544640541077, "num_tokens": 300113510.0, "step": 7905 }, { "epoch": 1.0057244625365729, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.530045509338379, "learning_rate": 5e-05, "loss": 0.9927, "mean_token_accuracy": 0.8685685992240906, "num_tokens": 300151524.0, "step": 7906 }, { "epoch": 1.0058516728151634, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.107199192047119, "learning_rate": 5e-05, "loss": 0.7547, "mean_token_accuracy": 0.873161792755127, "num_tokens": 300184281.0, "step": 7907 }, { "epoch": 1.005978883093754, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.116515159606934, "learning_rate": 5e-05, "loss": 1.0087, "mean_token_accuracy": 0.8598521947860718, "num_tokens": 300222772.0, "step": 7908 }, { "epoch": 1.0061060933723445, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.838999271392822, "learning_rate": 5e-05, "loss": 0.7841, "mean_token_accuracy": 0.866639256477356, "num_tokens": 300257347.0, "step": 7909 }, { "epoch": 1.006233303650935, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.952062606811523, "learning_rate": 5e-05, "loss": 0.9672, "mean_token_accuracy": 0.8684394359588623, "num_tokens": 300300678.0, "step": 7910 }, { "epoch": 1.0063605139295255, "ewc_loss": 0.5, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.484057426452637, "learning_rate": 5e-05, "loss": 0.8452, "mean_token_accuracy": 0.8781770467758179, "num_tokens": 300339851.0, "step": 7911 }, { "epoch": 1.006487724208116, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.023648262023926, "learning_rate": 5e-05, "loss": 0.8077, "mean_token_accuracy": 0.8835064172744751, "num_tokens": 300379753.0, "step": 7912 }, { "epoch": 1.0066149344867066, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.023531913757324, "learning_rate": 5e-05, "loss": 0.8898, "mean_token_accuracy": 0.8753820657730103, "num_tokens": 300418235.0, "step": 7913 }, { "epoch": 1.006742144765297, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.256526947021484, "learning_rate": 5e-05, "loss": 0.7022, "mean_token_accuracy": 0.8867013454437256, "num_tokens": 300450237.0, "step": 7914 }, { "epoch": 1.0068693550438876, "ewc_loss": 0.5, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.303744316101074, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.872785210609436, "num_tokens": 300489416.0, "step": 7915 }, { "epoch": 1.0069965653224782, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004482269287109375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.963253974914551, "learning_rate": 5e-05, "loss": 0.708, "mean_token_accuracy": 0.883086621761322, "num_tokens": 300526180.0, "step": 7916 }, { "epoch": 1.0071237756010685, "ewc_loss": 0.5, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.472325325012207, "learning_rate": 5e-05, "loss": 0.8645, "mean_token_accuracy": 0.8743618726730347, "num_tokens": 300566327.0, "step": 7917 }, { "epoch": 1.007250985879659, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.93485689163208, "learning_rate": 5e-05, "loss": 0.7698, "mean_token_accuracy": 0.8645414113998413, "num_tokens": 300610136.0, "step": 7918 }, { "epoch": 1.0073781961582495, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.440075874328613, "learning_rate": 5e-05, "loss": 0.8727, "mean_token_accuracy": 0.8719725012779236, "num_tokens": 300647238.0, "step": 7919 }, { "epoch": 1.00750540643684, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.057502746582031, "learning_rate": 5e-05, "loss": 0.7699, "mean_token_accuracy": 0.8772658109664917, "num_tokens": 300684291.0, "step": 7920 }, { "epoch": 1.0076326167154306, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.258340835571289, "learning_rate": 5e-05, "loss": 0.8577, "mean_token_accuracy": 0.8663500547409058, "num_tokens": 300718919.0, "step": 7921 }, { "epoch": 1.0077598269940211, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.130797863006592, "learning_rate": 5e-05, "loss": 0.7938, "mean_token_accuracy": 0.869423508644104, "num_tokens": 300752353.0, "step": 7922 }, { "epoch": 1.0078870372726116, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.14565372467041, "learning_rate": 5e-05, "loss": 0.8012, "mean_token_accuracy": 0.8786726593971252, "num_tokens": 300789214.0, "step": 7923 }, { "epoch": 1.0080142475512022, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.0336713790893555, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8659263849258423, "num_tokens": 300825629.0, "step": 7924 }, { "epoch": 1.0081414578297927, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.211549758911133, "learning_rate": 5e-05, "loss": 0.7831, "mean_token_accuracy": 0.8876333236694336, "num_tokens": 300863908.0, "step": 7925 }, { "epoch": 1.0082686681083832, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.84214448928833, "learning_rate": 5e-05, "loss": 0.7382, "mean_token_accuracy": 0.8683225512504578, "num_tokens": 300905458.0, "step": 7926 }, { "epoch": 1.0083958783869738, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.452598571777344, "learning_rate": 5e-05, "loss": 0.8053, "mean_token_accuracy": 0.8865261673927307, "num_tokens": 300943906.0, "step": 7927 }, { "epoch": 1.0085230886655643, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.588744163513184, "learning_rate": 5e-05, "loss": 0.6936, "mean_token_accuracy": 0.8784863948822021, "num_tokens": 300979371.0, "step": 7928 }, { "epoch": 1.0086502989441546, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.785638809204102, "learning_rate": 5e-05, "loss": 0.8848, "mean_token_accuracy": 0.8717122077941895, "num_tokens": 301017575.0, "step": 7929 }, { "epoch": 1.0087775092227451, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.5093913078308105, "learning_rate": 5e-05, "loss": 0.6968, "mean_token_accuracy": 0.8773292303085327, "num_tokens": 301053130.0, "step": 7930 }, { "epoch": 1.0089047195013356, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.947714805603027, "learning_rate": 5e-05, "loss": 0.8932, "mean_token_accuracy": 0.8859102725982666, "num_tokens": 301090740.0, "step": 7931 }, { "epoch": 1.0090319297799262, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.822775363922119, "learning_rate": 5e-05, "loss": 0.7698, "mean_token_accuracy": 0.8744072914123535, "num_tokens": 301133539.0, "step": 7932 }, { "epoch": 1.0091591400585167, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.650444030761719, "learning_rate": 5e-05, "loss": 0.9682, "mean_token_accuracy": 0.8519852161407471, "num_tokens": 301168243.0, "step": 7933 }, { "epoch": 1.0092863503371072, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.377275466918945, "learning_rate": 5e-05, "loss": 0.8264, "mean_token_accuracy": 0.8789952993392944, "num_tokens": 301204444.0, "step": 7934 }, { "epoch": 1.0094135606156978, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.962807655334473, "learning_rate": 5e-05, "loss": 0.7936, "mean_token_accuracy": 0.8898063898086548, "num_tokens": 301245768.0, "step": 7935 }, { "epoch": 1.0095407708942883, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00045013427734375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.7906174659729, "learning_rate": 5e-05, "loss": 0.8852, "mean_token_accuracy": 0.8587054014205933, "num_tokens": 301282381.0, "step": 7936 }, { "epoch": 1.0096679811728788, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000453948974609375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.364917755126953, "learning_rate": 5e-05, "loss": 0.7322, "mean_token_accuracy": 0.8821309804916382, "num_tokens": 301322538.0, "step": 7937 }, { "epoch": 1.0097951914514693, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000453948974609375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.932856559753418, "learning_rate": 5e-05, "loss": 0.8201, "mean_token_accuracy": 0.8764904737472534, "num_tokens": 301357527.0, "step": 7938 }, { "epoch": 1.0099224017300599, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0004520416259765625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.1856513023376465, "learning_rate": 5e-05, "loss": 0.7813, "mean_token_accuracy": 0.8604270219802856, "num_tokens": 301399609.0, "step": 7939 }, { "epoch": 1.0100496120086504, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0004520416259765625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.85990047454834, "learning_rate": 5e-05, "loss": 0.7953, "mean_token_accuracy": 0.873390793800354, "num_tokens": 301439271.0, "step": 7940 }, { "epoch": 1.0101768222872407, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0004520416259765625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.199511528015137, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.8739397525787354, "num_tokens": 301476895.0, "step": 7941 }, { "epoch": 1.0103040325658312, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004520416259765625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.776031017303467, "learning_rate": 5e-05, "loss": 0.79, "mean_token_accuracy": 0.8668241500854492, "num_tokens": 301519649.0, "step": 7942 }, { "epoch": 1.0104312428444218, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004520416259765625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.192144393920898, "learning_rate": 5e-05, "loss": 0.7543, "mean_token_accuracy": 0.8651344180107117, "num_tokens": 301557132.0, "step": 7943 }, { "epoch": 1.0105584531230123, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0004520416259765625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.75551176071167, "learning_rate": 5e-05, "loss": 0.7705, "mean_token_accuracy": 0.8706110715866089, "num_tokens": 301603604.0, "step": 7944 }, { "epoch": 1.0106856634016028, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0004520416259765625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.162835597991943, "learning_rate": 5e-05, "loss": 0.7447, "mean_token_accuracy": 0.865059494972229, "num_tokens": 301643797.0, "step": 7945 }, { "epoch": 1.0108128736801933, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0004558563232421875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.7802557945251465, "learning_rate": 5e-05, "loss": 0.7244, "mean_token_accuracy": 0.8820450901985168, "num_tokens": 301676658.0, "step": 7946 }, { "epoch": 1.0109400839587839, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0004558563232421875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.094254970550537, "learning_rate": 5e-05, "loss": 0.6935, "mean_token_accuracy": 0.8824008703231812, "num_tokens": 301714749.0, "step": 7947 }, { "epoch": 1.0110672942373744, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0004558563232421875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.8117570877075195, "learning_rate": 5e-05, "loss": 0.7372, "mean_token_accuracy": 0.878443717956543, "num_tokens": 301754116.0, "step": 7948 }, { "epoch": 1.011194504515965, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0004558563232421875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.030144214630127, "learning_rate": 5e-05, "loss": 0.6622, "mean_token_accuracy": 0.8849115967750549, "num_tokens": 301791441.0, "step": 7949 }, { "epoch": 1.0113217147945555, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0004558563232421875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.892484664916992, "learning_rate": 5e-05, "loss": 0.7858, "mean_token_accuracy": 0.8664214611053467, "num_tokens": 301830971.0, "step": 7950 }, { "epoch": 1.011448925073146, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0004558563232421875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.927227020263672, "learning_rate": 5e-05, "loss": 0.7026, "mean_token_accuracy": 0.8702155947685242, "num_tokens": 301867848.0, "step": 7951 }, { "epoch": 1.0115761353517365, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0004558563232421875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.017032623291016, "learning_rate": 5e-05, "loss": 0.7641, "mean_token_accuracy": 0.8736642003059387, "num_tokens": 301912700.0, "step": 7952 }, { "epoch": 1.0117033456303268, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004558563232421875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.768749237060547, "learning_rate": 5e-05, "loss": 0.6757, "mean_token_accuracy": 0.8794959187507629, "num_tokens": 301945890.0, "step": 7953 }, { "epoch": 1.0118305559089174, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000457763671875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.240147590637207, "learning_rate": 5e-05, "loss": 0.7538, "mean_token_accuracy": 0.88396155834198, "num_tokens": 301980095.0, "step": 7954 }, { "epoch": 1.0119577661875079, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000457763671875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.533571243286133, "learning_rate": 5e-05, "loss": 0.682, "mean_token_accuracy": 0.8747950792312622, "num_tokens": 302011855.0, "step": 7955 }, { "epoch": 1.0120849764660984, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000457763671875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.61910629272461, "learning_rate": 5e-05, "loss": 0.8409, "mean_token_accuracy": 0.8715629577636719, "num_tokens": 302048299.0, "step": 7956 }, { "epoch": 1.012212186744689, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000457763671875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.278054237365723, "learning_rate": 5e-05, "loss": 0.6734, "mean_token_accuracy": 0.8720552921295166, "num_tokens": 302082628.0, "step": 7957 }, { "epoch": 1.0123393970232795, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000457763671875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.990968704223633, "learning_rate": 5e-05, "loss": 0.8502, "mean_token_accuracy": 0.8813766241073608, "num_tokens": 302119832.0, "step": 7958 }, { "epoch": 1.01246660730187, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.315457820892334, "learning_rate": 5e-05, "loss": 0.68, "mean_token_accuracy": 0.880059540271759, "num_tokens": 302162930.0, "step": 7959 }, { "epoch": 1.0125938175804605, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.047271728515625, "learning_rate": 5e-05, "loss": 0.9281, "mean_token_accuracy": 0.8773293495178223, "num_tokens": 302202250.0, "step": 7960 }, { "epoch": 1.012721027859051, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 6.945101737976074, "learning_rate": 5e-05, "loss": 0.8148, "mean_token_accuracy": 0.8718651533126831, "num_tokens": 302239446.0, "step": 7961 }, { "epoch": 1.0128482381376416, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.229936599731445, "learning_rate": 5e-05, "loss": 0.8258, "mean_token_accuracy": 0.8870234489440918, "num_tokens": 302279949.0, "step": 7962 }, { "epoch": 1.012975448416232, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.899355411529541, "learning_rate": 5e-05, "loss": 0.892, "mean_token_accuracy": 0.8656207323074341, "num_tokens": 302318128.0, "step": 7963 }, { "epoch": 1.0131026586948226, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.94465970993042, "learning_rate": 5e-05, "loss": 0.7381, "mean_token_accuracy": 0.8721672296524048, "num_tokens": 302357353.0, "step": 7964 }, { "epoch": 1.0132298689734132, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.820239067077637, "learning_rate": 5e-05, "loss": 0.8914, "mean_token_accuracy": 0.8774228692054749, "num_tokens": 302399843.0, "step": 7965 }, { "epoch": 1.0133570792520035, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.1103339195251465, "learning_rate": 5e-05, "loss": 0.6966, "mean_token_accuracy": 0.8711708188056946, "num_tokens": 302440081.0, "step": 7966 }, { "epoch": 1.013484289530594, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.61718463897705, "learning_rate": 5e-05, "loss": 1.0156, "mean_token_accuracy": 0.8607000112533569, "num_tokens": 302479061.0, "step": 7967 }, { "epoch": 1.0136114998091845, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.284139633178711, "learning_rate": 5e-05, "loss": 0.697, "mean_token_accuracy": 0.884656548500061, "num_tokens": 302514490.0, "step": 7968 }, { "epoch": 1.013738710087775, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.773783683776855, "learning_rate": 5e-05, "loss": 1.0241, "mean_token_accuracy": 0.8773123025894165, "num_tokens": 302552555.0, "step": 7969 }, { "epoch": 1.0138659203663656, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 7.563824653625488, "learning_rate": 5e-05, "loss": 0.9692, "mean_token_accuracy": 0.8645332455635071, "num_tokens": 302589296.0, "step": 7970 }, { "epoch": 1.013993130644956, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.687222480773926, "learning_rate": 5e-05, "loss": 0.8876, "mean_token_accuracy": 0.890062689781189, "num_tokens": 302626085.0, "step": 7971 }, { "epoch": 1.0141203409235466, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.493084907531738, "learning_rate": 5e-05, "loss": 0.9654, "mean_token_accuracy": 0.8752723336219788, "num_tokens": 302662671.0, "step": 7972 }, { "epoch": 1.0142475512021372, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.3104681968688965, "learning_rate": 5e-05, "loss": 0.7861, "mean_token_accuracy": 0.87742680311203, "num_tokens": 302706112.0, "step": 7973 }, { "epoch": 1.0143747614807277, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.853830337524414, "learning_rate": 5e-05, "loss": 0.9847, "mean_token_accuracy": 0.862966775894165, "num_tokens": 302744942.0, "step": 7974 }, { "epoch": 1.0145019717593182, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.154294967651367, "learning_rate": 5e-05, "loss": 0.7742, "mean_token_accuracy": 0.8869903087615967, "num_tokens": 302784085.0, "step": 7975 }, { "epoch": 1.0146291820379088, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.25456428527832, "learning_rate": 5e-05, "loss": 0.8604, "mean_token_accuracy": 0.879331111907959, "num_tokens": 302821336.0, "step": 7976 }, { "epoch": 1.0147563923164993, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.763916492462158, "learning_rate": 5e-05, "loss": 0.8596, "mean_token_accuracy": 0.8747372627258301, "num_tokens": 302857245.0, "step": 7977 }, { "epoch": 1.0148836025950896, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.2685322761535645, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.8758403658866882, "num_tokens": 302894115.0, "step": 7978 }, { "epoch": 1.0150108128736801, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.331364631652832, "learning_rate": 5e-05, "loss": 0.8118, "mean_token_accuracy": 0.8951269388198853, "num_tokens": 302930162.0, "step": 7979 }, { "epoch": 1.0151380231522706, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.758752822875977, "learning_rate": 5e-05, "loss": 0.7012, "mean_token_accuracy": 0.8765752911567688, "num_tokens": 302965184.0, "step": 7980 }, { "epoch": 1.0152652334308612, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.684823036193848, "learning_rate": 5e-05, "loss": 0.9246, "mean_token_accuracy": 0.8688838481903076, "num_tokens": 303003895.0, "step": 7981 }, { "epoch": 1.0153924437094517, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.626918792724609, "learning_rate": 5e-05, "loss": 0.7161, "mean_token_accuracy": 0.8751167058944702, "num_tokens": 303038328.0, "step": 7982 }, { "epoch": 1.0155196539880422, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.962148666381836, "learning_rate": 5e-05, "loss": 1.0084, "mean_token_accuracy": 0.852418839931488, "num_tokens": 303075546.0, "step": 7983 }, { "epoch": 1.0156468642666328, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.751038551330566, "learning_rate": 5e-05, "loss": 0.7786, "mean_token_accuracy": 0.8703585863113403, "num_tokens": 303111825.0, "step": 7984 }, { "epoch": 1.0157740745452233, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.002049446105957, "learning_rate": 5e-05, "loss": 0.9079, "mean_token_accuracy": 0.8835063576698303, "num_tokens": 303147607.0, "step": 7985 }, { "epoch": 1.0159012848238138, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.063260078430176, "learning_rate": 5e-05, "loss": 0.8449, "mean_token_accuracy": 0.8694949150085449, "num_tokens": 303185646.0, "step": 7986 }, { "epoch": 1.0160284951024043, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.79128646850586, "learning_rate": 5e-05, "loss": 0.919, "mean_token_accuracy": 0.8789656162261963, "num_tokens": 303218723.0, "step": 7987 }, { "epoch": 1.0161557053809949, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.226360321044922, "learning_rate": 5e-05, "loss": 0.864, "mean_token_accuracy": 0.8697167634963989, "num_tokens": 303259012.0, "step": 7988 }, { "epoch": 1.0162829156595854, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.70281982421875, "learning_rate": 5e-05, "loss": 0.948, "mean_token_accuracy": 0.8718068599700928, "num_tokens": 303294868.0, "step": 7989 }, { "epoch": 1.0164101259381757, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.0932769775390625, "learning_rate": 5e-05, "loss": 0.8162, "mean_token_accuracy": 0.8749200701713562, "num_tokens": 303331987.0, "step": 7990 }, { "epoch": 1.0165373362167662, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.894193649291992, "learning_rate": 5e-05, "loss": 0.8751, "mean_token_accuracy": 0.8951739072799683, "num_tokens": 303364230.0, "step": 7991 }, { "epoch": 1.0166645464953568, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.020939350128174, "learning_rate": 5e-05, "loss": 0.7966, "mean_token_accuracy": 0.8811867237091064, "num_tokens": 303404169.0, "step": 7992 }, { "epoch": 1.0167917567739473, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.004262924194336, "learning_rate": 5e-05, "loss": 0.9579, "mean_token_accuracy": 0.8777170181274414, "num_tokens": 303445137.0, "step": 7993 }, { "epoch": 1.0169189670525378, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.330578804016113, "learning_rate": 5e-05, "loss": 0.822, "mean_token_accuracy": 0.8893833160400391, "num_tokens": 303486515.0, "step": 7994 }, { "epoch": 1.0170461773311283, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.618741989135742, "learning_rate": 5e-05, "loss": 0.9552, "mean_token_accuracy": 0.8724678754806519, "num_tokens": 303521514.0, "step": 7995 }, { "epoch": 1.0171733876097189, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.741787910461426, "learning_rate": 5e-05, "loss": 0.9124, "mean_token_accuracy": 0.872916579246521, "num_tokens": 303555606.0, "step": 7996 }, { "epoch": 1.0173005978883094, "ewc_loss": 0.5, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.938425064086914, "learning_rate": 5e-05, "loss": 0.8589, "mean_token_accuracy": 0.8721438646316528, "num_tokens": 303592075.0, "step": 7997 }, { "epoch": 1.0174278081669, "ewc_loss": 0.5, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.85067081451416, "learning_rate": 5e-05, "loss": 0.9198, "mean_token_accuracy": 0.8555791974067688, "num_tokens": 303632529.0, "step": 7998 }, { "epoch": 1.0175550184454905, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.585555076599121, "learning_rate": 5e-05, "loss": 0.784, "mean_token_accuracy": 0.880267858505249, "num_tokens": 303665419.0, "step": 7999 }, { "epoch": 1.017682228724081, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.660978317260742, "learning_rate": 5e-05, "loss": 0.8314, "mean_token_accuracy": 0.8709640502929688, "num_tokens": 303698778.0, "step": 8000 }, { "epoch": 1.0178094390026715, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.620954990386963, "learning_rate": 5e-05, "loss": 0.7965, "mean_token_accuracy": 0.8735276460647583, "num_tokens": 303738713.0, "step": 8001 }, { "epoch": 1.0179366492812618, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.395965099334717, "learning_rate": 5e-05, "loss": 0.7603, "mean_token_accuracy": 0.8742177486419678, "num_tokens": 303767821.0, "step": 8002 }, { "epoch": 1.0180638595598523, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.698736667633057, "learning_rate": 5e-05, "loss": 0.7782, "mean_token_accuracy": 0.8750561475753784, "num_tokens": 303813088.0, "step": 8003 }, { "epoch": 1.0181910698384429, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.31533145904541, "learning_rate": 5e-05, "loss": 0.8054, "mean_token_accuracy": 0.8566925525665283, "num_tokens": 303849926.0, "step": 8004 }, { "epoch": 1.0183182801170334, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.605172634124756, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.8786617517471313, "num_tokens": 303888681.0, "step": 8005 }, { "epoch": 1.018445490395624, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000461578369140625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.380815505981445, "learning_rate": 5e-05, "loss": 0.7378, "mean_token_accuracy": 0.8748617172241211, "num_tokens": 303926042.0, "step": 8006 }, { "epoch": 1.0185727006742145, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.514361381530762, "learning_rate": 5e-05, "loss": 0.7133, "mean_token_accuracy": 0.8838239908218384, "num_tokens": 303962014.0, "step": 8007 }, { "epoch": 1.018699910952805, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.370009422302246, "learning_rate": 5e-05, "loss": 0.7178, "mean_token_accuracy": 0.8768061995506287, "num_tokens": 303998856.0, "step": 8008 }, { "epoch": 1.0188271212313955, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.541621208190918, "learning_rate": 5e-05, "loss": 0.7147, "mean_token_accuracy": 0.8812968134880066, "num_tokens": 304034228.0, "step": 8009 }, { "epoch": 1.018954331509986, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.270593643188477, "learning_rate": 5e-05, "loss": 0.8027, "mean_token_accuracy": 0.8520970344543457, "num_tokens": 304070477.0, "step": 8010 }, { "epoch": 1.0190815417885766, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.60357141494751, "learning_rate": 5e-05, "loss": 0.7308, "mean_token_accuracy": 0.8782801032066345, "num_tokens": 304116073.0, "step": 8011 }, { "epoch": 1.019208752067167, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.172825336456299, "learning_rate": 5e-05, "loss": 0.7391, "mean_token_accuracy": 0.8676548004150391, "num_tokens": 304149950.0, "step": 8012 }, { "epoch": 1.0193359623457576, "ewc_loss": 0.375, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.708826065063477, "learning_rate": 5e-05, "loss": 0.6897, "mean_token_accuracy": 0.8901020288467407, "num_tokens": 304191824.0, "step": 8013 }, { "epoch": 1.0194631726243482, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.984158992767334, "learning_rate": 5e-05, "loss": 0.6469, "mean_token_accuracy": 0.8888487815856934, "num_tokens": 304223616.0, "step": 8014 }, { "epoch": 1.0195903829029385, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.9774370193481445, "learning_rate": 5e-05, "loss": 0.7721, "mean_token_accuracy": 0.871987521648407, "num_tokens": 304261120.0, "step": 8015 }, { "epoch": 1.019717593181529, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.701646327972412, "learning_rate": 5e-05, "loss": 0.732, "mean_token_accuracy": 0.8609930276870728, "num_tokens": 304294796.0, "step": 8016 }, { "epoch": 1.0198448034601195, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0004634857177734375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.372325897216797, "learning_rate": 5e-05, "loss": 0.8345, "mean_token_accuracy": 0.864967942237854, "num_tokens": 304333396.0, "step": 8017 }, { "epoch": 1.01997201373871, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00046539306640625, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.313885688781738, "learning_rate": 5e-05, "loss": 0.661, "mean_token_accuracy": 0.8753492832183838, "num_tokens": 304371246.0, "step": 8018 }, { "epoch": 1.0200992240173006, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00046539306640625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.97517204284668, "learning_rate": 5e-05, "loss": 0.8226, "mean_token_accuracy": 0.8858875632286072, "num_tokens": 304408179.0, "step": 8019 }, { "epoch": 1.020226434295891, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00046539306640625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.986876487731934, "learning_rate": 5e-05, "loss": 0.6524, "mean_token_accuracy": 0.8767716884613037, "num_tokens": 304442758.0, "step": 8020 }, { "epoch": 1.0203536445744816, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00046539306640625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.583070755004883, "learning_rate": 5e-05, "loss": 0.9405, "mean_token_accuracy": 0.8822873830795288, "num_tokens": 304479568.0, "step": 8021 }, { "epoch": 1.0204808548530722, "ewc_loss": 0.375, "ewc_loss_diag": 0.00046539306640625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.395448684692383, "learning_rate": 5e-05, "loss": 0.7347, "mean_token_accuracy": 0.8764792084693909, "num_tokens": 304521255.0, "step": 8022 }, { "epoch": 1.0206080651316627, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.423911094665527, "learning_rate": 5e-05, "loss": 0.9731, "mean_token_accuracy": 0.8765226602554321, "num_tokens": 304553569.0, "step": 8023 }, { "epoch": 1.0207352754102532, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.71170711517334, "learning_rate": 5e-05, "loss": 0.9378, "mean_token_accuracy": 0.8715898990631104, "num_tokens": 304594632.0, "step": 8024 }, { "epoch": 1.0208624856888437, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.843929767608643, "learning_rate": 5e-05, "loss": 0.8616, "mean_token_accuracy": 0.8709888458251953, "num_tokens": 304635255.0, "step": 8025 }, { "epoch": 1.0209896959674343, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.015368461608887, "learning_rate": 5e-05, "loss": 0.9696, "mean_token_accuracy": 0.8730416297912598, "num_tokens": 304674472.0, "step": 8026 }, { "epoch": 1.0211169062460246, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.114591598510742, "learning_rate": 5e-05, "loss": 0.6626, "mean_token_accuracy": 0.8827546834945679, "num_tokens": 304708950.0, "step": 8027 }, { "epoch": 1.021244116524615, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.613640785217285, "learning_rate": 5e-05, "loss": 1.099, "mean_token_accuracy": 0.8713847398757935, "num_tokens": 304752849.0, "step": 8028 }, { "epoch": 1.0213713268032056, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 6.579104900360107, "learning_rate": 5e-05, "loss": 0.8187, "mean_token_accuracy": 0.874373733997345, "num_tokens": 304791936.0, "step": 8029 }, { "epoch": 1.0214985370817962, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 11.389449119567871, "learning_rate": 5e-05, "loss": 1.2308, "mean_token_accuracy": 0.8765602707862854, "num_tokens": 304832228.0, "step": 8030 }, { "epoch": 1.0216257473603867, "ewc_loss": 0.87890625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 9.171130180358887, "learning_rate": 5e-05, "loss": 1.1816, "mean_token_accuracy": 0.891486406326294, "num_tokens": 304863169.0, "step": 8031 }, { "epoch": 1.0217529576389772, "ewc_loss": 0.625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.832698822021484, "learning_rate": 5e-05, "loss": 1.0716, "mean_token_accuracy": 0.8436974883079529, "num_tokens": 304904938.0, "step": 8032 }, { "epoch": 1.0218801679175677, "ewc_loss": 0.78125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 9.39677906036377, "learning_rate": 5e-05, "loss": 1.1466, "mean_token_accuracy": 0.8751943707466125, "num_tokens": 304940590.0, "step": 8033 }, { "epoch": 1.0220073781961583, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.54708194732666, "learning_rate": 5e-05, "loss": 0.9497, "mean_token_accuracy": 0.8706960082054138, "num_tokens": 304976945.0, "step": 8034 }, { "epoch": 1.0221345884747488, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.215324401855469, "learning_rate": 5e-05, "loss": 0.9983, "mean_token_accuracy": 0.8749674558639526, "num_tokens": 305021407.0, "step": 8035 }, { "epoch": 1.0222617987533393, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.130119323730469, "learning_rate": 5e-05, "loss": 0.927, "mean_token_accuracy": 0.8792812824249268, "num_tokens": 305053141.0, "step": 8036 }, { "epoch": 1.0223890090319299, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.42972469329834, "learning_rate": 5e-05, "loss": 0.87, "mean_token_accuracy": 0.8794419765472412, "num_tokens": 305086795.0, "step": 8037 }, { "epoch": 1.0225162193105204, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.33601188659668, "learning_rate": 5e-05, "loss": 0.9772, "mean_token_accuracy": 0.860092282295227, "num_tokens": 305125759.0, "step": 8038 }, { "epoch": 1.0226434295891107, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.414180755615234, "learning_rate": 5e-05, "loss": 0.7715, "mean_token_accuracy": 0.8742344975471497, "num_tokens": 305165169.0, "step": 8039 }, { "epoch": 1.0227706398677012, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.79979419708252, "learning_rate": 5e-05, "loss": 0.8831, "mean_token_accuracy": 0.8909493088722229, "num_tokens": 305203457.0, "step": 8040 }, { "epoch": 1.0228978501462918, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.953794479370117, "learning_rate": 5e-05, "loss": 0.7846, "mean_token_accuracy": 0.8647787570953369, "num_tokens": 305241021.0, "step": 8041 }, { "epoch": 1.0230250604248823, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.04161548614502, "learning_rate": 5e-05, "loss": 0.9842, "mean_token_accuracy": 0.8548177480697632, "num_tokens": 305277579.0, "step": 8042 }, { "epoch": 1.0231522707034728, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.02483606338501, "learning_rate": 5e-05, "loss": 0.8368, "mean_token_accuracy": 0.8661090731620789, "num_tokens": 305318027.0, "step": 8043 }, { "epoch": 1.0232794809820633, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.963109016418457, "learning_rate": 5e-05, "loss": 0.9432, "mean_token_accuracy": 0.8746896982192993, "num_tokens": 305351249.0, "step": 8044 }, { "epoch": 1.0234066912606539, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.269006252288818, "learning_rate": 5e-05, "loss": 0.8327, "mean_token_accuracy": 0.8763750791549683, "num_tokens": 305387906.0, "step": 8045 }, { "epoch": 1.0235339015392444, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.59650993347168, "learning_rate": 5e-05, "loss": 0.8711, "mean_token_accuracy": 0.8819530606269836, "num_tokens": 305419511.0, "step": 8046 }, { "epoch": 1.023661111817835, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.445281982421875, "learning_rate": 5e-05, "loss": 0.8471, "mean_token_accuracy": 0.8762121796607971, "num_tokens": 305457483.0, "step": 8047 }, { "epoch": 1.0237883220964255, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.199707984924316, "learning_rate": 5e-05, "loss": 0.8402, "mean_token_accuracy": 0.8782514333724976, "num_tokens": 305486498.0, "step": 8048 }, { "epoch": 1.023915532375016, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.394801616668701, "learning_rate": 5e-05, "loss": 0.8108, "mean_token_accuracy": 0.8754633069038391, "num_tokens": 305525250.0, "step": 8049 }, { "epoch": 1.0240427426536065, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.118462562561035, "learning_rate": 5e-05, "loss": 0.8579, "mean_token_accuracy": 0.8695855140686035, "num_tokens": 305566229.0, "step": 8050 }, { "epoch": 1.0241699529321968, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.0891218185424805, "learning_rate": 5e-05, "loss": 0.7243, "mean_token_accuracy": 0.8826698660850525, "num_tokens": 305601884.0, "step": 8051 }, { "epoch": 1.0242971632107873, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.455824851989746, "learning_rate": 5e-05, "loss": 0.8925, "mean_token_accuracy": 0.8697028160095215, "num_tokens": 305644495.0, "step": 8052 }, { "epoch": 1.0244243734893779, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.677730560302734, "learning_rate": 5e-05, "loss": 0.7016, "mean_token_accuracy": 0.878972053527832, "num_tokens": 305680662.0, "step": 8053 }, { "epoch": 1.0245515837679684, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.97470474243164, "learning_rate": 5e-05, "loss": 0.9148, "mean_token_accuracy": 0.8779066801071167, "num_tokens": 305720821.0, "step": 8054 }, { "epoch": 1.024678794046559, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.571563243865967, "learning_rate": 5e-05, "loss": 0.7381, "mean_token_accuracy": 0.8676164150238037, "num_tokens": 305757312.0, "step": 8055 }, { "epoch": 1.0248060043251495, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.324077606201172, "learning_rate": 5e-05, "loss": 1.0155, "mean_token_accuracy": 0.8650131225585938, "num_tokens": 305794372.0, "step": 8056 }, { "epoch": 1.02493321460374, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.079761505126953, "learning_rate": 5e-05, "loss": 0.8353, "mean_token_accuracy": 0.8737833499908447, "num_tokens": 305839975.0, "step": 8057 }, { "epoch": 1.0250604248823305, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.006270408630371, "learning_rate": 5e-05, "loss": 0.94, "mean_token_accuracy": 0.8822062015533447, "num_tokens": 305877964.0, "step": 8058 }, { "epoch": 1.025187635160921, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 7.9014129638671875, "learning_rate": 5e-05, "loss": 0.9358, "mean_token_accuracy": 0.8719377517700195, "num_tokens": 305911077.0, "step": 8059 }, { "epoch": 1.0253148454395116, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.838598251342773, "learning_rate": 5e-05, "loss": 0.8149, "mean_token_accuracy": 0.8806094527244568, "num_tokens": 305947741.0, "step": 8060 }, { "epoch": 1.025442055718102, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.412290573120117, "learning_rate": 5e-05, "loss": 0.9382, "mean_token_accuracy": 0.871107280254364, "num_tokens": 305986872.0, "step": 8061 }, { "epoch": 1.0255692659966926, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.917184352874756, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8660035729408264, "num_tokens": 306023437.0, "step": 8062 }, { "epoch": 1.0256964762752832, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.69411849975586, "learning_rate": 5e-05, "loss": 0.8828, "mean_token_accuracy": 0.882731020450592, "num_tokens": 306063026.0, "step": 8063 }, { "epoch": 1.0258236865538735, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.705513000488281, "learning_rate": 5e-05, "loss": 0.749, "mean_token_accuracy": 0.8668887615203857, "num_tokens": 306102574.0, "step": 8064 }, { "epoch": 1.025950896832464, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.775589942932129, "learning_rate": 5e-05, "loss": 0.9447, "mean_token_accuracy": 0.8658028244972229, "num_tokens": 306137194.0, "step": 8065 }, { "epoch": 1.0260781071110545, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.879453182220459, "learning_rate": 5e-05, "loss": 0.7617, "mean_token_accuracy": 0.8707420825958252, "num_tokens": 306171539.0, "step": 8066 }, { "epoch": 1.026205317389645, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.629378318786621, "learning_rate": 5e-05, "loss": 0.9309, "mean_token_accuracy": 0.8682337999343872, "num_tokens": 306212385.0, "step": 8067 }, { "epoch": 1.0263325276682356, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.1331305503845215, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.8762986063957214, "num_tokens": 306247624.0, "step": 8068 }, { "epoch": 1.026459737946826, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.32707691192627, "learning_rate": 5e-05, "loss": 0.9183, "mean_token_accuracy": 0.8590232133865356, "num_tokens": 306286445.0, "step": 8069 }, { "epoch": 1.0265869482254166, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.32253360748291, "learning_rate": 5e-05, "loss": 0.7857, "mean_token_accuracy": 0.8814841508865356, "num_tokens": 306327792.0, "step": 8070 }, { "epoch": 1.0267141585040072, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.0498628616333, "learning_rate": 5e-05, "loss": 0.8918, "mean_token_accuracy": 0.8578509092330933, "num_tokens": 306366811.0, "step": 8071 }, { "epoch": 1.0268413687825977, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.359780311584473, "learning_rate": 5e-05, "loss": 0.8159, "mean_token_accuracy": 0.868084192276001, "num_tokens": 306403938.0, "step": 8072 }, { "epoch": 1.0269685790611882, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.8931355476379395, "learning_rate": 5e-05, "loss": 0.767, "mean_token_accuracy": 0.8852542638778687, "num_tokens": 306440623.0, "step": 8073 }, { "epoch": 1.0270957893397787, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.296004295349121, "learning_rate": 5e-05, "loss": 0.7746, "mean_token_accuracy": 0.8718960881233215, "num_tokens": 306477623.0, "step": 8074 }, { "epoch": 1.0272229996183693, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.893772125244141, "learning_rate": 5e-05, "loss": 0.8075, "mean_token_accuracy": 0.870148241519928, "num_tokens": 306513586.0, "step": 8075 }, { "epoch": 1.0273502098969596, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.139256477355957, "learning_rate": 5e-05, "loss": 0.6919, "mean_token_accuracy": 0.8884928822517395, "num_tokens": 306548121.0, "step": 8076 }, { "epoch": 1.02747742017555, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.989193439483643, "learning_rate": 5e-05, "loss": 0.7851, "mean_token_accuracy": 0.8828518986701965, "num_tokens": 306596776.0, "step": 8077 }, { "epoch": 1.0276046304541406, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.014228343963623, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8658064007759094, "num_tokens": 306635497.0, "step": 8078 }, { "epoch": 1.0277318407327312, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.121723175048828, "learning_rate": 5e-05, "loss": 0.8861, "mean_token_accuracy": 0.860588788986206, "num_tokens": 306671933.0, "step": 8079 }, { "epoch": 1.0278590510113217, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.836202144622803, "learning_rate": 5e-05, "loss": 0.7003, "mean_token_accuracy": 0.8768550157546997, "num_tokens": 306709313.0, "step": 8080 }, { "epoch": 1.0279862612899122, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.317488670349121, "learning_rate": 5e-05, "loss": 0.8597, "mean_token_accuracy": 0.8583632707595825, "num_tokens": 306748181.0, "step": 8081 }, { "epoch": 1.0281134715685027, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.696990013122559, "learning_rate": 5e-05, "loss": 0.6787, "mean_token_accuracy": 0.8851367235183716, "num_tokens": 306786353.0, "step": 8082 }, { "epoch": 1.0282406818470933, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.49609375, "learning_rate": 5e-05, "loss": 0.8522, "mean_token_accuracy": 0.874401330947876, "num_tokens": 306821660.0, "step": 8083 }, { "epoch": 1.0283678921256838, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.708206653594971, "learning_rate": 5e-05, "loss": 0.6466, "mean_token_accuracy": 0.8952947854995728, "num_tokens": 306858479.0, "step": 8084 }, { "epoch": 1.0284951024042743, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.477815628051758, "learning_rate": 5e-05, "loss": 0.8627, "mean_token_accuracy": 0.8748591542243958, "num_tokens": 306890483.0, "step": 8085 }, { "epoch": 1.0286223126828649, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.932498455047607, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.8755756616592407, "num_tokens": 306931865.0, "step": 8086 }, { "epoch": 1.0287495229614554, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.170940399169922, "learning_rate": 5e-05, "loss": 0.8627, "mean_token_accuracy": 0.8700740337371826, "num_tokens": 306966907.0, "step": 8087 }, { "epoch": 1.0288767332400457, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.311708927154541, "learning_rate": 5e-05, "loss": 0.7699, "mean_token_accuracy": 0.8796085119247437, "num_tokens": 307006032.0, "step": 8088 }, { "epoch": 1.0290039435186362, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.646467208862305, "learning_rate": 5e-05, "loss": 0.7923, "mean_token_accuracy": 0.868211030960083, "num_tokens": 307045621.0, "step": 8089 }, { "epoch": 1.0291311537972267, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.705719947814941, "learning_rate": 5e-05, "loss": 0.8173, "mean_token_accuracy": 0.866998553276062, "num_tokens": 307078768.0, "step": 8090 }, { "epoch": 1.0292583640758173, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.138125896453857, "learning_rate": 5e-05, "loss": 0.7142, "mean_token_accuracy": 0.8779910206794739, "num_tokens": 307120089.0, "step": 8091 }, { "epoch": 1.0293855743544078, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.021539688110352, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8701404333114624, "num_tokens": 307157478.0, "step": 8092 }, { "epoch": 1.0295127846329983, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.801578521728516, "learning_rate": 5e-05, "loss": 0.6811, "mean_token_accuracy": 0.8767043352127075, "num_tokens": 307199842.0, "step": 8093 }, { "epoch": 1.0296399949115889, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.289576530456543, "learning_rate": 5e-05, "loss": 0.8166, "mean_token_accuracy": 0.8730183839797974, "num_tokens": 307240609.0, "step": 8094 }, { "epoch": 1.0297672051901794, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.599037170410156, "learning_rate": 5e-05, "loss": 0.7169, "mean_token_accuracy": 0.8647998571395874, "num_tokens": 307279112.0, "step": 8095 }, { "epoch": 1.02989441546877, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.512001037597656, "learning_rate": 5e-05, "loss": 0.8164, "mean_token_accuracy": 0.881565511226654, "num_tokens": 307313718.0, "step": 8096 }, { "epoch": 1.0300216257473604, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.494137287139893, "learning_rate": 5e-05, "loss": 0.6948, "mean_token_accuracy": 0.8739308714866638, "num_tokens": 307353660.0, "step": 8097 }, { "epoch": 1.030148836025951, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.689058303833008, "learning_rate": 5e-05, "loss": 0.8882, "mean_token_accuracy": 0.8657612800598145, "num_tokens": 307393974.0, "step": 8098 }, { "epoch": 1.0302760463045415, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.586744785308838, "learning_rate": 5e-05, "loss": 0.6847, "mean_token_accuracy": 0.8826606273651123, "num_tokens": 307434877.0, "step": 8099 }, { "epoch": 1.0304032565831318, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.614642143249512, "learning_rate": 5e-05, "loss": 0.8511, "mean_token_accuracy": 0.8859438300132751, "num_tokens": 307473167.0, "step": 8100 }, { "epoch": 1.0305304668617223, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.959371566772461, "learning_rate": 5e-05, "loss": 0.7169, "mean_token_accuracy": 0.8853933811187744, "num_tokens": 307513781.0, "step": 8101 }, { "epoch": 1.0306576771403129, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.131040573120117, "learning_rate": 5e-05, "loss": 0.7849, "mean_token_accuracy": 0.893564760684967, "num_tokens": 307549524.0, "step": 8102 }, { "epoch": 1.0307848874189034, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.449608325958252, "learning_rate": 5e-05, "loss": 0.7949, "mean_token_accuracy": 0.8734456896781921, "num_tokens": 307585196.0, "step": 8103 }, { "epoch": 1.030912097697494, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.545083522796631, "learning_rate": 5e-05, "loss": 0.8222, "mean_token_accuracy": 0.8640702962875366, "num_tokens": 307623002.0, "step": 8104 }, { "epoch": 1.0310393079760845, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.714265823364258, "learning_rate": 5e-05, "loss": 0.7925, "mean_token_accuracy": 0.8769087195396423, "num_tokens": 307667736.0, "step": 8105 }, { "epoch": 1.031166518254675, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.298235893249512, "learning_rate": 5e-05, "loss": 0.7102, "mean_token_accuracy": 0.8835495710372925, "num_tokens": 307706031.0, "step": 8106 }, { "epoch": 1.0312937285332655, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.655413627624512, "learning_rate": 5e-05, "loss": 0.8158, "mean_token_accuracy": 0.8614787459373474, "num_tokens": 307746333.0, "step": 8107 }, { "epoch": 1.031420938811856, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.41419792175293, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8607549667358398, "num_tokens": 307784479.0, "step": 8108 }, { "epoch": 1.0315481490904466, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.3659491539001465, "learning_rate": 5e-05, "loss": 0.7348, "mean_token_accuracy": 0.8719923496246338, "num_tokens": 307818245.0, "step": 8109 }, { "epoch": 1.031675359369037, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.660740852355957, "learning_rate": 5e-05, "loss": 0.7281, "mean_token_accuracy": 0.879316508769989, "num_tokens": 307858730.0, "step": 8110 }, { "epoch": 1.0318025696476276, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.052427291870117, "learning_rate": 5e-05, "loss": 0.6658, "mean_token_accuracy": 0.8881661891937256, "num_tokens": 307897067.0, "step": 8111 }, { "epoch": 1.0319297799262181, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.953629016876221, "learning_rate": 5e-05, "loss": 0.7719, "mean_token_accuracy": 0.8725490570068359, "num_tokens": 307935896.0, "step": 8112 }, { "epoch": 1.0320569902048085, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.728660583496094, "learning_rate": 5e-05, "loss": 0.7406, "mean_token_accuracy": 0.8555828928947449, "num_tokens": 307970930.0, "step": 8113 }, { "epoch": 1.032184200483399, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.335953712463379, "learning_rate": 5e-05, "loss": 0.8401, "mean_token_accuracy": 0.8625643253326416, "num_tokens": 308015857.0, "step": 8114 }, { "epoch": 1.0323114107619895, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.375551223754883, "learning_rate": 5e-05, "loss": 0.6786, "mean_token_accuracy": 0.8749871253967285, "num_tokens": 308055432.0, "step": 8115 }, { "epoch": 1.03243862104058, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.849270820617676, "learning_rate": 5e-05, "loss": 0.89, "mean_token_accuracy": 0.8619610071182251, "num_tokens": 308094922.0, "step": 8116 }, { "epoch": 1.0325658313191706, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.116425514221191, "learning_rate": 5e-05, "loss": 0.6673, "mean_token_accuracy": 0.8761190176010132, "num_tokens": 308132941.0, "step": 8117 }, { "epoch": 1.032693041597761, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.309708595275879, "learning_rate": 5e-05, "loss": 0.9986, "mean_token_accuracy": 0.8592214584350586, "num_tokens": 308176749.0, "step": 8118 }, { "epoch": 1.0328202518763516, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.400189399719238, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.8648218512535095, "num_tokens": 308215845.0, "step": 8119 }, { "epoch": 1.0329474621549422, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0004673004150390625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.141702651977539, "learning_rate": 5e-05, "loss": 0.9398, "mean_token_accuracy": 0.8795557022094727, "num_tokens": 308254220.0, "step": 8120 }, { "epoch": 1.0330746724335327, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.388900279998779, "learning_rate": 5e-05, "loss": 0.9027, "mean_token_accuracy": 0.8650804758071899, "num_tokens": 308290059.0, "step": 8121 }, { "epoch": 1.0332018827121232, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.943568706512451, "learning_rate": 5e-05, "loss": 0.9181, "mean_token_accuracy": 0.8535332083702087, "num_tokens": 308320478.0, "step": 8122 }, { "epoch": 1.0333290929907137, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.409865379333496, "learning_rate": 5e-05, "loss": 0.8997, "mean_token_accuracy": 0.8821178078651428, "num_tokens": 308356094.0, "step": 8123 }, { "epoch": 1.0334563032693043, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.579695701599121, "learning_rate": 5e-05, "loss": 0.7098, "mean_token_accuracy": 0.8747826814651489, "num_tokens": 308393120.0, "step": 8124 }, { "epoch": 1.0335835135478946, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.307658195495605, "learning_rate": 5e-05, "loss": 0.9336, "mean_token_accuracy": 0.8838091492652893, "num_tokens": 308424153.0, "step": 8125 }, { "epoch": 1.033710723826485, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.183588027954102, "learning_rate": 5e-05, "loss": 0.6958, "mean_token_accuracy": 0.877189576625824, "num_tokens": 308462884.0, "step": 8126 }, { "epoch": 1.0338379341050756, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.760334968566895, "learning_rate": 5e-05, "loss": 0.9476, "mean_token_accuracy": 0.8946918845176697, "num_tokens": 308501185.0, "step": 8127 }, { "epoch": 1.0339651443836662, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.055184841156006, "learning_rate": 5e-05, "loss": 0.9091, "mean_token_accuracy": 0.8625980615615845, "num_tokens": 308538770.0, "step": 8128 }, { "epoch": 1.0340923546622567, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.181462287902832, "learning_rate": 5e-05, "loss": 0.9731, "mean_token_accuracy": 0.8784162402153015, "num_tokens": 308575224.0, "step": 8129 }, { "epoch": 1.0342195649408472, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.119791984558105, "learning_rate": 5e-05, "loss": 0.9555, "mean_token_accuracy": 0.8853915929794312, "num_tokens": 308611859.0, "step": 8130 }, { "epoch": 1.0343467752194377, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.884234428405762, "learning_rate": 5e-05, "loss": 0.8939, "mean_token_accuracy": 0.8585847616195679, "num_tokens": 308648212.0, "step": 8131 }, { "epoch": 1.0344739854980283, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.518840789794922, "learning_rate": 5e-05, "loss": 0.962, "mean_token_accuracy": 0.8715497255325317, "num_tokens": 308689830.0, "step": 8132 }, { "epoch": 1.0346011957766188, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.266866683959961, "learning_rate": 5e-05, "loss": 0.8401, "mean_token_accuracy": 0.8580441474914551, "num_tokens": 308726238.0, "step": 8133 }, { "epoch": 1.0347284060552093, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.223649978637695, "learning_rate": 5e-05, "loss": 0.8836, "mean_token_accuracy": 0.8669210076332092, "num_tokens": 308755090.0, "step": 8134 }, { "epoch": 1.0348556163337999, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.555574417114258, "learning_rate": 5e-05, "loss": 0.8619, "mean_token_accuracy": 0.8605365753173828, "num_tokens": 308794025.0, "step": 8135 }, { "epoch": 1.0349828266123904, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.459778785705566, "learning_rate": 5e-05, "loss": 0.7917, "mean_token_accuracy": 0.8740990161895752, "num_tokens": 308831521.0, "step": 8136 }, { "epoch": 1.0351100368909807, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.007905006408691, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8775594830513, "num_tokens": 308868113.0, "step": 8137 }, { "epoch": 1.0352372471695712, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.962533473968506, "learning_rate": 5e-05, "loss": 0.7159, "mean_token_accuracy": 0.8733532428741455, "num_tokens": 308905070.0, "step": 8138 }, { "epoch": 1.0353644574481617, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.293022155761719, "learning_rate": 5e-05, "loss": 0.8725, "mean_token_accuracy": 0.8668147325515747, "num_tokens": 308942703.0, "step": 8139 }, { "epoch": 1.0354916677267523, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.66608190536499, "learning_rate": 5e-05, "loss": 0.6994, "mean_token_accuracy": 0.8759711980819702, "num_tokens": 308981328.0, "step": 8140 }, { "epoch": 1.0356188780053428, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.727348327636719, "learning_rate": 5e-05, "loss": 0.8486, "mean_token_accuracy": 0.8849518299102783, "num_tokens": 309017751.0, "step": 8141 }, { "epoch": 1.0357460882839333, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.421168804168701, "learning_rate": 5e-05, "loss": 0.7051, "mean_token_accuracy": 0.8732064962387085, "num_tokens": 309052216.0, "step": 8142 }, { "epoch": 1.0358732985625239, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.168972969055176, "learning_rate": 5e-05, "loss": 0.9655, "mean_token_accuracy": 0.8732444047927856, "num_tokens": 309091323.0, "step": 8143 }, { "epoch": 1.0360005088411144, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.56880521774292, "learning_rate": 5e-05, "loss": 0.7618, "mean_token_accuracy": 0.8754209876060486, "num_tokens": 309125324.0, "step": 8144 }, { "epoch": 1.036127719119705, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.236496925354004, "learning_rate": 5e-05, "loss": 0.9821, "mean_token_accuracy": 0.8731738328933716, "num_tokens": 309166009.0, "step": 8145 }, { "epoch": 1.0362549293982954, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.144872188568115, "learning_rate": 5e-05, "loss": 0.8227, "mean_token_accuracy": 0.8839576244354248, "num_tokens": 309205822.0, "step": 8146 }, { "epoch": 1.036382139676886, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.801719665527344, "learning_rate": 5e-05, "loss": 0.9501, "mean_token_accuracy": 0.8758852481842041, "num_tokens": 309243463.0, "step": 8147 }, { "epoch": 1.0365093499554765, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.590188980102539, "learning_rate": 5e-05, "loss": 0.912, "mean_token_accuracy": 0.8684619665145874, "num_tokens": 309276904.0, "step": 8148 }, { "epoch": 1.0366365602340668, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.274497985839844, "learning_rate": 5e-05, "loss": 0.905, "mean_token_accuracy": 0.8703185319900513, "num_tokens": 309314292.0, "step": 8149 }, { "epoch": 1.0367637705126573, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.573872089385986, "learning_rate": 5e-05, "loss": 0.8958, "mean_token_accuracy": 0.8643108010292053, "num_tokens": 309349231.0, "step": 8150 }, { "epoch": 1.0368909807912479, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.149552345275879, "learning_rate": 5e-05, "loss": 0.9543, "mean_token_accuracy": 0.8502690196037292, "num_tokens": 309387901.0, "step": 8151 }, { "epoch": 1.0370181910698384, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.254316806793213, "learning_rate": 5e-05, "loss": 0.7893, "mean_token_accuracy": 0.8758056163787842, "num_tokens": 309427167.0, "step": 8152 }, { "epoch": 1.037145401348429, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.298566818237305, "learning_rate": 5e-05, "loss": 0.8922, "mean_token_accuracy": 0.8733241558074951, "num_tokens": 309464121.0, "step": 8153 }, { "epoch": 1.0372726116270194, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.021382808685303, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8767497539520264, "num_tokens": 309506759.0, "step": 8154 }, { "epoch": 1.03739982190561, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.373516082763672, "learning_rate": 5e-05, "loss": 0.8546, "mean_token_accuracy": 0.8838975429534912, "num_tokens": 309546059.0, "step": 8155 }, { "epoch": 1.0375270321842005, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.092587471008301, "learning_rate": 5e-05, "loss": 0.7911, "mean_token_accuracy": 0.8665933609008789, "num_tokens": 309585076.0, "step": 8156 }, { "epoch": 1.037654242462791, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.129936218261719, "learning_rate": 5e-05, "loss": 0.84, "mean_token_accuracy": 0.881378173828125, "num_tokens": 309623354.0, "step": 8157 }, { "epoch": 1.0377814527413816, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.32020378112793, "learning_rate": 5e-05, "loss": 0.8262, "mean_token_accuracy": 0.8591341972351074, "num_tokens": 309668420.0, "step": 8158 }, { "epoch": 1.037908663019972, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.790895462036133, "learning_rate": 5e-05, "loss": 0.8196, "mean_token_accuracy": 0.8772196769714355, "num_tokens": 309711417.0, "step": 8159 }, { "epoch": 1.0380358732985626, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.436431884765625, "learning_rate": 5e-05, "loss": 0.7771, "mean_token_accuracy": 0.8759689331054688, "num_tokens": 309749938.0, "step": 8160 }, { "epoch": 1.0381630835771531, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.6182475090026855, "learning_rate": 5e-05, "loss": 0.7666, "mean_token_accuracy": 0.8769558668136597, "num_tokens": 309783328.0, "step": 8161 }, { "epoch": 1.0382902938557435, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.420562744140625, "learning_rate": 5e-05, "loss": 0.7514, "mean_token_accuracy": 0.8806313276290894, "num_tokens": 309821347.0, "step": 8162 }, { "epoch": 1.038417504134334, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.599365711212158, "learning_rate": 5e-05, "loss": 0.8152, "mean_token_accuracy": 0.8550706505775452, "num_tokens": 309858430.0, "step": 8163 }, { "epoch": 1.0385447144129245, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.3305277824401855, "learning_rate": 5e-05, "loss": 0.7478, "mean_token_accuracy": 0.871692419052124, "num_tokens": 309896096.0, "step": 8164 }, { "epoch": 1.038671924691515, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.6363139152526855, "learning_rate": 5e-05, "loss": 0.7611, "mean_token_accuracy": 0.8720219135284424, "num_tokens": 309929069.0, "step": 8165 }, { "epoch": 1.0387991349701056, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.232176303863525, "learning_rate": 5e-05, "loss": 0.7576, "mean_token_accuracy": 0.8605279326438904, "num_tokens": 309969706.0, "step": 8166 }, { "epoch": 1.038926345248696, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.675273895263672, "learning_rate": 5e-05, "loss": 0.766, "mean_token_accuracy": 0.867917537689209, "num_tokens": 310003990.0, "step": 8167 }, { "epoch": 1.0390535555272866, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.181509971618652, "learning_rate": 5e-05, "loss": 0.7738, "mean_token_accuracy": 0.8536821007728577, "num_tokens": 310036253.0, "step": 8168 }, { "epoch": 1.0391807658058771, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.701570510864258, "learning_rate": 5e-05, "loss": 0.6989, "mean_token_accuracy": 0.8882694244384766, "num_tokens": 310073189.0, "step": 8169 }, { "epoch": 1.0393079760844677, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.128057956695557, "learning_rate": 5e-05, "loss": 0.7305, "mean_token_accuracy": 0.8666940927505493, "num_tokens": 310110432.0, "step": 8170 }, { "epoch": 1.0394351863630582, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.740896701812744, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8657596111297607, "num_tokens": 310154830.0, "step": 8171 }, { "epoch": 1.0395623966416487, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.047963619232178, "learning_rate": 5e-05, "loss": 0.7277, "mean_token_accuracy": 0.8589552640914917, "num_tokens": 310187270.0, "step": 8172 }, { "epoch": 1.0396896069202393, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.84535026550293, "learning_rate": 5e-05, "loss": 0.7544, "mean_token_accuracy": 0.8744990825653076, "num_tokens": 310223736.0, "step": 8173 }, { "epoch": 1.0398168171988296, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.945250988006592, "learning_rate": 5e-05, "loss": 0.7713, "mean_token_accuracy": 0.8555434942245483, "num_tokens": 310259262.0, "step": 8174 }, { "epoch": 1.03994402747742, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.938258171081543, "learning_rate": 5e-05, "loss": 0.7647, "mean_token_accuracy": 0.8736695051193237, "num_tokens": 310300438.0, "step": 8175 }, { "epoch": 1.0400712377560106, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.839788913726807, "learning_rate": 5e-05, "loss": 0.6611, "mean_token_accuracy": 0.8826299905776978, "num_tokens": 310336630.0, "step": 8176 }, { "epoch": 1.0401984480346012, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.084639549255371, "learning_rate": 5e-05, "loss": 0.7865, "mean_token_accuracy": 0.8683886528015137, "num_tokens": 310368755.0, "step": 8177 }, { "epoch": 1.0403256583131917, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.705416202545166, "learning_rate": 5e-05, "loss": 0.6487, "mean_token_accuracy": 0.8834356069564819, "num_tokens": 310403520.0, "step": 8178 }, { "epoch": 1.0404528685917822, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.288948059082031, "learning_rate": 5e-05, "loss": 0.7921, "mean_token_accuracy": 0.8701422214508057, "num_tokens": 310433805.0, "step": 8179 }, { "epoch": 1.0405800788703727, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.573408603668213, "learning_rate": 5e-05, "loss": 0.6958, "mean_token_accuracy": 0.8690313100814819, "num_tokens": 310469801.0, "step": 8180 }, { "epoch": 1.0407072891489633, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.427560806274414, "learning_rate": 5e-05, "loss": 0.7743, "mean_token_accuracy": 0.8889845609664917, "num_tokens": 310504972.0, "step": 8181 }, { "epoch": 1.0408344994275538, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.61620569229126, "learning_rate": 5e-05, "loss": 0.6995, "mean_token_accuracy": 0.8689250946044922, "num_tokens": 310537851.0, "step": 8182 }, { "epoch": 1.0409617097061443, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.327667236328125, "learning_rate": 5e-05, "loss": 0.8308, "mean_token_accuracy": 0.8699017763137817, "num_tokens": 310572647.0, "step": 8183 }, { "epoch": 1.0410889199847349, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.93350887298584, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.8661995530128479, "num_tokens": 310612239.0, "step": 8184 }, { "epoch": 1.0412161302633254, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.826720714569092, "learning_rate": 5e-05, "loss": 0.8035, "mean_token_accuracy": 0.8673031330108643, "num_tokens": 310650442.0, "step": 8185 }, { "epoch": 1.0413433405419157, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.516702175140381, "learning_rate": 5e-05, "loss": 0.7804, "mean_token_accuracy": 0.8707403540611267, "num_tokens": 310687687.0, "step": 8186 }, { "epoch": 1.0414705508205062, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.02977180480957, "learning_rate": 5e-05, "loss": 0.751, "mean_token_accuracy": 0.8635797500610352, "num_tokens": 310725165.0, "step": 8187 }, { "epoch": 1.0415977610990967, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.266833305358887, "learning_rate": 5e-05, "loss": 0.7657, "mean_token_accuracy": 0.885854184627533, "num_tokens": 310762532.0, "step": 8188 }, { "epoch": 1.0417249713776873, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.222215175628662, "learning_rate": 5e-05, "loss": 0.6945, "mean_token_accuracy": 0.8620463609695435, "num_tokens": 310800903.0, "step": 8189 }, { "epoch": 1.0418521816562778, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 9.219459533691406, "learning_rate": 5e-05, "loss": 0.9399, "mean_token_accuracy": 0.8537095785140991, "num_tokens": 310841391.0, "step": 8190 }, { "epoch": 1.0419793919348683, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00023555755615234375, "grad_norm": 5.613845348358154, "learning_rate": 5e-05, "loss": 0.6655, "mean_token_accuracy": 0.8665368556976318, "num_tokens": 310879027.0, "step": 8191 }, { "epoch": 1.0421066022134589, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.324793815612793, "learning_rate": 5e-05, "loss": 1.021, "mean_token_accuracy": 0.8749233484268188, "num_tokens": 310918448.0, "step": 8192 }, { "epoch": 1.0422338124920494, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 6.290380954742432, "learning_rate": 5e-05, "loss": 0.7923, "mean_token_accuracy": 0.8687152862548828, "num_tokens": 310955641.0, "step": 8193 }, { "epoch": 1.04236102277064, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.60374641418457, "learning_rate": 5e-05, "loss": 1.2116, "mean_token_accuracy": 0.8490414023399353, "num_tokens": 310996487.0, "step": 8194 }, { "epoch": 1.0424882330492304, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 8.442822456359863, "learning_rate": 5e-05, "loss": 1.1326, "mean_token_accuracy": 0.8645997047424316, "num_tokens": 311038208.0, "step": 8195 }, { "epoch": 1.042615443327821, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.818229675292969, "learning_rate": 5e-05, "loss": 0.9251, "mean_token_accuracy": 0.871989905834198, "num_tokens": 311073720.0, "step": 8196 }, { "epoch": 1.0427426536064115, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 9.515786170959473, "learning_rate": 5e-05, "loss": 1.1021, "mean_token_accuracy": 0.8862698078155518, "num_tokens": 311115488.0, "step": 8197 }, { "epoch": 1.0428698638850018, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.97715425491333, "learning_rate": 5e-05, "loss": 0.7891, "mean_token_accuracy": 0.8705914616584778, "num_tokens": 311159217.0, "step": 8198 }, { "epoch": 1.0429970741635923, "ewc_loss": 0.828125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 11.055302619934082, "learning_rate": 5e-05, "loss": 1.2281, "mean_token_accuracy": 0.8629014492034912, "num_tokens": 311196058.0, "step": 8199 }, { "epoch": 1.0431242844421829, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 7.912237167358398, "learning_rate": 5e-05, "loss": 0.9828, "mean_token_accuracy": 0.876911997795105, "num_tokens": 311231024.0, "step": 8200 }, { "epoch": 1.0432514947207734, "ewc_loss": 0.75, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 11.141082763671875, "learning_rate": 5e-05, "loss": 1.1144, "mean_token_accuracy": 0.8715980052947998, "num_tokens": 311266883.0, "step": 8201 }, { "epoch": 1.043378704999364, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 8.99241828918457, "learning_rate": 5e-05, "loss": 1.224, "mean_token_accuracy": 0.8556696772575378, "num_tokens": 311302594.0, "step": 8202 }, { "epoch": 1.0435059152779544, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.34764289855957, "learning_rate": 5e-05, "loss": 0.9749, "mean_token_accuracy": 0.8915663957595825, "num_tokens": 311334384.0, "step": 8203 }, { "epoch": 1.043633125556545, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 8.515982627868652, "learning_rate": 5e-05, "loss": 1.0242, "mean_token_accuracy": 0.8742225170135498, "num_tokens": 311381350.0, "step": 8204 }, { "epoch": 1.0437603358351355, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.912561416625977, "learning_rate": 5e-05, "loss": 1.0242, "mean_token_accuracy": 0.8789443969726562, "num_tokens": 311418107.0, "step": 8205 }, { "epoch": 1.043887546113726, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.87827205657959, "learning_rate": 5e-05, "loss": 0.9112, "mean_token_accuracy": 0.8711689710617065, "num_tokens": 311457881.0, "step": 8206 }, { "epoch": 1.0440147563923166, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.930563926696777, "learning_rate": 5e-05, "loss": 1.0442, "mean_token_accuracy": 0.8668093085289001, "num_tokens": 311492821.0, "step": 8207 }, { "epoch": 1.044141966670907, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.608299255371094, "learning_rate": 5e-05, "loss": 0.8959, "mean_token_accuracy": 0.872627854347229, "num_tokens": 311527981.0, "step": 8208 }, { "epoch": 1.0442691769494976, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.240089416503906, "learning_rate": 5e-05, "loss": 1.0808, "mean_token_accuracy": 0.8663108944892883, "num_tokens": 311568853.0, "step": 8209 }, { "epoch": 1.0443963872280881, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 7.944091796875, "learning_rate": 5e-05, "loss": 0.917, "mean_token_accuracy": 0.8830361366271973, "num_tokens": 311613321.0, "step": 8210 }, { "epoch": 1.0445235975066784, "ewc_loss": 0.71875, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 10.242277145385742, "learning_rate": 5e-05, "loss": 1.0948, "mean_token_accuracy": 0.8668572306632996, "num_tokens": 311656849.0, "step": 8211 }, { "epoch": 1.044650807785269, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.390335083007812, "learning_rate": 5e-05, "loss": 1.0383, "mean_token_accuracy": 0.8593230247497559, "num_tokens": 311692231.0, "step": 8212 }, { "epoch": 1.0447780180638595, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.81329345703125, "learning_rate": 5e-05, "loss": 1.0015, "mean_token_accuracy": 0.8830519318580627, "num_tokens": 311732793.0, "step": 8213 }, { "epoch": 1.04490522834245, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.457381248474121, "learning_rate": 5e-05, "loss": 0.9919, "mean_token_accuracy": 0.8694942593574524, "num_tokens": 311773320.0, "step": 8214 }, { "epoch": 1.0450324386210406, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.253938674926758, "learning_rate": 5e-05, "loss": 0.9862, "mean_token_accuracy": 0.8697657585144043, "num_tokens": 311822187.0, "step": 8215 }, { "epoch": 1.045159648899631, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.228779792785645, "learning_rate": 5e-05, "loss": 0.9133, "mean_token_accuracy": 0.8701207637786865, "num_tokens": 311860272.0, "step": 8216 }, { "epoch": 1.0452868591782216, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.903755187988281, "learning_rate": 5e-05, "loss": 0.9543, "mean_token_accuracy": 0.8718441724777222, "num_tokens": 311894479.0, "step": 8217 }, { "epoch": 1.0454140694568121, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.77971076965332, "learning_rate": 5e-05, "loss": 0.8167, "mean_token_accuracy": 0.883002519607544, "num_tokens": 311929401.0, "step": 8218 }, { "epoch": 1.0455412797354027, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.01828384399414, "learning_rate": 5e-05, "loss": 0.9845, "mean_token_accuracy": 0.8635179400444031, "num_tokens": 311969401.0, "step": 8219 }, { "epoch": 1.0456684900139932, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.417880535125732, "learning_rate": 5e-05, "loss": 0.8625, "mean_token_accuracy": 0.8617749810218811, "num_tokens": 312004295.0, "step": 8220 }, { "epoch": 1.0457957002925837, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.33414363861084, "learning_rate": 5e-05, "loss": 0.9815, "mean_token_accuracy": 0.8705282211303711, "num_tokens": 312038487.0, "step": 8221 }, { "epoch": 1.0459229105711743, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.538847923278809, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.8736083507537842, "num_tokens": 312076011.0, "step": 8222 }, { "epoch": 1.0460501208497646, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.317395210266113, "learning_rate": 5e-05, "loss": 1.0212, "mean_token_accuracy": 0.8665198087692261, "num_tokens": 312114489.0, "step": 8223 }, { "epoch": 1.046177331128355, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.960338592529297, "learning_rate": 5e-05, "loss": 0.8829, "mean_token_accuracy": 0.8702153563499451, "num_tokens": 312152388.0, "step": 8224 }, { "epoch": 1.0463045414069456, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.760427474975586, "learning_rate": 5e-05, "loss": 0.9426, "mean_token_accuracy": 0.8785029649734497, "num_tokens": 312190269.0, "step": 8225 }, { "epoch": 1.0464317516855361, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.059595108032227, "learning_rate": 5e-05, "loss": 0.8587, "mean_token_accuracy": 0.8807149529457092, "num_tokens": 312229909.0, "step": 8226 }, { "epoch": 1.0465589619641267, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.235486030578613, "learning_rate": 5e-05, "loss": 0.964, "mean_token_accuracy": 0.8589528799057007, "num_tokens": 312267527.0, "step": 8227 }, { "epoch": 1.0466861722427172, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.768397331237793, "learning_rate": 5e-05, "loss": 0.8156, "mean_token_accuracy": 0.875280499458313, "num_tokens": 312310632.0, "step": 8228 }, { "epoch": 1.0468133825213077, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.134478569030762, "learning_rate": 5e-05, "loss": 0.9091, "mean_token_accuracy": 0.86164391040802, "num_tokens": 312351570.0, "step": 8229 }, { "epoch": 1.0469405927998983, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000469207763671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.342288017272949, "learning_rate": 5e-05, "loss": 0.7133, "mean_token_accuracy": 0.8929609060287476, "num_tokens": 312382765.0, "step": 8230 }, { "epoch": 1.0470678030784888, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.39242935180664, "learning_rate": 5e-05, "loss": 0.8562, "mean_token_accuracy": 0.8828352689743042, "num_tokens": 312415517.0, "step": 8231 }, { "epoch": 1.0471950133570793, "ewc_loss": 0.375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.047022342681885, "learning_rate": 5e-05, "loss": 0.7173, "mean_token_accuracy": 0.881517231464386, "num_tokens": 312450574.0, "step": 8232 }, { "epoch": 1.0473222236356698, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.604049682617188, "learning_rate": 5e-05, "loss": 0.9262, "mean_token_accuracy": 0.8632421493530273, "num_tokens": 312483054.0, "step": 8233 }, { "epoch": 1.0474494339142604, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.069589614868164, "learning_rate": 5e-05, "loss": 0.7545, "mean_token_accuracy": 0.8697109222412109, "num_tokens": 312527825.0, "step": 8234 }, { "epoch": 1.0475766441928507, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.596198081970215, "learning_rate": 5e-05, "loss": 0.9322, "mean_token_accuracy": 0.8629308938980103, "num_tokens": 312566971.0, "step": 8235 }, { "epoch": 1.0477038544714412, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.2715373039245605, "learning_rate": 5e-05, "loss": 0.7796, "mean_token_accuracy": 0.873350977897644, "num_tokens": 312603527.0, "step": 8236 }, { "epoch": 1.0478310647500317, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.347259521484375, "learning_rate": 5e-05, "loss": 0.896, "mean_token_accuracy": 0.8723639249801636, "num_tokens": 312634281.0, "step": 8237 }, { "epoch": 1.0479582750286223, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.512142181396484, "learning_rate": 5e-05, "loss": 0.8176, "mean_token_accuracy": 0.8675936460494995, "num_tokens": 312668036.0, "step": 8238 }, { "epoch": 1.0480854853072128, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.915910243988037, "learning_rate": 5e-05, "loss": 0.8618, "mean_token_accuracy": 0.864682137966156, "num_tokens": 312707422.0, "step": 8239 }, { "epoch": 1.0482126955858033, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.641546249389648, "learning_rate": 5e-05, "loss": 0.7899, "mean_token_accuracy": 0.8716956973075867, "num_tokens": 312740715.0, "step": 8240 }, { "epoch": 1.0483399058643939, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.580766677856445, "learning_rate": 5e-05, "loss": 0.7795, "mean_token_accuracy": 0.8765592575073242, "num_tokens": 312773196.0, "step": 8241 }, { "epoch": 1.0484671161429844, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.635066509246826, "learning_rate": 5e-05, "loss": 0.7891, "mean_token_accuracy": 0.8627431988716125, "num_tokens": 312817502.0, "step": 8242 }, { "epoch": 1.048594326421575, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.510865688323975, "learning_rate": 5e-05, "loss": 0.7668, "mean_token_accuracy": 0.8685380220413208, "num_tokens": 312852810.0, "step": 8243 }, { "epoch": 1.0487215367001654, "ewc_loss": 0.375, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.50961446762085, "learning_rate": 5e-05, "loss": 0.708, "mean_token_accuracy": 0.8802794218063354, "num_tokens": 312889069.0, "step": 8244 }, { "epoch": 1.048848746978756, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0004711151123046875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.524362087249756, "learning_rate": 5e-05, "loss": 0.769, "mean_token_accuracy": 0.8693574070930481, "num_tokens": 312929032.0, "step": 8245 }, { "epoch": 1.0489759572573465, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.4355573654174805, "learning_rate": 5e-05, "loss": 0.7199, "mean_token_accuracy": 0.8755859136581421, "num_tokens": 312963792.0, "step": 8246 }, { "epoch": 1.0491031675359368, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.514023303985596, "learning_rate": 5e-05, "loss": 0.6952, "mean_token_accuracy": 0.8878962993621826, "num_tokens": 312999532.0, "step": 8247 }, { "epoch": 1.0492303778145273, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.394101142883301, "learning_rate": 5e-05, "loss": 0.7379, "mean_token_accuracy": 0.8707207441329956, "num_tokens": 313040807.0, "step": 8248 }, { "epoch": 1.0493575880931179, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.527719497680664, "learning_rate": 5e-05, "loss": 0.7037, "mean_token_accuracy": 0.8791208267211914, "num_tokens": 313074895.0, "step": 8249 }, { "epoch": 1.0494847983717084, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.295830726623535, "learning_rate": 5e-05, "loss": 0.705, "mean_token_accuracy": 0.8763606548309326, "num_tokens": 313115299.0, "step": 8250 }, { "epoch": 1.049612008650299, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.650245189666748, "learning_rate": 5e-05, "loss": 0.7436, "mean_token_accuracy": 0.8698376417160034, "num_tokens": 313157578.0, "step": 8251 }, { "epoch": 1.0497392189288894, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.12745475769043, "learning_rate": 5e-05, "loss": 0.7114, "mean_token_accuracy": 0.8712019920349121, "num_tokens": 313202219.0, "step": 8252 }, { "epoch": 1.04986642920748, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.814788341522217, "learning_rate": 5e-05, "loss": 0.731, "mean_token_accuracy": 0.8783867359161377, "num_tokens": 313242345.0, "step": 8253 }, { "epoch": 1.0499936394860705, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.943609714508057, "learning_rate": 5e-05, "loss": 0.68, "mean_token_accuracy": 0.8780288696289062, "num_tokens": 313277052.0, "step": 8254 }, { "epoch": 1.050120849764661, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 8.022211074829102, "learning_rate": 5e-05, "loss": 0.7646, "mean_token_accuracy": 0.8710561394691467, "num_tokens": 313311291.0, "step": 8255 }, { "epoch": 1.0502480600432516, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.718564987182617, "learning_rate": 5e-05, "loss": 0.6653, "mean_token_accuracy": 0.8773362040519714, "num_tokens": 313349141.0, "step": 8256 }, { "epoch": 1.050375270321842, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.352157592773438, "learning_rate": 5e-05, "loss": 0.8418, "mean_token_accuracy": 0.8579633831977844, "num_tokens": 313387575.0, "step": 8257 }, { "epoch": 1.0505024806004326, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.442249298095703, "learning_rate": 5e-05, "loss": 0.674, "mean_token_accuracy": 0.8682059645652771, "num_tokens": 313424087.0, "step": 8258 }, { "epoch": 1.0506296908790231, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.725153923034668, "learning_rate": 5e-05, "loss": 0.8335, "mean_token_accuracy": 0.8757356405258179, "num_tokens": 313464517.0, "step": 8259 }, { "epoch": 1.0507569011576134, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.318914413452148, "learning_rate": 5e-05, "loss": 0.7079, "mean_token_accuracy": 0.8622289299964905, "num_tokens": 313504146.0, "step": 8260 }, { "epoch": 1.050884111436204, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.900367736816406, "learning_rate": 5e-05, "loss": 0.8873, "mean_token_accuracy": 0.8680545091629028, "num_tokens": 313541842.0, "step": 8261 }, { "epoch": 1.0510113217147945, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.617711067199707, "learning_rate": 5e-05, "loss": 0.715, "mean_token_accuracy": 0.867647647857666, "num_tokens": 313580695.0, "step": 8262 }, { "epoch": 1.051138531993385, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.465377807617188, "learning_rate": 5e-05, "loss": 0.8957, "mean_token_accuracy": 0.8615957498550415, "num_tokens": 313619494.0, "step": 8263 }, { "epoch": 1.0512657422719756, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.415823936462402, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8797882795333862, "num_tokens": 313653234.0, "step": 8264 }, { "epoch": 1.051392952550566, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0004730224609375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.35045862197876, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.870587170124054, "num_tokens": 313694395.0, "step": 8265 }, { "epoch": 1.0515201628291566, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.344438552856445, "learning_rate": 5e-05, "loss": 0.8851, "mean_token_accuracy": 0.8593451976776123, "num_tokens": 313733370.0, "step": 8266 }, { "epoch": 1.0516473731077471, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.326255798339844, "learning_rate": 5e-05, "loss": 0.7232, "mean_token_accuracy": 0.8593862056732178, "num_tokens": 313770611.0, "step": 8267 }, { "epoch": 1.0517745833863377, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.328083992004395, "learning_rate": 5e-05, "loss": 0.8557, "mean_token_accuracy": 0.8907073736190796, "num_tokens": 313807262.0, "step": 8268 }, { "epoch": 1.0519017936649282, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 5.860938549041748, "learning_rate": 5e-05, "loss": 0.6915, "mean_token_accuracy": 0.86466383934021, "num_tokens": 313844294.0, "step": 8269 }, { "epoch": 1.0520290039435187, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.09495735168457, "learning_rate": 5e-05, "loss": 0.9778, "mean_token_accuracy": 0.8823840022087097, "num_tokens": 313878701.0, "step": 8270 }, { "epoch": 1.0521562142221093, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 6.789161682128906, "learning_rate": 5e-05, "loss": 0.7842, "mean_token_accuracy": 0.8793855905532837, "num_tokens": 313917780.0, "step": 8271 }, { "epoch": 1.0522834245006996, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.68559741973877, "learning_rate": 5e-05, "loss": 0.984, "mean_token_accuracy": 0.8855273723602295, "num_tokens": 313948611.0, "step": 8272 }, { "epoch": 1.05241063477929, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.388190269470215, "learning_rate": 5e-05, "loss": 0.9823, "mean_token_accuracy": 0.877830445766449, "num_tokens": 313981999.0, "step": 8273 }, { "epoch": 1.0525378450578806, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.811190605163574, "learning_rate": 5e-05, "loss": 0.8125, "mean_token_accuracy": 0.8795795440673828, "num_tokens": 314021199.0, "step": 8274 }, { "epoch": 1.0526650553364711, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.387682914733887, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.8866117596626282, "num_tokens": 314058187.0, "step": 8275 }, { "epoch": 1.0527922656150617, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 6.733407497406006, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.8739863634109497, "num_tokens": 314095612.0, "step": 8276 }, { "epoch": 1.0529194758936522, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.827938079833984, "learning_rate": 5e-05, "loss": 1.0228, "mean_token_accuracy": 0.8676387667655945, "num_tokens": 314124895.0, "step": 8277 }, { "epoch": 1.0530466861722427, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.504450798034668, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.8816776275634766, "num_tokens": 314167787.0, "step": 8278 }, { "epoch": 1.0531738964508333, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.802332878112793, "learning_rate": 5e-05, "loss": 0.9112, "mean_token_accuracy": 0.8756307363510132, "num_tokens": 314203389.0, "step": 8279 }, { "epoch": 1.0533011067294238, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.378738403320312, "learning_rate": 5e-05, "loss": 0.9188, "mean_token_accuracy": 0.8829140663146973, "num_tokens": 314245161.0, "step": 8280 }, { "epoch": 1.0534283170080143, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.503449440002441, "learning_rate": 5e-05, "loss": 0.8557, "mean_token_accuracy": 0.8495409488677979, "num_tokens": 314288097.0, "step": 8281 }, { "epoch": 1.0535555272866048, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.774737358093262, "learning_rate": 5e-05, "loss": 0.9468, "mean_token_accuracy": 0.8701149821281433, "num_tokens": 314324066.0, "step": 8282 }, { "epoch": 1.0536827375651954, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.242047309875488, "learning_rate": 5e-05, "loss": 0.7476, "mean_token_accuracy": 0.882105827331543, "num_tokens": 314358430.0, "step": 8283 }, { "epoch": 1.0538099478437857, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.499492645263672, "learning_rate": 5e-05, "loss": 0.865, "mean_token_accuracy": 0.8727072477340698, "num_tokens": 314394199.0, "step": 8284 }, { "epoch": 1.0539371581223762, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.538812637329102, "learning_rate": 5e-05, "loss": 0.8202, "mean_token_accuracy": 0.8737003207206726, "num_tokens": 314432851.0, "step": 8285 }, { "epoch": 1.0540643684009667, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.800037860870361, "learning_rate": 5e-05, "loss": 0.8035, "mean_token_accuracy": 0.872367262840271, "num_tokens": 314473843.0, "step": 8286 }, { "epoch": 1.0541915786795573, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.884940147399902, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8858230113983154, "num_tokens": 314518782.0, "step": 8287 }, { "epoch": 1.0543187889581478, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.265864372253418, "learning_rate": 5e-05, "loss": 0.7443, "mean_token_accuracy": 0.8700754642486572, "num_tokens": 314561140.0, "step": 8288 }, { "epoch": 1.0544459992367383, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.125524520874023, "learning_rate": 5e-05, "loss": 0.819, "mean_token_accuracy": 0.8716384768486023, "num_tokens": 314602605.0, "step": 8289 }, { "epoch": 1.0545732095153288, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.021908760070801, "learning_rate": 5e-05, "loss": 0.765, "mean_token_accuracy": 0.8592463731765747, "num_tokens": 314647719.0, "step": 8290 }, { "epoch": 1.0547004197939194, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.281929016113281, "learning_rate": 5e-05, "loss": 0.8779, "mean_token_accuracy": 0.8590438365936279, "num_tokens": 314690016.0, "step": 8291 }, { "epoch": 1.05482763007251, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.858985900878906, "learning_rate": 5e-05, "loss": 0.7124, "mean_token_accuracy": 0.8767621517181396, "num_tokens": 314730432.0, "step": 8292 }, { "epoch": 1.0549548403511004, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.406529426574707, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.880544900894165, "num_tokens": 314764957.0, "step": 8293 }, { "epoch": 1.055082050629691, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.753928184509277, "learning_rate": 5e-05, "loss": 0.6987, "mean_token_accuracy": 0.8767859935760498, "num_tokens": 314803787.0, "step": 8294 }, { "epoch": 1.0552092609082815, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.504939079284668, "learning_rate": 5e-05, "loss": 0.8283, "mean_token_accuracy": 0.8763118982315063, "num_tokens": 314841732.0, "step": 8295 }, { "epoch": 1.0553364711868718, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.659488201141357, "learning_rate": 5e-05, "loss": 0.6802, "mean_token_accuracy": 0.8786834478378296, "num_tokens": 314878212.0, "step": 8296 }, { "epoch": 1.0554636814654623, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.642229080200195, "learning_rate": 5e-05, "loss": 0.8722, "mean_token_accuracy": 0.8695180416107178, "num_tokens": 314917130.0, "step": 8297 }, { "epoch": 1.0555908917440529, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.567721843719482, "learning_rate": 5e-05, "loss": 0.6875, "mean_token_accuracy": 0.8815325498580933, "num_tokens": 314961804.0, "step": 8298 }, { "epoch": 1.0557181020226434, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.803772926330566, "learning_rate": 5e-05, "loss": 0.8383, "mean_token_accuracy": 0.8897092342376709, "num_tokens": 314997157.0, "step": 8299 }, { "epoch": 1.055845312301234, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.603304386138916, "learning_rate": 5e-05, "loss": 0.7677, "mean_token_accuracy": 0.8559378385543823, "num_tokens": 315038370.0, "step": 8300 }, { "epoch": 1.0559725225798244, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.829812049865723, "learning_rate": 5e-05, "loss": 0.8709, "mean_token_accuracy": 0.8862581849098206, "num_tokens": 315081716.0, "step": 8301 }, { "epoch": 1.056099732858415, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.851134777069092, "learning_rate": 5e-05, "loss": 0.8325, "mean_token_accuracy": 0.8539479970932007, "num_tokens": 315124528.0, "step": 8302 }, { "epoch": 1.0562269431370055, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.590094566345215, "learning_rate": 5e-05, "loss": 0.8515, "mean_token_accuracy": 0.8858560919761658, "num_tokens": 315163893.0, "step": 8303 }, { "epoch": 1.056354153415596, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.216354846954346, "learning_rate": 5e-05, "loss": 0.8878, "mean_token_accuracy": 0.8509504199028015, "num_tokens": 315199462.0, "step": 8304 }, { "epoch": 1.0564813636941865, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.122947692871094, "learning_rate": 5e-05, "loss": 0.8611, "mean_token_accuracy": 0.8722320795059204, "num_tokens": 315245075.0, "step": 8305 }, { "epoch": 1.056608573972777, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.494836807250977, "learning_rate": 5e-05, "loss": 0.816, "mean_token_accuracy": 0.8763421773910522, "num_tokens": 315289291.0, "step": 8306 }, { "epoch": 1.0567357842513676, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.7190937995910645, "learning_rate": 5e-05, "loss": 0.777, "mean_token_accuracy": 0.8784817457199097, "num_tokens": 315327750.0, "step": 8307 }, { "epoch": 1.0568629945299581, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.590694427490234, "learning_rate": 5e-05, "loss": 0.7974, "mean_token_accuracy": 0.8706717491149902, "num_tokens": 315366863.0, "step": 8308 }, { "epoch": 1.0569902048085484, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.556947708129883, "learning_rate": 5e-05, "loss": 0.726, "mean_token_accuracy": 0.8865443468093872, "num_tokens": 315406935.0, "step": 8309 }, { "epoch": 1.057117415087139, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.495351791381836, "learning_rate": 5e-05, "loss": 0.7705, "mean_token_accuracy": 0.8699436187744141, "num_tokens": 315446483.0, "step": 8310 }, { "epoch": 1.0572446253657295, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.601789951324463, "learning_rate": 5e-05, "loss": 0.7681, "mean_token_accuracy": 0.8716480731964111, "num_tokens": 315482862.0, "step": 8311 }, { "epoch": 1.05737183564432, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.31451940536499, "learning_rate": 5e-05, "loss": 0.7263, "mean_token_accuracy": 0.8758707046508789, "num_tokens": 315522649.0, "step": 8312 }, { "epoch": 1.0574990459229106, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.731459617614746, "learning_rate": 5e-05, "loss": 0.7853, "mean_token_accuracy": 0.8683009147644043, "num_tokens": 315562642.0, "step": 8313 }, { "epoch": 1.057626256201501, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.139396667480469, "learning_rate": 5e-05, "loss": 0.6504, "mean_token_accuracy": 0.8945931196212769, "num_tokens": 315594056.0, "step": 8314 }, { "epoch": 1.0577534664800916, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.870931148529053, "learning_rate": 5e-05, "loss": 0.7787, "mean_token_accuracy": 0.870361328125, "num_tokens": 315632474.0, "step": 8315 }, { "epoch": 1.0578806767586821, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.987979412078857, "learning_rate": 5e-05, "loss": 0.705, "mean_token_accuracy": 0.8727101683616638, "num_tokens": 315662036.0, "step": 8316 }, { "epoch": 1.0580078870372727, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.036039352416992, "learning_rate": 5e-05, "loss": 0.7683, "mean_token_accuracy": 0.877018928527832, "num_tokens": 315701974.0, "step": 8317 }, { "epoch": 1.0581350973158632, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.820501327514648, "learning_rate": 5e-05, "loss": 0.6993, "mean_token_accuracy": 0.8700757026672363, "num_tokens": 315745876.0, "step": 8318 }, { "epoch": 1.0582623075944537, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.208772659301758, "learning_rate": 5e-05, "loss": 0.7844, "mean_token_accuracy": 0.8728340864181519, "num_tokens": 315787386.0, "step": 8319 }, { "epoch": 1.058389517873044, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.7052130699157715, "learning_rate": 5e-05, "loss": 0.6588, "mean_token_accuracy": 0.8812010884284973, "num_tokens": 315824047.0, "step": 8320 }, { "epoch": 1.0585167281516346, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.346464157104492, "learning_rate": 5e-05, "loss": 0.8515, "mean_token_accuracy": 0.8567886352539062, "num_tokens": 315863890.0, "step": 8321 }, { "epoch": 1.058643938430225, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.661944389343262, "learning_rate": 5e-05, "loss": 0.7095, "mean_token_accuracy": 0.8711695671081543, "num_tokens": 315902855.0, "step": 8322 }, { "epoch": 1.0587711487088156, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.417242050170898, "learning_rate": 5e-05, "loss": 0.8233, "mean_token_accuracy": 0.8758115768432617, "num_tokens": 315942112.0, "step": 8323 }, { "epoch": 1.0588983589874061, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.7609710693359375, "learning_rate": 5e-05, "loss": 0.692, "mean_token_accuracy": 0.8778777122497559, "num_tokens": 315981256.0, "step": 8324 }, { "epoch": 1.0590255692659967, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.232762336730957, "learning_rate": 5e-05, "loss": 0.8427, "mean_token_accuracy": 0.8667391538619995, "num_tokens": 316021192.0, "step": 8325 }, { "epoch": 1.0591527795445872, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.079496383666992, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.868956983089447, "num_tokens": 316056261.0, "step": 8326 }, { "epoch": 1.0592799898231777, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.775374412536621, "learning_rate": 5e-05, "loss": 0.7244, "mean_token_accuracy": 0.8929849863052368, "num_tokens": 316093390.0, "step": 8327 }, { "epoch": 1.0594072001017683, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.51711368560791, "learning_rate": 5e-05, "loss": 0.7945, "mean_token_accuracy": 0.8610373139381409, "num_tokens": 316133771.0, "step": 8328 }, { "epoch": 1.0595344103803588, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.198307991027832, "learning_rate": 5e-05, "loss": 0.719, "mean_token_accuracy": 0.8744195699691772, "num_tokens": 316172123.0, "step": 8329 }, { "epoch": 1.0596616206589493, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.015914916992188, "learning_rate": 5e-05, "loss": 0.7423, "mean_token_accuracy": 0.8855987787246704, "num_tokens": 316211866.0, "step": 8330 }, { "epoch": 1.0597888309375398, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.6829118728637695, "learning_rate": 5e-05, "loss": 0.6704, "mean_token_accuracy": 0.8808801174163818, "num_tokens": 316248838.0, "step": 8331 }, { "epoch": 1.0599160412161304, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.520529747009277, "learning_rate": 5e-05, "loss": 0.8086, "mean_token_accuracy": 0.8753362894058228, "num_tokens": 316286551.0, "step": 8332 }, { "epoch": 1.0600432514947207, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.252010822296143, "learning_rate": 5e-05, "loss": 0.6893, "mean_token_accuracy": 0.8699893951416016, "num_tokens": 316321278.0, "step": 8333 }, { "epoch": 1.0601704617733112, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 9.044046401977539, "learning_rate": 5e-05, "loss": 0.8777, "mean_token_accuracy": 0.872300922870636, "num_tokens": 316356282.0, "step": 8334 }, { "epoch": 1.0602976720519017, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.087597846984863, "learning_rate": 5e-05, "loss": 0.6849, "mean_token_accuracy": 0.8708928823471069, "num_tokens": 316398858.0, "step": 8335 }, { "epoch": 1.0604248823304923, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0004749298095703125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.400126457214355, "learning_rate": 5e-05, "loss": 0.9553, "mean_token_accuracy": 0.8729161024093628, "num_tokens": 316432883.0, "step": 8336 }, { "epoch": 1.0605520926090828, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.572570323944092, "learning_rate": 5e-05, "loss": 0.7244, "mean_token_accuracy": 0.8856511116027832, "num_tokens": 316469305.0, "step": 8337 }, { "epoch": 1.0606793028876733, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.025447845458984, "learning_rate": 5e-05, "loss": 0.9023, "mean_token_accuracy": 0.8907074928283691, "num_tokens": 316509072.0, "step": 8338 }, { "epoch": 1.0608065131662638, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.610414981842041, "learning_rate": 5e-05, "loss": 0.9011, "mean_token_accuracy": 0.8706764578819275, "num_tokens": 316549391.0, "step": 8339 }, { "epoch": 1.0609337234448544, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.739543437957764, "learning_rate": 5e-05, "loss": 0.782, "mean_token_accuracy": 0.8919499516487122, "num_tokens": 316583696.0, "step": 8340 }, { "epoch": 1.061060933723445, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.476598739624023, "learning_rate": 5e-05, "loss": 0.9318, "mean_token_accuracy": 0.8662130832672119, "num_tokens": 316618421.0, "step": 8341 }, { "epoch": 1.0611881440020354, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.7264604568481445, "learning_rate": 5e-05, "loss": 0.7256, "mean_token_accuracy": 0.8735843300819397, "num_tokens": 316660284.0, "step": 8342 }, { "epoch": 1.061315354280626, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.852131843566895, "learning_rate": 5e-05, "loss": 0.9822, "mean_token_accuracy": 0.8569137454032898, "num_tokens": 316699753.0, "step": 8343 }, { "epoch": 1.0614425645592165, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.792171001434326, "learning_rate": 5e-05, "loss": 0.7497, "mean_token_accuracy": 0.8717800378799438, "num_tokens": 316735336.0, "step": 8344 }, { "epoch": 1.0615697748378068, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.34529972076416, "learning_rate": 5e-05, "loss": 0.8608, "mean_token_accuracy": 0.8783325552940369, "num_tokens": 316775327.0, "step": 8345 }, { "epoch": 1.0616969851163973, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.433046817779541, "learning_rate": 5e-05, "loss": 0.8678, "mean_token_accuracy": 0.8559194803237915, "num_tokens": 316817027.0, "step": 8346 }, { "epoch": 1.0618241953949878, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.514151096343994, "learning_rate": 5e-05, "loss": 0.7762, "mean_token_accuracy": 0.8749610185623169, "num_tokens": 316850980.0, "step": 8347 }, { "epoch": 1.0619514056735784, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.928122043609619, "learning_rate": 5e-05, "loss": 0.775, "mean_token_accuracy": 0.8889099359512329, "num_tokens": 316886533.0, "step": 8348 }, { "epoch": 1.062078615952169, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.023528575897217, "learning_rate": 5e-05, "loss": 0.6859, "mean_token_accuracy": 0.8823664784431458, "num_tokens": 316927447.0, "step": 8349 }, { "epoch": 1.0622058262307594, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.145547866821289, "learning_rate": 5e-05, "loss": 0.8528, "mean_token_accuracy": 0.8628711700439453, "num_tokens": 316965983.0, "step": 8350 }, { "epoch": 1.06233303650935, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.915648937225342, "learning_rate": 5e-05, "loss": 0.7005, "mean_token_accuracy": 0.8754969835281372, "num_tokens": 316999678.0, "step": 8351 }, { "epoch": 1.0624602467879405, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.143034934997559, "learning_rate": 5e-05, "loss": 0.8393, "mean_token_accuracy": 0.8649070262908936, "num_tokens": 317038650.0, "step": 8352 }, { "epoch": 1.062587457066531, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.927850723266602, "learning_rate": 5e-05, "loss": 0.7087, "mean_token_accuracy": 0.8764551877975464, "num_tokens": 317075507.0, "step": 8353 }, { "epoch": 1.0627146673451215, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.112588882446289, "learning_rate": 5e-05, "loss": 0.7609, "mean_token_accuracy": 0.8889285326004028, "num_tokens": 317109364.0, "step": 8354 }, { "epoch": 1.062841877623712, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.883002281188965, "learning_rate": 5e-05, "loss": 0.7083, "mean_token_accuracy": 0.8760223984718323, "num_tokens": 317149959.0, "step": 8355 }, { "epoch": 1.0629690879023026, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.237432479858398, "learning_rate": 5e-05, "loss": 0.8207, "mean_token_accuracy": 0.872003436088562, "num_tokens": 317182604.0, "step": 8356 }, { "epoch": 1.0630962981808931, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.6952033042907715, "learning_rate": 5e-05, "loss": 0.6796, "mean_token_accuracy": 0.8818016648292542, "num_tokens": 317224190.0, "step": 8357 }, { "epoch": 1.0632235084594834, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.512151718139648, "learning_rate": 5e-05, "loss": 0.857, "mean_token_accuracy": 0.8713892698287964, "num_tokens": 317261607.0, "step": 8358 }, { "epoch": 1.063350718738074, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.473652362823486, "learning_rate": 5e-05, "loss": 0.741, "mean_token_accuracy": 0.8592360615730286, "num_tokens": 317300772.0, "step": 8359 }, { "epoch": 1.0634779290166645, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.832372665405273, "learning_rate": 5e-05, "loss": 0.9023, "mean_token_accuracy": 0.8732225894927979, "num_tokens": 317336981.0, "step": 8360 }, { "epoch": 1.063605139295255, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.444559097290039, "learning_rate": 5e-05, "loss": 0.749, "mean_token_accuracy": 0.8634042739868164, "num_tokens": 317376522.0, "step": 8361 }, { "epoch": 1.0637323495738455, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.982217788696289, "learning_rate": 5e-05, "loss": 0.9399, "mean_token_accuracy": 0.8739772439002991, "num_tokens": 317409924.0, "step": 8362 }, { "epoch": 1.063859559852436, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.748617172241211, "learning_rate": 5e-05, "loss": 0.733, "mean_token_accuracy": 0.8852962255477905, "num_tokens": 317440762.0, "step": 8363 }, { "epoch": 1.0639867701310266, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.74060344696045, "learning_rate": 5e-05, "loss": 0.9385, "mean_token_accuracy": 0.8764969706535339, "num_tokens": 317471606.0, "step": 8364 }, { "epoch": 1.0641139804096171, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.312537670135498, "learning_rate": 5e-05, "loss": 0.8858, "mean_token_accuracy": 0.864061713218689, "num_tokens": 317508383.0, "step": 8365 }, { "epoch": 1.0642411906882077, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.045307159423828, "learning_rate": 5e-05, "loss": 0.8233, "mean_token_accuracy": 0.8929524421691895, "num_tokens": 317546690.0, "step": 8366 }, { "epoch": 1.0643684009667982, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.769592761993408, "learning_rate": 5e-05, "loss": 0.8407, "mean_token_accuracy": 0.8868089914321899, "num_tokens": 317584926.0, "step": 8367 }, { "epoch": 1.0644956112453887, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.406798362731934, "learning_rate": 5e-05, "loss": 0.8147, "mean_token_accuracy": 0.8638654947280884, "num_tokens": 317619941.0, "step": 8368 }, { "epoch": 1.064622821523979, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.891809463500977, "learning_rate": 5e-05, "loss": 0.8059, "mean_token_accuracy": 0.8855156898498535, "num_tokens": 317651006.0, "step": 8369 }, { "epoch": 1.0647500318025696, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.2665276527404785, "learning_rate": 5e-05, "loss": 0.7488, "mean_token_accuracy": 0.8764921426773071, "num_tokens": 317688051.0, "step": 8370 }, { "epoch": 1.06487724208116, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.6484293937683105, "learning_rate": 5e-05, "loss": 0.7608, "mean_token_accuracy": 0.8783456087112427, "num_tokens": 317727447.0, "step": 8371 }, { "epoch": 1.0650044523597506, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.490434646606445, "learning_rate": 5e-05, "loss": 0.7358, "mean_token_accuracy": 0.8837829828262329, "num_tokens": 317771495.0, "step": 8372 }, { "epoch": 1.0651316626383411, "ewc_loss": 0.375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.309907913208008, "learning_rate": 5e-05, "loss": 0.706, "mean_token_accuracy": 0.8838516473770142, "num_tokens": 317808166.0, "step": 8373 }, { "epoch": 1.0652588729169317, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.71507453918457, "learning_rate": 5e-05, "loss": 0.7635, "mean_token_accuracy": 0.8781676888465881, "num_tokens": 317850928.0, "step": 8374 }, { "epoch": 1.0653860831955222, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.055835247039795, "learning_rate": 5e-05, "loss": 0.6971, "mean_token_accuracy": 0.8785333037376404, "num_tokens": 317893156.0, "step": 8375 }, { "epoch": 1.0655132934741127, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.906924247741699, "learning_rate": 5e-05, "loss": 0.8565, "mean_token_accuracy": 0.848872721195221, "num_tokens": 317937146.0, "step": 8376 }, { "epoch": 1.0656405037527032, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.8675856590271, "learning_rate": 5e-05, "loss": 0.6699, "mean_token_accuracy": 0.8833863735198975, "num_tokens": 317977429.0, "step": 8377 }, { "epoch": 1.0657677140312938, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.130257606506348, "learning_rate": 5e-05, "loss": 0.8003, "mean_token_accuracy": 0.8693206906318665, "num_tokens": 318013445.0, "step": 8378 }, { "epoch": 1.0658949243098843, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.638277053833008, "learning_rate": 5e-05, "loss": 0.7183, "mean_token_accuracy": 0.867067277431488, "num_tokens": 318054082.0, "step": 8379 }, { "epoch": 1.0660221345884748, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.442501068115234, "learning_rate": 5e-05, "loss": 0.8687, "mean_token_accuracy": 0.8588230013847351, "num_tokens": 318096125.0, "step": 8380 }, { "epoch": 1.0661493448670654, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.404177188873291, "learning_rate": 5e-05, "loss": 0.6753, "mean_token_accuracy": 0.8763284087181091, "num_tokens": 318130859.0, "step": 8381 }, { "epoch": 1.0662765551456557, "ewc_loss": 0.5, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.784664154052734, "learning_rate": 5e-05, "loss": 0.839, "mean_token_accuracy": 0.8803501129150391, "num_tokens": 318171616.0, "step": 8382 }, { "epoch": 1.0664037654242462, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.382210731506348, "learning_rate": 5e-05, "loss": 0.7223, "mean_token_accuracy": 0.8676043748855591, "num_tokens": 318206741.0, "step": 8383 }, { "epoch": 1.0665309757028367, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.876570701599121, "learning_rate": 5e-05, "loss": 0.9266, "mean_token_accuracy": 0.8695515394210815, "num_tokens": 318252513.0, "step": 8384 }, { "epoch": 1.0666581859814273, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 6.8012471199035645, "learning_rate": 5e-05, "loss": 0.7848, "mean_token_accuracy": 0.8639754056930542, "num_tokens": 318286632.0, "step": 8385 }, { "epoch": 1.0667853962600178, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.354158401489258, "learning_rate": 5e-05, "loss": 0.8741, "mean_token_accuracy": 0.8765355944633484, "num_tokens": 318326923.0, "step": 8386 }, { "epoch": 1.0669126065386083, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000476837158203125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.5445966720581055, "learning_rate": 5e-05, "loss": 0.8925, "mean_token_accuracy": 0.8527998328208923, "num_tokens": 318364159.0, "step": 8387 }, { "epoch": 1.0670398168171988, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.348790168762207, "learning_rate": 5e-05, "loss": 0.8382, "mean_token_accuracy": 0.8512250781059265, "num_tokens": 318402715.0, "step": 8388 }, { "epoch": 1.0671670270957894, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.254956245422363, "learning_rate": 5e-05, "loss": 0.8737, "mean_token_accuracy": 0.8702347874641418, "num_tokens": 318445925.0, "step": 8389 }, { "epoch": 1.06729423737438, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.57110071182251, "learning_rate": 5e-05, "loss": 0.7214, "mean_token_accuracy": 0.8662881851196289, "num_tokens": 318481756.0, "step": 8390 }, { "epoch": 1.0674214476529704, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.791595458984375, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.873536229133606, "num_tokens": 318526885.0, "step": 8391 }, { "epoch": 1.067548657931561, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.29702615737915, "learning_rate": 5e-05, "loss": 0.7135, "mean_token_accuracy": 0.8692365288734436, "num_tokens": 318569396.0, "step": 8392 }, { "epoch": 1.0676758682101515, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.110234260559082, "learning_rate": 5e-05, "loss": 0.9078, "mean_token_accuracy": 0.88700270652771, "num_tokens": 318606404.0, "step": 8393 }, { "epoch": 1.0678030784887418, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.549854278564453, "learning_rate": 5e-05, "loss": 0.7819, "mean_token_accuracy": 0.8601595759391785, "num_tokens": 318642457.0, "step": 8394 }, { "epoch": 1.0679302887673323, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.965163230895996, "learning_rate": 5e-05, "loss": 0.971, "mean_token_accuracy": 0.8635607957839966, "num_tokens": 318680498.0, "step": 8395 }, { "epoch": 1.0680574990459228, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.265104293823242, "learning_rate": 5e-05, "loss": 0.8857, "mean_token_accuracy": 0.8653225302696228, "num_tokens": 318720493.0, "step": 8396 }, { "epoch": 1.0681847093245134, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.204971313476562, "learning_rate": 5e-05, "loss": 0.918, "mean_token_accuracy": 0.8650925159454346, "num_tokens": 318755297.0, "step": 8397 }, { "epoch": 1.068311919603104, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.839743137359619, "learning_rate": 5e-05, "loss": 0.8828, "mean_token_accuracy": 0.8762640953063965, "num_tokens": 318788716.0, "step": 8398 }, { "epoch": 1.0684391298816944, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.44133186340332, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8726941347122192, "num_tokens": 318825130.0, "step": 8399 }, { "epoch": 1.068566340160285, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.951659679412842, "learning_rate": 5e-05, "loss": 0.8491, "mean_token_accuracy": 0.8742822408676147, "num_tokens": 318861216.0, "step": 8400 }, { "epoch": 1.0686935504388755, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.385249137878418, "learning_rate": 5e-05, "loss": 0.7741, "mean_token_accuracy": 0.8768237233161926, "num_tokens": 318896410.0, "step": 8401 }, { "epoch": 1.068820760717466, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.4601969718933105, "learning_rate": 5e-05, "loss": 0.7369, "mean_token_accuracy": 0.8857738375663757, "num_tokens": 318933834.0, "step": 8402 }, { "epoch": 1.0689479709960565, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.897548198699951, "learning_rate": 5e-05, "loss": 0.862, "mean_token_accuracy": 0.8552346229553223, "num_tokens": 318973752.0, "step": 8403 }, { "epoch": 1.069075181274647, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.7903265953063965, "learning_rate": 5e-05, "loss": 0.6559, "mean_token_accuracy": 0.8888795375823975, "num_tokens": 319009376.0, "step": 8404 }, { "epoch": 1.0692023915532376, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.497848510742188, "learning_rate": 5e-05, "loss": 0.8523, "mean_token_accuracy": 0.8741848468780518, "num_tokens": 319050175.0, "step": 8405 }, { "epoch": 1.0693296018318281, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.344975471496582, "learning_rate": 5e-05, "loss": 0.6804, "mean_token_accuracy": 0.8743546605110168, "num_tokens": 319080808.0, "step": 8406 }, { "epoch": 1.0694568121104184, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0004787445068359375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.000770568847656, "learning_rate": 5e-05, "loss": 0.8947, "mean_token_accuracy": 0.8791930675506592, "num_tokens": 319123123.0, "step": 8407 }, { "epoch": 1.069584022389009, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.309003829956055, "learning_rate": 5e-05, "loss": 0.6857, "mean_token_accuracy": 0.8797277212142944, "num_tokens": 319155116.0, "step": 8408 }, { "epoch": 1.0697112326675995, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.247692108154297, "learning_rate": 5e-05, "loss": 0.9374, "mean_token_accuracy": 0.8828716278076172, "num_tokens": 319188087.0, "step": 8409 }, { "epoch": 1.06983844294619, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 6.859976291656494, "learning_rate": 5e-05, "loss": 0.8544, "mean_token_accuracy": 0.8610195517539978, "num_tokens": 319228820.0, "step": 8410 }, { "epoch": 1.0699656532247805, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.831609725952148, "learning_rate": 5e-05, "loss": 1.034, "mean_token_accuracy": 0.8499060869216919, "num_tokens": 319266046.0, "step": 8411 }, { "epoch": 1.070092863503371, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.700993537902832, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.8764753341674805, "num_tokens": 319305209.0, "step": 8412 }, { "epoch": 1.0702200737819616, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.738009929656982, "learning_rate": 5e-05, "loss": 0.8662, "mean_token_accuracy": 0.8682471513748169, "num_tokens": 319347463.0, "step": 8413 }, { "epoch": 1.0703472840605521, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.265264511108398, "learning_rate": 5e-05, "loss": 0.8751, "mean_token_accuracy": 0.8818094730377197, "num_tokens": 319383965.0, "step": 8414 }, { "epoch": 1.0704744943391427, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.9565749168396, "learning_rate": 5e-05, "loss": 0.7227, "mean_token_accuracy": 0.8799859881401062, "num_tokens": 319418330.0, "step": 8415 }, { "epoch": 1.0706017046177332, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.440722465515137, "learning_rate": 5e-05, "loss": 0.9037, "mean_token_accuracy": 0.8700703978538513, "num_tokens": 319460518.0, "step": 8416 }, { "epoch": 1.0707289148963237, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.835145473480225, "learning_rate": 5e-05, "loss": 0.7585, "mean_token_accuracy": 0.8693807125091553, "num_tokens": 319504044.0, "step": 8417 }, { "epoch": 1.070856125174914, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.392911911010742, "learning_rate": 5e-05, "loss": 0.8741, "mean_token_accuracy": 0.8737558126449585, "num_tokens": 319540707.0, "step": 8418 }, { "epoch": 1.0709833354535045, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.858196258544922, "learning_rate": 5e-05, "loss": 0.7305, "mean_token_accuracy": 0.877895712852478, "num_tokens": 319580968.0, "step": 8419 }, { "epoch": 1.071110545732095, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.489413261413574, "learning_rate": 5e-05, "loss": 0.9277, "mean_token_accuracy": 0.8641320466995239, "num_tokens": 319619271.0, "step": 8420 }, { "epoch": 1.0712377560106856, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.767648696899414, "learning_rate": 5e-05, "loss": 0.7725, "mean_token_accuracy": 0.867728590965271, "num_tokens": 319657728.0, "step": 8421 }, { "epoch": 1.0713649662892761, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.689997673034668, "learning_rate": 5e-05, "loss": 0.9401, "mean_token_accuracy": 0.8653384447097778, "num_tokens": 319696807.0, "step": 8422 }, { "epoch": 1.0714921765678667, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.819738864898682, "learning_rate": 5e-05, "loss": 0.798, "mean_token_accuracy": 0.8645520806312561, "num_tokens": 319734397.0, "step": 8423 }, { "epoch": 1.0716193868464572, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.674705505371094, "learning_rate": 5e-05, "loss": 0.9644, "mean_token_accuracy": 0.8599058985710144, "num_tokens": 319769247.0, "step": 8424 }, { "epoch": 1.0717465971250477, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.101172924041748, "learning_rate": 5e-05, "loss": 0.8257, "mean_token_accuracy": 0.8687283992767334, "num_tokens": 319801147.0, "step": 8425 }, { "epoch": 1.0718738074036382, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.318868637084961, "learning_rate": 5e-05, "loss": 0.8803, "mean_token_accuracy": 0.8790833353996277, "num_tokens": 319834083.0, "step": 8426 }, { "epoch": 1.0720010176822288, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00048065185546875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.3719258308410645, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.885141134262085, "num_tokens": 319872586.0, "step": 8427 }, { "epoch": 1.0721282279608193, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.965213775634766, "learning_rate": 5e-05, "loss": 0.8532, "mean_token_accuracy": 0.871792733669281, "num_tokens": 319902756.0, "step": 8428 }, { "epoch": 1.0722554382394098, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.409817218780518, "learning_rate": 5e-05, "loss": 0.8087, "mean_token_accuracy": 0.8753329515457153, "num_tokens": 319943610.0, "step": 8429 }, { "epoch": 1.0723826485180004, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.863588809967041, "learning_rate": 5e-05, "loss": 0.8527, "mean_token_accuracy": 0.8662204146385193, "num_tokens": 319983762.0, "step": 8430 }, { "epoch": 1.0725098587965907, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.206252574920654, "learning_rate": 5e-05, "loss": 0.7722, "mean_token_accuracy": 0.8722037076950073, "num_tokens": 320019560.0, "step": 8431 }, { "epoch": 1.0726370690751812, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.031438827514648, "learning_rate": 5e-05, "loss": 0.8425, "mean_token_accuracy": 0.8653998374938965, "num_tokens": 320052907.0, "step": 8432 }, { "epoch": 1.0727642793537717, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.950345039367676, "learning_rate": 5e-05, "loss": 0.7882, "mean_token_accuracy": 0.8557512760162354, "num_tokens": 320087670.0, "step": 8433 }, { "epoch": 1.0728914896323622, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.232293128967285, "learning_rate": 5e-05, "loss": 0.8585, "mean_token_accuracy": 0.8733262419700623, "num_tokens": 320125768.0, "step": 8434 }, { "epoch": 1.0730186999109528, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.814713001251221, "learning_rate": 5e-05, "loss": 0.7018, "mean_token_accuracy": 0.8802680969238281, "num_tokens": 320163498.0, "step": 8435 }, { "epoch": 1.0731459101895433, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.330873489379883, "learning_rate": 5e-05, "loss": 0.8759, "mean_token_accuracy": 0.8677340745925903, "num_tokens": 320200951.0, "step": 8436 }, { "epoch": 1.0732731204681338, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.868743419647217, "learning_rate": 5e-05, "loss": 0.7137, "mean_token_accuracy": 0.879795491695404, "num_tokens": 320239258.0, "step": 8437 }, { "epoch": 1.0734003307467244, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.211840629577637, "learning_rate": 5e-05, "loss": 0.8452, "mean_token_accuracy": 0.8697391748428345, "num_tokens": 320276872.0, "step": 8438 }, { "epoch": 1.073527541025315, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.082926273345947, "learning_rate": 5e-05, "loss": 0.7717, "mean_token_accuracy": 0.8672703504562378, "num_tokens": 320315340.0, "step": 8439 }, { "epoch": 1.0736547513039054, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.8910627365112305, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.8792238235473633, "num_tokens": 320351213.0, "step": 8440 }, { "epoch": 1.073781961582496, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.358884811401367, "learning_rate": 5e-05, "loss": 0.7922, "mean_token_accuracy": 0.8688493371009827, "num_tokens": 320391456.0, "step": 8441 }, { "epoch": 1.0739091718610865, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.524528503417969, "learning_rate": 5e-05, "loss": 0.7565, "mean_token_accuracy": 0.8781492114067078, "num_tokens": 320428193.0, "step": 8442 }, { "epoch": 1.0740363821396768, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.573357582092285, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8773339986801147, "num_tokens": 320466685.0, "step": 8443 }, { "epoch": 1.0741635924182673, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.256450176239014, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8669484853744507, "num_tokens": 320508936.0, "step": 8444 }, { "epoch": 1.0742908026968578, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.705890655517578, "learning_rate": 5e-05, "loss": 0.8417, "mean_token_accuracy": 0.8523197770118713, "num_tokens": 320546644.0, "step": 8445 }, { "epoch": 1.0744180129754484, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.092615604400635, "learning_rate": 5e-05, "loss": 0.643, "mean_token_accuracy": 0.8914608359336853, "num_tokens": 320571074.0, "step": 8446 }, { "epoch": 1.074545223254039, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.7804646492004395, "learning_rate": 5e-05, "loss": 0.7698, "mean_token_accuracy": 0.8694319725036621, "num_tokens": 320605412.0, "step": 8447 }, { "epoch": 1.0746724335326294, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.9876508712768555, "learning_rate": 5e-05, "loss": 0.7257, "mean_token_accuracy": 0.8672839403152466, "num_tokens": 320643077.0, "step": 8448 }, { "epoch": 1.07479964381122, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.871820449829102, "learning_rate": 5e-05, "loss": 0.7445, "mean_token_accuracy": 0.8796373009681702, "num_tokens": 320680630.0, "step": 8449 }, { "epoch": 1.0749268540898105, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.858163833618164, "learning_rate": 5e-05, "loss": 0.6699, "mean_token_accuracy": 0.8818832635879517, "num_tokens": 320720405.0, "step": 8450 }, { "epoch": 1.075054064368401, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.027871131896973, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8781067132949829, "num_tokens": 320753453.0, "step": 8451 }, { "epoch": 1.0751812746469915, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.724116325378418, "learning_rate": 5e-05, "loss": 0.6806, "mean_token_accuracy": 0.8735474348068237, "num_tokens": 320795316.0, "step": 8452 }, { "epoch": 1.075308484925582, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.192673683166504, "learning_rate": 5e-05, "loss": 0.7691, "mean_token_accuracy": 0.8777639865875244, "num_tokens": 320834022.0, "step": 8453 }, { "epoch": 1.0754356952041726, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.6015305519104, "learning_rate": 5e-05, "loss": 0.7986, "mean_token_accuracy": 0.8385680317878723, "num_tokens": 320878410.0, "step": 8454 }, { "epoch": 1.0755629054827631, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.374735832214355, "learning_rate": 5e-05, "loss": 0.7605, "mean_token_accuracy": 0.8900320529937744, "num_tokens": 320918767.0, "step": 8455 }, { "epoch": 1.0756901157613534, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.509382724761963, "learning_rate": 5e-05, "loss": 0.6948, "mean_token_accuracy": 0.8731487989425659, "num_tokens": 320963845.0, "step": 8456 }, { "epoch": 1.075817326039944, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.518065452575684, "learning_rate": 5e-05, "loss": 0.8248, "mean_token_accuracy": 0.8791419267654419, "num_tokens": 321004811.0, "step": 8457 }, { "epoch": 1.0759445363185345, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.575868129730225, "learning_rate": 5e-05, "loss": 0.7051, "mean_token_accuracy": 0.8701236248016357, "num_tokens": 321040894.0, "step": 8458 }, { "epoch": 1.076071746597125, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.418179512023926, "learning_rate": 5e-05, "loss": 0.9058, "mean_token_accuracy": 0.8559257984161377, "num_tokens": 321083602.0, "step": 8459 }, { "epoch": 1.0761989568757155, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.903164386749268, "learning_rate": 5e-05, "loss": 0.7648, "mean_token_accuracy": 0.8697153329849243, "num_tokens": 321119444.0, "step": 8460 }, { "epoch": 1.076326167154306, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.983726501464844, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8774886131286621, "num_tokens": 321155125.0, "step": 8461 }, { "epoch": 1.0764533774328966, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.388280868530273, "learning_rate": 5e-05, "loss": 0.7584, "mean_token_accuracy": 0.8787030577659607, "num_tokens": 321193645.0, "step": 8462 }, { "epoch": 1.0765805877114871, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.3704514503479, "learning_rate": 5e-05, "loss": 0.7993, "mean_token_accuracy": 0.8584750890731812, "num_tokens": 321227118.0, "step": 8463 }, { "epoch": 1.0767077979900777, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.8271002769470215, "learning_rate": 5e-05, "loss": 0.7909, "mean_token_accuracy": 0.8705393075942993, "num_tokens": 321263245.0, "step": 8464 }, { "epoch": 1.0768350082686682, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.87652063369751, "learning_rate": 5e-05, "loss": 0.7676, "mean_token_accuracy": 0.8505114912986755, "num_tokens": 321301476.0, "step": 8465 }, { "epoch": 1.0769622185472587, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.146683692932129, "learning_rate": 5e-05, "loss": 0.7951, "mean_token_accuracy": 0.8732852339744568, "num_tokens": 321333464.0, "step": 8466 }, { "epoch": 1.077089428825849, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.643702030181885, "learning_rate": 5e-05, "loss": 0.6867, "mean_token_accuracy": 0.8718473315238953, "num_tokens": 321367807.0, "step": 8467 }, { "epoch": 1.0772166391044395, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.286528587341309, "learning_rate": 5e-05, "loss": 0.7767, "mean_token_accuracy": 0.8830891847610474, "num_tokens": 321408638.0, "step": 8468 }, { "epoch": 1.07734384938303, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.623640537261963, "learning_rate": 5e-05, "loss": 0.7135, "mean_token_accuracy": 0.8692305088043213, "num_tokens": 321451036.0, "step": 8469 }, { "epoch": 1.0774710596616206, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.260175704956055, "learning_rate": 5e-05, "loss": 0.7797, "mean_token_accuracy": 0.8825314044952393, "num_tokens": 321491423.0, "step": 8470 }, { "epoch": 1.0775982699402111, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.744528770446777, "learning_rate": 5e-05, "loss": 0.735, "mean_token_accuracy": 0.8613153696060181, "num_tokens": 321527956.0, "step": 8471 }, { "epoch": 1.0777254802188017, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.114973068237305, "learning_rate": 5e-05, "loss": 0.779, "mean_token_accuracy": 0.8820620775222778, "num_tokens": 321560185.0, "step": 8472 }, { "epoch": 1.0778526904973922, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.928752422332764, "learning_rate": 5e-05, "loss": 0.7067, "mean_token_accuracy": 0.8773364424705505, "num_tokens": 321602073.0, "step": 8473 }, { "epoch": 1.0779799007759827, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.901064395904541, "learning_rate": 5e-05, "loss": 0.7863, "mean_token_accuracy": 0.8720607757568359, "num_tokens": 321636722.0, "step": 8474 }, { "epoch": 1.0781071110545732, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.118016242980957, "learning_rate": 5e-05, "loss": 0.7402, "mean_token_accuracy": 0.8734273910522461, "num_tokens": 321675348.0, "step": 8475 }, { "epoch": 1.0782343213331638, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.682508945465088, "learning_rate": 5e-05, "loss": 0.7677, "mean_token_accuracy": 0.872124433517456, "num_tokens": 321711983.0, "step": 8476 }, { "epoch": 1.0783615316117543, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.222264766693115, "learning_rate": 5e-05, "loss": 0.6808, "mean_token_accuracy": 0.8878788948059082, "num_tokens": 321748268.0, "step": 8477 }, { "epoch": 1.0784887418903448, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.589454650878906, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8679651021957397, "num_tokens": 321783082.0, "step": 8478 }, { "epoch": 1.0786159521689354, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.177776336669922, "learning_rate": 5e-05, "loss": 0.7392, "mean_token_accuracy": 0.8646308183670044, "num_tokens": 321823634.0, "step": 8479 }, { "epoch": 1.0787431624475257, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.663328647613525, "learning_rate": 5e-05, "loss": 0.7334, "mean_token_accuracy": 0.8787367939949036, "num_tokens": 321859033.0, "step": 8480 }, { "epoch": 1.0788703727261162, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.007009983062744, "learning_rate": 5e-05, "loss": 0.7116, "mean_token_accuracy": 0.870746374130249, "num_tokens": 321896227.0, "step": 8481 }, { "epoch": 1.0789975830047067, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.90087890625, "learning_rate": 5e-05, "loss": 0.768, "mean_token_accuracy": 0.8728359937667847, "num_tokens": 321934329.0, "step": 8482 }, { "epoch": 1.0791247932832972, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.6824846267700195, "learning_rate": 5e-05, "loss": 0.7045, "mean_token_accuracy": 0.8664648532867432, "num_tokens": 321977426.0, "step": 8483 }, { "epoch": 1.0792520035618878, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.346874237060547, "learning_rate": 5e-05, "loss": 0.7925, "mean_token_accuracy": 0.8728652596473694, "num_tokens": 322014603.0, "step": 8484 }, { "epoch": 1.0793792138404783, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.263315200805664, "learning_rate": 5e-05, "loss": 0.645, "mean_token_accuracy": 0.8826238512992859, "num_tokens": 322051809.0, "step": 8485 }, { "epoch": 1.0795064241190688, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.9392728805542, "learning_rate": 5e-05, "loss": 0.8895, "mean_token_accuracy": 0.8706591129302979, "num_tokens": 322091628.0, "step": 8486 }, { "epoch": 1.0796336343976594, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.000775337219238, "learning_rate": 5e-05, "loss": 0.7149, "mean_token_accuracy": 0.865106463432312, "num_tokens": 322130616.0, "step": 8487 }, { "epoch": 1.0797608446762499, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.426168441772461, "learning_rate": 5e-05, "loss": 0.9159, "mean_token_accuracy": 0.8852181434631348, "num_tokens": 322172525.0, "step": 8488 }, { "epoch": 1.0798880549548404, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.471932411193848, "learning_rate": 5e-05, "loss": 0.7526, "mean_token_accuracy": 0.8760338425636292, "num_tokens": 322213438.0, "step": 8489 }, { "epoch": 1.080015265233431, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.020898818969727, "learning_rate": 5e-05, "loss": 0.9369, "mean_token_accuracy": 0.8784279227256775, "num_tokens": 322252492.0, "step": 8490 }, { "epoch": 1.0801424755120215, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.7064528465271, "learning_rate": 5e-05, "loss": 0.9231, "mean_token_accuracy": 0.8727642893791199, "num_tokens": 322295320.0, "step": 8491 }, { "epoch": 1.0802696857906118, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.340818405151367, "learning_rate": 5e-05, "loss": 0.7448, "mean_token_accuracy": 0.88463294506073, "num_tokens": 322335475.0, "step": 8492 }, { "epoch": 1.0803968960692023, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.174349784851074, "learning_rate": 5e-05, "loss": 0.9584, "mean_token_accuracy": 0.8790810704231262, "num_tokens": 322373559.0, "step": 8493 }, { "epoch": 1.0805241063477928, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.723161697387695, "learning_rate": 5e-05, "loss": 0.6956, "mean_token_accuracy": 0.8617831468582153, "num_tokens": 322407978.0, "step": 8494 }, { "epoch": 1.0806513166263834, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00075531005859375, "grad_norm": 11.170636177062988, "learning_rate": 5e-05, "loss": 1.1686, "mean_token_accuracy": 0.8711889982223511, "num_tokens": 322443657.0, "step": 8495 }, { "epoch": 1.080778526904974, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 6.847747325897217, "learning_rate": 5e-05, "loss": 0.8446, "mean_token_accuracy": 0.8892265558242798, "num_tokens": 322479424.0, "step": 8496 }, { "epoch": 1.0809057371835644, "ewc_loss": 0.9609375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 12.19768238067627, "learning_rate": 5e-05, "loss": 1.368, "mean_token_accuracy": 0.8585571050643921, "num_tokens": 322523112.0, "step": 8497 }, { "epoch": 1.081032947462155, "ewc_loss": 1.0390625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0009918212890625, "grad_norm": 9.911767959594727, "learning_rate": 5e-05, "loss": 1.3915, "mean_token_accuracy": 0.8777023553848267, "num_tokens": 322560444.0, "step": 8498 }, { "epoch": 1.0811601577407455, "ewc_loss": 0.6875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.707494735717773, "learning_rate": 5e-05, "loss": 1.0399, "mean_token_accuracy": 0.8728943467140198, "num_tokens": 322591211.0, "step": 8499 }, { "epoch": 1.081287368019336, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 9.46090030670166, "learning_rate": 5e-05, "loss": 1.2811, "mean_token_accuracy": 0.8548042178153992, "num_tokens": 322625129.0, "step": 8500 }, { "epoch": 1.0814145782979265, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 9.905564308166504, "learning_rate": 5e-05, "loss": 1.0611, "mean_token_accuracy": 0.8738546371459961, "num_tokens": 322663428.0, "step": 8501 }, { "epoch": 1.081541788576517, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.826423645019531, "learning_rate": 5e-05, "loss": 0.98, "mean_token_accuracy": 0.8808121085166931, "num_tokens": 322707936.0, "step": 8502 }, { "epoch": 1.0816689988551076, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.020050048828125, "learning_rate": 5e-05, "loss": 0.9656, "mean_token_accuracy": 0.884594202041626, "num_tokens": 322748605.0, "step": 8503 }, { "epoch": 1.0817962091336981, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.375548362731934, "learning_rate": 5e-05, "loss": 0.9717, "mean_token_accuracy": 0.8555907607078552, "num_tokens": 322788009.0, "step": 8504 }, { "epoch": 1.0819234194122884, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.544576644897461, "learning_rate": 5e-05, "loss": 0.9367, "mean_token_accuracy": 0.870718240737915, "num_tokens": 322825159.0, "step": 8505 }, { "epoch": 1.082050629690879, "ewc_loss": 0.5, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.9189300537109375, "learning_rate": 5e-05, "loss": 0.9441, "mean_token_accuracy": 0.845842719078064, "num_tokens": 322866769.0, "step": 8506 }, { "epoch": 1.0821778399694695, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.284943580627441, "learning_rate": 5e-05, "loss": 0.8314, "mean_token_accuracy": 0.8806315064430237, "num_tokens": 322901851.0, "step": 8507 }, { "epoch": 1.08230505024806, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.598318576812744, "learning_rate": 5e-05, "loss": 0.8279, "mean_token_accuracy": 0.8746552467346191, "num_tokens": 322938527.0, "step": 8508 }, { "epoch": 1.0824322605266505, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.127612113952637, "learning_rate": 5e-05, "loss": 0.8218, "mean_token_accuracy": 0.8765262365341187, "num_tokens": 322972135.0, "step": 8509 }, { "epoch": 1.082559470805241, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.308104515075684, "learning_rate": 5e-05, "loss": 0.8105, "mean_token_accuracy": 0.8602627515792847, "num_tokens": 323010451.0, "step": 8510 }, { "epoch": 1.0826866810838316, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.179840087890625, "learning_rate": 5e-05, "loss": 0.8379, "mean_token_accuracy": 0.8761971592903137, "num_tokens": 323047583.0, "step": 8511 }, { "epoch": 1.0828138913624221, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.94304895401001, "learning_rate": 5e-05, "loss": 0.7497, "mean_token_accuracy": 0.8648196458816528, "num_tokens": 323088754.0, "step": 8512 }, { "epoch": 1.0829411016410126, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.577452659606934, "learning_rate": 5e-05, "loss": 0.8623, "mean_token_accuracy": 0.8780426979064941, "num_tokens": 323128510.0, "step": 8513 }, { "epoch": 1.0830683119196032, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.514218330383301, "learning_rate": 5e-05, "loss": 0.7192, "mean_token_accuracy": 0.8688749670982361, "num_tokens": 323172557.0, "step": 8514 }, { "epoch": 1.0831955221981937, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.264543533325195, "learning_rate": 5e-05, "loss": 0.9047, "mean_token_accuracy": 0.8821899890899658, "num_tokens": 323208298.0, "step": 8515 }, { "epoch": 1.083322732476784, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.402622699737549, "learning_rate": 5e-05, "loss": 0.7603, "mean_token_accuracy": 0.8629629611968994, "num_tokens": 323250824.0, "step": 8516 }, { "epoch": 1.0834499427553745, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.900691032409668, "learning_rate": 5e-05, "loss": 1.0315, "mean_token_accuracy": 0.8764008283615112, "num_tokens": 323290792.0, "step": 8517 }, { "epoch": 1.083577153033965, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.25408411026001, "learning_rate": 5e-05, "loss": 0.8853, "mean_token_accuracy": 0.8682106733322144, "num_tokens": 323327806.0, "step": 8518 }, { "epoch": 1.0837043633125556, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.703849792480469, "learning_rate": 5e-05, "loss": 1.0457, "mean_token_accuracy": 0.8738185167312622, "num_tokens": 323364275.0, "step": 8519 }, { "epoch": 1.0838315735911461, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 8.37182331085205, "learning_rate": 5e-05, "loss": 0.9951, "mean_token_accuracy": 0.8808530569076538, "num_tokens": 323395671.0, "step": 8520 }, { "epoch": 1.0839587838697367, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.281206130981445, "learning_rate": 5e-05, "loss": 0.8882, "mean_token_accuracy": 0.8750331401824951, "num_tokens": 323432719.0, "step": 8521 }, { "epoch": 1.0840859941483272, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.86882209777832, "learning_rate": 5e-05, "loss": 1.0295, "mean_token_accuracy": 0.863990306854248, "num_tokens": 323474431.0, "step": 8522 }, { "epoch": 1.0842132044269177, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.011938571929932, "learning_rate": 5e-05, "loss": 0.7765, "mean_token_accuracy": 0.8760432004928589, "num_tokens": 323515439.0, "step": 8523 }, { "epoch": 1.0843404147055082, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.312485694885254, "learning_rate": 5e-05, "loss": 1.0245, "mean_token_accuracy": 0.8617023825645447, "num_tokens": 323552613.0, "step": 8524 }, { "epoch": 1.0844676249840988, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.905026912689209, "learning_rate": 5e-05, "loss": 0.7344, "mean_token_accuracy": 0.8922542929649353, "num_tokens": 323590037.0, "step": 8525 }, { "epoch": 1.0845948352626893, "ewc_loss": 0.625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.270543098449707, "learning_rate": 5e-05, "loss": 0.9598, "mean_token_accuracy": 0.8827779293060303, "num_tokens": 323634540.0, "step": 8526 }, { "epoch": 1.0847220455412798, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.516420841217041, "learning_rate": 5e-05, "loss": 0.8749, "mean_token_accuracy": 0.8764681816101074, "num_tokens": 323666772.0, "step": 8527 }, { "epoch": 1.0848492558198704, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0004825592041015625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.629185676574707, "learning_rate": 5e-05, "loss": 0.9114, "mean_token_accuracy": 0.8730031251907349, "num_tokens": 323707847.0, "step": 8528 }, { "epoch": 1.0849764660984607, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.806769371032715, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8891031742095947, "num_tokens": 323737648.0, "step": 8529 }, { "epoch": 1.0851036763770512, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.001729965209961, "learning_rate": 5e-05, "loss": 0.8208, "mean_token_accuracy": 0.8767130970954895, "num_tokens": 323772621.0, "step": 8530 }, { "epoch": 1.0852308866556417, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.743565559387207, "learning_rate": 5e-05, "loss": 0.8932, "mean_token_accuracy": 0.8622646331787109, "num_tokens": 323806384.0, "step": 8531 }, { "epoch": 1.0853580969342322, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.751243591308594, "learning_rate": 5e-05, "loss": 0.7377, "mean_token_accuracy": 0.892785906791687, "num_tokens": 323845287.0, "step": 8532 }, { "epoch": 1.0854853072128228, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.498332500457764, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.8799473643302917, "num_tokens": 323884182.0, "step": 8533 }, { "epoch": 1.0856125174914133, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.744087219238281, "learning_rate": 5e-05, "loss": 0.847, "mean_token_accuracy": 0.8545893430709839, "num_tokens": 323919357.0, "step": 8534 }, { "epoch": 1.0857397277700038, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.338099479675293, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8801394701004028, "num_tokens": 323957993.0, "step": 8535 }, { "epoch": 1.0858669380485944, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.670189380645752, "learning_rate": 5e-05, "loss": 0.7272, "mean_token_accuracy": 0.8831402659416199, "num_tokens": 323991726.0, "step": 8536 }, { "epoch": 1.0859941483271849, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.340392112731934, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.872785747051239, "num_tokens": 324031639.0, "step": 8537 }, { "epoch": 1.0861213586057754, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.566348552703857, "learning_rate": 5e-05, "loss": 0.7743, "mean_token_accuracy": 0.8645908832550049, "num_tokens": 324063359.0, "step": 8538 }, { "epoch": 1.086248568884366, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.321264743804932, "learning_rate": 5e-05, "loss": 0.7391, "mean_token_accuracy": 0.8694438338279724, "num_tokens": 324105058.0, "step": 8539 }, { "epoch": 1.0863757791629565, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.583199977874756, "learning_rate": 5e-05, "loss": 0.7862, "mean_token_accuracy": 0.860260009765625, "num_tokens": 324144209.0, "step": 8540 }, { "epoch": 1.0865029894415468, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.221054553985596, "learning_rate": 5e-05, "loss": 0.6974, "mean_token_accuracy": 0.8783448934555054, "num_tokens": 324179194.0, "step": 8541 }, { "epoch": 1.0866301997201373, "ewc_loss": 0.375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.677912712097168, "learning_rate": 5e-05, "loss": 0.7141, "mean_token_accuracy": 0.8814974427223206, "num_tokens": 324218405.0, "step": 8542 }, { "epoch": 1.0867574099987278, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.080400466918945, "learning_rate": 5e-05, "loss": 0.7832, "mean_token_accuracy": 0.8478682637214661, "num_tokens": 324258298.0, "step": 8543 }, { "epoch": 1.0868846202773184, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.802484035491943, "learning_rate": 5e-05, "loss": 0.7221, "mean_token_accuracy": 0.8819804191589355, "num_tokens": 324296518.0, "step": 8544 }, { "epoch": 1.0870118305559089, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.901772975921631, "learning_rate": 5e-05, "loss": 0.7021, "mean_token_accuracy": 0.8712976574897766, "num_tokens": 324336724.0, "step": 8545 }, { "epoch": 1.0871390408344994, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.034734725952148, "learning_rate": 5e-05, "loss": 0.7818, "mean_token_accuracy": 0.866946816444397, "num_tokens": 324373464.0, "step": 8546 }, { "epoch": 1.08726625111309, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.626255035400391, "learning_rate": 5e-05, "loss": 0.6662, "mean_token_accuracy": 0.8777742981910706, "num_tokens": 324410893.0, "step": 8547 }, { "epoch": 1.0873934613916805, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.466636657714844, "learning_rate": 5e-05, "loss": 0.7964, "mean_token_accuracy": 0.8784979581832886, "num_tokens": 324446765.0, "step": 8548 }, { "epoch": 1.087520671670271, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.205010414123535, "learning_rate": 5e-05, "loss": 0.6485, "mean_token_accuracy": 0.8751113414764404, "num_tokens": 324482745.0, "step": 8549 }, { "epoch": 1.0876478819488615, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 9.127431869506836, "learning_rate": 5e-05, "loss": 0.8745, "mean_token_accuracy": 0.8791103959083557, "num_tokens": 324513764.0, "step": 8550 }, { "epoch": 1.087775092227452, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 5.901495456695557, "learning_rate": 5e-05, "loss": 0.6719, "mean_token_accuracy": 0.8723983764648438, "num_tokens": 324549502.0, "step": 8551 }, { "epoch": 1.0879023025060426, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.780069351196289, "learning_rate": 5e-05, "loss": 0.9774, "mean_token_accuracy": 0.8738484382629395, "num_tokens": 324584456.0, "step": 8552 }, { "epoch": 1.0880295127846331, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.497171878814697, "learning_rate": 5e-05, "loss": 0.7992, "mean_token_accuracy": 0.8647343516349792, "num_tokens": 324624070.0, "step": 8553 }, { "epoch": 1.0881567230632234, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.620521545410156, "learning_rate": 5e-05, "loss": 1.0071, "mean_token_accuracy": 0.8736199140548706, "num_tokens": 324665971.0, "step": 8554 }, { "epoch": 1.088283933341814, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 7.899153709411621, "learning_rate": 5e-05, "loss": 0.9377, "mean_token_accuracy": 0.8876746892929077, "num_tokens": 324706705.0, "step": 8555 }, { "epoch": 1.0884111436204045, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.035436630249023, "learning_rate": 5e-05, "loss": 0.8659, "mean_token_accuracy": 0.8653340339660645, "num_tokens": 324744701.0, "step": 8556 }, { "epoch": 1.088538353898995, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.022622108459473, "learning_rate": 5e-05, "loss": 1.0089, "mean_token_accuracy": 0.8685379028320312, "num_tokens": 324781447.0, "step": 8557 }, { "epoch": 1.0886655641775855, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.459023952484131, "learning_rate": 5e-05, "loss": 0.7142, "mean_token_accuracy": 0.8745605945587158, "num_tokens": 324816753.0, "step": 8558 }, { "epoch": 1.088792774456176, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.275286674499512, "learning_rate": 5e-05, "loss": 1.084, "mean_token_accuracy": 0.866667628288269, "num_tokens": 324855093.0, "step": 8559 }, { "epoch": 1.0889199847347666, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 6.878767490386963, "learning_rate": 5e-05, "loss": 0.8516, "mean_token_accuracy": 0.8719519376754761, "num_tokens": 324893107.0, "step": 8560 }, { "epoch": 1.0890471950133571, "ewc_loss": 0.734375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 10.555814743041992, "learning_rate": 5e-05, "loss": 1.0976, "mean_token_accuracy": 0.8722435235977173, "num_tokens": 324927589.0, "step": 8561 }, { "epoch": 1.0891744052919476, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 8.610544204711914, "learning_rate": 5e-05, "loss": 1.0939, "mean_token_accuracy": 0.8808532357215881, "num_tokens": 324958801.0, "step": 8562 }, { "epoch": 1.0893016155705382, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.146805763244629, "learning_rate": 5e-05, "loss": 0.9221, "mean_token_accuracy": 0.8768056035041809, "num_tokens": 324995617.0, "step": 8563 }, { "epoch": 1.0894288258491287, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 8.954450607299805, "learning_rate": 5e-05, "loss": 1.0636, "mean_token_accuracy": 0.8673684597015381, "num_tokens": 325035702.0, "step": 8564 }, { "epoch": 1.089556036127719, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.035540580749512, "learning_rate": 5e-05, "loss": 0.9182, "mean_token_accuracy": 0.8606210350990295, "num_tokens": 325078063.0, "step": 8565 }, { "epoch": 1.0896832464063095, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.743505477905273, "learning_rate": 5e-05, "loss": 0.9592, "mean_token_accuracy": 0.871397078037262, "num_tokens": 325113127.0, "step": 8566 }, { "epoch": 1.0898104566849, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.9557390213012695, "learning_rate": 5e-05, "loss": 0.9077, "mean_token_accuracy": 0.8693962097167969, "num_tokens": 325149451.0, "step": 8567 }, { "epoch": 1.0899376669634906, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.819035053253174, "learning_rate": 5e-05, "loss": 0.8231, "mean_token_accuracy": 0.8696184754371643, "num_tokens": 325193265.0, "step": 8568 }, { "epoch": 1.0900648772420811, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.405072212219238, "learning_rate": 5e-05, "loss": 0.911, "mean_token_accuracy": 0.8746752738952637, "num_tokens": 325235160.0, "step": 8569 }, { "epoch": 1.0901920875206716, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.882571220397949, "learning_rate": 5e-05, "loss": 0.7431, "mean_token_accuracy": 0.8680754899978638, "num_tokens": 325276299.0, "step": 8570 }, { "epoch": 1.0903192977992622, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.146361351013184, "learning_rate": 5e-05, "loss": 0.987, "mean_token_accuracy": 0.8586401343345642, "num_tokens": 325313509.0, "step": 8571 }, { "epoch": 1.0904465080778527, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.5226569175720215, "learning_rate": 5e-05, "loss": 0.7672, "mean_token_accuracy": 0.8611212968826294, "num_tokens": 325355883.0, "step": 8572 }, { "epoch": 1.0905737183564432, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.824142456054688, "learning_rate": 5e-05, "loss": 1.0076, "mean_token_accuracy": 0.8770243525505066, "num_tokens": 325392937.0, "step": 8573 }, { "epoch": 1.0907009286350338, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.112340927124023, "learning_rate": 5e-05, "loss": 0.8566, "mean_token_accuracy": 0.8758202791213989, "num_tokens": 325422132.0, "step": 8574 }, { "epoch": 1.0908281389136243, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.851397514343262, "learning_rate": 5e-05, "loss": 1.0293, "mean_token_accuracy": 0.8706227540969849, "num_tokens": 325460383.0, "step": 8575 }, { "epoch": 1.0909553491922148, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.9126410484313965, "learning_rate": 5e-05, "loss": 0.9969, "mean_token_accuracy": 0.8623982071876526, "num_tokens": 325499242.0, "step": 8576 }, { "epoch": 1.0910825594708053, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.234933853149414, "learning_rate": 5e-05, "loss": 0.9529, "mean_token_accuracy": 0.8781497478485107, "num_tokens": 325535389.0, "step": 8577 }, { "epoch": 1.0912097697493957, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.00418472290039, "learning_rate": 5e-05, "loss": 0.9248, "mean_token_accuracy": 0.8794336318969727, "num_tokens": 325574459.0, "step": 8578 }, { "epoch": 1.0913369800279862, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.720440864562988, "learning_rate": 5e-05, "loss": 0.9778, "mean_token_accuracy": 0.8578193187713623, "num_tokens": 325616296.0, "step": 8579 }, { "epoch": 1.0914641903065767, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.632299423217773, "learning_rate": 5e-05, "loss": 0.8748, "mean_token_accuracy": 0.8617666959762573, "num_tokens": 325652707.0, "step": 8580 }, { "epoch": 1.0915914005851672, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.715577125549316, "learning_rate": 5e-05, "loss": 0.9091, "mean_token_accuracy": 0.88290935754776, "num_tokens": 325692423.0, "step": 8581 }, { "epoch": 1.0917186108637578, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.239065647125244, "learning_rate": 5e-05, "loss": 0.7994, "mean_token_accuracy": 0.8708314895629883, "num_tokens": 325731638.0, "step": 8582 }, { "epoch": 1.0918458211423483, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.847606658935547, "learning_rate": 5e-05, "loss": 0.923, "mean_token_accuracy": 0.8799375295639038, "num_tokens": 325771482.0, "step": 8583 }, { "epoch": 1.0919730314209388, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.238502025604248, "learning_rate": 5e-05, "loss": 0.8233, "mean_token_accuracy": 0.8677230477333069, "num_tokens": 325807447.0, "step": 8584 }, { "epoch": 1.0921002416995294, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.739312171936035, "learning_rate": 5e-05, "loss": 0.9307, "mean_token_accuracy": 0.8720477819442749, "num_tokens": 325848070.0, "step": 8585 }, { "epoch": 1.0922274519781199, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.445049285888672, "learning_rate": 5e-05, "loss": 0.8022, "mean_token_accuracy": 0.8813847899436951, "num_tokens": 325889027.0, "step": 8586 }, { "epoch": 1.0923546622567104, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.422636032104492, "learning_rate": 5e-05, "loss": 0.9113, "mean_token_accuracy": 0.8673638701438904, "num_tokens": 325926547.0, "step": 8587 }, { "epoch": 1.092481872535301, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.542230606079102, "learning_rate": 5e-05, "loss": 0.82, "mean_token_accuracy": 0.8770079016685486, "num_tokens": 325968968.0, "step": 8588 }, { "epoch": 1.0926090828138915, "ewc_loss": 0.5, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.159382820129395, "learning_rate": 5e-05, "loss": 0.8758, "mean_token_accuracy": 0.8732929229736328, "num_tokens": 326005060.0, "step": 8589 }, { "epoch": 1.0927362930924818, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.470961093902588, "learning_rate": 5e-05, "loss": 0.8336, "mean_token_accuracy": 0.8623697757720947, "num_tokens": 326049895.0, "step": 8590 }, { "epoch": 1.0928635033710723, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.04695987701416, "learning_rate": 5e-05, "loss": 0.8906, "mean_token_accuracy": 0.864812433719635, "num_tokens": 326087772.0, "step": 8591 }, { "epoch": 1.0929907136496628, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.330013275146484, "learning_rate": 5e-05, "loss": 0.7502, "mean_token_accuracy": 0.8766801357269287, "num_tokens": 326126518.0, "step": 8592 }, { "epoch": 1.0931179239282534, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.101378440856934, "learning_rate": 5e-05, "loss": 0.8773, "mean_token_accuracy": 0.8641035556793213, "num_tokens": 326166672.0, "step": 8593 }, { "epoch": 1.0932451342068439, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.172327518463135, "learning_rate": 5e-05, "loss": 0.7468, "mean_token_accuracy": 0.8703840970993042, "num_tokens": 326209099.0, "step": 8594 }, { "epoch": 1.0933723444854344, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.177684783935547, "learning_rate": 5e-05, "loss": 0.8877, "mean_token_accuracy": 0.8630762696266174, "num_tokens": 326244101.0, "step": 8595 }, { "epoch": 1.093499554764025, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.119275093078613, "learning_rate": 5e-05, "loss": 0.7337, "mean_token_accuracy": 0.8725267052650452, "num_tokens": 326279997.0, "step": 8596 }, { "epoch": 1.0936267650426155, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.136484146118164, "learning_rate": 5e-05, "loss": 0.859, "mean_token_accuracy": 0.8651463985443115, "num_tokens": 326321241.0, "step": 8597 }, { "epoch": 1.093753975321206, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.174264430999756, "learning_rate": 5e-05, "loss": 0.74, "mean_token_accuracy": 0.8717303276062012, "num_tokens": 326359662.0, "step": 8598 }, { "epoch": 1.0938811855997965, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.010939598083496, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.8808675408363342, "num_tokens": 326397072.0, "step": 8599 }, { "epoch": 1.094008395878387, "ewc_loss": 0.375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.279731273651123, "learning_rate": 5e-05, "loss": 0.7486, "mean_token_accuracy": 0.8677206635475159, "num_tokens": 326440029.0, "step": 8600 }, { "epoch": 1.0941356061569776, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.800479412078857, "learning_rate": 5e-05, "loss": 0.769, "mean_token_accuracy": 0.8762413263320923, "num_tokens": 326479586.0, "step": 8601 }, { "epoch": 1.094262816435568, "ewc_loss": 0.375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.369256019592285, "learning_rate": 5e-05, "loss": 0.7486, "mean_token_accuracy": 0.8698305487632751, "num_tokens": 326518235.0, "step": 8602 }, { "epoch": 1.0943900267141584, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.647190093994141, "learning_rate": 5e-05, "loss": 0.7415, "mean_token_accuracy": 0.8774283528327942, "num_tokens": 326559420.0, "step": 8603 }, { "epoch": 1.094517236992749, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.402103424072266, "learning_rate": 5e-05, "loss": 0.7602, "mean_token_accuracy": 0.8675718307495117, "num_tokens": 326597020.0, "step": 8604 }, { "epoch": 1.0946444472713395, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.558187484741211, "learning_rate": 5e-05, "loss": 0.8343, "mean_token_accuracy": 0.8455250263214111, "num_tokens": 326632248.0, "step": 8605 }, { "epoch": 1.09477165754993, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.405730247497559, "learning_rate": 5e-05, "loss": 0.7213, "mean_token_accuracy": 0.877069354057312, "num_tokens": 326666185.0, "step": 8606 }, { "epoch": 1.0948988678285205, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.517187595367432, "learning_rate": 5e-05, "loss": 0.7845, "mean_token_accuracy": 0.8534883856773376, "num_tokens": 326700356.0, "step": 8607 }, { "epoch": 1.095026078107111, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.358396053314209, "learning_rate": 5e-05, "loss": 0.7193, "mean_token_accuracy": 0.8755083084106445, "num_tokens": 326739448.0, "step": 8608 }, { "epoch": 1.0951532883857016, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.539479732513428, "learning_rate": 5e-05, "loss": 0.7257, "mean_token_accuracy": 0.8731772899627686, "num_tokens": 326776290.0, "step": 8609 }, { "epoch": 1.0952804986642921, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.282044410705566, "learning_rate": 5e-05, "loss": 0.7006, "mean_token_accuracy": 0.8771395683288574, "num_tokens": 326817975.0, "step": 8610 }, { "epoch": 1.0954077089428826, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.609984397888184, "learning_rate": 5e-05, "loss": 0.726, "mean_token_accuracy": 0.872922956943512, "num_tokens": 326852404.0, "step": 8611 }, { "epoch": 1.0955349192214732, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.206507682800293, "learning_rate": 5e-05, "loss": 0.7232, "mean_token_accuracy": 0.8641495108604431, "num_tokens": 326888519.0, "step": 8612 }, { "epoch": 1.0956621295000637, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.649028301239014, "learning_rate": 5e-05, "loss": 0.7314, "mean_token_accuracy": 0.8676899671554565, "num_tokens": 326925644.0, "step": 8613 }, { "epoch": 1.095789339778654, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.125887870788574, "learning_rate": 5e-05, "loss": 0.708, "mean_token_accuracy": 0.8701232075691223, "num_tokens": 326961490.0, "step": 8614 }, { "epoch": 1.0959165500572445, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.748458385467529, "learning_rate": 5e-05, "loss": 0.7736, "mean_token_accuracy": 0.8608893752098083, "num_tokens": 326998726.0, "step": 8615 }, { "epoch": 1.096043760335835, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 7.016924858093262, "learning_rate": 5e-05, "loss": 0.7444, "mean_token_accuracy": 0.8552290201187134, "num_tokens": 327040717.0, "step": 8616 }, { "epoch": 1.0961709706144256, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.89596700668335, "learning_rate": 5e-05, "loss": 0.7592, "mean_token_accuracy": 0.8686124086380005, "num_tokens": 327070597.0, "step": 8617 }, { "epoch": 1.0962981808930161, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.824519157409668, "learning_rate": 5e-05, "loss": 0.7384, "mean_token_accuracy": 0.8566726446151733, "num_tokens": 327108598.0, "step": 8618 }, { "epoch": 1.0964253911716066, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.139963150024414, "learning_rate": 5e-05, "loss": 0.7388, "mean_token_accuracy": 0.8799841403961182, "num_tokens": 327142750.0, "step": 8619 }, { "epoch": 1.0965526014501972, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.596187114715576, "learning_rate": 5e-05, "loss": 0.7061, "mean_token_accuracy": 0.8599106669425964, "num_tokens": 327182536.0, "step": 8620 }, { "epoch": 1.0966798117287877, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.454413414001465, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.8699666261672974, "num_tokens": 327218135.0, "step": 8621 }, { "epoch": 1.0968070220073782, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.370707988739014, "learning_rate": 5e-05, "loss": 0.6391, "mean_token_accuracy": 0.8833255767822266, "num_tokens": 327251581.0, "step": 8622 }, { "epoch": 1.0969342322859688, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.776162147521973, "learning_rate": 5e-05, "loss": 0.8413, "mean_token_accuracy": 0.8720296621322632, "num_tokens": 327288466.0, "step": 8623 }, { "epoch": 1.0970614425645593, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.332752227783203, "learning_rate": 5e-05, "loss": 0.6744, "mean_token_accuracy": 0.8749613761901855, "num_tokens": 327325646.0, "step": 8624 }, { "epoch": 1.0971886528431498, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.798004150390625, "learning_rate": 5e-05, "loss": 0.8968, "mean_token_accuracy": 0.8645820021629333, "num_tokens": 327358244.0, "step": 8625 }, { "epoch": 1.0973158631217403, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.8003973960876465, "learning_rate": 5e-05, "loss": 0.691, "mean_token_accuracy": 0.879578709602356, "num_tokens": 327397315.0, "step": 8626 }, { "epoch": 1.0974430734003306, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.115741729736328, "learning_rate": 5e-05, "loss": 0.817, "mean_token_accuracy": 0.8724155426025391, "num_tokens": 327432642.0, "step": 8627 }, { "epoch": 1.0975702836789212, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.703254222869873, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.8596533536911011, "num_tokens": 327478563.0, "step": 8628 }, { "epoch": 1.0976974939575117, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.835772514343262, "learning_rate": 5e-05, "loss": 0.6904, "mean_token_accuracy": 0.8785889148712158, "num_tokens": 327514295.0, "step": 8629 }, { "epoch": 1.0978247042361022, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.91710376739502, "learning_rate": 5e-05, "loss": 0.8589, "mean_token_accuracy": 0.8777270317077637, "num_tokens": 327550496.0, "step": 8630 }, { "epoch": 1.0979519145146928, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000225067138671875, "grad_norm": 5.6971001625061035, "learning_rate": 5e-05, "loss": 0.5962, "mean_token_accuracy": 0.8880409002304077, "num_tokens": 327584692.0, "step": 8631 }, { "epoch": 1.0980791247932833, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.402983665466309, "learning_rate": 5e-05, "loss": 0.9869, "mean_token_accuracy": 0.8798770904541016, "num_tokens": 327627710.0, "step": 8632 }, { "epoch": 1.0982063350718738, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 5.848862171173096, "learning_rate": 5e-05, "loss": 0.7309, "mean_token_accuracy": 0.8627743721008301, "num_tokens": 327666958.0, "step": 8633 }, { "epoch": 1.0983335453504643, "ewc_loss": 0.88671875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 11.522143363952637, "learning_rate": 5e-05, "loss": 1.2366, "mean_token_accuracy": 0.8767822980880737, "num_tokens": 327700829.0, "step": 8634 }, { "epoch": 1.0984607556290549, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 8.685355186462402, "learning_rate": 5e-05, "loss": 1.1932, "mean_token_accuracy": 0.8630716800689697, "num_tokens": 327741410.0, "step": 8635 }, { "epoch": 1.0985879659076454, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.156144142150879, "learning_rate": 5e-05, "loss": 1.0613, "mean_token_accuracy": 0.8578438758850098, "num_tokens": 327778305.0, "step": 8636 }, { "epoch": 1.098715176186236, "ewc_loss": 0.83203125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 9.483351707458496, "learning_rate": 5e-05, "loss": 1.2256, "mean_token_accuracy": 0.8634852766990662, "num_tokens": 327813384.0, "step": 8637 }, { "epoch": 1.0988423864648265, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.421782493591309, "learning_rate": 5e-05, "loss": 0.8956, "mean_token_accuracy": 0.8735759258270264, "num_tokens": 327851928.0, "step": 8638 }, { "epoch": 1.0989695967434168, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.105327606201172, "learning_rate": 5e-05, "loss": 1.0603, "mean_token_accuracy": 0.8817317485809326, "num_tokens": 327890053.0, "step": 8639 }, { "epoch": 1.0990968070220073, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 7.768190383911133, "learning_rate": 5e-05, "loss": 0.916, "mean_token_accuracy": 0.8714591264724731, "num_tokens": 327928947.0, "step": 8640 }, { "epoch": 1.0992240173005978, "ewc_loss": 0.703125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.448528289794922, "learning_rate": 5e-05, "loss": 1.0592, "mean_token_accuracy": 0.8762556910514832, "num_tokens": 327966286.0, "step": 8641 }, { "epoch": 1.0993512275791884, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.211618423461914, "learning_rate": 5e-05, "loss": 1.0024, "mean_token_accuracy": 0.869604229927063, "num_tokens": 328005480.0, "step": 8642 }, { "epoch": 1.0994784378577789, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.835661888122559, "learning_rate": 5e-05, "loss": 0.9922, "mean_token_accuracy": 0.8771873712539673, "num_tokens": 328040587.0, "step": 8643 }, { "epoch": 1.0996056481363694, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.290986061096191, "learning_rate": 5e-05, "loss": 1.0003, "mean_token_accuracy": 0.8666463494300842, "num_tokens": 328075856.0, "step": 8644 }, { "epoch": 1.09973285841496, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.974209785461426, "learning_rate": 5e-05, "loss": 0.9725, "mean_token_accuracy": 0.8617115020751953, "num_tokens": 328116243.0, "step": 8645 }, { "epoch": 1.0998600686935505, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.173223495483398, "learning_rate": 5e-05, "loss": 0.9373, "mean_token_accuracy": 0.8720400333404541, "num_tokens": 328151854.0, "step": 8646 }, { "epoch": 1.099987278972141, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.175010681152344, "learning_rate": 5e-05, "loss": 0.8737, "mean_token_accuracy": 0.8664283752441406, "num_tokens": 328191455.0, "step": 8647 }, { "epoch": 1.1001144892507315, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.220645904541016, "learning_rate": 5e-05, "loss": 0.8724, "mean_token_accuracy": 0.8761429190635681, "num_tokens": 328232034.0, "step": 8648 }, { "epoch": 1.100241699529322, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.454712390899658, "learning_rate": 5e-05, "loss": 0.7904, "mean_token_accuracy": 0.8735746741294861, "num_tokens": 328265581.0, "step": 8649 }, { "epoch": 1.1003689098079126, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.484658241271973, "learning_rate": 5e-05, "loss": 0.956, "mean_token_accuracy": 0.8534634113311768, "num_tokens": 328304354.0, "step": 8650 }, { "epoch": 1.100496120086503, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.89193868637085, "learning_rate": 5e-05, "loss": 0.7571, "mean_token_accuracy": 0.8642638921737671, "num_tokens": 328347924.0, "step": 8651 }, { "epoch": 1.1006233303650934, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.991242408752441, "learning_rate": 5e-05, "loss": 0.8726, "mean_token_accuracy": 0.885433554649353, "num_tokens": 328385603.0, "step": 8652 }, { "epoch": 1.100750540643684, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.553586959838867, "learning_rate": 5e-05, "loss": 0.703, "mean_token_accuracy": 0.8799845576286316, "num_tokens": 328422881.0, "step": 8653 }, { "epoch": 1.1008777509222745, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.562556266784668, "learning_rate": 5e-05, "loss": 0.9958, "mean_token_accuracy": 0.8672580718994141, "num_tokens": 328461191.0, "step": 8654 }, { "epoch": 1.101004961200865, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 6.863095283508301, "learning_rate": 5e-05, "loss": 0.8267, "mean_token_accuracy": 0.8698713779449463, "num_tokens": 328499824.0, "step": 8655 }, { "epoch": 1.1011321714794555, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.596346855163574, "learning_rate": 5e-05, "loss": 0.9971, "mean_token_accuracy": 0.8769434094429016, "num_tokens": 328538456.0, "step": 8656 }, { "epoch": 1.101259381758046, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.714741230010986, "learning_rate": 5e-05, "loss": 0.9782, "mean_token_accuracy": 0.8575031757354736, "num_tokens": 328574722.0, "step": 8657 }, { "epoch": 1.1013865920366366, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.753754615783691, "learning_rate": 5e-05, "loss": 0.9081, "mean_token_accuracy": 0.8736173510551453, "num_tokens": 328608248.0, "step": 8658 }, { "epoch": 1.101513802315227, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.222380638122559, "learning_rate": 5e-05, "loss": 0.9663, "mean_token_accuracy": 0.8717817068099976, "num_tokens": 328651011.0, "step": 8659 }, { "epoch": 1.1016410125938176, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.67077112197876, "learning_rate": 5e-05, "loss": 0.8088, "mean_token_accuracy": 0.8682594895362854, "num_tokens": 328694438.0, "step": 8660 }, { "epoch": 1.1017682228724082, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.492009162902832, "learning_rate": 5e-05, "loss": 0.9436, "mean_token_accuracy": 0.8707908987998962, "num_tokens": 328731851.0, "step": 8661 }, { "epoch": 1.1018954331509987, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.162078857421875, "learning_rate": 5e-05, "loss": 0.7704, "mean_token_accuracy": 0.8705053329467773, "num_tokens": 328772159.0, "step": 8662 }, { "epoch": 1.102022643429589, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.541996955871582, "learning_rate": 5e-05, "loss": 0.8575, "mean_token_accuracy": 0.8872768878936768, "num_tokens": 328814712.0, "step": 8663 }, { "epoch": 1.1021498537081795, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.1674885749816895, "learning_rate": 5e-05, "loss": 0.792, "mean_token_accuracy": 0.866588294506073, "num_tokens": 328854243.0, "step": 8664 }, { "epoch": 1.10227706398677, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.372819900512695, "learning_rate": 5e-05, "loss": 0.9087, "mean_token_accuracy": 0.859230637550354, "num_tokens": 328894109.0, "step": 8665 }, { "epoch": 1.1024042742653606, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.265382289886475, "learning_rate": 5e-05, "loss": 0.7562, "mean_token_accuracy": 0.8844704031944275, "num_tokens": 328937275.0, "step": 8666 }, { "epoch": 1.1025314845439511, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.185194969177246, "learning_rate": 5e-05, "loss": 0.8203, "mean_token_accuracy": 0.8815103769302368, "num_tokens": 328974572.0, "step": 8667 }, { "epoch": 1.1026586948225416, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.312264442443848, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8678387403488159, "num_tokens": 329013871.0, "step": 8668 }, { "epoch": 1.1027859051011322, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.959372520446777, "learning_rate": 5e-05, "loss": 0.842, "mean_token_accuracy": 0.8593797087669373, "num_tokens": 329050333.0, "step": 8669 }, { "epoch": 1.1029131153797227, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.4081878662109375, "learning_rate": 5e-05, "loss": 0.8031, "mean_token_accuracy": 0.8672277927398682, "num_tokens": 329091817.0, "step": 8670 }, { "epoch": 1.1030403256583132, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.658325672149658, "learning_rate": 5e-05, "loss": 0.7482, "mean_token_accuracy": 0.8740394711494446, "num_tokens": 329128523.0, "step": 8671 }, { "epoch": 1.1031675359369038, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.56131649017334, "learning_rate": 5e-05, "loss": 0.731, "mean_token_accuracy": 0.8884454965591431, "num_tokens": 329166320.0, "step": 8672 }, { "epoch": 1.1032947462154943, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.400424957275391, "learning_rate": 5e-05, "loss": 0.722, "mean_token_accuracy": 0.8737898468971252, "num_tokens": 329195140.0, "step": 8673 }, { "epoch": 1.1034219564940848, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.7265496253967285, "learning_rate": 5e-05, "loss": 0.7506, "mean_token_accuracy": 0.8822906613349915, "num_tokens": 329230814.0, "step": 8674 }, { "epoch": 1.1035491667726753, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.156864643096924, "learning_rate": 5e-05, "loss": 0.6903, "mean_token_accuracy": 0.8793665170669556, "num_tokens": 329270695.0, "step": 8675 }, { "epoch": 1.1036763770512656, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.97733736038208, "learning_rate": 5e-05, "loss": 0.8233, "mean_token_accuracy": 0.8584648966789246, "num_tokens": 329308061.0, "step": 8676 }, { "epoch": 1.1038035873298562, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.857012748718262, "learning_rate": 5e-05, "loss": 0.696, "mean_token_accuracy": 0.8692452311515808, "num_tokens": 329347093.0, "step": 8677 }, { "epoch": 1.1039307976084467, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.279020309448242, "learning_rate": 5e-05, "loss": 0.8233, "mean_token_accuracy": 0.867661714553833, "num_tokens": 329385009.0, "step": 8678 }, { "epoch": 1.1040580078870372, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.634424209594727, "learning_rate": 5e-05, "loss": 0.7084, "mean_token_accuracy": 0.8712000846862793, "num_tokens": 329425317.0, "step": 8679 }, { "epoch": 1.1041852181656278, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.550053596496582, "learning_rate": 5e-05, "loss": 0.8133, "mean_token_accuracy": 0.8791736960411072, "num_tokens": 329461581.0, "step": 8680 }, { "epoch": 1.1043124284442183, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.53354549407959, "learning_rate": 5e-05, "loss": 0.665, "mean_token_accuracy": 0.8838636875152588, "num_tokens": 329494064.0, "step": 8681 }, { "epoch": 1.1044396387228088, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.733220100402832, "learning_rate": 5e-05, "loss": 0.8606, "mean_token_accuracy": 0.8782767653465271, "num_tokens": 329530807.0, "step": 8682 }, { "epoch": 1.1045668490013993, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.6156325340271, "learning_rate": 5e-05, "loss": 0.7797, "mean_token_accuracy": 0.8528122901916504, "num_tokens": 329575689.0, "step": 8683 }, { "epoch": 1.1046940592799899, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.657422065734863, "learning_rate": 5e-05, "loss": 0.8958, "mean_token_accuracy": 0.8662521839141846, "num_tokens": 329616855.0, "step": 8684 }, { "epoch": 1.1048212695585804, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000484466552734375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.991955280303955, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.8738092184066772, "num_tokens": 329654085.0, "step": 8685 }, { "epoch": 1.104948479837171, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.181106567382812, "learning_rate": 5e-05, "loss": 0.8461, "mean_token_accuracy": 0.8711370229721069, "num_tokens": 329688375.0, "step": 8686 }, { "epoch": 1.1050756901157615, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.436195373535156, "learning_rate": 5e-05, "loss": 0.8255, "mean_token_accuracy": 0.862018346786499, "num_tokens": 329734258.0, "step": 8687 }, { "epoch": 1.1052029003943518, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0004863739013671875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.605616092681885, "learning_rate": 5e-05, "loss": 0.7875, "mean_token_accuracy": 0.8722205758094788, "num_tokens": 329772080.0, "step": 8688 }, { "epoch": 1.1053301106729423, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.7081193923950195, "learning_rate": 5e-05, "loss": 0.7652, "mean_token_accuracy": 0.8821507096290588, "num_tokens": 329817809.0, "step": 8689 }, { "epoch": 1.1054573209515328, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.315186500549316, "learning_rate": 5e-05, "loss": 0.7144, "mean_token_accuracy": 0.8774548768997192, "num_tokens": 329856063.0, "step": 8690 }, { "epoch": 1.1055845312301233, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.728750705718994, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8610790967941284, "num_tokens": 329892548.0, "step": 8691 }, { "epoch": 1.1057117415087139, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.294673919677734, "learning_rate": 5e-05, "loss": 0.682, "mean_token_accuracy": 0.886019766330719, "num_tokens": 329931897.0, "step": 8692 }, { "epoch": 1.1058389517873044, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.589694023132324, "learning_rate": 5e-05, "loss": 0.7687, "mean_token_accuracy": 0.8659776449203491, "num_tokens": 329967477.0, "step": 8693 }, { "epoch": 1.105966162065895, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.427794933319092, "learning_rate": 5e-05, "loss": 0.7416, "mean_token_accuracy": 0.8700648546218872, "num_tokens": 330004840.0, "step": 8694 }, { "epoch": 1.1060933723444855, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.3495659828186035, "learning_rate": 5e-05, "loss": 0.7179, "mean_token_accuracy": 0.8726319670677185, "num_tokens": 330048928.0, "step": 8695 }, { "epoch": 1.106220582623076, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.6288251876831055, "learning_rate": 5e-05, "loss": 0.7503, "mean_token_accuracy": 0.8671956658363342, "num_tokens": 330084112.0, "step": 8696 }, { "epoch": 1.1063477929016665, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.117254734039307, "learning_rate": 5e-05, "loss": 0.7408, "mean_token_accuracy": 0.8616108298301697, "num_tokens": 330127533.0, "step": 8697 }, { "epoch": 1.106475003180257, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.8341169357299805, "learning_rate": 5e-05, "loss": 0.8238, "mean_token_accuracy": 0.8516529202461243, "num_tokens": 330169379.0, "step": 8698 }, { "epoch": 1.1066022134588476, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.88141393661499, "learning_rate": 5e-05, "loss": 0.6653, "mean_token_accuracy": 0.8808913230895996, "num_tokens": 330205553.0, "step": 8699 }, { "epoch": 1.106729423737438, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.112716674804688, "learning_rate": 5e-05, "loss": 0.7629, "mean_token_accuracy": 0.8765367865562439, "num_tokens": 330241088.0, "step": 8700 }, { "epoch": 1.1068566340160284, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.599040508270264, "learning_rate": 5e-05, "loss": 0.675, "mean_token_accuracy": 0.8719633221626282, "num_tokens": 330282520.0, "step": 8701 }, { "epoch": 1.106983844294619, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.5048246383667, "learning_rate": 5e-05, "loss": 0.8139, "mean_token_accuracy": 0.8673861622810364, "num_tokens": 330326231.0, "step": 8702 }, { "epoch": 1.1071110545732095, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.256968021392822, "learning_rate": 5e-05, "loss": 0.6358, "mean_token_accuracy": 0.8811741471290588, "num_tokens": 330363674.0, "step": 8703 }, { "epoch": 1.1072382648518, "ewc_loss": 0.5, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 9.00037670135498, "learning_rate": 5e-05, "loss": 0.7952, "mean_token_accuracy": 0.8947933316230774, "num_tokens": 330398890.0, "step": 8704 }, { "epoch": 1.1073654751303905, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.111269474029541, "learning_rate": 5e-05, "loss": 0.6556, "mean_token_accuracy": 0.8792579770088196, "num_tokens": 330434564.0, "step": 8705 }, { "epoch": 1.107492685408981, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.295681953430176, "learning_rate": 5e-05, "loss": 0.9083, "mean_token_accuracy": 0.8792967796325684, "num_tokens": 330475580.0, "step": 8706 }, { "epoch": 1.1076198956875716, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.583073139190674, "learning_rate": 5e-05, "loss": 0.7037, "mean_token_accuracy": 0.881629467010498, "num_tokens": 330513255.0, "step": 8707 }, { "epoch": 1.107747105966162, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.796116828918457, "learning_rate": 5e-05, "loss": 0.9503, "mean_token_accuracy": 0.8629308938980103, "num_tokens": 330554101.0, "step": 8708 }, { "epoch": 1.1078743162447526, "ewc_loss": 0.5, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.609612464904785, "learning_rate": 5e-05, "loss": 0.9203, "mean_token_accuracy": 0.856941819190979, "num_tokens": 330595312.0, "step": 8709 }, { "epoch": 1.1080015265233432, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.412065029144287, "learning_rate": 5e-05, "loss": 0.7525, "mean_token_accuracy": 0.8800781965255737, "num_tokens": 330633510.0, "step": 8710 }, { "epoch": 1.1081287368019337, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.675362586975098, "learning_rate": 5e-05, "loss": 0.8987, "mean_token_accuracy": 0.8767920136451721, "num_tokens": 330673626.0, "step": 8711 }, { "epoch": 1.108255947080524, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.262556552886963, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8697100877761841, "num_tokens": 330715051.0, "step": 8712 }, { "epoch": 1.1083831573591145, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.6150541305542, "learning_rate": 5e-05, "loss": 0.9661, "mean_token_accuracy": 0.8711452484130859, "num_tokens": 330765603.0, "step": 8713 }, { "epoch": 1.108510367637705, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.317410469055176, "learning_rate": 5e-05, "loss": 0.7464, "mean_token_accuracy": 0.8683942556381226, "num_tokens": 330799427.0, "step": 8714 }, { "epoch": 1.1086375779162956, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.702306747436523, "learning_rate": 5e-05, "loss": 1.0017, "mean_token_accuracy": 0.874786913394928, "num_tokens": 330840810.0, "step": 8715 }, { "epoch": 1.108764788194886, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.581168174743652, "learning_rate": 5e-05, "loss": 0.9888, "mean_token_accuracy": 0.8628337383270264, "num_tokens": 330878325.0, "step": 8716 }, { "epoch": 1.1088919984734766, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.490326881408691, "learning_rate": 5e-05, "loss": 0.8735, "mean_token_accuracy": 0.8834378719329834, "num_tokens": 330914817.0, "step": 8717 }, { "epoch": 1.1090192087520672, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.51368236541748, "learning_rate": 5e-05, "loss": 0.9482, "mean_token_accuracy": 0.8794836401939392, "num_tokens": 330951890.0, "step": 8718 }, { "epoch": 1.1091464190306577, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.171712875366211, "learning_rate": 5e-05, "loss": 0.8127, "mean_token_accuracy": 0.8597123622894287, "num_tokens": 330986807.0, "step": 8719 }, { "epoch": 1.1092736293092482, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.93822956085205, "learning_rate": 5e-05, "loss": 0.9629, "mean_token_accuracy": 0.8639580011367798, "num_tokens": 331027132.0, "step": 8720 }, { "epoch": 1.1094008395878387, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.107203006744385, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8773894906044006, "num_tokens": 331062669.0, "step": 8721 }, { "epoch": 1.1095280498664293, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.35746955871582, "learning_rate": 5e-05, "loss": 0.8623, "mean_token_accuracy": 0.8736429214477539, "num_tokens": 331097376.0, "step": 8722 }, { "epoch": 1.1096552601450198, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.754483222961426, "learning_rate": 5e-05, "loss": 0.822, "mean_token_accuracy": 0.8822253346443176, "num_tokens": 331130930.0, "step": 8723 }, { "epoch": 1.1097824704236103, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.325554847717285, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8698632121086121, "num_tokens": 331172012.0, "step": 8724 }, { "epoch": 1.1099096807022006, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.345379829406738, "learning_rate": 5e-05, "loss": 0.9444, "mean_token_accuracy": 0.8507681488990784, "num_tokens": 331210016.0, "step": 8725 }, { "epoch": 1.1100368909807912, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.810661315917969, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.8646848201751709, "num_tokens": 331251818.0, "step": 8726 }, { "epoch": 1.1101641012593817, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.646085739135742, "learning_rate": 5e-05, "loss": 0.9159, "mean_token_accuracy": 0.8596042394638062, "num_tokens": 331283034.0, "step": 8727 }, { "epoch": 1.1102913115379722, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.774130344390869, "learning_rate": 5e-05, "loss": 0.7481, "mean_token_accuracy": 0.8677968978881836, "num_tokens": 331324985.0, "step": 8728 }, { "epoch": 1.1104185218165628, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.7244291305542, "learning_rate": 5e-05, "loss": 0.8902, "mean_token_accuracy": 0.8742832541465759, "num_tokens": 331358595.0, "step": 8729 }, { "epoch": 1.1105457320951533, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.891148567199707, "learning_rate": 5e-05, "loss": 0.77, "mean_token_accuracy": 0.8678761124610901, "num_tokens": 331387388.0, "step": 8730 }, { "epoch": 1.1106729423737438, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.639084815979004, "learning_rate": 5e-05, "loss": 0.9241, "mean_token_accuracy": 0.8672904968261719, "num_tokens": 331423562.0, "step": 8731 }, { "epoch": 1.1108001526523343, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 6.95573091506958, "learning_rate": 5e-05, "loss": 0.767, "mean_token_accuracy": 0.875806450843811, "num_tokens": 331458835.0, "step": 8732 }, { "epoch": 1.1109273629309249, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.621477127075195, "learning_rate": 5e-05, "loss": 0.9016, "mean_token_accuracy": 0.8649101257324219, "num_tokens": 331489889.0, "step": 8733 }, { "epoch": 1.1110545732095154, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.861551761627197, "learning_rate": 5e-05, "loss": 0.7182, "mean_token_accuracy": 0.8849471211433411, "num_tokens": 331529103.0, "step": 8734 }, { "epoch": 1.111181783488106, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.809956550598145, "learning_rate": 5e-05, "loss": 0.9066, "mean_token_accuracy": 0.8762389421463013, "num_tokens": 331566100.0, "step": 8735 }, { "epoch": 1.1113089937666965, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.69947624206543, "learning_rate": 5e-05, "loss": 0.7512, "mean_token_accuracy": 0.8773442506790161, "num_tokens": 331606704.0, "step": 8736 }, { "epoch": 1.1114362040452868, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.132759094238281, "learning_rate": 5e-05, "loss": 0.9312, "mean_token_accuracy": 0.8816447257995605, "num_tokens": 331642019.0, "step": 8737 }, { "epoch": 1.1115634143238773, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 6.88115119934082, "learning_rate": 5e-05, "loss": 0.7911, "mean_token_accuracy": 0.872612714767456, "num_tokens": 331672888.0, "step": 8738 }, { "epoch": 1.1116906246024678, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.070242881774902, "learning_rate": 5e-05, "loss": 0.9846, "mean_token_accuracy": 0.8731998205184937, "num_tokens": 331710383.0, "step": 8739 }, { "epoch": 1.1118178348810583, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.468711853027344, "learning_rate": 5e-05, "loss": 0.8662, "mean_token_accuracy": 0.8757444620132446, "num_tokens": 331754299.0, "step": 8740 }, { "epoch": 1.1119450451596489, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.36012077331543, "learning_rate": 5e-05, "loss": 0.9046, "mean_token_accuracy": 0.8732383251190186, "num_tokens": 331789329.0, "step": 8741 }, { "epoch": 1.1120722554382394, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 7.934582233428955, "learning_rate": 5e-05, "loss": 0.8776, "mean_token_accuracy": 0.8803180456161499, "num_tokens": 331822727.0, "step": 8742 }, { "epoch": 1.11219946571683, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.5637898445129395, "learning_rate": 5e-05, "loss": 0.7965, "mean_token_accuracy": 0.8753942251205444, "num_tokens": 331863091.0, "step": 8743 }, { "epoch": 1.1123266759954205, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.988234996795654, "learning_rate": 5e-05, "loss": 0.8576, "mean_token_accuracy": 0.8751447200775146, "num_tokens": 331900866.0, "step": 8744 }, { "epoch": 1.112453886274011, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.369045257568359, "learning_rate": 5e-05, "loss": 0.7397, "mean_token_accuracy": 0.8855022192001343, "num_tokens": 331936774.0, "step": 8745 }, { "epoch": 1.1125810965526015, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.689681053161621, "learning_rate": 5e-05, "loss": 0.7861, "mean_token_accuracy": 0.8763411045074463, "num_tokens": 331974629.0, "step": 8746 }, { "epoch": 1.112708306831192, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.569076061248779, "learning_rate": 5e-05, "loss": 0.7498, "mean_token_accuracy": 0.8808712363243103, "num_tokens": 332008707.0, "step": 8747 }, { "epoch": 1.1128355171097826, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.328323841094971, "learning_rate": 5e-05, "loss": 0.7104, "mean_token_accuracy": 0.8863598704338074, "num_tokens": 332046973.0, "step": 8748 }, { "epoch": 1.112962727388373, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.747770309448242, "learning_rate": 5e-05, "loss": 0.8066, "mean_token_accuracy": 0.8643197417259216, "num_tokens": 332092658.0, "step": 8749 }, { "epoch": 1.1130899376669634, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.200893878936768, "learning_rate": 5e-05, "loss": 0.7124, "mean_token_accuracy": 0.8756676316261292, "num_tokens": 332129705.0, "step": 8750 }, { "epoch": 1.113217147945554, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.725564479827881, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8719966411590576, "num_tokens": 332163129.0, "step": 8751 }, { "epoch": 1.1133443582241445, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.24343729019165, "learning_rate": 5e-05, "loss": 0.7274, "mean_token_accuracy": 0.8718453645706177, "num_tokens": 332198148.0, "step": 8752 }, { "epoch": 1.113471568502735, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.611865520477295, "learning_rate": 5e-05, "loss": 0.7436, "mean_token_accuracy": 0.8728685975074768, "num_tokens": 332238068.0, "step": 8753 }, { "epoch": 1.1135987787813255, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.276320934295654, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8627681732177734, "num_tokens": 332277455.0, "step": 8754 }, { "epoch": 1.113725989059916, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.576194763183594, "learning_rate": 5e-05, "loss": 0.7582, "mean_token_accuracy": 0.8680959939956665, "num_tokens": 332314778.0, "step": 8755 }, { "epoch": 1.1138531993385066, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.245828151702881, "learning_rate": 5e-05, "loss": 0.7069, "mean_token_accuracy": 0.8767890930175781, "num_tokens": 332347863.0, "step": 8756 }, { "epoch": 1.113980409617097, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.626345157623291, "learning_rate": 5e-05, "loss": 0.6857, "mean_token_accuracy": 0.885584831237793, "num_tokens": 332389577.0, "step": 8757 }, { "epoch": 1.1141076198956876, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.159857273101807, "learning_rate": 5e-05, "loss": 0.7228, "mean_token_accuracy": 0.8668526411056519, "num_tokens": 332434140.0, "step": 8758 }, { "epoch": 1.1142348301742782, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.752981662750244, "learning_rate": 5e-05, "loss": 0.7202, "mean_token_accuracy": 0.8763431310653687, "num_tokens": 332468843.0, "step": 8759 }, { "epoch": 1.1143620404528687, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.9305500984191895, "learning_rate": 5e-05, "loss": 0.6683, "mean_token_accuracy": 0.8800394535064697, "num_tokens": 332509863.0, "step": 8760 }, { "epoch": 1.114489250731459, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.059355735778809, "learning_rate": 5e-05, "loss": 0.7777, "mean_token_accuracy": 0.8699156045913696, "num_tokens": 332550242.0, "step": 8761 }, { "epoch": 1.1146164610100495, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.597032070159912, "learning_rate": 5e-05, "loss": 0.6688, "mean_token_accuracy": 0.8722465634346008, "num_tokens": 332588606.0, "step": 8762 }, { "epoch": 1.11474367128864, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.499664306640625, "learning_rate": 5e-05, "loss": 0.7868, "mean_token_accuracy": 0.8765566349029541, "num_tokens": 332623787.0, "step": 8763 }, { "epoch": 1.1148708815672306, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.206000804901123, "learning_rate": 5e-05, "loss": 0.6908, "mean_token_accuracy": 0.8629747033119202, "num_tokens": 332665358.0, "step": 8764 }, { "epoch": 1.114998091845821, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 9.128100395202637, "learning_rate": 5e-05, "loss": 0.8486, "mean_token_accuracy": 0.8816106915473938, "num_tokens": 332700541.0, "step": 8765 }, { "epoch": 1.1151253021244116, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.952770709991455, "learning_rate": 5e-05, "loss": 0.7039, "mean_token_accuracy": 0.8646188378334045, "num_tokens": 332737555.0, "step": 8766 }, { "epoch": 1.1152525124030022, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.660730361938477, "learning_rate": 5e-05, "loss": 0.9891, "mean_token_accuracy": 0.8644869327545166, "num_tokens": 332775957.0, "step": 8767 }, { "epoch": 1.1153797226815927, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.546757221221924, "learning_rate": 5e-05, "loss": 0.7494, "mean_token_accuracy": 0.8752652406692505, "num_tokens": 332812000.0, "step": 8768 }, { "epoch": 1.1155069329601832, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.240955352783203, "learning_rate": 5e-05, "loss": 0.9367, "mean_token_accuracy": 0.8859161138534546, "num_tokens": 332848974.0, "step": 8769 }, { "epoch": 1.1156341432387737, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.9058003425598145, "learning_rate": 5e-05, "loss": 0.9458, "mean_token_accuracy": 0.8752368688583374, "num_tokens": 332890362.0, "step": 8770 }, { "epoch": 1.1157613535173643, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.445259094238281, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.8679746389389038, "num_tokens": 332924547.0, "step": 8771 }, { "epoch": 1.1158885637959548, "ewc_loss": 0.625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.331226348876953, "learning_rate": 5e-05, "loss": 0.9535, "mean_token_accuracy": 0.8826953172683716, "num_tokens": 332957697.0, "step": 8772 }, { "epoch": 1.1160157740745453, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 5.922394275665283, "learning_rate": 5e-05, "loss": 0.6772, "mean_token_accuracy": 0.8722543716430664, "num_tokens": 332992671.0, "step": 8773 }, { "epoch": 1.1161429843531356, "ewc_loss": 0.78125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000732421875, "grad_norm": 11.0525541305542, "learning_rate": 5e-05, "loss": 1.127, "mean_token_accuracy": 0.8794209957122803, "num_tokens": 333033092.0, "step": 8774 }, { "epoch": 1.1162701946317262, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.153841018676758, "learning_rate": 5e-05, "loss": 0.9251, "mean_token_accuracy": 0.868009090423584, "num_tokens": 333071631.0, "step": 8775 }, { "epoch": 1.1163974049103167, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 11.41936206817627, "learning_rate": 5e-05, "loss": 1.1689, "mean_token_accuracy": 0.880662739276886, "num_tokens": 333100931.0, "step": 8776 }, { "epoch": 1.1165246151889072, "ewc_loss": 0.890625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000843048095703125, "grad_norm": 9.430099487304688, "learning_rate": 5e-05, "loss": 1.2067, "mean_token_accuracy": 0.893279492855072, "num_tokens": 333141063.0, "step": 8777 }, { "epoch": 1.1166518254674977, "ewc_loss": 0.625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.873392105102539, "learning_rate": 5e-05, "loss": 0.9994, "mean_token_accuracy": 0.8721555471420288, "num_tokens": 333180322.0, "step": 8778 }, { "epoch": 1.1167790357460883, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 9.321457862854004, "learning_rate": 5e-05, "loss": 1.1322, "mean_token_accuracy": 0.8662398457527161, "num_tokens": 333219836.0, "step": 8779 }, { "epoch": 1.1169062460246788, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.766450881958008, "learning_rate": 5e-05, "loss": 0.9888, "mean_token_accuracy": 0.8664449453353882, "num_tokens": 333263748.0, "step": 8780 }, { "epoch": 1.1170334563032693, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.992311477661133, "learning_rate": 5e-05, "loss": 0.9563, "mean_token_accuracy": 0.8730542063713074, "num_tokens": 333296110.0, "step": 8781 }, { "epoch": 1.1171606665818599, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.27679443359375, "learning_rate": 5e-05, "loss": 0.9517, "mean_token_accuracy": 0.8693931102752686, "num_tokens": 333328372.0, "step": 8782 }, { "epoch": 1.1172878768604504, "ewc_loss": 0.5, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.240005493164062, "learning_rate": 5e-05, "loss": 0.8474, "mean_token_accuracy": 0.8794839382171631, "num_tokens": 333367680.0, "step": 8783 }, { "epoch": 1.117415087139041, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.343006134033203, "learning_rate": 5e-05, "loss": 0.9012, "mean_token_accuracy": 0.8739559650421143, "num_tokens": 333406761.0, "step": 8784 }, { "epoch": 1.1175422974176314, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.468527793884277, "learning_rate": 5e-05, "loss": 0.7722, "mean_token_accuracy": 0.8770002126693726, "num_tokens": 333443268.0, "step": 8785 }, { "epoch": 1.1176695076962218, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.584836959838867, "learning_rate": 5e-05, "loss": 0.9171, "mean_token_accuracy": 0.8609031438827515, "num_tokens": 333478770.0, "step": 8786 }, { "epoch": 1.1177967179748123, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.084655284881592, "learning_rate": 5e-05, "loss": 0.7085, "mean_token_accuracy": 0.8889228105545044, "num_tokens": 333514971.0, "step": 8787 }, { "epoch": 1.1179239282534028, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.733078002929688, "learning_rate": 5e-05, "loss": 0.8957, "mean_token_accuracy": 0.8721742630004883, "num_tokens": 333553610.0, "step": 8788 }, { "epoch": 1.1180511385319933, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.966978549957275, "learning_rate": 5e-05, "loss": 0.7569, "mean_token_accuracy": 0.8731396198272705, "num_tokens": 333604283.0, "step": 8789 }, { "epoch": 1.1181783488105839, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.791790008544922, "learning_rate": 5e-05, "loss": 0.868, "mean_token_accuracy": 0.8836749792098999, "num_tokens": 333641621.0, "step": 8790 }, { "epoch": 1.1183055590891744, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.030152797698975, "learning_rate": 5e-05, "loss": 0.749, "mean_token_accuracy": 0.882935643196106, "num_tokens": 333676435.0, "step": 8791 }, { "epoch": 1.118432769367765, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.69894027709961, "learning_rate": 5e-05, "loss": 0.8884, "mean_token_accuracy": 0.8846160173416138, "num_tokens": 333715308.0, "step": 8792 }, { "epoch": 1.1185599796463555, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.163969993591309, "learning_rate": 5e-05, "loss": 0.8448, "mean_token_accuracy": 0.8581249713897705, "num_tokens": 333752086.0, "step": 8793 }, { "epoch": 1.118687189924946, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.436361312866211, "learning_rate": 5e-05, "loss": 0.8825, "mean_token_accuracy": 0.8700025081634521, "num_tokens": 333789726.0, "step": 8794 }, { "epoch": 1.1188144002035365, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.248612880706787, "learning_rate": 5e-05, "loss": 0.8433, "mean_token_accuracy": 0.8642327785491943, "num_tokens": 333828518.0, "step": 8795 }, { "epoch": 1.118941610482127, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.309589385986328, "learning_rate": 5e-05, "loss": 0.8982, "mean_token_accuracy": 0.866244912147522, "num_tokens": 333870520.0, "step": 8796 }, { "epoch": 1.1190688207607176, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.125995635986328, "learning_rate": 5e-05, "loss": 0.7543, "mean_token_accuracy": 0.8756915926933289, "num_tokens": 333914676.0, "step": 8797 }, { "epoch": 1.119196031039308, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.446693420410156, "learning_rate": 5e-05, "loss": 0.8604, "mean_token_accuracy": 0.8783450126647949, "num_tokens": 333953381.0, "step": 8798 }, { "epoch": 1.1193232413178984, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.927089214324951, "learning_rate": 5e-05, "loss": 0.7933, "mean_token_accuracy": 0.8598564863204956, "num_tokens": 333983275.0, "step": 8799 }, { "epoch": 1.119450451596489, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.668448448181152, "learning_rate": 5e-05, "loss": 0.8914, "mean_token_accuracy": 0.8793210983276367, "num_tokens": 334018530.0, "step": 8800 }, { "epoch": 1.1195776618750795, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.885548114776611, "learning_rate": 5e-05, "loss": 0.7491, "mean_token_accuracy": 0.879734992980957, "num_tokens": 334056606.0, "step": 8801 }, { "epoch": 1.11970487215367, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.728948593139648, "learning_rate": 5e-05, "loss": 0.9176, "mean_token_accuracy": 0.8753983378410339, "num_tokens": 334098541.0, "step": 8802 }, { "epoch": 1.1198320824322605, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.117300987243652, "learning_rate": 5e-05, "loss": 0.8105, "mean_token_accuracy": 0.8626018166542053, "num_tokens": 334140243.0, "step": 8803 }, { "epoch": 1.119959292710851, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.466390609741211, "learning_rate": 5e-05, "loss": 0.8764, "mean_token_accuracy": 0.8827193975448608, "num_tokens": 334177126.0, "step": 8804 }, { "epoch": 1.1200865029894416, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.487176418304443, "learning_rate": 5e-05, "loss": 0.8169, "mean_token_accuracy": 0.8795788288116455, "num_tokens": 334217058.0, "step": 8805 }, { "epoch": 1.120213713268032, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.8461503982543945, "learning_rate": 5e-05, "loss": 0.7832, "mean_token_accuracy": 0.891814112663269, "num_tokens": 334255060.0, "step": 8806 }, { "epoch": 1.1203409235466226, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.824553966522217, "learning_rate": 5e-05, "loss": 0.8393, "mean_token_accuracy": 0.8699891567230225, "num_tokens": 334290661.0, "step": 8807 }, { "epoch": 1.1204681338252132, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.306083679199219, "learning_rate": 5e-05, "loss": 0.7228, "mean_token_accuracy": 0.8865426778793335, "num_tokens": 334328661.0, "step": 8808 }, { "epoch": 1.1205953441038037, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.903205871582031, "learning_rate": 5e-05, "loss": 0.7952, "mean_token_accuracy": 0.8734420537948608, "num_tokens": 334367535.0, "step": 8809 }, { "epoch": 1.120722554382394, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.213809013366699, "learning_rate": 5e-05, "loss": 0.7249, "mean_token_accuracy": 0.8774906396865845, "num_tokens": 334407366.0, "step": 8810 }, { "epoch": 1.1208497646609845, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.715735912322998, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8746792078018188, "num_tokens": 334444049.0, "step": 8811 }, { "epoch": 1.120976974939575, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.399784564971924, "learning_rate": 5e-05, "loss": 0.7601, "mean_token_accuracy": 0.8688219785690308, "num_tokens": 334485405.0, "step": 8812 }, { "epoch": 1.1211041852181656, "ewc_loss": 0.375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.432430267333984, "learning_rate": 5e-05, "loss": 0.6671, "mean_token_accuracy": 0.8963720798492432, "num_tokens": 334525209.0, "step": 8813 }, { "epoch": 1.121231395496756, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.618525505065918, "learning_rate": 5e-05, "loss": 0.7489, "mean_token_accuracy": 0.8742183446884155, "num_tokens": 334569156.0, "step": 8814 }, { "epoch": 1.1213586057753466, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.188745021820068, "learning_rate": 5e-05, "loss": 0.6894, "mean_token_accuracy": 0.8813278079032898, "num_tokens": 334602692.0, "step": 8815 }, { "epoch": 1.1214858160539372, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.813404560089111, "learning_rate": 5e-05, "loss": 0.7718, "mean_token_accuracy": 0.8687360286712646, "num_tokens": 334644926.0, "step": 8816 }, { "epoch": 1.1216130263325277, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.976720333099365, "learning_rate": 5e-05, "loss": 0.6855, "mean_token_accuracy": 0.8785368204116821, "num_tokens": 334690116.0, "step": 8817 }, { "epoch": 1.1217402366111182, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.9994120597839355, "learning_rate": 5e-05, "loss": 0.7304, "mean_token_accuracy": 0.8870214819908142, "num_tokens": 334728403.0, "step": 8818 }, { "epoch": 1.1218674468897087, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.781925678253174, "learning_rate": 5e-05, "loss": 0.6938, "mean_token_accuracy": 0.8751389980316162, "num_tokens": 334763394.0, "step": 8819 }, { "epoch": 1.1219946571682993, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.2283935546875, "learning_rate": 5e-05, "loss": 0.773, "mean_token_accuracy": 0.8831003904342651, "num_tokens": 334795029.0, "step": 8820 }, { "epoch": 1.1221218674468898, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.578132629394531, "learning_rate": 5e-05, "loss": 0.6895, "mean_token_accuracy": 0.8713397979736328, "num_tokens": 334833298.0, "step": 8821 }, { "epoch": 1.1222490777254803, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.544482231140137, "learning_rate": 5e-05, "loss": 0.7836, "mean_token_accuracy": 0.8845828771591187, "num_tokens": 334865146.0, "step": 8822 }, { "epoch": 1.1223762880040706, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.381539821624756, "learning_rate": 5e-05, "loss": 0.7226, "mean_token_accuracy": 0.8610697984695435, "num_tokens": 334899392.0, "step": 8823 }, { "epoch": 1.1225034982826612, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.856178283691406, "learning_rate": 5e-05, "loss": 0.9117, "mean_token_accuracy": 0.8612315058708191, "num_tokens": 334936501.0, "step": 8824 }, { "epoch": 1.1226307085612517, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.399685382843018, "learning_rate": 5e-05, "loss": 0.7226, "mean_token_accuracy": 0.8669183254241943, "num_tokens": 334975920.0, "step": 8825 }, { "epoch": 1.1227579188398422, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.8850736618042, "learning_rate": 5e-05, "loss": 0.8987, "mean_token_accuracy": 0.8729145526885986, "num_tokens": 335012104.0, "step": 8826 }, { "epoch": 1.1228851291184327, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.8846259117126465, "learning_rate": 5e-05, "loss": 0.7962, "mean_token_accuracy": 0.8596220016479492, "num_tokens": 335051302.0, "step": 8827 }, { "epoch": 1.1230123393970233, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.2616605758667, "learning_rate": 5e-05, "loss": 0.8696, "mean_token_accuracy": 0.8710960149765015, "num_tokens": 335091778.0, "step": 8828 }, { "epoch": 1.1231395496756138, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.665403842926025, "learning_rate": 5e-05, "loss": 0.8778, "mean_token_accuracy": 0.8601812124252319, "num_tokens": 335120687.0, "step": 8829 }, { "epoch": 1.1232667599542043, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.189706802368164, "learning_rate": 5e-05, "loss": 0.7648, "mean_token_accuracy": 0.8711729049682617, "num_tokens": 335158166.0, "step": 8830 }, { "epoch": 1.1233939702327949, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.4417085647583, "learning_rate": 5e-05, "loss": 0.8403, "mean_token_accuracy": 0.8803223371505737, "num_tokens": 335194674.0, "step": 8831 }, { "epoch": 1.1235211805113854, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.401054382324219, "learning_rate": 5e-05, "loss": 0.6682, "mean_token_accuracy": 0.879919171333313, "num_tokens": 335229768.0, "step": 8832 }, { "epoch": 1.123648390789976, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.060551643371582, "learning_rate": 5e-05, "loss": 0.8955, "mean_token_accuracy": 0.8785825967788696, "num_tokens": 335270500.0, "step": 8833 }, { "epoch": 1.1237756010685664, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.210105895996094, "learning_rate": 5e-05, "loss": 0.6603, "mean_token_accuracy": 0.8840198516845703, "num_tokens": 335309672.0, "step": 8834 }, { "epoch": 1.1239028113471567, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.354150772094727, "learning_rate": 5e-05, "loss": 0.9202, "mean_token_accuracy": 0.88468337059021, "num_tokens": 335341505.0, "step": 8835 }, { "epoch": 1.1240300216257473, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.757935047149658, "learning_rate": 5e-05, "loss": 0.8201, "mean_token_accuracy": 0.8595476150512695, "num_tokens": 335379013.0, "step": 8836 }, { "epoch": 1.1241572319043378, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.894125938415527, "learning_rate": 5e-05, "loss": 0.9558, "mean_token_accuracy": 0.865729033946991, "num_tokens": 335415170.0, "step": 8837 }, { "epoch": 1.1242844421829283, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.6957268714904785, "learning_rate": 5e-05, "loss": 0.8875, "mean_token_accuracy": 0.881866455078125, "num_tokens": 335455476.0, "step": 8838 }, { "epoch": 1.1244116524615189, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.7417778968811035, "learning_rate": 5e-05, "loss": 0.8334, "mean_token_accuracy": 0.8682122230529785, "num_tokens": 335497185.0, "step": 8839 }, { "epoch": 1.1245388627401094, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.309290885925293, "learning_rate": 5e-05, "loss": 0.9694, "mean_token_accuracy": 0.8554329872131348, "num_tokens": 335531754.0, "step": 8840 }, { "epoch": 1.1246660730187, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.006277084350586, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.8786784410476685, "num_tokens": 335570486.0, "step": 8841 }, { "epoch": 1.1247932832972904, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.324700355529785, "learning_rate": 5e-05, "loss": 0.8924, "mean_token_accuracy": 0.8634795546531677, "num_tokens": 335610179.0, "step": 8842 }, { "epoch": 1.124920493575881, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.21297025680542, "learning_rate": 5e-05, "loss": 0.8311, "mean_token_accuracy": 0.856884241104126, "num_tokens": 335639785.0, "step": 8843 }, { "epoch": 1.1250477038544715, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00048828125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.639092445373535, "learning_rate": 5e-05, "loss": 0.7853, "mean_token_accuracy": 0.8714674711227417, "num_tokens": 335679012.0, "step": 8844 }, { "epoch": 1.125174914133062, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.821516513824463, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.8845043778419495, "num_tokens": 335717495.0, "step": 8845 }, { "epoch": 1.1253021244116526, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.947517395019531, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.8740509748458862, "num_tokens": 335751617.0, "step": 8846 }, { "epoch": 1.125429334690243, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.3179349899292, "learning_rate": 5e-05, "loss": 0.8662, "mean_token_accuracy": 0.8643398284912109, "num_tokens": 335782698.0, "step": 8847 }, { "epoch": 1.1255565449688334, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.542507648468018, "learning_rate": 5e-05, "loss": 0.7156, "mean_token_accuracy": 0.8656002283096313, "num_tokens": 335823091.0, "step": 8848 }, { "epoch": 1.125683755247424, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.764362335205078, "learning_rate": 5e-05, "loss": 0.884, "mean_token_accuracy": 0.8709907531738281, "num_tokens": 335863318.0, "step": 8849 }, { "epoch": 1.1258109655260145, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.3402299880981445, "learning_rate": 5e-05, "loss": 0.689, "mean_token_accuracy": 0.8751823902130127, "num_tokens": 335900467.0, "step": 8850 }, { "epoch": 1.125938175804605, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.182464599609375, "learning_rate": 5e-05, "loss": 0.97, "mean_token_accuracy": 0.8648488521575928, "num_tokens": 335942274.0, "step": 8851 }, { "epoch": 1.1260653860831955, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.5003180503845215, "learning_rate": 5e-05, "loss": 0.7403, "mean_token_accuracy": 0.8708337545394897, "num_tokens": 335975073.0, "step": 8852 }, { "epoch": 1.126192596361786, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.23742389678955, "learning_rate": 5e-05, "loss": 1.0099, "mean_token_accuracy": 0.8575475215911865, "num_tokens": 336011846.0, "step": 8853 }, { "epoch": 1.1263198066403766, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.132826328277588, "learning_rate": 5e-05, "loss": 0.8584, "mean_token_accuracy": 0.8671007752418518, "num_tokens": 336050969.0, "step": 8854 }, { "epoch": 1.126447016918967, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.720473289489746, "learning_rate": 5e-05, "loss": 0.8988, "mean_token_accuracy": 0.8867067694664001, "num_tokens": 336087714.0, "step": 8855 }, { "epoch": 1.1265742271975576, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.6326117515563965, "learning_rate": 5e-05, "loss": 0.9207, "mean_token_accuracy": 0.8652210831642151, "num_tokens": 336124937.0, "step": 8856 }, { "epoch": 1.1267014374761481, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.11352825164795, "learning_rate": 5e-05, "loss": 0.8416, "mean_token_accuracy": 0.8879565596580505, "num_tokens": 336167666.0, "step": 8857 }, { "epoch": 1.1268286477547387, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.612548828125, "learning_rate": 5e-05, "loss": 0.8236, "mean_token_accuracy": 0.8848944902420044, "num_tokens": 336202170.0, "step": 8858 }, { "epoch": 1.126955858033329, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.04104232788086, "learning_rate": 5e-05, "loss": 0.8245, "mean_token_accuracy": 0.882300078868866, "num_tokens": 336235613.0, "step": 8859 }, { "epoch": 1.1270830683119195, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.188335418701172, "learning_rate": 5e-05, "loss": 0.7938, "mean_token_accuracy": 0.8690816164016724, "num_tokens": 336272027.0, "step": 8860 }, { "epoch": 1.12721027859051, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.38637638092041, "learning_rate": 5e-05, "loss": 0.9349, "mean_token_accuracy": 0.8578115701675415, "num_tokens": 336312554.0, "step": 8861 }, { "epoch": 1.1273374888691006, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.852553367614746, "learning_rate": 5e-05, "loss": 0.7458, "mean_token_accuracy": 0.8736410737037659, "num_tokens": 336356719.0, "step": 8862 }, { "epoch": 1.127464699147691, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.603279113769531, "learning_rate": 5e-05, "loss": 0.8868, "mean_token_accuracy": 0.8801349401473999, "num_tokens": 336395242.0, "step": 8863 }, { "epoch": 1.1275919094262816, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.975643157958984, "learning_rate": 5e-05, "loss": 0.7879, "mean_token_accuracy": 0.8684046864509583, "num_tokens": 336431738.0, "step": 8864 }, { "epoch": 1.1277191197048722, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.319714546203613, "learning_rate": 5e-05, "loss": 0.8984, "mean_token_accuracy": 0.8703576326370239, "num_tokens": 336467799.0, "step": 8865 }, { "epoch": 1.1278463299834627, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.366560459136963, "learning_rate": 5e-05, "loss": 0.8872, "mean_token_accuracy": 0.8525593280792236, "num_tokens": 336506628.0, "step": 8866 }, { "epoch": 1.1279735402620532, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.785281658172607, "learning_rate": 5e-05, "loss": 0.8129, "mean_token_accuracy": 0.880474328994751, "num_tokens": 336546034.0, "step": 8867 }, { "epoch": 1.1281007505406437, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.619395732879639, "learning_rate": 5e-05, "loss": 0.8134, "mean_token_accuracy": 0.8692995309829712, "num_tokens": 336578561.0, "step": 8868 }, { "epoch": 1.1282279608192343, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.48464822769165, "learning_rate": 5e-05, "loss": 0.7391, "mean_token_accuracy": 0.8844581842422485, "num_tokens": 336610500.0, "step": 8869 }, { "epoch": 1.1283551710978248, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.590488433837891, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8825259804725647, "num_tokens": 336653889.0, "step": 8870 }, { "epoch": 1.1284823813764153, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.485278606414795, "learning_rate": 5e-05, "loss": 0.7779, "mean_token_accuracy": 0.870303750038147, "num_tokens": 336690605.0, "step": 8871 }, { "epoch": 1.1286095916550056, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.423061847686768, "learning_rate": 5e-05, "loss": 0.7764, "mean_token_accuracy": 0.8660516738891602, "num_tokens": 336729159.0, "step": 8872 }, { "epoch": 1.1287368019335962, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.5730977058410645, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.8783057928085327, "num_tokens": 336762376.0, "step": 8873 }, { "epoch": 1.1288640122121867, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.297750473022461, "learning_rate": 5e-05, "loss": 0.7216, "mean_token_accuracy": 0.8796383142471313, "num_tokens": 336797060.0, "step": 8874 }, { "epoch": 1.1289912224907772, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.618382453918457, "learning_rate": 5e-05, "loss": 0.8015, "mean_token_accuracy": 0.8550309538841248, "num_tokens": 336838686.0, "step": 8875 }, { "epoch": 1.1291184327693677, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.244240760803223, "learning_rate": 5e-05, "loss": 0.7768, "mean_token_accuracy": 0.8569878339767456, "num_tokens": 336887535.0, "step": 8876 }, { "epoch": 1.1292456430479583, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.612018585205078, "learning_rate": 5e-05, "loss": 0.722, "mean_token_accuracy": 0.879550576210022, "num_tokens": 336920167.0, "step": 8877 }, { "epoch": 1.1293728533265488, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.223781108856201, "learning_rate": 5e-05, "loss": 0.7087, "mean_token_accuracy": 0.8732620477676392, "num_tokens": 336953925.0, "step": 8878 }, { "epoch": 1.1295000636051393, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.616589069366455, "learning_rate": 5e-05, "loss": 0.7203, "mean_token_accuracy": 0.8790845274925232, "num_tokens": 336992407.0, "step": 8879 }, { "epoch": 1.1296272738837299, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.178627014160156, "learning_rate": 5e-05, "loss": 0.7462, "mean_token_accuracy": 0.8611375093460083, "num_tokens": 337029495.0, "step": 8880 }, { "epoch": 1.1297544841623204, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.6671624183654785, "learning_rate": 5e-05, "loss": 0.7424, "mean_token_accuracy": 0.870328426361084, "num_tokens": 337067533.0, "step": 8881 }, { "epoch": 1.129881694440911, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.069614410400391, "learning_rate": 5e-05, "loss": 0.743, "mean_token_accuracy": 0.860471785068512, "num_tokens": 337105112.0, "step": 8882 }, { "epoch": 1.1300089047195012, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.7918381690979, "learning_rate": 5e-05, "loss": 0.7771, "mean_token_accuracy": 0.8620548248291016, "num_tokens": 337140807.0, "step": 8883 }, { "epoch": 1.1301361149980917, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.884875297546387, "learning_rate": 5e-05, "loss": 0.6307, "mean_token_accuracy": 0.8923244476318359, "num_tokens": 337182644.0, "step": 8884 }, { "epoch": 1.1302633252766823, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.063138961791992, "learning_rate": 5e-05, "loss": 0.7597, "mean_token_accuracy": 0.8755483627319336, "num_tokens": 337226868.0, "step": 8885 }, { "epoch": 1.1303905355552728, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.60430908203125, "learning_rate": 5e-05, "loss": 0.7157, "mean_token_accuracy": 0.8611681461334229, "num_tokens": 337262118.0, "step": 8886 }, { "epoch": 1.1305177458338633, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.461421966552734, "learning_rate": 5e-05, "loss": 0.8593, "mean_token_accuracy": 0.8541771173477173, "num_tokens": 337299575.0, "step": 8887 }, { "epoch": 1.1306449561124539, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.251527309417725, "learning_rate": 5e-05, "loss": 0.6497, "mean_token_accuracy": 0.8780730962753296, "num_tokens": 337335178.0, "step": 8888 }, { "epoch": 1.1307721663910444, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 9.012858390808105, "learning_rate": 5e-05, "loss": 0.9501, "mean_token_accuracy": 0.8560216426849365, "num_tokens": 337368982.0, "step": 8889 }, { "epoch": 1.130899376669635, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.049973487854004, "learning_rate": 5e-05, "loss": 0.6962, "mean_token_accuracy": 0.8670459985733032, "num_tokens": 337403708.0, "step": 8890 }, { "epoch": 1.1310265869482254, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.382973670959473, "learning_rate": 5e-05, "loss": 0.9513, "mean_token_accuracy": 0.8760557770729065, "num_tokens": 337441265.0, "step": 8891 }, { "epoch": 1.131153797226816, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.612655162811279, "learning_rate": 5e-05, "loss": 0.7452, "mean_token_accuracy": 0.8756881952285767, "num_tokens": 337475600.0, "step": 8892 }, { "epoch": 1.1312810075054065, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.768336296081543, "learning_rate": 5e-05, "loss": 0.9009, "mean_token_accuracy": 0.8836348056793213, "num_tokens": 337513051.0, "step": 8893 }, { "epoch": 1.131408217783997, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.852635383605957, "learning_rate": 5e-05, "loss": 0.9609, "mean_token_accuracy": 0.8612865209579468, "num_tokens": 337552187.0, "step": 8894 }, { "epoch": 1.1315354280625876, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.9958391189575195, "learning_rate": 5e-05, "loss": 0.6925, "mean_token_accuracy": 0.885695219039917, "num_tokens": 337586877.0, "step": 8895 }, { "epoch": 1.131662638341178, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.378682136535645, "learning_rate": 5e-05, "loss": 0.9639, "mean_token_accuracy": 0.87321537733078, "num_tokens": 337627979.0, "step": 8896 }, { "epoch": 1.1317898486197684, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.609622955322266, "learning_rate": 5e-05, "loss": 0.6583, "mean_token_accuracy": 0.8741928935050964, "num_tokens": 337670064.0, "step": 8897 }, { "epoch": 1.131917058898359, "ewc_loss": 0.8125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 11.31735897064209, "learning_rate": 5e-05, "loss": 1.2731, "mean_token_accuracy": 0.847907304763794, "num_tokens": 337709670.0, "step": 8898 }, { "epoch": 1.1320442691769494, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.063615322113037, "learning_rate": 5e-05, "loss": 0.9177, "mean_token_accuracy": 0.8681989908218384, "num_tokens": 337747507.0, "step": 8899 }, { "epoch": 1.13217147945554, "ewc_loss": 0.92578125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000873565673828125, "grad_norm": 12.012198448181152, "learning_rate": 5e-05, "loss": 1.3027, "mean_token_accuracy": 0.8708215951919556, "num_tokens": 337789542.0, "step": 8900 }, { "epoch": 1.1322986897341305, "ewc_loss": 0.984375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000934600830078125, "grad_norm": 9.843401908874512, "learning_rate": 5e-05, "loss": 1.3057, "mean_token_accuracy": 0.8842220306396484, "num_tokens": 337832100.0, "step": 8901 }, { "epoch": 1.132425900012721, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.472126007080078, "learning_rate": 5e-05, "loss": 1.0208, "mean_token_accuracy": 0.8743435740470886, "num_tokens": 337869395.0, "step": 8902 }, { "epoch": 1.1325531102913116, "ewc_loss": 0.796875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 9.47024917602539, "learning_rate": 5e-05, "loss": 1.2028, "mean_token_accuracy": 0.8590937852859497, "num_tokens": 337909644.0, "step": 8903 }, { "epoch": 1.132680320569902, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.450677871704102, "learning_rate": 5e-05, "loss": 1.051, "mean_token_accuracy": 0.8598574995994568, "num_tokens": 337947255.0, "step": 8904 }, { "epoch": 1.1328075308484926, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.053179740905762, "learning_rate": 5e-05, "loss": 0.9673, "mean_token_accuracy": 0.8851552605628967, "num_tokens": 337986377.0, "step": 8905 }, { "epoch": 1.1329347411270831, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.52538013458252, "learning_rate": 5e-05, "loss": 0.9321, "mean_token_accuracy": 0.8794790506362915, "num_tokens": 338025234.0, "step": 8906 }, { "epoch": 1.1330619514056737, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.642972946166992, "learning_rate": 5e-05, "loss": 0.9698, "mean_token_accuracy": 0.861137866973877, "num_tokens": 338069070.0, "step": 8907 }, { "epoch": 1.133189161684264, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.037725448608398, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.876907229423523, "num_tokens": 338111529.0, "step": 8908 }, { "epoch": 1.1333163719628545, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.25708293914795, "learning_rate": 5e-05, "loss": 0.9131, "mean_token_accuracy": 0.8609786033630371, "num_tokens": 338155881.0, "step": 8909 }, { "epoch": 1.133443582241445, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.685024738311768, "learning_rate": 5e-05, "loss": 0.8343, "mean_token_accuracy": 0.8657107353210449, "num_tokens": 338192570.0, "step": 8910 }, { "epoch": 1.1335707925200356, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.111742973327637, "learning_rate": 5e-05, "loss": 0.8469, "mean_token_accuracy": 0.869916558265686, "num_tokens": 338225088.0, "step": 8911 }, { "epoch": 1.133698002798626, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.302483081817627, "learning_rate": 5e-05, "loss": 0.8062, "mean_token_accuracy": 0.8589338660240173, "num_tokens": 338267301.0, "step": 8912 }, { "epoch": 1.1338252130772166, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.320375442504883, "learning_rate": 5e-05, "loss": 0.8235, "mean_token_accuracy": 0.8780754804611206, "num_tokens": 338305130.0, "step": 8913 }, { "epoch": 1.1339524233558071, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.839836120605469, "learning_rate": 5e-05, "loss": 0.7592, "mean_token_accuracy": 0.8562272191047668, "num_tokens": 338347556.0, "step": 8914 }, { "epoch": 1.1340796336343977, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.797054290771484, "learning_rate": 5e-05, "loss": 0.8973, "mean_token_accuracy": 0.8711093068122864, "num_tokens": 338384994.0, "step": 8915 }, { "epoch": 1.1342068439129882, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.461404800415039, "learning_rate": 5e-05, "loss": 0.7385, "mean_token_accuracy": 0.864107608795166, "num_tokens": 338427311.0, "step": 8916 }, { "epoch": 1.1343340541915787, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.419185638427734, "learning_rate": 5e-05, "loss": 0.9255, "mean_token_accuracy": 0.8812531232833862, "num_tokens": 338464091.0, "step": 8917 }, { "epoch": 1.1344612644701693, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.515780448913574, "learning_rate": 5e-05, "loss": 0.7461, "mean_token_accuracy": 0.8731194734573364, "num_tokens": 338507961.0, "step": 8918 }, { "epoch": 1.1345884747487598, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.92870044708252, "learning_rate": 5e-05, "loss": 1.0226, "mean_token_accuracy": 0.8743807673454285, "num_tokens": 338544028.0, "step": 8919 }, { "epoch": 1.1347156850273503, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.463286399841309, "learning_rate": 5e-05, "loss": 0.984, "mean_token_accuracy": 0.8573651313781738, "num_tokens": 338587827.0, "step": 8920 }, { "epoch": 1.1348428953059406, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.411681175231934, "learning_rate": 5e-05, "loss": 0.9967, "mean_token_accuracy": 0.8769069910049438, "num_tokens": 338627729.0, "step": 8921 }, { "epoch": 1.1349701055845312, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.374128341674805, "learning_rate": 5e-05, "loss": 0.952, "mean_token_accuracy": 0.8859541416168213, "num_tokens": 338663805.0, "step": 8922 }, { "epoch": 1.1350973158631217, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.993989944458008, "learning_rate": 5e-05, "loss": 0.863, "mean_token_accuracy": 0.8717319369316101, "num_tokens": 338702204.0, "step": 8923 }, { "epoch": 1.1352245261417122, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.828990936279297, "learning_rate": 5e-05, "loss": 0.976, "mean_token_accuracy": 0.8759890794754028, "num_tokens": 338748769.0, "step": 8924 }, { "epoch": 1.1353517364203027, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.988062858581543, "learning_rate": 5e-05, "loss": 0.7423, "mean_token_accuracy": 0.8801735043525696, "num_tokens": 338793617.0, "step": 8925 }, { "epoch": 1.1354789466988933, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.184444427490234, "learning_rate": 5e-05, "loss": 0.9696, "mean_token_accuracy": 0.8739616870880127, "num_tokens": 338827893.0, "step": 8926 }, { "epoch": 1.1356061569774838, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 6.936920166015625, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8775608539581299, "num_tokens": 338866036.0, "step": 8927 }, { "epoch": 1.1357333672560743, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.179767608642578, "learning_rate": 5e-05, "loss": 0.9961, "mean_token_accuracy": 0.8655806183815002, "num_tokens": 338900595.0, "step": 8928 }, { "epoch": 1.1358605775346649, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.391678333282471, "learning_rate": 5e-05, "loss": 0.8061, "mean_token_accuracy": 0.8882954120635986, "num_tokens": 338930268.0, "step": 8929 }, { "epoch": 1.1359877878132554, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.727349281311035, "learning_rate": 5e-05, "loss": 0.9362, "mean_token_accuracy": 0.8746452927589417, "num_tokens": 338974633.0, "step": 8930 }, { "epoch": 1.136114998091846, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.723209381103516, "learning_rate": 5e-05, "loss": 0.923, "mean_token_accuracy": 0.8628979921340942, "num_tokens": 339008884.0, "step": 8931 }, { "epoch": 1.1362422083704362, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.16926097869873, "learning_rate": 5e-05, "loss": 0.9267, "mean_token_accuracy": 0.848564624786377, "num_tokens": 339043923.0, "step": 8932 }, { "epoch": 1.1363694186490267, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.720686912536621, "learning_rate": 5e-05, "loss": 0.8458, "mean_token_accuracy": 0.8787491321563721, "num_tokens": 339080545.0, "step": 8933 }, { "epoch": 1.1364966289276173, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.841796875, "learning_rate": 5e-05, "loss": 0.8208, "mean_token_accuracy": 0.8727966547012329, "num_tokens": 339114243.0, "step": 8934 }, { "epoch": 1.1366238392062078, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.5764875411987305, "learning_rate": 5e-05, "loss": 0.8104, "mean_token_accuracy": 0.8797858953475952, "num_tokens": 339147609.0, "step": 8935 }, { "epoch": 1.1367510494847983, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.7354817390441895, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8785310387611389, "num_tokens": 339186295.0, "step": 8936 }, { "epoch": 1.1368782597633889, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.44790506362915, "learning_rate": 5e-05, "loss": 0.7609, "mean_token_accuracy": 0.8773184418678284, "num_tokens": 339217572.0, "step": 8937 }, { "epoch": 1.1370054700419794, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.66265344619751, "learning_rate": 5e-05, "loss": 0.769, "mean_token_accuracy": 0.874608039855957, "num_tokens": 339257305.0, "step": 8938 }, { "epoch": 1.13713268032057, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.399826526641846, "learning_rate": 5e-05, "loss": 0.7664, "mean_token_accuracy": 0.8660794496536255, "num_tokens": 339295117.0, "step": 8939 }, { "epoch": 1.1372598905991604, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.5936970710754395, "learning_rate": 5e-05, "loss": 0.7524, "mean_token_accuracy": 0.8740631341934204, "num_tokens": 339330543.0, "step": 8940 }, { "epoch": 1.137387100877751, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.383553504943848, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8727766871452332, "num_tokens": 339375040.0, "step": 8941 }, { "epoch": 1.1375143111563415, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.563918113708496, "learning_rate": 5e-05, "loss": 0.752, "mean_token_accuracy": 0.8663647174835205, "num_tokens": 339410862.0, "step": 8942 }, { "epoch": 1.137641521434932, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.348167896270752, "learning_rate": 5e-05, "loss": 0.7131, "mean_token_accuracy": 0.8762593269348145, "num_tokens": 339441929.0, "step": 8943 }, { "epoch": 1.1377687317135226, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.5433526039123535, "learning_rate": 5e-05, "loss": 0.7321, "mean_token_accuracy": 0.8757053017616272, "num_tokens": 339486156.0, "step": 8944 }, { "epoch": 1.137895941992113, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.321285247802734, "learning_rate": 5e-05, "loss": 0.7532, "mean_token_accuracy": 0.8598672151565552, "num_tokens": 339531676.0, "step": 8945 }, { "epoch": 1.1380231522707034, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.525440216064453, "learning_rate": 5e-05, "loss": 0.7572, "mean_token_accuracy": 0.8630341291427612, "num_tokens": 339571003.0, "step": 8946 }, { "epoch": 1.138150362549294, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000492095947265625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.330276966094971, "learning_rate": 5e-05, "loss": 0.7157, "mean_token_accuracy": 0.8668932318687439, "num_tokens": 339604389.0, "step": 8947 }, { "epoch": 1.1382775728278844, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.500247955322266, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8600717782974243, "num_tokens": 339649267.0, "step": 8948 }, { "epoch": 1.138404783106475, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.304328918457031, "learning_rate": 5e-05, "loss": 0.7293, "mean_token_accuracy": 0.8637546300888062, "num_tokens": 339685564.0, "step": 8949 }, { "epoch": 1.1385319933850655, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.536940574645996, "learning_rate": 5e-05, "loss": 0.6814, "mean_token_accuracy": 0.8868973255157471, "num_tokens": 339716312.0, "step": 8950 }, { "epoch": 1.138659203663656, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.221587181091309, "learning_rate": 5e-05, "loss": 0.7121, "mean_token_accuracy": 0.8717867136001587, "num_tokens": 339760666.0, "step": 8951 }, { "epoch": 1.1387864139422466, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.629894733428955, "learning_rate": 5e-05, "loss": 0.7005, "mean_token_accuracy": 0.8808516263961792, "num_tokens": 339799432.0, "step": 8952 }, { "epoch": 1.138913624220837, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.080889701843262, "learning_rate": 5e-05, "loss": 0.7097, "mean_token_accuracy": 0.8679600358009338, "num_tokens": 339839787.0, "step": 8953 }, { "epoch": 1.1390408344994276, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.840601921081543, "learning_rate": 5e-05, "loss": 0.7299, "mean_token_accuracy": 0.8748281598091125, "num_tokens": 339876193.0, "step": 8954 }, { "epoch": 1.1391680447780181, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.800014019012451, "learning_rate": 5e-05, "loss": 0.6816, "mean_token_accuracy": 0.8703837394714355, "num_tokens": 339919791.0, "step": 8955 }, { "epoch": 1.1392952550566087, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.263816833496094, "learning_rate": 5e-05, "loss": 0.792, "mean_token_accuracy": 0.8650915622711182, "num_tokens": 339954432.0, "step": 8956 }, { "epoch": 1.139422465335199, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00022983551025390625, "grad_norm": 6.266703128814697, "learning_rate": 5e-05, "loss": 0.6426, "mean_token_accuracy": 0.8767778873443604, "num_tokens": 339996327.0, "step": 8957 }, { "epoch": 1.1395496756137895, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 9.106399536132812, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.8995507955551147, "num_tokens": 340031604.0, "step": 8958 }, { "epoch": 1.13967688589238, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00020885467529296875, "grad_norm": 5.546255588531494, "learning_rate": 5e-05, "loss": 0.6189, "mean_token_accuracy": 0.8725457191467285, "num_tokens": 340072433.0, "step": 8959 }, { "epoch": 1.1398040961709706, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.492107391357422, "learning_rate": 5e-05, "loss": 1.0328, "mean_token_accuracy": 0.8728348016738892, "num_tokens": 340114306.0, "step": 8960 }, { "epoch": 1.139931306449561, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 5.767152309417725, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.8511897325515747, "num_tokens": 340152826.0, "step": 8961 }, { "epoch": 1.1400585167281516, "ewc_loss": 0.921875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0008697509765625, "grad_norm": 11.772579193115234, "learning_rate": 5e-05, "loss": 1.3207, "mean_token_accuracy": 0.8646020293235779, "num_tokens": 340191110.0, "step": 8962 }, { "epoch": 1.1401857270067421, "ewc_loss": 0.8125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 8.733189582824707, "learning_rate": 5e-05, "loss": 1.1811, "mean_token_accuracy": 0.872097373008728, "num_tokens": 340230606.0, "step": 8963 }, { "epoch": 1.1403129372853327, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.523224830627441, "learning_rate": 5e-05, "loss": 1.0606, "mean_token_accuracy": 0.864335298538208, "num_tokens": 340274831.0, "step": 8964 }, { "epoch": 1.1404401475639232, "ewc_loss": 0.90625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 9.62069034576416, "learning_rate": 5e-05, "loss": 1.2556, "mean_token_accuracy": 0.8751075863838196, "num_tokens": 340308245.0, "step": 8965 }, { "epoch": 1.1405673578425137, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.783984184265137, "learning_rate": 5e-05, "loss": 0.8664, "mean_token_accuracy": 0.8918540477752686, "num_tokens": 340343091.0, "step": 8966 }, { "epoch": 1.1406945681211043, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.192365646362305, "learning_rate": 5e-05, "loss": 1.1495, "mean_token_accuracy": 0.8748339414596558, "num_tokens": 340382968.0, "step": 8967 }, { "epoch": 1.1408217783996948, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.982297897338867, "learning_rate": 5e-05, "loss": 0.9183, "mean_token_accuracy": 0.8809577226638794, "num_tokens": 340421740.0, "step": 8968 }, { "epoch": 1.1409489886782853, "ewc_loss": 0.75, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.737351417541504, "learning_rate": 5e-05, "loss": 1.1296, "mean_token_accuracy": 0.8679829239845276, "num_tokens": 340460156.0, "step": 8969 }, { "epoch": 1.1410761989568756, "ewc_loss": 0.703125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 8.63025188446045, "learning_rate": 5e-05, "loss": 1.0755, "mean_token_accuracy": 0.8735661506652832, "num_tokens": 340496106.0, "step": 8970 }, { "epoch": 1.1412034092354661, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 10.063311576843262, "learning_rate": 5e-05, "loss": 0.9994, "mean_token_accuracy": 0.8752848505973816, "num_tokens": 340530338.0, "step": 8971 }, { "epoch": 1.1413306195140567, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 8.723286628723145, "learning_rate": 5e-05, "loss": 1.0368, "mean_token_accuracy": 0.8762373328208923, "num_tokens": 340565590.0, "step": 8972 }, { "epoch": 1.1414578297926472, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.98769474029541, "learning_rate": 5e-05, "loss": 0.9149, "mean_token_accuracy": 0.8761483430862427, "num_tokens": 340609651.0, "step": 8973 }, { "epoch": 1.1415850400712377, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.674521446228027, "learning_rate": 5e-05, "loss": 0.9803, "mean_token_accuracy": 0.8729293942451477, "num_tokens": 340650119.0, "step": 8974 }, { "epoch": 1.1417122503498283, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.923312187194824, "learning_rate": 5e-05, "loss": 0.8176, "mean_token_accuracy": 0.8739013671875, "num_tokens": 340686734.0, "step": 8975 }, { "epoch": 1.1418394606284188, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.064863204956055, "learning_rate": 5e-05, "loss": 0.9356, "mean_token_accuracy": 0.8803351521492004, "num_tokens": 340722405.0, "step": 8976 }, { "epoch": 1.1419666709070093, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.192306041717529, "learning_rate": 5e-05, "loss": 0.7712, "mean_token_accuracy": 0.8820526599884033, "num_tokens": 340763402.0, "step": 8977 }, { "epoch": 1.1420938811855998, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.622200012207031, "learning_rate": 5e-05, "loss": 0.9116, "mean_token_accuracy": 0.8987235426902771, "num_tokens": 340799410.0, "step": 8978 }, { "epoch": 1.1422210914641904, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.257905960083008, "learning_rate": 5e-05, "loss": 0.8742, "mean_token_accuracy": 0.8632020950317383, "num_tokens": 340832970.0, "step": 8979 }, { "epoch": 1.142348301742781, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.686769485473633, "learning_rate": 5e-05, "loss": 0.9651, "mean_token_accuracy": 0.8798413872718811, "num_tokens": 340872500.0, "step": 8980 }, { "epoch": 1.1424755120213712, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 7.751911163330078, "learning_rate": 5e-05, "loss": 0.9603, "mean_token_accuracy": 0.8676585555076599, "num_tokens": 340913784.0, "step": 8981 }, { "epoch": 1.1426027222999617, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.043527603149414, "learning_rate": 5e-05, "loss": 0.881, "mean_token_accuracy": 0.8847757577896118, "num_tokens": 340945660.0, "step": 8982 }, { "epoch": 1.1427299325785523, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.093050956726074, "learning_rate": 5e-05, "loss": 0.9679, "mean_token_accuracy": 0.8717812299728394, "num_tokens": 340979528.0, "step": 8983 }, { "epoch": 1.1428571428571428, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.019743919372559, "learning_rate": 5e-05, "loss": 0.8011, "mean_token_accuracy": 0.8705340623855591, "num_tokens": 341016256.0, "step": 8984 }, { "epoch": 1.1429843531357333, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.468301773071289, "learning_rate": 5e-05, "loss": 0.9621, "mean_token_accuracy": 0.8649309873580933, "num_tokens": 341056339.0, "step": 8985 }, { "epoch": 1.1431115634143239, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.104295253753662, "learning_rate": 5e-05, "loss": 0.7191, "mean_token_accuracy": 0.8811652660369873, "num_tokens": 341093060.0, "step": 8986 }, { "epoch": 1.1432387736929144, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.21279239654541, "learning_rate": 5e-05, "loss": 1.0015, "mean_token_accuracy": 0.8664140701293945, "num_tokens": 341128634.0, "step": 8987 }, { "epoch": 1.143365983971505, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.83435583114624, "learning_rate": 5e-05, "loss": 0.8023, "mean_token_accuracy": 0.8568649291992188, "num_tokens": 341160649.0, "step": 8988 }, { "epoch": 1.1434931942500954, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.85989761352539, "learning_rate": 5e-05, "loss": 1.0105, "mean_token_accuracy": 0.8722368478775024, "num_tokens": 341196660.0, "step": 8989 }, { "epoch": 1.143620404528686, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 7.594663143157959, "learning_rate": 5e-05, "loss": 0.934, "mean_token_accuracy": 0.8630080223083496, "num_tokens": 341236704.0, "step": 8990 }, { "epoch": 1.1437476148072765, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.515254974365234, "learning_rate": 5e-05, "loss": 0.985, "mean_token_accuracy": 0.869051992893219, "num_tokens": 341272874.0, "step": 8991 }, { "epoch": 1.143874825085867, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.29694652557373, "learning_rate": 5e-05, "loss": 0.994, "mean_token_accuracy": 0.8668006658554077, "num_tokens": 341310209.0, "step": 8992 }, { "epoch": 1.1440020353644575, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.511003494262695, "learning_rate": 5e-05, "loss": 0.812, "mean_token_accuracy": 0.8859373331069946, "num_tokens": 341341717.0, "step": 8993 }, { "epoch": 1.144129245643048, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.262019157409668, "learning_rate": 5e-05, "loss": 0.8774, "mean_token_accuracy": 0.8895412087440491, "num_tokens": 341374298.0, "step": 8994 }, { "epoch": 1.1442564559216384, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.9730730056762695, "learning_rate": 5e-05, "loss": 0.8132, "mean_token_accuracy": 0.8717097043991089, "num_tokens": 341416214.0, "step": 8995 }, { "epoch": 1.144383666200229, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.8024678230285645, "learning_rate": 5e-05, "loss": 0.8307, "mean_token_accuracy": 0.8718600273132324, "num_tokens": 341457419.0, "step": 8996 }, { "epoch": 1.1445108764788194, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.029041290283203, "learning_rate": 5e-05, "loss": 0.7902, "mean_token_accuracy": 0.8800031542778015, "num_tokens": 341493251.0, "step": 8997 }, { "epoch": 1.14463808675741, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.2316575050354, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8582258224487305, "num_tokens": 341531501.0, "step": 8998 }, { "epoch": 1.1447652970360005, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.204309463500977, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8874010443687439, "num_tokens": 341572114.0, "step": 8999 }, { "epoch": 1.144892507314591, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.967503547668457, "learning_rate": 5e-05, "loss": 0.7361, "mean_token_accuracy": 0.8663235902786255, "num_tokens": 341617446.0, "step": 9000 }, { "epoch": 1.1450197175931816, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.35933780670166, "learning_rate": 5e-05, "loss": 0.8234, "mean_token_accuracy": 0.8710698485374451, "num_tokens": 341653636.0, "step": 9001 }, { "epoch": 1.145146927871772, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.818897724151611, "learning_rate": 5e-05, "loss": 0.8267, "mean_token_accuracy": 0.8389069437980652, "num_tokens": 341692174.0, "step": 9002 }, { "epoch": 1.1452741381503626, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.594015121459961, "learning_rate": 5e-05, "loss": 0.8573, "mean_token_accuracy": 0.8648784160614014, "num_tokens": 341737215.0, "step": 9003 }, { "epoch": 1.1454013484289531, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.729108810424805, "learning_rate": 5e-05, "loss": 0.7216, "mean_token_accuracy": 0.8693199157714844, "num_tokens": 341778571.0, "step": 9004 }, { "epoch": 1.1455285587075437, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.733049392700195, "learning_rate": 5e-05, "loss": 0.8989, "mean_token_accuracy": 0.8604352474212646, "num_tokens": 341814725.0, "step": 9005 }, { "epoch": 1.145655768986134, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.978343486785889, "learning_rate": 5e-05, "loss": 0.6917, "mean_token_accuracy": 0.8846002817153931, "num_tokens": 341854982.0, "step": 9006 }, { "epoch": 1.1457829792647245, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.301097869873047, "learning_rate": 5e-05, "loss": 0.8195, "mean_token_accuracy": 0.880394458770752, "num_tokens": 341888952.0, "step": 9007 }, { "epoch": 1.145910189543315, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.558037281036377, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8771819472312927, "num_tokens": 341927985.0, "step": 9008 }, { "epoch": 1.1460373998219056, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.431674957275391, "learning_rate": 5e-05, "loss": 0.7403, "mean_token_accuracy": 0.8727987408638, "num_tokens": 341959813.0, "step": 9009 }, { "epoch": 1.146164610100496, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.172813415527344, "learning_rate": 5e-05, "loss": 0.8059, "mean_token_accuracy": 0.8766260147094727, "num_tokens": 341997231.0, "step": 9010 }, { "epoch": 1.1462918203790866, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.7334113121032715, "learning_rate": 5e-05, "loss": 0.6643, "mean_token_accuracy": 0.8832494020462036, "num_tokens": 342029277.0, "step": 9011 }, { "epoch": 1.1464190306576771, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.785581588745117, "learning_rate": 5e-05, "loss": 0.8378, "mean_token_accuracy": 0.8820303678512573, "num_tokens": 342069297.0, "step": 9012 }, { "epoch": 1.1465462409362677, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.397877216339111, "learning_rate": 5e-05, "loss": 0.6973, "mean_token_accuracy": 0.8700217008590698, "num_tokens": 342111525.0, "step": 9013 }, { "epoch": 1.1466734512148582, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.255941390991211, "learning_rate": 5e-05, "loss": 0.923, "mean_token_accuracy": 0.872367262840271, "num_tokens": 342148445.0, "step": 9014 }, { "epoch": 1.1468006614934487, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.697737693786621, "learning_rate": 5e-05, "loss": 0.7235, "mean_token_accuracy": 0.8729574680328369, "num_tokens": 342186855.0, "step": 9015 }, { "epoch": 1.1469278717720393, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.011149406433105, "learning_rate": 5e-05, "loss": 0.9224, "mean_token_accuracy": 0.8707853555679321, "num_tokens": 342224619.0, "step": 9016 }, { "epoch": 1.1470550820506298, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.483962535858154, "learning_rate": 5e-05, "loss": 0.783, "mean_token_accuracy": 0.8822149634361267, "num_tokens": 342261775.0, "step": 9017 }, { "epoch": 1.1471822923292203, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.156839370727539, "learning_rate": 5e-05, "loss": 0.8112, "mean_token_accuracy": 0.8764107823371887, "num_tokens": 342296355.0, "step": 9018 }, { "epoch": 1.1473095026078106, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.880100727081299, "learning_rate": 5e-05, "loss": 0.8882, "mean_token_accuracy": 0.8544023036956787, "num_tokens": 342337363.0, "step": 9019 }, { "epoch": 1.1474367128864011, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.70798921585083, "learning_rate": 5e-05, "loss": 0.7806, "mean_token_accuracy": 0.8732017874717712, "num_tokens": 342376462.0, "step": 9020 }, { "epoch": 1.1475639231649917, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.508280277252197, "learning_rate": 5e-05, "loss": 0.7513, "mean_token_accuracy": 0.8759326934814453, "num_tokens": 342417341.0, "step": 9021 }, { "epoch": 1.1476911334435822, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.174321174621582, "learning_rate": 5e-05, "loss": 0.7827, "mean_token_accuracy": 0.8836499452590942, "num_tokens": 342455489.0, "step": 9022 }, { "epoch": 1.1478183437221727, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.650909423828125, "learning_rate": 5e-05, "loss": 0.6709, "mean_token_accuracy": 0.8803925514221191, "num_tokens": 342489594.0, "step": 9023 }, { "epoch": 1.1479455540007633, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 9.111075401306152, "learning_rate": 5e-05, "loss": 0.8997, "mean_token_accuracy": 0.8648576736450195, "num_tokens": 342525473.0, "step": 9024 }, { "epoch": 1.1480727642793538, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.194363117218018, "learning_rate": 5e-05, "loss": 0.6495, "mean_token_accuracy": 0.8861639499664307, "num_tokens": 342557472.0, "step": 9025 }, { "epoch": 1.1481999745579443, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.687204360961914, "learning_rate": 5e-05, "loss": 0.9628, "mean_token_accuracy": 0.8778179883956909, "num_tokens": 342593386.0, "step": 9026 }, { "epoch": 1.1483271848365348, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.016498565673828, "learning_rate": 5e-05, "loss": 0.8129, "mean_token_accuracy": 0.8703444600105286, "num_tokens": 342631370.0, "step": 9027 }, { "epoch": 1.1484543951151254, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.012749671936035, "learning_rate": 5e-05, "loss": 0.9397, "mean_token_accuracy": 0.8730531334877014, "num_tokens": 342675570.0, "step": 9028 }, { "epoch": 1.148581605393716, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.195707321166992, "learning_rate": 5e-05, "loss": 0.9276, "mean_token_accuracy": 0.8676680326461792, "num_tokens": 342717519.0, "step": 9029 }, { "epoch": 1.1487088156723062, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.469442844390869, "learning_rate": 5e-05, "loss": 0.7311, "mean_token_accuracy": 0.886163592338562, "num_tokens": 342754104.0, "step": 9030 }, { "epoch": 1.1488360259508967, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.89136791229248, "learning_rate": 5e-05, "loss": 0.9789, "mean_token_accuracy": 0.8597450256347656, "num_tokens": 342796059.0, "step": 9031 }, { "epoch": 1.1489632362294873, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.637269496917725, "learning_rate": 5e-05, "loss": 0.7307, "mean_token_accuracy": 0.8672646284103394, "num_tokens": 342833731.0, "step": 9032 }, { "epoch": 1.1490904465080778, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.25617504119873, "learning_rate": 5e-05, "loss": 0.9133, "mean_token_accuracy": 0.8827031254768372, "num_tokens": 342871918.0, "step": 9033 }, { "epoch": 1.1492176567866683, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.792804718017578, "learning_rate": 5e-05, "loss": 0.78, "mean_token_accuracy": 0.8572949767112732, "num_tokens": 342903129.0, "step": 9034 }, { "epoch": 1.1493448670652588, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.0794038772583, "learning_rate": 5e-05, "loss": 0.9735, "mean_token_accuracy": 0.8642709851264954, "num_tokens": 342941503.0, "step": 9035 }, { "epoch": 1.1494720773438494, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.303179740905762, "learning_rate": 5e-05, "loss": 0.8279, "mean_token_accuracy": 0.8673276901245117, "num_tokens": 342977766.0, "step": 9036 }, { "epoch": 1.14959928762244, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.622300148010254, "learning_rate": 5e-05, "loss": 0.8986, "mean_token_accuracy": 0.8752576112747192, "num_tokens": 343019865.0, "step": 9037 }, { "epoch": 1.1497264979010304, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.543306350708008, "learning_rate": 5e-05, "loss": 0.8711, "mean_token_accuracy": 0.862622857093811, "num_tokens": 343056500.0, "step": 9038 }, { "epoch": 1.149853708179621, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.251718521118164, "learning_rate": 5e-05, "loss": 0.9017, "mean_token_accuracy": 0.8598313331604004, "num_tokens": 343098404.0, "step": 9039 }, { "epoch": 1.1499809184582115, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.499055862426758, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8869882225990295, "num_tokens": 343137360.0, "step": 9040 }, { "epoch": 1.150108128736802, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.052611351013184, "learning_rate": 5e-05, "loss": 0.8018, "mean_token_accuracy": 0.8819493651390076, "num_tokens": 343176038.0, "step": 9041 }, { "epoch": 1.1502353390153925, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.3604230880737305, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.8760685920715332, "num_tokens": 343217888.0, "step": 9042 }, { "epoch": 1.150362549293983, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.98173713684082, "learning_rate": 5e-05, "loss": 0.7999, "mean_token_accuracy": 0.8769327402114868, "num_tokens": 343257217.0, "step": 9043 }, { "epoch": 1.1504897595725734, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.264523029327393, "learning_rate": 5e-05, "loss": 0.7281, "mean_token_accuracy": 0.8780808448791504, "num_tokens": 343290607.0, "step": 9044 }, { "epoch": 1.150616969851164, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.9528422355651855, "learning_rate": 5e-05, "loss": 0.7779, "mean_token_accuracy": 0.8739802837371826, "num_tokens": 343329608.0, "step": 9045 }, { "epoch": 1.1507441801297544, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.197022914886475, "learning_rate": 5e-05, "loss": 0.7062, "mean_token_accuracy": 0.8828216791152954, "num_tokens": 343365459.0, "step": 9046 }, { "epoch": 1.150871390408345, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.91236686706543, "learning_rate": 5e-05, "loss": 0.7611, "mean_token_accuracy": 0.8773305416107178, "num_tokens": 343403909.0, "step": 9047 }, { "epoch": 1.1509986006869355, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.12017297744751, "learning_rate": 5e-05, "loss": 0.6617, "mean_token_accuracy": 0.8932434916496277, "num_tokens": 343438506.0, "step": 9048 }, { "epoch": 1.151125810965526, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.013081550598145, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.8678587675094604, "num_tokens": 343475059.0, "step": 9049 }, { "epoch": 1.1512530212441165, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.938751697540283, "learning_rate": 5e-05, "loss": 0.711, "mean_token_accuracy": 0.8753012418746948, "num_tokens": 343511286.0, "step": 9050 }, { "epoch": 1.151380231522707, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.226446151733398, "learning_rate": 5e-05, "loss": 0.7706, "mean_token_accuracy": 0.8846533298492432, "num_tokens": 343548092.0, "step": 9051 }, { "epoch": 1.1515074418012976, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.752928733825684, "learning_rate": 5e-05, "loss": 0.6457, "mean_token_accuracy": 0.8881278038024902, "num_tokens": 343586487.0, "step": 9052 }, { "epoch": 1.1516346520798881, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.420924186706543, "learning_rate": 5e-05, "loss": 0.7937, "mean_token_accuracy": 0.88652503490448, "num_tokens": 343627747.0, "step": 9053 }, { "epoch": 1.1517618623584787, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00049591064453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.6756415367126465, "learning_rate": 5e-05, "loss": 0.7309, "mean_token_accuracy": 0.8600151538848877, "num_tokens": 343672725.0, "step": 9054 }, { "epoch": 1.151889072637069, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.556513786315918, "learning_rate": 5e-05, "loss": 0.8584, "mean_token_accuracy": 0.869985044002533, "num_tokens": 343714426.0, "step": 9055 }, { "epoch": 1.1520162829156595, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.732990741729736, "learning_rate": 5e-05, "loss": 0.7286, "mean_token_accuracy": 0.8679654002189636, "num_tokens": 343754613.0, "step": 9056 }, { "epoch": 1.15214349319425, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.476297378540039, "learning_rate": 5e-05, "loss": 0.8603, "mean_token_accuracy": 0.8736259937286377, "num_tokens": 343796187.0, "step": 9057 }, { "epoch": 1.1522707034728406, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.0245232582092285, "learning_rate": 5e-05, "loss": 0.748, "mean_token_accuracy": 0.8676092028617859, "num_tokens": 343833323.0, "step": 9058 }, { "epoch": 1.152397913751431, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.056659698486328, "learning_rate": 5e-05, "loss": 0.8518, "mean_token_accuracy": 0.8647816777229309, "num_tokens": 343872642.0, "step": 9059 }, { "epoch": 1.1525251240300216, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.4856038093566895, "learning_rate": 5e-05, "loss": 0.7813, "mean_token_accuracy": 0.8719271421432495, "num_tokens": 343909104.0, "step": 9060 }, { "epoch": 1.1526523343086121, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.4202070236206055, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.867353081703186, "num_tokens": 343947102.0, "step": 9061 }, { "epoch": 1.1527795445872027, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.9182209968566895, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8804643750190735, "num_tokens": 343985887.0, "step": 9062 }, { "epoch": 1.1529067548657932, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.956684589385986, "learning_rate": 5e-05, "loss": 0.735, "mean_token_accuracy": 0.8668148517608643, "num_tokens": 344019427.0, "step": 9063 }, { "epoch": 1.1530339651443837, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.224738121032715, "learning_rate": 5e-05, "loss": 0.8226, "mean_token_accuracy": 0.865032434463501, "num_tokens": 344061706.0, "step": 9064 }, { "epoch": 1.1531611754229742, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.687117576599121, "learning_rate": 5e-05, "loss": 0.6535, "mean_token_accuracy": 0.8865548372268677, "num_tokens": 344100912.0, "step": 9065 }, { "epoch": 1.1532883857015648, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.50827693939209, "learning_rate": 5e-05, "loss": 0.843, "mean_token_accuracy": 0.8691699504852295, "num_tokens": 344135645.0, "step": 9066 }, { "epoch": 1.1534155959801553, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.47974157333374, "learning_rate": 5e-05, "loss": 0.6884, "mean_token_accuracy": 0.8740134835243225, "num_tokens": 344176467.0, "step": 9067 }, { "epoch": 1.1535428062587456, "ewc_loss": 0.5, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.786429405212402, "learning_rate": 5e-05, "loss": 0.8841, "mean_token_accuracy": 0.8667936325073242, "num_tokens": 344213498.0, "step": 9068 }, { "epoch": 1.1536700165373361, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.464820861816406, "learning_rate": 5e-05, "loss": 0.6587, "mean_token_accuracy": 0.8888916969299316, "num_tokens": 344250105.0, "step": 9069 }, { "epoch": 1.1537972268159267, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.88900089263916, "learning_rate": 5e-05, "loss": 0.9093, "mean_token_accuracy": 0.8709481358528137, "num_tokens": 344284017.0, "step": 9070 }, { "epoch": 1.1539244370945172, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.794821739196777, "learning_rate": 5e-05, "loss": 0.7318, "mean_token_accuracy": 0.8758149147033691, "num_tokens": 344317729.0, "step": 9071 }, { "epoch": 1.1540516473731077, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.497578620910645, "learning_rate": 5e-05, "loss": 0.9294, "mean_token_accuracy": 0.8565031290054321, "num_tokens": 344356032.0, "step": 9072 }, { "epoch": 1.1541788576516983, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.382519721984863, "learning_rate": 5e-05, "loss": 0.8518, "mean_token_accuracy": 0.8642287254333496, "num_tokens": 344385694.0, "step": 9073 }, { "epoch": 1.1543060679302888, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.775546073913574, "learning_rate": 5e-05, "loss": 0.81, "mean_token_accuracy": 0.8694177865982056, "num_tokens": 344420435.0, "step": 9074 }, { "epoch": 1.1544332782088793, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.792469501495361, "learning_rate": 5e-05, "loss": 0.7973, "mean_token_accuracy": 0.8819476366043091, "num_tokens": 344461065.0, "step": 9075 }, { "epoch": 1.1545604884874698, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.3278489112854, "learning_rate": 5e-05, "loss": 0.7191, "mean_token_accuracy": 0.8798382878303528, "num_tokens": 344495419.0, "step": 9076 }, { "epoch": 1.1546876987660604, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.803017616271973, "learning_rate": 5e-05, "loss": 0.8401, "mean_token_accuracy": 0.8608153462409973, "num_tokens": 344536304.0, "step": 9077 }, { "epoch": 1.154814909044651, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.398143291473389, "learning_rate": 5e-05, "loss": 0.7608, "mean_token_accuracy": 0.8745088577270508, "num_tokens": 344577118.0, "step": 9078 }, { "epoch": 1.1549421193232412, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.452450275421143, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8635956645011902, "num_tokens": 344609919.0, "step": 9079 }, { "epoch": 1.1550693296018317, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.753732204437256, "learning_rate": 5e-05, "loss": 0.7617, "mean_token_accuracy": 0.8743689656257629, "num_tokens": 344651543.0, "step": 9080 }, { "epoch": 1.1551965398804223, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.010409832000732, "learning_rate": 5e-05, "loss": 0.6973, "mean_token_accuracy": 0.8774063587188721, "num_tokens": 344691477.0, "step": 9081 }, { "epoch": 1.1553237501590128, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.139764785766602, "learning_rate": 5e-05, "loss": 0.794, "mean_token_accuracy": 0.8720580339431763, "num_tokens": 344731533.0, "step": 9082 }, { "epoch": 1.1554509604376033, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.658085346221924, "learning_rate": 5e-05, "loss": 0.6777, "mean_token_accuracy": 0.8837769031524658, "num_tokens": 344771628.0, "step": 9083 }, { "epoch": 1.1555781707161938, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.523945808410645, "learning_rate": 5e-05, "loss": 0.8668, "mean_token_accuracy": 0.8634270429611206, "num_tokens": 344812079.0, "step": 9084 }, { "epoch": 1.1557053809947844, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.3813347816467285, "learning_rate": 5e-05, "loss": 0.6651, "mean_token_accuracy": 0.879283607006073, "num_tokens": 344856395.0, "step": 9085 }, { "epoch": 1.155832591273375, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.931591033935547, "learning_rate": 5e-05, "loss": 0.8639, "mean_token_accuracy": 0.8811414241790771, "num_tokens": 344894578.0, "step": 9086 }, { "epoch": 1.1559598015519654, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.313508987426758, "learning_rate": 5e-05, "loss": 0.7518, "mean_token_accuracy": 0.8579444289207458, "num_tokens": 344937896.0, "step": 9087 }, { "epoch": 1.156087011830556, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.111413955688477, "learning_rate": 5e-05, "loss": 0.8891, "mean_token_accuracy": 0.8870056867599487, "num_tokens": 344972852.0, "step": 9088 }, { "epoch": 1.1562142221091465, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.769074440002441, "learning_rate": 5e-05, "loss": 0.7754, "mean_token_accuracy": 0.8691891431808472, "num_tokens": 345010322.0, "step": 9089 }, { "epoch": 1.156341432387737, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.600906372070312, "learning_rate": 5e-05, "loss": 0.857, "mean_token_accuracy": 0.8909027576446533, "num_tokens": 345046268.0, "step": 9090 }, { "epoch": 1.1564686426663275, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.6151299476623535, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.888904333114624, "num_tokens": 345078601.0, "step": 9091 }, { "epoch": 1.156595852944918, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.483901500701904, "learning_rate": 5e-05, "loss": 0.7899, "mean_token_accuracy": 0.8732868432998657, "num_tokens": 345121863.0, "step": 9092 }, { "epoch": 1.1567230632235084, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.375550270080566, "learning_rate": 5e-05, "loss": 0.8084, "mean_token_accuracy": 0.8925204277038574, "num_tokens": 345153884.0, "step": 9093 }, { "epoch": 1.156850273502099, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.614033222198486, "learning_rate": 5e-05, "loss": 0.6914, "mean_token_accuracy": 0.8792404532432556, "num_tokens": 345188793.0, "step": 9094 }, { "epoch": 1.1569774837806894, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.887550354003906, "learning_rate": 5e-05, "loss": 0.8817, "mean_token_accuracy": 0.8805780410766602, "num_tokens": 345222537.0, "step": 9095 }, { "epoch": 1.15710469405928, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.4049835205078125, "learning_rate": 5e-05, "loss": 0.7116, "mean_token_accuracy": 0.870675802230835, "num_tokens": 345262006.0, "step": 9096 }, { "epoch": 1.1572319043378705, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.065178871154785, "learning_rate": 5e-05, "loss": 0.9937, "mean_token_accuracy": 0.8598288893699646, "num_tokens": 345295388.0, "step": 9097 }, { "epoch": 1.157359114616461, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.748343467712402, "learning_rate": 5e-05, "loss": 0.7375, "mean_token_accuracy": 0.8856743574142456, "num_tokens": 345331144.0, "step": 9098 }, { "epoch": 1.1574863248950515, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.828208923339844, "learning_rate": 5e-05, "loss": 0.9912, "mean_token_accuracy": 0.8563767075538635, "num_tokens": 345373009.0, "step": 9099 }, { "epoch": 1.157613535173642, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.323838233947754, "learning_rate": 5e-05, "loss": 0.8169, "mean_token_accuracy": 0.8848565816879272, "num_tokens": 345411931.0, "step": 9100 }, { "epoch": 1.1577407454522326, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.205049514770508, "learning_rate": 5e-05, "loss": 0.8828, "mean_token_accuracy": 0.8710025548934937, "num_tokens": 345448818.0, "step": 9101 }, { "epoch": 1.1578679557308231, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.682883262634277, "learning_rate": 5e-05, "loss": 0.8273, "mean_token_accuracy": 0.8813177943229675, "num_tokens": 345484867.0, "step": 9102 }, { "epoch": 1.1579951660094137, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.751631736755371, "learning_rate": 5e-05, "loss": 0.8295, "mean_token_accuracy": 0.8733292818069458, "num_tokens": 345525949.0, "step": 9103 }, { "epoch": 1.158122376288004, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.56036376953125, "learning_rate": 5e-05, "loss": 0.7741, "mean_token_accuracy": 0.8850327730178833, "num_tokens": 345563027.0, "step": 9104 }, { "epoch": 1.1582495865665945, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.907881736755371, "learning_rate": 5e-05, "loss": 0.7969, "mean_token_accuracy": 0.8861572742462158, "num_tokens": 345605920.0, "step": 9105 }, { "epoch": 1.158376796845185, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000499725341796875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.025528907775879, "learning_rate": 5e-05, "loss": 0.7649, "mean_token_accuracy": 0.8683534860610962, "num_tokens": 345645843.0, "step": 9106 }, { "epoch": 1.1585040071237755, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.429497718811035, "learning_rate": 5e-05, "loss": 0.8581, "mean_token_accuracy": 0.8730674386024475, "num_tokens": 345684207.0, "step": 9107 }, { "epoch": 1.158631217402366, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.542189121246338, "learning_rate": 5e-05, "loss": 0.7269, "mean_token_accuracy": 0.8676400780677795, "num_tokens": 345721628.0, "step": 9108 }, { "epoch": 1.1587584276809566, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.922809600830078, "learning_rate": 5e-05, "loss": 0.8942, "mean_token_accuracy": 0.8800852298736572, "num_tokens": 345761864.0, "step": 9109 }, { "epoch": 1.1588856379595471, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.520168781280518, "learning_rate": 5e-05, "loss": 0.7148, "mean_token_accuracy": 0.8754826784133911, "num_tokens": 345802478.0, "step": 9110 }, { "epoch": 1.1590128482381377, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.001227378845215, "learning_rate": 5e-05, "loss": 0.9338, "mean_token_accuracy": 0.8774743676185608, "num_tokens": 345846598.0, "step": 9111 }, { "epoch": 1.1591400585167282, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.065827369689941, "learning_rate": 5e-05, "loss": 0.8491, "mean_token_accuracy": 0.86104416847229, "num_tokens": 345885050.0, "step": 9112 }, { "epoch": 1.1592672687953187, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.416899681091309, "learning_rate": 5e-05, "loss": 0.9911, "mean_token_accuracy": 0.8443019986152649, "num_tokens": 345927460.0, "step": 9113 }, { "epoch": 1.1593944790739092, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.74875545501709, "learning_rate": 5e-05, "loss": 0.8669, "mean_token_accuracy": 0.8727877736091614, "num_tokens": 345960855.0, "step": 9114 }, { "epoch": 1.1595216893524998, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.451994895935059, "learning_rate": 5e-05, "loss": 0.7619, "mean_token_accuracy": 0.880157470703125, "num_tokens": 345997086.0, "step": 9115 }, { "epoch": 1.1596488996310903, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.220898628234863, "learning_rate": 5e-05, "loss": 0.8515, "mean_token_accuracy": 0.8782212734222412, "num_tokens": 346034317.0, "step": 9116 }, { "epoch": 1.1597761099096806, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.871055603027344, "learning_rate": 5e-05, "loss": 0.7449, "mean_token_accuracy": 0.8690493106842041, "num_tokens": 346072626.0, "step": 9117 }, { "epoch": 1.1599033201882711, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.365979194641113, "learning_rate": 5e-05, "loss": 0.8809, "mean_token_accuracy": 0.8658697605133057, "num_tokens": 346109792.0, "step": 9118 }, { "epoch": 1.1600305304668617, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.806670188903809, "learning_rate": 5e-05, "loss": 0.6866, "mean_token_accuracy": 0.8820300102233887, "num_tokens": 346147696.0, "step": 9119 }, { "epoch": 1.1601577407454522, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.312607765197754, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.886218786239624, "num_tokens": 346183944.0, "step": 9120 }, { "epoch": 1.1602849510240427, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.880655288696289, "learning_rate": 5e-05, "loss": 0.6978, "mean_token_accuracy": 0.8805253505706787, "num_tokens": 346223352.0, "step": 9121 }, { "epoch": 1.1604121613026332, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.291584014892578, "learning_rate": 5e-05, "loss": 0.8469, "mean_token_accuracy": 0.870284378528595, "num_tokens": 346255008.0, "step": 9122 }, { "epoch": 1.1605393715812238, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.903410911560059, "learning_rate": 5e-05, "loss": 0.8336, "mean_token_accuracy": 0.8478254675865173, "num_tokens": 346285188.0, "step": 9123 }, { "epoch": 1.1606665818598143, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.245953559875488, "learning_rate": 5e-05, "loss": 0.8319, "mean_token_accuracy": 0.8781822919845581, "num_tokens": 346318691.0, "step": 9124 }, { "epoch": 1.1607937921384048, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.014294147491455, "learning_rate": 5e-05, "loss": 0.6992, "mean_token_accuracy": 0.8867028951644897, "num_tokens": 346353800.0, "step": 9125 }, { "epoch": 1.1609210024169954, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.073189735412598, "learning_rate": 5e-05, "loss": 0.8705, "mean_token_accuracy": 0.8582468628883362, "num_tokens": 346396616.0, "step": 9126 }, { "epoch": 1.161048212695586, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.162900447845459, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.8728271722793579, "num_tokens": 346438244.0, "step": 9127 }, { "epoch": 1.1611754229741762, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.867875576019287, "learning_rate": 5e-05, "loss": 0.8047, "mean_token_accuracy": 0.8726086020469666, "num_tokens": 346475081.0, "step": 9128 }, { "epoch": 1.1613026332527667, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.2910566329956055, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.8787076473236084, "num_tokens": 346513435.0, "step": 9129 }, { "epoch": 1.1614298435313573, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.702408313751221, "learning_rate": 5e-05, "loss": 0.8004, "mean_token_accuracy": 0.8672882318496704, "num_tokens": 346553216.0, "step": 9130 }, { "epoch": 1.1615570538099478, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.332076549530029, "learning_rate": 5e-05, "loss": 0.7518, "mean_token_accuracy": 0.869297444820404, "num_tokens": 346589271.0, "step": 9131 }, { "epoch": 1.1616842640885383, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.649759769439697, "learning_rate": 5e-05, "loss": 0.7574, "mean_token_accuracy": 0.8742949962615967, "num_tokens": 346631806.0, "step": 9132 }, { "epoch": 1.1618114743671288, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.256758213043213, "learning_rate": 5e-05, "loss": 0.6986, "mean_token_accuracy": 0.8872525095939636, "num_tokens": 346670805.0, "step": 9133 }, { "epoch": 1.1619386846457194, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.724084854125977, "learning_rate": 5e-05, "loss": 0.7268, "mean_token_accuracy": 0.8832157850265503, "num_tokens": 346710506.0, "step": 9134 }, { "epoch": 1.16206589492431, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.119376182556152, "learning_rate": 5e-05, "loss": 0.749, "mean_token_accuracy": 0.8631826639175415, "num_tokens": 346753410.0, "step": 9135 }, { "epoch": 1.1621931052029004, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.875367641448975, "learning_rate": 5e-05, "loss": 0.8139, "mean_token_accuracy": 0.8587667346000671, "num_tokens": 346795104.0, "step": 9136 }, { "epoch": 1.162320315481491, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.899857997894287, "learning_rate": 5e-05, "loss": 0.7466, "mean_token_accuracy": 0.8581491708755493, "num_tokens": 346836888.0, "step": 9137 }, { "epoch": 1.1624475257600815, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.152289390563965, "learning_rate": 5e-05, "loss": 0.7828, "mean_token_accuracy": 0.8701638579368591, "num_tokens": 346874449.0, "step": 9138 }, { "epoch": 1.162574736038672, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.614624977111816, "learning_rate": 5e-05, "loss": 0.6078, "mean_token_accuracy": 0.8968862295150757, "num_tokens": 346906706.0, "step": 9139 }, { "epoch": 1.1627019463172625, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.531694412231445, "learning_rate": 5e-05, "loss": 0.8223, "mean_token_accuracy": 0.8727396130561829, "num_tokens": 346947250.0, "step": 9140 }, { "epoch": 1.162829156595853, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.341927528381348, "learning_rate": 5e-05, "loss": 0.6896, "mean_token_accuracy": 0.8714128732681274, "num_tokens": 346984159.0, "step": 9141 }, { "epoch": 1.1629563668744434, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.934415817260742, "learning_rate": 5e-05, "loss": 0.9263, "mean_token_accuracy": 0.8545553684234619, "num_tokens": 347026449.0, "step": 9142 }, { "epoch": 1.163083577153034, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.325939655303955, "learning_rate": 5e-05, "loss": 0.6535, "mean_token_accuracy": 0.8881552219390869, "num_tokens": 347057903.0, "step": 9143 }, { "epoch": 1.1632107874316244, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.971807479858398, "learning_rate": 5e-05, "loss": 0.9223, "mean_token_accuracy": 0.8741772174835205, "num_tokens": 347095805.0, "step": 9144 }, { "epoch": 1.163337997710215, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 6.914037227630615, "learning_rate": 5e-05, "loss": 0.7724, "mean_token_accuracy": 0.8737279176712036, "num_tokens": 347133357.0, "step": 9145 }, { "epoch": 1.1634652079888055, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.215569496154785, "learning_rate": 5e-05, "loss": 0.8934, "mean_token_accuracy": 0.8629621863365173, "num_tokens": 347169679.0, "step": 9146 }, { "epoch": 1.163592418267396, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.842778205871582, "learning_rate": 5e-05, "loss": 0.8821, "mean_token_accuracy": 0.8672808408737183, "num_tokens": 347207990.0, "step": 9147 }, { "epoch": 1.1637196285459865, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.900566577911377, "learning_rate": 5e-05, "loss": 0.6873, "mean_token_accuracy": 0.8848059773445129, "num_tokens": 347244427.0, "step": 9148 }, { "epoch": 1.163846838824577, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.889177322387695, "learning_rate": 5e-05, "loss": 0.8459, "mean_token_accuracy": 0.8937177062034607, "num_tokens": 347276303.0, "step": 9149 }, { "epoch": 1.1639740491031676, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.949639320373535, "learning_rate": 5e-05, "loss": 0.6616, "mean_token_accuracy": 0.8776177167892456, "num_tokens": 347318682.0, "step": 9150 }, { "epoch": 1.1641012593817581, "ewc_loss": 0.625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.943931579589844, "learning_rate": 5e-05, "loss": 1.0047, "mean_token_accuracy": 0.8653104901313782, "num_tokens": 347359355.0, "step": 9151 }, { "epoch": 1.1642284696603487, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.1238250732421875, "learning_rate": 5e-05, "loss": 0.7511, "mean_token_accuracy": 0.8686316013336182, "num_tokens": 347397617.0, "step": 9152 }, { "epoch": 1.164355679938939, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 10.357046127319336, "learning_rate": 5e-05, "loss": 1.1533, "mean_token_accuracy": 0.8529748320579529, "num_tokens": 347433067.0, "step": 9153 }, { "epoch": 1.1644828902175295, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 7.875720024108887, "learning_rate": 5e-05, "loss": 0.9551, "mean_token_accuracy": 0.8861522674560547, "num_tokens": 347467091.0, "step": 9154 }, { "epoch": 1.16461010049612, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.071697235107422, "learning_rate": 5e-05, "loss": 0.9547, "mean_token_accuracy": 0.8806695938110352, "num_tokens": 347503921.0, "step": 9155 }, { "epoch": 1.1647373107747105, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 8.984977722167969, "learning_rate": 5e-05, "loss": 1.0714, "mean_token_accuracy": 0.8767647743225098, "num_tokens": 347547061.0, "step": 9156 }, { "epoch": 1.164864521053301, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.235931396484375, "learning_rate": 5e-05, "loss": 0.817, "mean_token_accuracy": 0.8672057390213013, "num_tokens": 347583845.0, "step": 9157 }, { "epoch": 1.1649917313318916, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.928604125976562, "learning_rate": 5e-05, "loss": 1.0444, "mean_token_accuracy": 0.8732988834381104, "num_tokens": 347624146.0, "step": 9158 }, { "epoch": 1.1651189416104821, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.071503162384033, "learning_rate": 5e-05, "loss": 0.8641, "mean_token_accuracy": 0.8711473941802979, "num_tokens": 347666171.0, "step": 9159 }, { "epoch": 1.1652461518890727, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.872605323791504, "learning_rate": 5e-05, "loss": 1.0353, "mean_token_accuracy": 0.8707519769668579, "num_tokens": 347706960.0, "step": 9160 }, { "epoch": 1.1653733621676632, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.06224536895752, "learning_rate": 5e-05, "loss": 0.9982, "mean_token_accuracy": 0.8712416887283325, "num_tokens": 347744830.0, "step": 9161 }, { "epoch": 1.1655005724462537, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.606266021728516, "learning_rate": 5e-05, "loss": 0.8979, "mean_token_accuracy": 0.8781745433807373, "num_tokens": 347791033.0, "step": 9162 }, { "epoch": 1.1656277827248442, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.59472942352295, "learning_rate": 5e-05, "loss": 0.9921, "mean_token_accuracy": 0.871721625328064, "num_tokens": 347832579.0, "step": 9163 }, { "epoch": 1.1657549930034348, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.550962448120117, "learning_rate": 5e-05, "loss": 0.769, "mean_token_accuracy": 0.8894419074058533, "num_tokens": 347863404.0, "step": 9164 }, { "epoch": 1.1658822032820253, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.799378395080566, "learning_rate": 5e-05, "loss": 0.963, "mean_token_accuracy": 0.8685144186019897, "num_tokens": 347900882.0, "step": 9165 }, { "epoch": 1.1660094135606156, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.37147331237793, "learning_rate": 5e-05, "loss": 0.8544, "mean_token_accuracy": 0.8659784197807312, "num_tokens": 347943105.0, "step": 9166 }, { "epoch": 1.1661366238392061, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.489914894104004, "learning_rate": 5e-05, "loss": 0.9809, "mean_token_accuracy": 0.8403292894363403, "num_tokens": 347978783.0, "step": 9167 }, { "epoch": 1.1662638341177967, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.568631172180176, "learning_rate": 5e-05, "loss": 0.8358, "mean_token_accuracy": 0.8820538520812988, "num_tokens": 348016169.0, "step": 9168 }, { "epoch": 1.1663910443963872, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.87109899520874, "learning_rate": 5e-05, "loss": 0.8357, "mean_token_accuracy": 0.8654556274414062, "num_tokens": 348052431.0, "step": 9169 }, { "epoch": 1.1665182546749777, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.824369430541992, "learning_rate": 5e-05, "loss": 0.8963, "mean_token_accuracy": 0.8685530424118042, "num_tokens": 348091412.0, "step": 9170 }, { "epoch": 1.1666454649535682, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.321535110473633, "learning_rate": 5e-05, "loss": 0.7624, "mean_token_accuracy": 0.8676910400390625, "num_tokens": 348128370.0, "step": 9171 }, { "epoch": 1.1667726752321588, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.143848419189453, "learning_rate": 5e-05, "loss": 0.8769, "mean_token_accuracy": 0.867443859577179, "num_tokens": 348167076.0, "step": 9172 }, { "epoch": 1.1668998855107493, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.91961669921875, "learning_rate": 5e-05, "loss": 0.6976, "mean_token_accuracy": 0.8790147304534912, "num_tokens": 348206926.0, "step": 9173 }, { "epoch": 1.1670270957893398, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.56216812133789, "learning_rate": 5e-05, "loss": 0.8774, "mean_token_accuracy": 0.8670768737792969, "num_tokens": 348247209.0, "step": 9174 }, { "epoch": 1.1671543060679304, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.637568473815918, "learning_rate": 5e-05, "loss": 0.7202, "mean_token_accuracy": 0.8721498250961304, "num_tokens": 348286613.0, "step": 9175 }, { "epoch": 1.1672815163465209, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.88271713256836, "learning_rate": 5e-05, "loss": 0.9068, "mean_token_accuracy": 0.8691813945770264, "num_tokens": 348322246.0, "step": 9176 }, { "epoch": 1.1674087266251112, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.644448757171631, "learning_rate": 5e-05, "loss": 0.7293, "mean_token_accuracy": 0.8788930177688599, "num_tokens": 348359435.0, "step": 9177 }, { "epoch": 1.1675359369037017, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.924294471740723, "learning_rate": 5e-05, "loss": 0.8657, "mean_token_accuracy": 0.8897451758384705, "num_tokens": 348397330.0, "step": 9178 }, { "epoch": 1.1676631471822922, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 6.891220569610596, "learning_rate": 5e-05, "loss": 0.842, "mean_token_accuracy": 0.8635295033454895, "num_tokens": 348435413.0, "step": 9179 }, { "epoch": 1.1677903574608828, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.705827713012695, "learning_rate": 5e-05, "loss": 0.9262, "mean_token_accuracy": 0.8621729612350464, "num_tokens": 348474963.0, "step": 9180 }, { "epoch": 1.1679175677394733, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.124105453491211, "learning_rate": 5e-05, "loss": 0.8364, "mean_token_accuracy": 0.8728281259536743, "num_tokens": 348511306.0, "step": 9181 }, { "epoch": 1.1680447780180638, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.443309783935547, "learning_rate": 5e-05, "loss": 0.8603, "mean_token_accuracy": 0.8795929551124573, "num_tokens": 348552257.0, "step": 9182 }, { "epoch": 1.1681719882966544, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.191301345825195, "learning_rate": 5e-05, "loss": 0.8241, "mean_token_accuracy": 0.868281364440918, "num_tokens": 348588143.0, "step": 9183 }, { "epoch": 1.168299198575245, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.298513412475586, "learning_rate": 5e-05, "loss": 0.8639, "mean_token_accuracy": 0.8724298477172852, "num_tokens": 348633525.0, "step": 9184 }, { "epoch": 1.1684264088538354, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.081501007080078, "learning_rate": 5e-05, "loss": 0.799, "mean_token_accuracy": 0.8708659410476685, "num_tokens": 348676024.0, "step": 9185 }, { "epoch": 1.168553619132426, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.378534317016602, "learning_rate": 5e-05, "loss": 0.8406, "mean_token_accuracy": 0.8812161684036255, "num_tokens": 348715704.0, "step": 9186 }, { "epoch": 1.1686808294110165, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.924924850463867, "learning_rate": 5e-05, "loss": 0.7611, "mean_token_accuracy": 0.8681843876838684, "num_tokens": 348753160.0, "step": 9187 }, { "epoch": 1.168808039689607, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.497406959533691, "learning_rate": 5e-05, "loss": 0.8602, "mean_token_accuracy": 0.8739246129989624, "num_tokens": 348796423.0, "step": 9188 }, { "epoch": 1.1689352499681975, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.945111274719238, "learning_rate": 5e-05, "loss": 0.7059, "mean_token_accuracy": 0.8912345170974731, "num_tokens": 348832813.0, "step": 9189 }, { "epoch": 1.169062460246788, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.365531921386719, "learning_rate": 5e-05, "loss": 0.8589, "mean_token_accuracy": 0.8739898800849915, "num_tokens": 348868273.0, "step": 9190 }, { "epoch": 1.1691896705253784, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.186868667602539, "learning_rate": 5e-05, "loss": 0.7545, "mean_token_accuracy": 0.8798922300338745, "num_tokens": 348908008.0, "step": 9191 }, { "epoch": 1.169316880803969, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.960576057434082, "learning_rate": 5e-05, "loss": 0.8258, "mean_token_accuracy": 0.8738428950309753, "num_tokens": 348947701.0, "step": 9192 }, { "epoch": 1.1694440910825594, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.501315116882324, "learning_rate": 5e-05, "loss": 0.7851, "mean_token_accuracy": 0.8726292252540588, "num_tokens": 348992221.0, "step": 9193 }, { "epoch": 1.16957130136115, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.517250061035156, "learning_rate": 5e-05, "loss": 0.7425, "mean_token_accuracy": 0.8820267915725708, "num_tokens": 349031982.0, "step": 9194 }, { "epoch": 1.1696985116397405, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.700603485107422, "learning_rate": 5e-05, "loss": 0.7572, "mean_token_accuracy": 0.8836977481842041, "num_tokens": 349070905.0, "step": 9195 }, { "epoch": 1.169825721918331, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.304103374481201, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.871527910232544, "num_tokens": 349114736.0, "step": 9196 }, { "epoch": 1.1699529321969215, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.704049110412598, "learning_rate": 5e-05, "loss": 0.7906, "mean_token_accuracy": 0.8649953007698059, "num_tokens": 349153760.0, "step": 9197 }, { "epoch": 1.170080142475512, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.31498908996582, "learning_rate": 5e-05, "loss": 0.6832, "mean_token_accuracy": 0.8886345028877258, "num_tokens": 349192464.0, "step": 9198 }, { "epoch": 1.1702073527541026, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.553503036499023, "learning_rate": 5e-05, "loss": 0.7403, "mean_token_accuracy": 0.8735055327415466, "num_tokens": 349234444.0, "step": 9199 }, { "epoch": 1.1703345630326931, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.433497905731201, "learning_rate": 5e-05, "loss": 0.7629, "mean_token_accuracy": 0.8609329462051392, "num_tokens": 349276777.0, "step": 9200 }, { "epoch": 1.1704617733112836, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.3668718338012695, "learning_rate": 5e-05, "loss": 0.6913, "mean_token_accuracy": 0.882678210735321, "num_tokens": 349317158.0, "step": 9201 }, { "epoch": 1.170588983589874, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.610989093780518, "learning_rate": 5e-05, "loss": 0.7626, "mean_token_accuracy": 0.8677879571914673, "num_tokens": 349357043.0, "step": 9202 }, { "epoch": 1.1707161938684645, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.174951553344727, "learning_rate": 5e-05, "loss": 0.7586, "mean_token_accuracy": 0.8568315505981445, "num_tokens": 349395843.0, "step": 9203 }, { "epoch": 1.170843404147055, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.784238338470459, "learning_rate": 5e-05, "loss": 0.751, "mean_token_accuracy": 0.8689913153648376, "num_tokens": 349429410.0, "step": 9204 }, { "epoch": 1.1709706144256455, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.968810558319092, "learning_rate": 5e-05, "loss": 0.7346, "mean_token_accuracy": 0.8595640659332275, "num_tokens": 349466507.0, "step": 9205 }, { "epoch": 1.171097824704236, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.96447229385376, "learning_rate": 5e-05, "loss": 0.7183, "mean_token_accuracy": 0.8867583870887756, "num_tokens": 349502918.0, "step": 9206 }, { "epoch": 1.1712250349828266, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.7969136238098145, "learning_rate": 5e-05, "loss": 0.6898, "mean_token_accuracy": 0.8728642463684082, "num_tokens": 349546740.0, "step": 9207 }, { "epoch": 1.1713522452614171, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.207271575927734, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8769875764846802, "num_tokens": 349590479.0, "step": 9208 }, { "epoch": 1.1714794555400077, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.549706935882568, "learning_rate": 5e-05, "loss": 0.6156, "mean_token_accuracy": 0.889663577079773, "num_tokens": 349627726.0, "step": 9209 }, { "epoch": 1.1716066658185982, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.569093704223633, "learning_rate": 5e-05, "loss": 0.8167, "mean_token_accuracy": 0.8741521835327148, "num_tokens": 349663547.0, "step": 9210 }, { "epoch": 1.1717338760971887, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.239677906036377, "learning_rate": 5e-05, "loss": 0.6848, "mean_token_accuracy": 0.868975818157196, "num_tokens": 349696927.0, "step": 9211 }, { "epoch": 1.1718610863757792, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 9.053818702697754, "learning_rate": 5e-05, "loss": 0.8688, "mean_token_accuracy": 0.8800086379051208, "num_tokens": 349736027.0, "step": 9212 }, { "epoch": 1.1719882966543698, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.124017238616943, "learning_rate": 5e-05, "loss": 0.7065, "mean_token_accuracy": 0.8670761585235596, "num_tokens": 349778225.0, "step": 9213 }, { "epoch": 1.1721155069329603, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.318873405456543, "learning_rate": 5e-05, "loss": 0.9567, "mean_token_accuracy": 0.8731193542480469, "num_tokens": 349818093.0, "step": 9214 }, { "epoch": 1.1722427172115506, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.670176029205322, "learning_rate": 5e-05, "loss": 0.7677, "mean_token_accuracy": 0.8701594471931458, "num_tokens": 349856433.0, "step": 9215 }, { "epoch": 1.1723699274901411, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.724325180053711, "learning_rate": 5e-05, "loss": 0.9547, "mean_token_accuracy": 0.8667072057723999, "num_tokens": 349898829.0, "step": 9216 }, { "epoch": 1.1724971377687317, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.7296600341796875, "learning_rate": 5e-05, "loss": 0.8773, "mean_token_accuracy": 0.8767796754837036, "num_tokens": 349937059.0, "step": 9217 }, { "epoch": 1.1726243480473222, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.270264625549316, "learning_rate": 5e-05, "loss": 0.7934, "mean_token_accuracy": 0.8638520240783691, "num_tokens": 349973433.0, "step": 9218 }, { "epoch": 1.1727515583259127, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.865143775939941, "learning_rate": 5e-05, "loss": 0.9904, "mean_token_accuracy": 0.8542303442955017, "num_tokens": 350012859.0, "step": 9219 }, { "epoch": 1.1728787686045032, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.097376346588135, "learning_rate": 5e-05, "loss": 0.7236, "mean_token_accuracy": 0.8612360954284668, "num_tokens": 350055128.0, "step": 9220 }, { "epoch": 1.1730059788830938, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.909782409667969, "learning_rate": 5e-05, "loss": 1.0278, "mean_token_accuracy": 0.8598732948303223, "num_tokens": 350090058.0, "step": 9221 }, { "epoch": 1.1731331891616843, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.2943010330200195, "learning_rate": 5e-05, "loss": 0.7231, "mean_token_accuracy": 0.8801017999649048, "num_tokens": 350129441.0, "step": 9222 }, { "epoch": 1.1732603994402748, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.131247520446777, "learning_rate": 5e-05, "loss": 1.0279, "mean_token_accuracy": 0.8805540800094604, "num_tokens": 350168866.0, "step": 9223 }, { "epoch": 1.1733876097188654, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 7.8517961502075195, "learning_rate": 5e-05, "loss": 0.9933, "mean_token_accuracy": 0.8703545331954956, "num_tokens": 350207498.0, "step": 9224 }, { "epoch": 1.1735148199974559, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.834038734436035, "learning_rate": 5e-05, "loss": 0.9742, "mean_token_accuracy": 0.8629981875419617, "num_tokens": 350248729.0, "step": 9225 }, { "epoch": 1.1736420302760462, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 8.789544105529785, "learning_rate": 5e-05, "loss": 1.0215, "mean_token_accuracy": 0.8725300431251526, "num_tokens": 350279399.0, "step": 9226 }, { "epoch": 1.1737692405546367, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.340615272521973, "learning_rate": 5e-05, "loss": 0.8595, "mean_token_accuracy": 0.8562172651290894, "num_tokens": 350321135.0, "step": 9227 }, { "epoch": 1.1738964508332272, "ewc_loss": 0.625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.254623413085938, "learning_rate": 5e-05, "loss": 0.9982, "mean_token_accuracy": 0.870322585105896, "num_tokens": 350359348.0, "step": 9228 }, { "epoch": 1.1740236611118178, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.224239349365234, "learning_rate": 5e-05, "loss": 0.8256, "mean_token_accuracy": 0.8769074082374573, "num_tokens": 350401140.0, "step": 9229 }, { "epoch": 1.1741508713904083, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.70354175567627, "learning_rate": 5e-05, "loss": 0.9437, "mean_token_accuracy": 0.8685263395309448, "num_tokens": 350443936.0, "step": 9230 }, { "epoch": 1.1742780816689988, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.825880527496338, "learning_rate": 5e-05, "loss": 0.8901, "mean_token_accuracy": 0.8827182054519653, "num_tokens": 350486110.0, "step": 9231 }, { "epoch": 1.1744052919475894, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.6422600746154785, "learning_rate": 5e-05, "loss": 0.8699, "mean_token_accuracy": 0.8524205684661865, "num_tokens": 350521964.0, "step": 9232 }, { "epoch": 1.1745325022261799, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.2942476272583, "learning_rate": 5e-05, "loss": 0.874, "mean_token_accuracy": 0.8854174613952637, "num_tokens": 350557578.0, "step": 9233 }, { "epoch": 1.1746597125047704, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.0758795738220215, "learning_rate": 5e-05, "loss": 0.7505, "mean_token_accuracy": 0.8724584579467773, "num_tokens": 350595693.0, "step": 9234 }, { "epoch": 1.174786922783361, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.472267150878906, "learning_rate": 5e-05, "loss": 0.8652, "mean_token_accuracy": 0.8798015713691711, "num_tokens": 350633247.0, "step": 9235 }, { "epoch": 1.1749141330619515, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.967592716217041, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8808860778808594, "num_tokens": 350673048.0, "step": 9236 }, { "epoch": 1.175041343340542, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.526500701904297, "learning_rate": 5e-05, "loss": 0.913, "mean_token_accuracy": 0.8605079054832458, "num_tokens": 350708840.0, "step": 9237 }, { "epoch": 1.1751685536191325, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.955138683319092, "learning_rate": 5e-05, "loss": 0.7904, "mean_token_accuracy": 0.8728014230728149, "num_tokens": 350748875.0, "step": 9238 }, { "epoch": 1.175295763897723, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.493874549865723, "learning_rate": 5e-05, "loss": 0.8608, "mean_token_accuracy": 0.8704947829246521, "num_tokens": 350783200.0, "step": 9239 }, { "epoch": 1.1754229741763134, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 6.975098133087158, "learning_rate": 5e-05, "loss": 0.7756, "mean_token_accuracy": 0.8736798763275146, "num_tokens": 350819514.0, "step": 9240 }, { "epoch": 1.175550184454904, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.406826972961426, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.8938521146774292, "num_tokens": 350854261.0, "step": 9241 }, { "epoch": 1.1756773947334944, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.993773937225342, "learning_rate": 5e-05, "loss": 0.7767, "mean_token_accuracy": 0.8732364177703857, "num_tokens": 350897847.0, "step": 9242 }, { "epoch": 1.175804605012085, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.332146644592285, "learning_rate": 5e-05, "loss": 0.8664, "mean_token_accuracy": 0.866775393486023, "num_tokens": 350931926.0, "step": 9243 }, { "epoch": 1.1759318152906755, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.954896926879883, "learning_rate": 5e-05, "loss": 0.7289, "mean_token_accuracy": 0.8844988942146301, "num_tokens": 350972352.0, "step": 9244 }, { "epoch": 1.176059025569266, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.36598014831543, "learning_rate": 5e-05, "loss": 0.8506, "mean_token_accuracy": 0.8701563477516174, "num_tokens": 351012245.0, "step": 9245 }, { "epoch": 1.1761862358478565, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.871774673461914, "learning_rate": 5e-05, "loss": 0.7135, "mean_token_accuracy": 0.8794254064559937, "num_tokens": 351043005.0, "step": 9246 }, { "epoch": 1.176313446126447, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.41830062866211, "learning_rate": 5e-05, "loss": 0.9058, "mean_token_accuracy": 0.8579721450805664, "num_tokens": 351084810.0, "step": 9247 }, { "epoch": 1.1764406564050376, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.85809850692749, "learning_rate": 5e-05, "loss": 0.7232, "mean_token_accuracy": 0.8776466250419617, "num_tokens": 351124243.0, "step": 9248 }, { "epoch": 1.1765678666836281, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.41642951965332, "learning_rate": 5e-05, "loss": 0.8344, "mean_token_accuracy": 0.8818602561950684, "num_tokens": 351163650.0, "step": 9249 }, { "epoch": 1.1766950769622184, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.935367584228516, "learning_rate": 5e-05, "loss": 0.8482, "mean_token_accuracy": 0.8475674390792847, "num_tokens": 351205912.0, "step": 9250 }, { "epoch": 1.176822287240809, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.314077377319336, "learning_rate": 5e-05, "loss": 0.8629, "mean_token_accuracy": 0.871951699256897, "num_tokens": 351250515.0, "step": 9251 }, { "epoch": 1.1769494975193995, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.079006195068359, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.870824933052063, "num_tokens": 351284222.0, "step": 9252 }, { "epoch": 1.17707670779799, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.072168350219727, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.8631468415260315, "num_tokens": 351327799.0, "step": 9253 }, { "epoch": 1.1772039180765805, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.3140668869018555, "learning_rate": 5e-05, "loss": 0.7748, "mean_token_accuracy": 0.8734930157661438, "num_tokens": 351362194.0, "step": 9254 }, { "epoch": 1.177331128355171, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.71795654296875, "learning_rate": 5e-05, "loss": 0.7677, "mean_token_accuracy": 0.8793151378631592, "num_tokens": 351395186.0, "step": 9255 }, { "epoch": 1.1774583386337616, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.504951477050781, "learning_rate": 5e-05, "loss": 0.7864, "mean_token_accuracy": 0.8742243051528931, "num_tokens": 351435484.0, "step": 9256 }, { "epoch": 1.1775855489123521, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.467111587524414, "learning_rate": 5e-05, "loss": 0.7745, "mean_token_accuracy": 0.8647363781929016, "num_tokens": 351474613.0, "step": 9257 }, { "epoch": 1.1777127591909426, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.611035346984863, "learning_rate": 5e-05, "loss": 0.7506, "mean_token_accuracy": 0.8798219561576843, "num_tokens": 351508612.0, "step": 9258 }, { "epoch": 1.1778399694695332, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.317594528198242, "learning_rate": 5e-05, "loss": 0.7956, "mean_token_accuracy": 0.8577005863189697, "num_tokens": 351554811.0, "step": 9259 }, { "epoch": 1.1779671797481237, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.645040512084961, "learning_rate": 5e-05, "loss": 0.7271, "mean_token_accuracy": 0.8749227523803711, "num_tokens": 351591540.0, "step": 9260 }, { "epoch": 1.1780943900267142, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.258544445037842, "learning_rate": 5e-05, "loss": 0.7278, "mean_token_accuracy": 0.872383713722229, "num_tokens": 351630529.0, "step": 9261 }, { "epoch": 1.1782216003053048, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.602293014526367, "learning_rate": 5e-05, "loss": 0.7788, "mean_token_accuracy": 0.8603838682174683, "num_tokens": 351669412.0, "step": 9262 }, { "epoch": 1.1783488105838953, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.303487300872803, "learning_rate": 5e-05, "loss": 0.74, "mean_token_accuracy": 0.8685196042060852, "num_tokens": 351711825.0, "step": 9263 }, { "epoch": 1.1784760208624856, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.507688045501709, "learning_rate": 5e-05, "loss": 0.7315, "mean_token_accuracy": 0.8775290250778198, "num_tokens": 351747967.0, "step": 9264 }, { "epoch": 1.1786032311410761, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.363519191741943, "learning_rate": 5e-05, "loss": 0.6972, "mean_token_accuracy": 0.8801732659339905, "num_tokens": 351786208.0, "step": 9265 }, { "epoch": 1.1787304414196667, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.417236328125, "learning_rate": 5e-05, "loss": 0.7326, "mean_token_accuracy": 0.8711639642715454, "num_tokens": 351825711.0, "step": 9266 }, { "epoch": 1.1788576516982572, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.408384323120117, "learning_rate": 5e-05, "loss": 0.6834, "mean_token_accuracy": 0.8829258680343628, "num_tokens": 351866772.0, "step": 9267 }, { "epoch": 1.1789848619768477, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.3731608390808105, "learning_rate": 5e-05, "loss": 0.71, "mean_token_accuracy": 0.8738698959350586, "num_tokens": 351905812.0, "step": 9268 }, { "epoch": 1.1791120722554382, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.4387407302856445, "learning_rate": 5e-05, "loss": 0.6946, "mean_token_accuracy": 0.8789764642715454, "num_tokens": 351950295.0, "step": 9269 }, { "epoch": 1.1792392825340288, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.3226847648620605, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8605138063430786, "num_tokens": 351990980.0, "step": 9270 }, { "epoch": 1.1793664928126193, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.459987163543701, "learning_rate": 5e-05, "loss": 0.751, "mean_token_accuracy": 0.860357403755188, "num_tokens": 352034483.0, "step": 9271 }, { "epoch": 1.1794937030912098, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.274048328399658, "learning_rate": 5e-05, "loss": 0.6664, "mean_token_accuracy": 0.8815467953681946, "num_tokens": 352069195.0, "step": 9272 }, { "epoch": 1.1796209133698004, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.50950813293457, "learning_rate": 5e-05, "loss": 0.6608, "mean_token_accuracy": 0.884128987789154, "num_tokens": 352103393.0, "step": 9273 }, { "epoch": 1.1797481236483909, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.183073043823242, "learning_rate": 5e-05, "loss": 0.663, "mean_token_accuracy": 0.8804032802581787, "num_tokens": 352138032.0, "step": 9274 }, { "epoch": 1.1798753339269812, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.630311489105225, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8638232946395874, "num_tokens": 352177243.0, "step": 9275 }, { "epoch": 1.1800025442055717, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 7.02716064453125, "learning_rate": 5e-05, "loss": 0.7276, "mean_token_accuracy": 0.8570065498352051, "num_tokens": 352212461.0, "step": 9276 }, { "epoch": 1.1801297544841622, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.8646416664123535, "learning_rate": 5e-05, "loss": 0.7637, "mean_token_accuracy": 0.863347589969635, "num_tokens": 352250001.0, "step": 9277 }, { "epoch": 1.1802569647627528, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.706445217132568, "learning_rate": 5e-05, "loss": 0.7065, "mean_token_accuracy": 0.8591022491455078, "num_tokens": 352288893.0, "step": 9278 }, { "epoch": 1.1803841750413433, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.346733093261719, "learning_rate": 5e-05, "loss": 0.773, "mean_token_accuracy": 0.874382734298706, "num_tokens": 352328255.0, "step": 9279 }, { "epoch": 1.1805113853199338, "ewc_loss": 0.265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 6.106831073760986, "learning_rate": 5e-05, "loss": 0.6379, "mean_token_accuracy": 0.8692747354507446, "num_tokens": 352364871.0, "step": 9280 }, { "epoch": 1.1806385955985244, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 9.340888023376465, "learning_rate": 5e-05, "loss": 0.8469, "mean_token_accuracy": 0.8809612989425659, "num_tokens": 352398894.0, "step": 9281 }, { "epoch": 1.1807658058771149, "ewc_loss": 0.2421875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00019168853759765625, "grad_norm": 5.259457588195801, "learning_rate": 5e-05, "loss": 0.6355, "mean_token_accuracy": 0.8646527528762817, "num_tokens": 352438327.0, "step": 9282 }, { "epoch": 1.1808930161557054, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 11.122772216796875, "learning_rate": 5e-05, "loss": 1.1283, "mean_token_accuracy": 0.8611962795257568, "num_tokens": 352475732.0, "step": 9283 }, { "epoch": 1.181020226434296, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 5.90915584564209, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.8565343618392944, "num_tokens": 352509405.0, "step": 9284 }, { "epoch": 1.1811474367128865, "ewc_loss": 1.09375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00104522705078125, "grad_norm": 12.974138259887695, "learning_rate": 5e-05, "loss": 1.4978, "mean_token_accuracy": 0.8631659746170044, "num_tokens": 352547172.0, "step": 9285 }, { "epoch": 1.181274646991477, "ewc_loss": 1.0703125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0010223388671875, "grad_norm": 9.936322212219238, "learning_rate": 5e-05, "loss": 1.4549, "mean_token_accuracy": 0.8693315982818604, "num_tokens": 352585705.0, "step": 9286 }, { "epoch": 1.1814018572700675, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 11.644306182861328, "learning_rate": 5e-05, "loss": 1.0787, "mean_token_accuracy": 0.8808534741401672, "num_tokens": 352629246.0, "step": 9287 }, { "epoch": 1.181529067548658, "ewc_loss": 0.953125, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000904083251953125, "grad_norm": 9.82613754272461, "learning_rate": 5e-05, "loss": 1.3514, "mean_token_accuracy": 0.8643454313278198, "num_tokens": 352669504.0, "step": 9288 }, { "epoch": 1.1816562778272484, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 11.187971115112305, "learning_rate": 5e-05, "loss": 1.1677, "mean_token_accuracy": 0.8684719800949097, "num_tokens": 352708981.0, "step": 9289 }, { "epoch": 1.1817834881058389, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 8.755011558532715, "learning_rate": 5e-05, "loss": 1.0641, "mean_token_accuracy": 0.8676285743713379, "num_tokens": 352751646.0, "step": 9290 }, { "epoch": 1.1819106983844294, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 10.64068603515625, "learning_rate": 5e-05, "loss": 1.1231, "mean_token_accuracy": 0.8715288639068604, "num_tokens": 352789959.0, "step": 9291 }, { "epoch": 1.18203790866302, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.342598915100098, "learning_rate": 5e-05, "loss": 1.0162, "mean_token_accuracy": 0.8571335077285767, "num_tokens": 352828725.0, "step": 9292 }, { "epoch": 1.1821651189416105, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.36469841003418, "learning_rate": 5e-05, "loss": 1.0143, "mean_token_accuracy": 0.8902154564857483, "num_tokens": 352866373.0, "step": 9293 }, { "epoch": 1.182292329220201, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.235468864440918, "learning_rate": 5e-05, "loss": 0.9724, "mean_token_accuracy": 0.8740818500518799, "num_tokens": 352908964.0, "step": 9294 }, { "epoch": 1.1824195394987915, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.234540939331055, "learning_rate": 5e-05, "loss": 1.0319, "mean_token_accuracy": 0.879757821559906, "num_tokens": 352946292.0, "step": 9295 }, { "epoch": 1.182546749777382, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.316588401794434, "learning_rate": 5e-05, "loss": 0.9652, "mean_token_accuracy": 0.8746745586395264, "num_tokens": 352985875.0, "step": 9296 }, { "epoch": 1.1826739600559726, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.948521614074707, "learning_rate": 5e-05, "loss": 1.0508, "mean_token_accuracy": 0.8679820895195007, "num_tokens": 353032019.0, "step": 9297 }, { "epoch": 1.1828011703345631, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.289576530456543, "learning_rate": 5e-05, "loss": 0.9631, "mean_token_accuracy": 0.8729457259178162, "num_tokens": 353062624.0, "step": 9298 }, { "epoch": 1.1829283806131534, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.663406372070312, "learning_rate": 5e-05, "loss": 0.9954, "mean_token_accuracy": 0.8757058382034302, "num_tokens": 353105401.0, "step": 9299 }, { "epoch": 1.183055590891744, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.111085891723633, "learning_rate": 5e-05, "loss": 0.9716, "mean_token_accuracy": 0.8550747036933899, "num_tokens": 353143180.0, "step": 9300 }, { "epoch": 1.1831828011703345, "ewc_loss": 0.625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.540863990783691, "learning_rate": 5e-05, "loss": 0.9847, "mean_token_accuracy": 0.8740164041519165, "num_tokens": 353177800.0, "step": 9301 }, { "epoch": 1.183310011448925, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.94904899597168, "learning_rate": 5e-05, "loss": 0.8815, "mean_token_accuracy": 0.8718384504318237, "num_tokens": 353221491.0, "step": 9302 }, { "epoch": 1.1834372217275155, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.517934799194336, "learning_rate": 5e-05, "loss": 1.0304, "mean_token_accuracy": 0.8599636554718018, "num_tokens": 353263851.0, "step": 9303 }, { "epoch": 1.183564432006106, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.927200794219971, "learning_rate": 5e-05, "loss": 0.9333, "mean_token_accuracy": 0.8587611317634583, "num_tokens": 353299197.0, "step": 9304 }, { "epoch": 1.1836916422846966, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.480040550231934, "learning_rate": 5e-05, "loss": 1.0718, "mean_token_accuracy": 0.8529983758926392, "num_tokens": 353339008.0, "step": 9305 }, { "epoch": 1.1838188525632871, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.027633666992188, "learning_rate": 5e-05, "loss": 0.9, "mean_token_accuracy": 0.8706315755844116, "num_tokens": 353378143.0, "step": 9306 }, { "epoch": 1.1839460628418776, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.247944831848145, "learning_rate": 5e-05, "loss": 0.9384, "mean_token_accuracy": 0.8912903070449829, "num_tokens": 353412659.0, "step": 9307 }, { "epoch": 1.1840732731204682, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.100390434265137, "learning_rate": 5e-05, "loss": 0.9369, "mean_token_accuracy": 0.8612165451049805, "num_tokens": 353454459.0, "step": 9308 }, { "epoch": 1.1842004833990587, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.827692031860352, "learning_rate": 5e-05, "loss": 0.9903, "mean_token_accuracy": 0.8629475831985474, "num_tokens": 353501446.0, "step": 9309 }, { "epoch": 1.1843276936776492, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.966046333312988, "learning_rate": 5e-05, "loss": 0.8974, "mean_token_accuracy": 0.8662559986114502, "num_tokens": 353538129.0, "step": 9310 }, { "epoch": 1.1844549039562398, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.60571575164795, "learning_rate": 5e-05, "loss": 0.9197, "mean_token_accuracy": 0.8814209699630737, "num_tokens": 353581488.0, "step": 9311 }, { "epoch": 1.1845821142348303, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.585020065307617, "learning_rate": 5e-05, "loss": 0.8161, "mean_token_accuracy": 0.8722937107086182, "num_tokens": 353622729.0, "step": 9312 }, { "epoch": 1.1847093245134206, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.776764869689941, "learning_rate": 5e-05, "loss": 0.9217, "mean_token_accuracy": 0.8803333640098572, "num_tokens": 353663696.0, "step": 9313 }, { "epoch": 1.1848365347920111, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.344419956207275, "learning_rate": 5e-05, "loss": 0.8248, "mean_token_accuracy": 0.8630657196044922, "num_tokens": 353705934.0, "step": 9314 }, { "epoch": 1.1849637450706016, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.991616249084473, "learning_rate": 5e-05, "loss": 0.9266, "mean_token_accuracy": 0.8896046876907349, "num_tokens": 353741535.0, "step": 9315 }, { "epoch": 1.1850909553491922, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.453958988189697, "learning_rate": 5e-05, "loss": 0.8128, "mean_token_accuracy": 0.8712878227233887, "num_tokens": 353784446.0, "step": 9316 }, { "epoch": 1.1852181656277827, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.911450386047363, "learning_rate": 5e-05, "loss": 0.97, "mean_token_accuracy": 0.8710360527038574, "num_tokens": 353821955.0, "step": 9317 }, { "epoch": 1.1853453759063732, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.752892971038818, "learning_rate": 5e-05, "loss": 0.8669, "mean_token_accuracy": 0.8731415271759033, "num_tokens": 353859994.0, "step": 9318 }, { "epoch": 1.1854725861849638, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.429890632629395, "learning_rate": 5e-05, "loss": 0.9199, "mean_token_accuracy": 0.8780146837234497, "num_tokens": 353903301.0, "step": 9319 }, { "epoch": 1.1855997964635543, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.8567376136779785, "learning_rate": 5e-05, "loss": 0.9148, "mean_token_accuracy": 0.8493410348892212, "num_tokens": 353944527.0, "step": 9320 }, { "epoch": 1.1857270067421448, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.9849772453308105, "learning_rate": 5e-05, "loss": 0.8326, "mean_token_accuracy": 0.8797985315322876, "num_tokens": 353977402.0, "step": 9321 }, { "epoch": 1.1858542170207353, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.662408828735352, "learning_rate": 5e-05, "loss": 0.786, "mean_token_accuracy": 0.8778629899024963, "num_tokens": 354016681.0, "step": 9322 }, { "epoch": 1.1859814272993259, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.9530744552612305, "learning_rate": 5e-05, "loss": 0.8721, "mean_token_accuracy": 0.8619898557662964, "num_tokens": 354057923.0, "step": 9323 }, { "epoch": 1.1861086375779162, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.33558988571167, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8717930316925049, "num_tokens": 354097787.0, "step": 9324 }, { "epoch": 1.1862358478565067, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.133890151977539, "learning_rate": 5e-05, "loss": 0.8269, "mean_token_accuracy": 0.8776578903198242, "num_tokens": 354136694.0, "step": 9325 }, { "epoch": 1.1863630581350972, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.108262538909912, "learning_rate": 5e-05, "loss": 0.6867, "mean_token_accuracy": 0.8876110911369324, "num_tokens": 354176326.0, "step": 9326 }, { "epoch": 1.1864902684136878, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.28248119354248, "learning_rate": 5e-05, "loss": 0.8659, "mean_token_accuracy": 0.8696147203445435, "num_tokens": 354214350.0, "step": 9327 }, { "epoch": 1.1866174786922783, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.032812118530273, "learning_rate": 5e-05, "loss": 0.7295, "mean_token_accuracy": 0.8719656467437744, "num_tokens": 354252910.0, "step": 9328 }, { "epoch": 1.1867446889708688, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.306321144104004, "learning_rate": 5e-05, "loss": 0.8424, "mean_token_accuracy": 0.873731255531311, "num_tokens": 354291105.0, "step": 9329 }, { "epoch": 1.1868718992494594, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.104099750518799, "learning_rate": 5e-05, "loss": 0.7304, "mean_token_accuracy": 0.8757432699203491, "num_tokens": 354329767.0, "step": 9330 }, { "epoch": 1.1869991095280499, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.200536727905273, "learning_rate": 5e-05, "loss": 0.8287, "mean_token_accuracy": 0.8799201250076294, "num_tokens": 354369826.0, "step": 9331 }, { "epoch": 1.1871263198066404, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.2605085372924805, "learning_rate": 5e-05, "loss": 0.7137, "mean_token_accuracy": 0.8840996026992798, "num_tokens": 354406088.0, "step": 9332 }, { "epoch": 1.187253530085231, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.9478888511657715, "learning_rate": 5e-05, "loss": 0.7685, "mean_token_accuracy": 0.8842247128486633, "num_tokens": 354445675.0, "step": 9333 }, { "epoch": 1.1873807403638215, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.458450794219971, "learning_rate": 5e-05, "loss": 0.7998, "mean_token_accuracy": 0.8544459939002991, "num_tokens": 354482733.0, "step": 9334 }, { "epoch": 1.187507950642412, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.59360933303833, "learning_rate": 5e-05, "loss": 0.7889, "mean_token_accuracy": 0.8667927980422974, "num_tokens": 354515483.0, "step": 9335 }, { "epoch": 1.1876351609210025, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.6145830154418945, "learning_rate": 5e-05, "loss": 0.7239, "mean_token_accuracy": 0.8805661797523499, "num_tokens": 354551972.0, "step": 9336 }, { "epoch": 1.187762371199593, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.352147579193115, "learning_rate": 5e-05, "loss": 0.743, "mean_token_accuracy": 0.8717012405395508, "num_tokens": 354592089.0, "step": 9337 }, { "epoch": 1.1878895814781834, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.697735786437988, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.8846144676208496, "num_tokens": 354629582.0, "step": 9338 }, { "epoch": 1.1880167917567739, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.225090026855469, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.8503222465515137, "num_tokens": 354677114.0, "step": 9339 }, { "epoch": 1.1881440020353644, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.754796504974365, "learning_rate": 5e-05, "loss": 0.7734, "mean_token_accuracy": 0.8711540699005127, "num_tokens": 354723125.0, "step": 9340 }, { "epoch": 1.188271212313955, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.15214729309082, "learning_rate": 5e-05, "loss": 0.7481, "mean_token_accuracy": 0.8630778789520264, "num_tokens": 354761032.0, "step": 9341 }, { "epoch": 1.1883984225925455, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.7606587409973145, "learning_rate": 5e-05, "loss": 0.7295, "mean_token_accuracy": 0.8805787563323975, "num_tokens": 354798522.0, "step": 9342 }, { "epoch": 1.188525632871136, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.116091728210449, "learning_rate": 5e-05, "loss": 0.708, "mean_token_accuracy": 0.8731889724731445, "num_tokens": 354833752.0, "step": 9343 }, { "epoch": 1.1886528431497265, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.776419639587402, "learning_rate": 5e-05, "loss": 0.7433, "mean_token_accuracy": 0.8758547306060791, "num_tokens": 354869901.0, "step": 9344 }, { "epoch": 1.188780053428317, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.090869426727295, "learning_rate": 5e-05, "loss": 0.6965, "mean_token_accuracy": 0.8727120757102966, "num_tokens": 354902487.0, "step": 9345 }, { "epoch": 1.1889072637069076, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.803066253662109, "learning_rate": 5e-05, "loss": 0.8096, "mean_token_accuracy": 0.8496531248092651, "num_tokens": 354948778.0, "step": 9346 }, { "epoch": 1.189034473985498, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.054177284240723, "learning_rate": 5e-05, "loss": 0.7023, "mean_token_accuracy": 0.8691520094871521, "num_tokens": 354990242.0, "step": 9347 }, { "epoch": 1.1891616842640884, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.802095413208008, "learning_rate": 5e-05, "loss": 0.6997, "mean_token_accuracy": 0.8853626251220703, "num_tokens": 355025462.0, "step": 9348 }, { "epoch": 1.189288894542679, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.060200214385986, "learning_rate": 5e-05, "loss": 0.741, "mean_token_accuracy": 0.8647408485412598, "num_tokens": 355067488.0, "step": 9349 }, { "epoch": 1.1894161048212695, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.789490222930908, "learning_rate": 5e-05, "loss": 0.76, "mean_token_accuracy": 0.869110107421875, "num_tokens": 355101734.0, "step": 9350 }, { "epoch": 1.18954331509986, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.055538654327393, "learning_rate": 5e-05, "loss": 0.7275, "mean_token_accuracy": 0.8630010485649109, "num_tokens": 355140209.0, "step": 9351 }, { "epoch": 1.1896705253784505, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.8217315673828125, "learning_rate": 5e-05, "loss": 0.7544, "mean_token_accuracy": 0.8687838912010193, "num_tokens": 355177424.0, "step": 9352 }, { "epoch": 1.189797735657041, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.983776092529297, "learning_rate": 5e-05, "loss": 0.6816, "mean_token_accuracy": 0.8734918832778931, "num_tokens": 355215112.0, "step": 9353 }, { "epoch": 1.1899249459356316, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.928181171417236, "learning_rate": 5e-05, "loss": 0.7243, "mean_token_accuracy": 0.8829846382141113, "num_tokens": 355250402.0, "step": 9354 }, { "epoch": 1.1900521562142221, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.846132278442383, "learning_rate": 5e-05, "loss": 0.6689, "mean_token_accuracy": 0.8775506019592285, "num_tokens": 355289887.0, "step": 9355 }, { "epoch": 1.1901793664928126, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.118529319763184, "learning_rate": 5e-05, "loss": 0.7609, "mean_token_accuracy": 0.8743205070495605, "num_tokens": 355330946.0, "step": 9356 }, { "epoch": 1.1903065767714032, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.632811546325684, "learning_rate": 5e-05, "loss": 0.6632, "mean_token_accuracy": 0.8739107847213745, "num_tokens": 355369029.0, "step": 9357 }, { "epoch": 1.1904337870499937, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.425546646118164, "learning_rate": 5e-05, "loss": 0.869, "mean_token_accuracy": 0.8533251285552979, "num_tokens": 355406145.0, "step": 9358 }, { "epoch": 1.1905609973285842, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.371286392211914, "learning_rate": 5e-05, "loss": 0.6075, "mean_token_accuracy": 0.8918718099594116, "num_tokens": 355445702.0, "step": 9359 }, { "epoch": 1.1906882076071748, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.830209732055664, "learning_rate": 5e-05, "loss": 0.8527, "mean_token_accuracy": 0.8699244856834412, "num_tokens": 355485592.0, "step": 9360 }, { "epoch": 1.1908154178857653, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.225385665893555, "learning_rate": 5e-05, "loss": 0.643, "mean_token_accuracy": 0.8869106769561768, "num_tokens": 355523208.0, "step": 9361 }, { "epoch": 1.1909426281643556, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 9.047847747802734, "learning_rate": 5e-05, "loss": 0.8996, "mean_token_accuracy": 0.873141884803772, "num_tokens": 355557729.0, "step": 9362 }, { "epoch": 1.1910698384429461, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.59395694732666, "learning_rate": 5e-05, "loss": 0.689, "mean_token_accuracy": 0.8830930590629578, "num_tokens": 355597222.0, "step": 9363 }, { "epoch": 1.1911970487215366, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.541279792785645, "learning_rate": 5e-05, "loss": 0.8336, "mean_token_accuracy": 0.888190746307373, "num_tokens": 355630477.0, "step": 9364 }, { "epoch": 1.1913242590001272, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.548447132110596, "learning_rate": 5e-05, "loss": 0.8214, "mean_token_accuracy": 0.8710496425628662, "num_tokens": 355664506.0, "step": 9365 }, { "epoch": 1.1914514692787177, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.169455528259277, "learning_rate": 5e-05, "loss": 0.7447, "mean_token_accuracy": 0.8725878596305847, "num_tokens": 355710268.0, "step": 9366 }, { "epoch": 1.1915786795573082, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.787825584411621, "learning_rate": 5e-05, "loss": 0.9564, "mean_token_accuracy": 0.8559989929199219, "num_tokens": 355746396.0, "step": 9367 }, { "epoch": 1.1917058898358988, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00023937225341796875, "grad_norm": 5.863797664642334, "learning_rate": 5e-05, "loss": 0.6817, "mean_token_accuracy": 0.8639365434646606, "num_tokens": 355787483.0, "step": 9368 }, { "epoch": 1.1918331001144893, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.210805892944336, "learning_rate": 5e-05, "loss": 1.0287, "mean_token_accuracy": 0.8646599054336548, "num_tokens": 355824138.0, "step": 9369 }, { "epoch": 1.1919603103930798, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 5.797837257385254, "learning_rate": 5e-05, "loss": 0.7259, "mean_token_accuracy": 0.8631714582443237, "num_tokens": 355854258.0, "step": 9370 }, { "epoch": 1.1920875206716703, "ewc_loss": 0.83203125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 11.199338912963867, "learning_rate": 5e-05, "loss": 1.2003, "mean_token_accuracy": 0.8703507781028748, "num_tokens": 355889750.0, "step": 9371 }, { "epoch": 1.1922147309502609, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 8.203482627868652, "learning_rate": 5e-05, "loss": 1.054, "mean_token_accuracy": 0.8810326457023621, "num_tokens": 355923950.0, "step": 9372 }, { "epoch": 1.1923419412288512, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.037975311279297, "learning_rate": 5e-05, "loss": 1.003, "mean_token_accuracy": 0.8792010545730591, "num_tokens": 355959807.0, "step": 9373 }, { "epoch": 1.1924691515074417, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.00075531005859375, "grad_norm": 9.31486988067627, "learning_rate": 5e-05, "loss": 1.1455, "mean_token_accuracy": 0.8793584704399109, "num_tokens": 355992574.0, "step": 9374 }, { "epoch": 1.1925963617860322, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.130712509155273, "learning_rate": 5e-05, "loss": 0.9057, "mean_token_accuracy": 0.8644426465034485, "num_tokens": 356033726.0, "step": 9375 }, { "epoch": 1.1927235720646228, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.175558090209961, "learning_rate": 5e-05, "loss": 1.0759, "mean_token_accuracy": 0.8822333216667175, "num_tokens": 356066959.0, "step": 9376 }, { "epoch": 1.1928507823432133, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.618719577789307, "learning_rate": 5e-05, "loss": 0.8978, "mean_token_accuracy": 0.8733471632003784, "num_tokens": 356102316.0, "step": 9377 }, { "epoch": 1.1929779926218038, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.516388893127441, "learning_rate": 5e-05, "loss": 1.027, "mean_token_accuracy": 0.8832798600196838, "num_tokens": 356146203.0, "step": 9378 }, { "epoch": 1.1931052029003943, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 8.331539154052734, "learning_rate": 5e-05, "loss": 1.0487, "mean_token_accuracy": 0.861583948135376, "num_tokens": 356184673.0, "step": 9379 }, { "epoch": 1.1932324131789849, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.706563949584961, "learning_rate": 5e-05, "loss": 1.0037, "mean_token_accuracy": 0.8691547513008118, "num_tokens": 356224693.0, "step": 9380 }, { "epoch": 1.1933596234575754, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.410889625549316, "learning_rate": 5e-05, "loss": 1.0119, "mean_token_accuracy": 0.8656446933746338, "num_tokens": 356265454.0, "step": 9381 }, { "epoch": 1.193486833736166, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.762228012084961, "learning_rate": 5e-05, "loss": 0.9223, "mean_token_accuracy": 0.8714680671691895, "num_tokens": 356302866.0, "step": 9382 }, { "epoch": 1.1936140440147565, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000507354736328125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.196368217468262, "learning_rate": 5e-05, "loss": 0.9053, "mean_token_accuracy": 0.875750720500946, "num_tokens": 356341897.0, "step": 9383 }, { "epoch": 1.193741254293347, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.188457489013672, "learning_rate": 5e-05, "loss": 0.8577, "mean_token_accuracy": 0.8789986968040466, "num_tokens": 356380637.0, "step": 9384 }, { "epoch": 1.1938684645719375, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.934749126434326, "learning_rate": 5e-05, "loss": 0.8789, "mean_token_accuracy": 0.8623355627059937, "num_tokens": 356422584.0, "step": 9385 }, { "epoch": 1.193995674850528, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.9083991050720215, "learning_rate": 5e-05, "loss": 0.8095, "mean_token_accuracy": 0.8778362274169922, "num_tokens": 356457007.0, "step": 9386 }, { "epoch": 1.1941228851291183, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.673252105712891, "learning_rate": 5e-05, "loss": 0.7905, "mean_token_accuracy": 0.8765720129013062, "num_tokens": 356491837.0, "step": 9387 }, { "epoch": 1.1942500954077089, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.792972564697266, "learning_rate": 5e-05, "loss": 0.7918, "mean_token_accuracy": 0.8724594116210938, "num_tokens": 356527892.0, "step": 9388 }, { "epoch": 1.1943773056862994, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.550072193145752, "learning_rate": 5e-05, "loss": 0.7778, "mean_token_accuracy": 0.8666243553161621, "num_tokens": 356564310.0, "step": 9389 }, { "epoch": 1.19450451596489, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.591006755828857, "learning_rate": 5e-05, "loss": 0.7335, "mean_token_accuracy": 0.8805390000343323, "num_tokens": 356606476.0, "step": 9390 }, { "epoch": 1.1946317262434805, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.653726577758789, "learning_rate": 5e-05, "loss": 0.7511, "mean_token_accuracy": 0.8701686859130859, "num_tokens": 356646104.0, "step": 9391 }, { "epoch": 1.194758936522071, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.314113616943359, "learning_rate": 5e-05, "loss": 0.6733, "mean_token_accuracy": 0.8883262276649475, "num_tokens": 356687196.0, "step": 9392 }, { "epoch": 1.1948861468006615, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.866805553436279, "learning_rate": 5e-05, "loss": 0.7797, "mean_token_accuracy": 0.8715592622756958, "num_tokens": 356717634.0, "step": 9393 }, { "epoch": 1.195013357079252, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.99643087387085, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.858910322189331, "num_tokens": 356755494.0, "step": 9394 }, { "epoch": 1.1951405673578426, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.209410667419434, "learning_rate": 5e-05, "loss": 0.7726, "mean_token_accuracy": 0.8793920278549194, "num_tokens": 356800798.0, "step": 9395 }, { "epoch": 1.195267777636433, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.588581562042236, "learning_rate": 5e-05, "loss": 0.6302, "mean_token_accuracy": 0.8919435143470764, "num_tokens": 356837819.0, "step": 9396 }, { "epoch": 1.1953949879150234, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.777400970458984, "learning_rate": 5e-05, "loss": 0.8307, "mean_token_accuracy": 0.8795075416564941, "num_tokens": 356876371.0, "step": 9397 }, { "epoch": 1.195522198193614, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.16566801071167, "learning_rate": 5e-05, "loss": 0.7053, "mean_token_accuracy": 0.8734920024871826, "num_tokens": 356911427.0, "step": 9398 }, { "epoch": 1.1956494084722045, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.433684349060059, "learning_rate": 5e-05, "loss": 0.9034, "mean_token_accuracy": 0.8810209035873413, "num_tokens": 356951229.0, "step": 9399 }, { "epoch": 1.195776618750795, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.164029598236084, "learning_rate": 5e-05, "loss": 0.6705, "mean_token_accuracy": 0.8792737126350403, "num_tokens": 356988954.0, "step": 9400 }, { "epoch": 1.1959038290293855, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.809305191040039, "learning_rate": 5e-05, "loss": 1.042, "mean_token_accuracy": 0.8595840930938721, "num_tokens": 357027982.0, "step": 9401 }, { "epoch": 1.196031039307976, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.077539920806885, "learning_rate": 5e-05, "loss": 0.8651, "mean_token_accuracy": 0.8716492652893066, "num_tokens": 357061434.0, "step": 9402 }, { "epoch": 1.1961582495865666, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.2020902633667, "learning_rate": 5e-05, "loss": 0.9795, "mean_token_accuracy": 0.8700885772705078, "num_tokens": 357099255.0, "step": 9403 }, { "epoch": 1.196285459865157, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.153115272521973, "learning_rate": 5e-05, "loss": 0.9488, "mean_token_accuracy": 0.8805983066558838, "num_tokens": 357131917.0, "step": 9404 }, { "epoch": 1.1964126701437476, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.658972263336182, "learning_rate": 5e-05, "loss": 0.7767, "mean_token_accuracy": 0.8818714022636414, "num_tokens": 357171520.0, "step": 9405 }, { "epoch": 1.1965398804223382, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.017078399658203, "learning_rate": 5e-05, "loss": 0.9727, "mean_token_accuracy": 0.8739489912986755, "num_tokens": 357208517.0, "step": 9406 }, { "epoch": 1.1966670907009287, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.618268966674805, "learning_rate": 5e-05, "loss": 0.6949, "mean_token_accuracy": 0.8812160491943359, "num_tokens": 357245300.0, "step": 9407 }, { "epoch": 1.1967943009795192, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.790763854980469, "learning_rate": 5e-05, "loss": 0.9746, "mean_token_accuracy": 0.8876371383666992, "num_tokens": 357284765.0, "step": 9408 }, { "epoch": 1.1969215112581097, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 6.93086576461792, "learning_rate": 5e-05, "loss": 0.8565, "mean_token_accuracy": 0.8631983995437622, "num_tokens": 357322968.0, "step": 9409 }, { "epoch": 1.1970487215367003, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.793170928955078, "learning_rate": 5e-05, "loss": 1.0029, "mean_token_accuracy": 0.8779690265655518, "num_tokens": 357360546.0, "step": 9410 }, { "epoch": 1.1971759318152906, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 7.995601177215576, "learning_rate": 5e-05, "loss": 0.9232, "mean_token_accuracy": 0.8866750001907349, "num_tokens": 357394433.0, "step": 9411 }, { "epoch": 1.1973031420938811, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.76986312866211, "learning_rate": 5e-05, "loss": 0.9061, "mean_token_accuracy": 0.8708577156066895, "num_tokens": 357431498.0, "step": 9412 }, { "epoch": 1.1974303523724716, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.355751037597656, "learning_rate": 5e-05, "loss": 0.976, "mean_token_accuracy": 0.869823694229126, "num_tokens": 357466930.0, "step": 9413 }, { "epoch": 1.1975575626510622, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.940712928771973, "learning_rate": 5e-05, "loss": 0.8787, "mean_token_accuracy": 0.862881064414978, "num_tokens": 357509343.0, "step": 9414 }, { "epoch": 1.1976847729296527, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.121033668518066, "learning_rate": 5e-05, "loss": 0.8939, "mean_token_accuracy": 0.8726670742034912, "num_tokens": 357541980.0, "step": 9415 }, { "epoch": 1.1978119832082432, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.8763251304626465, "learning_rate": 5e-05, "loss": 0.8239, "mean_token_accuracy": 0.8778558969497681, "num_tokens": 357579380.0, "step": 9416 }, { "epoch": 1.1979391934868338, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.499528408050537, "learning_rate": 5e-05, "loss": 0.7541, "mean_token_accuracy": 0.8843241333961487, "num_tokens": 357614666.0, "step": 9417 }, { "epoch": 1.1980664037654243, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.132859230041504, "learning_rate": 5e-05, "loss": 0.82, "mean_token_accuracy": 0.8785421252250671, "num_tokens": 357648532.0, "step": 9418 }, { "epoch": 1.1981936140440148, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.034665584564209, "learning_rate": 5e-05, "loss": 0.7124, "mean_token_accuracy": 0.8772099018096924, "num_tokens": 357683585.0, "step": 9419 }, { "epoch": 1.1983208243226053, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.388528823852539, "learning_rate": 5e-05, "loss": 0.9, "mean_token_accuracy": 0.8529691696166992, "num_tokens": 357725580.0, "step": 9420 }, { "epoch": 1.1984480346011959, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.767948150634766, "learning_rate": 5e-05, "loss": 0.7115, "mean_token_accuracy": 0.869564950466156, "num_tokens": 357763017.0, "step": 9421 }, { "epoch": 1.1985752448797862, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.71226978302002, "learning_rate": 5e-05, "loss": 0.8471, "mean_token_accuracy": 0.8825797438621521, "num_tokens": 357797433.0, "step": 9422 }, { "epoch": 1.1987024551583767, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.566102504730225, "learning_rate": 5e-05, "loss": 0.7166, "mean_token_accuracy": 0.8702675700187683, "num_tokens": 357835348.0, "step": 9423 }, { "epoch": 1.1988296654369672, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.101350784301758, "learning_rate": 5e-05, "loss": 0.9403, "mean_token_accuracy": 0.8676707148551941, "num_tokens": 357877376.0, "step": 9424 }, { "epoch": 1.1989568757155578, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.6795430183410645, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.8735659122467041, "num_tokens": 357912983.0, "step": 9425 }, { "epoch": 1.1990840859941483, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.146773338317871, "learning_rate": 5e-05, "loss": 0.9121, "mean_token_accuracy": 0.8818646669387817, "num_tokens": 357948068.0, "step": 9426 }, { "epoch": 1.1992112962727388, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.222949981689453, "learning_rate": 5e-05, "loss": 0.8241, "mean_token_accuracy": 0.8778637647628784, "num_tokens": 357984019.0, "step": 9427 }, { "epoch": 1.1993385065513293, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.577066421508789, "learning_rate": 5e-05, "loss": 0.8797, "mean_token_accuracy": 0.8803043365478516, "num_tokens": 358022440.0, "step": 9428 }, { "epoch": 1.1994657168299199, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.732958793640137, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.8716537952423096, "num_tokens": 358061615.0, "step": 9429 }, { "epoch": 1.1995929271085104, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.894128799438477, "learning_rate": 5e-05, "loss": 0.8308, "mean_token_accuracy": 0.8711612820625305, "num_tokens": 358097042.0, "step": 9430 }, { "epoch": 1.199720137387101, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.853911399841309, "learning_rate": 5e-05, "loss": 0.8823, "mean_token_accuracy": 0.8672841787338257, "num_tokens": 358134067.0, "step": 9431 }, { "epoch": 1.1998473476656915, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.537709712982178, "learning_rate": 5e-05, "loss": 0.7826, "mean_token_accuracy": 0.8721449971199036, "num_tokens": 358178975.0, "step": 9432 }, { "epoch": 1.199974557944282, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.81735897064209, "learning_rate": 5e-05, "loss": 0.7959, "mean_token_accuracy": 0.8750119209289551, "num_tokens": 358212879.0, "step": 9433 }, { "epoch": 1.2001017682228725, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.391894340515137, "learning_rate": 5e-05, "loss": 0.7251, "mean_token_accuracy": 0.886397659778595, "num_tokens": 358245711.0, "step": 9434 }, { "epoch": 1.200228978501463, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.788812637329102, "learning_rate": 5e-05, "loss": 0.7409, "mean_token_accuracy": 0.884957492351532, "num_tokens": 358283073.0, "step": 9435 }, { "epoch": 1.2003561887800533, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.286891937255859, "learning_rate": 5e-05, "loss": 0.7205, "mean_token_accuracy": 0.8794512152671814, "num_tokens": 358320338.0, "step": 9436 }, { "epoch": 1.2004833990586439, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.8424859046936035, "learning_rate": 5e-05, "loss": 0.7743, "mean_token_accuracy": 0.8746328353881836, "num_tokens": 358358704.0, "step": 9437 }, { "epoch": 1.2006106093372344, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.122003078460693, "learning_rate": 5e-05, "loss": 0.7045, "mean_token_accuracy": 0.8782986402511597, "num_tokens": 358393512.0, "step": 9438 }, { "epoch": 1.200737819615825, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.9898529052734375, "learning_rate": 5e-05, "loss": 0.8171, "mean_token_accuracy": 0.8650574684143066, "num_tokens": 358435843.0, "step": 9439 }, { "epoch": 1.2008650298944155, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.94819974899292, "learning_rate": 5e-05, "loss": 0.6563, "mean_token_accuracy": 0.8884254693984985, "num_tokens": 358468918.0, "step": 9440 }, { "epoch": 1.200992240173006, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.160314559936523, "learning_rate": 5e-05, "loss": 0.8693, "mean_token_accuracy": 0.8517041206359863, "num_tokens": 358511275.0, "step": 9441 }, { "epoch": 1.2011194504515965, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.780274868011475, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.8644084930419922, "num_tokens": 358552665.0, "step": 9442 }, { "epoch": 1.201246660730187, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0005035400390625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.348671913146973, "learning_rate": 5e-05, "loss": 0.7918, "mean_token_accuracy": 0.8803608417510986, "num_tokens": 358584725.0, "step": 9443 }, { "epoch": 1.2013738710087776, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.630024433135986, "learning_rate": 5e-05, "loss": 0.6793, "mean_token_accuracy": 0.8792819976806641, "num_tokens": 358620026.0, "step": 9444 }, { "epoch": 1.201501081287368, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.619834899902344, "learning_rate": 5e-05, "loss": 0.8522, "mean_token_accuracy": 0.8710106611251831, "num_tokens": 358657386.0, "step": 9445 }, { "epoch": 1.2016282915659584, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.512400150299072, "learning_rate": 5e-05, "loss": 0.6985, "mean_token_accuracy": 0.8746682405471802, "num_tokens": 358695069.0, "step": 9446 }, { "epoch": 1.201755501844549, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.818825721740723, "learning_rate": 5e-05, "loss": 0.8655, "mean_token_accuracy": 0.8802909851074219, "num_tokens": 358728091.0, "step": 9447 }, { "epoch": 1.2018827121231395, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.669595241546631, "learning_rate": 5e-05, "loss": 0.7592, "mean_token_accuracy": 0.8615518808364868, "num_tokens": 358767930.0, "step": 9448 }, { "epoch": 1.20200992240173, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.662100791931152, "learning_rate": 5e-05, "loss": 0.9042, "mean_token_accuracy": 0.8692294359207153, "num_tokens": 358803705.0, "step": 9449 }, { "epoch": 1.2021371326803205, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.154580593109131, "learning_rate": 5e-05, "loss": 0.7837, "mean_token_accuracy": 0.8733394145965576, "num_tokens": 358840730.0, "step": 9450 }, { "epoch": 1.202264342958911, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.031990051269531, "learning_rate": 5e-05, "loss": 0.8529, "mean_token_accuracy": 0.867194414138794, "num_tokens": 358881457.0, "step": 9451 }, { "epoch": 1.2023915532375016, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.704379558563232, "learning_rate": 5e-05, "loss": 0.8437, "mean_token_accuracy": 0.8679067492485046, "num_tokens": 358917034.0, "step": 9452 }, { "epoch": 1.202518763516092, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.313846111297607, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.8658870458602905, "num_tokens": 358957851.0, "step": 9453 }, { "epoch": 1.2026459737946826, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.070435523986816, "learning_rate": 5e-05, "loss": 0.8436, "mean_token_accuracy": 0.8646260499954224, "num_tokens": 358999816.0, "step": 9454 }, { "epoch": 1.2027731840732732, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.958177089691162, "learning_rate": 5e-05, "loss": 0.7724, "mean_token_accuracy": 0.8527749180793762, "num_tokens": 359041067.0, "step": 9455 }, { "epoch": 1.2029003943518637, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.154273986816406, "learning_rate": 5e-05, "loss": 0.8178, "mean_token_accuracy": 0.8701357245445251, "num_tokens": 359080495.0, "step": 9456 }, { "epoch": 1.2030276046304542, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.953860759735107, "learning_rate": 5e-05, "loss": 0.6811, "mean_token_accuracy": 0.8841339349746704, "num_tokens": 359118245.0, "step": 9457 }, { "epoch": 1.2031548149090447, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.039459228515625, "learning_rate": 5e-05, "loss": 0.8368, "mean_token_accuracy": 0.862716555595398, "num_tokens": 359159840.0, "step": 9458 }, { "epoch": 1.2032820251876353, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.072681427001953, "learning_rate": 5e-05, "loss": 0.7302, "mean_token_accuracy": 0.8676952123641968, "num_tokens": 359198917.0, "step": 9459 }, { "epoch": 1.2034092354662256, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.885256290435791, "learning_rate": 5e-05, "loss": 0.7739, "mean_token_accuracy": 0.8712978959083557, "num_tokens": 359231904.0, "step": 9460 }, { "epoch": 1.203536445744816, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.170904636383057, "learning_rate": 5e-05, "loss": 0.766, "mean_token_accuracy": 0.8593022227287292, "num_tokens": 359271713.0, "step": 9461 }, { "epoch": 1.2036636560234066, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.764925003051758, "learning_rate": 5e-05, "loss": 0.7819, "mean_token_accuracy": 0.8697487711906433, "num_tokens": 359312548.0, "step": 9462 }, { "epoch": 1.2037908663019972, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.229098796844482, "learning_rate": 5e-05, "loss": 0.6938, "mean_token_accuracy": 0.8814975023269653, "num_tokens": 359348171.0, "step": 9463 }, { "epoch": 1.2039180765805877, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.7031660079956055, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.8628454208374023, "num_tokens": 359383863.0, "step": 9464 }, { "epoch": 1.2040452868591782, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.249213218688965, "learning_rate": 5e-05, "loss": 0.7313, "mean_token_accuracy": 0.8681262135505676, "num_tokens": 359421595.0, "step": 9465 }, { "epoch": 1.2041724971377687, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.652885913848877, "learning_rate": 5e-05, "loss": 0.7094, "mean_token_accuracy": 0.8812724351882935, "num_tokens": 359457889.0, "step": 9466 }, { "epoch": 1.2042997074163593, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.220725059509277, "learning_rate": 5e-05, "loss": 0.7347, "mean_token_accuracy": 0.8655166625976562, "num_tokens": 359496938.0, "step": 9467 }, { "epoch": 1.2044269176949498, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.705944538116455, "learning_rate": 5e-05, "loss": 0.7722, "mean_token_accuracy": 0.8637512922286987, "num_tokens": 359536124.0, "step": 9468 }, { "epoch": 1.2045541279735403, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.110401630401611, "learning_rate": 5e-05, "loss": 0.7425, "mean_token_accuracy": 0.863048791885376, "num_tokens": 359573525.0, "step": 9469 }, { "epoch": 1.2046813382521309, "ewc_loss": 0.375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.839997291564941, "learning_rate": 5e-05, "loss": 0.7346, "mean_token_accuracy": 0.8743658661842346, "num_tokens": 359614960.0, "step": 9470 }, { "epoch": 1.2048085485307212, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.910775661468506, "learning_rate": 5e-05, "loss": 0.7116, "mean_token_accuracy": 0.8660072088241577, "num_tokens": 359653367.0, "step": 9471 }, { "epoch": 1.2049357588093117, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.125733375549316, "learning_rate": 5e-05, "loss": 0.8349, "mean_token_accuracy": 0.8524839878082275, "num_tokens": 359692983.0, "step": 9472 }, { "epoch": 1.2050629690879022, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.5697340965271, "learning_rate": 5e-05, "loss": 0.707, "mean_token_accuracy": 0.8630839586257935, "num_tokens": 359729487.0, "step": 9473 }, { "epoch": 1.2051901793664928, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.623000144958496, "learning_rate": 5e-05, "loss": 0.7878, "mean_token_accuracy": 0.8801558017730713, "num_tokens": 359764805.0, "step": 9474 }, { "epoch": 1.2053173896450833, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.10264778137207, "learning_rate": 5e-05, "loss": 0.6933, "mean_token_accuracy": 0.861083984375, "num_tokens": 359803014.0, "step": 9475 }, { "epoch": 1.2054445999236738, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.358389854431152, "learning_rate": 5e-05, "loss": 0.9114, "mean_token_accuracy": 0.8746803998947144, "num_tokens": 359837866.0, "step": 9476 }, { "epoch": 1.2055718102022643, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 5.850075721740723, "learning_rate": 5e-05, "loss": 0.6525, "mean_token_accuracy": 0.885307788848877, "num_tokens": 359874052.0, "step": 9477 }, { "epoch": 1.2056990204808549, "ewc_loss": 0.625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.991560935974121, "learning_rate": 5e-05, "loss": 0.9425, "mean_token_accuracy": 0.8883672952651978, "num_tokens": 359915746.0, "step": 9478 }, { "epoch": 1.2058262307594454, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 6.707077980041504, "learning_rate": 5e-05, "loss": 0.7849, "mean_token_accuracy": 0.8848527669906616, "num_tokens": 359945857.0, "step": 9479 }, { "epoch": 1.205953441038036, "ewc_loss": 0.625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.450959205627441, "learning_rate": 5e-05, "loss": 0.9797, "mean_token_accuracy": 0.8783565163612366, "num_tokens": 359988455.0, "step": 9480 }, { "epoch": 1.2060806513166265, "ewc_loss": 0.625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.294917106628418, "learning_rate": 5e-05, "loss": 0.9732, "mean_token_accuracy": 0.8758406639099121, "num_tokens": 360030610.0, "step": 9481 }, { "epoch": 1.206207861595217, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.3080363273620605, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.8637000322341919, "num_tokens": 360068919.0, "step": 9482 }, { "epoch": 1.2063350718738075, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.044133186340332, "learning_rate": 5e-05, "loss": 1.0434, "mean_token_accuracy": 0.8711745738983154, "num_tokens": 360099073.0, "step": 9483 }, { "epoch": 1.206462282152398, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 5.954458236694336, "learning_rate": 5e-05, "loss": 0.7215, "mean_token_accuracy": 0.873593807220459, "num_tokens": 360141947.0, "step": 9484 }, { "epoch": 1.2065894924309883, "ewc_loss": 0.9921875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000942230224609375, "grad_norm": 12.34195327758789, "learning_rate": 5e-05, "loss": 1.3852, "mean_token_accuracy": 0.8691704869270325, "num_tokens": 360179435.0, "step": 9485 }, { "epoch": 1.2067167027095789, "ewc_loss": 0.828125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0007781982421875, "grad_norm": 8.98508358001709, "learning_rate": 5e-05, "loss": 1.1793, "mean_token_accuracy": 0.8777874708175659, "num_tokens": 360221254.0, "step": 9486 }, { "epoch": 1.2068439129881694, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 11.75992202758789, "learning_rate": 5e-05, "loss": 1.1387, "mean_token_accuracy": 0.8809335231781006, "num_tokens": 360261657.0, "step": 9487 }, { "epoch": 1.20697112326676, "ewc_loss": 0.92578125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000873565673828125, "grad_norm": 9.868919372558594, "learning_rate": 5e-05, "loss": 1.2786, "mean_token_accuracy": 0.8756604194641113, "num_tokens": 360294408.0, "step": 9488 }, { "epoch": 1.2070983335453505, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 11.12851619720459, "learning_rate": 5e-05, "loss": 1.109, "mean_token_accuracy": 0.8810553550720215, "num_tokens": 360332275.0, "step": 9489 }, { "epoch": 1.207225543823941, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 8.95198917388916, "learning_rate": 5e-05, "loss": 1.0035, "mean_token_accuracy": 0.879608154296875, "num_tokens": 360372381.0, "step": 9490 }, { "epoch": 1.2073527541025315, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.787571907043457, "learning_rate": 5e-05, "loss": 1.1564, "mean_token_accuracy": 0.8676424026489258, "num_tokens": 360404066.0, "step": 9491 }, { "epoch": 1.207479964381122, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.601861953735352, "learning_rate": 5e-05, "loss": 0.9278, "mean_token_accuracy": 0.8800644874572754, "num_tokens": 360434457.0, "step": 9492 }, { "epoch": 1.2076071746597126, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 10.58641529083252, "learning_rate": 5e-05, "loss": 1.1302, "mean_token_accuracy": 0.873864471912384, "num_tokens": 360467905.0, "step": 9493 }, { "epoch": 1.207734384938303, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.683974266052246, "learning_rate": 5e-05, "loss": 0.9795, "mean_token_accuracy": 0.875312089920044, "num_tokens": 360505190.0, "step": 9494 }, { "epoch": 1.2078615952168934, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.286592483520508, "learning_rate": 5e-05, "loss": 1.0478, "mean_token_accuracy": 0.8738833665847778, "num_tokens": 360539191.0, "step": 9495 }, { "epoch": 1.207988805495484, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.722405433654785, "learning_rate": 5e-05, "loss": 1.0312, "mean_token_accuracy": 0.8606705665588379, "num_tokens": 360569508.0, "step": 9496 }, { "epoch": 1.2081160157740745, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.838220596313477, "learning_rate": 5e-05, "loss": 0.995, "mean_token_accuracy": 0.8814284205436707, "num_tokens": 360604669.0, "step": 9497 }, { "epoch": 1.208243226052665, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.408278465270996, "learning_rate": 5e-05, "loss": 0.8548, "mean_token_accuracy": 0.8949576020240784, "num_tokens": 360639009.0, "step": 9498 }, { "epoch": 1.2083704363312555, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.655135154724121, "learning_rate": 5e-05, "loss": 0.9444, "mean_token_accuracy": 0.8874193429946899, "num_tokens": 360673857.0, "step": 9499 }, { "epoch": 1.208497646609846, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.00415325164795, "learning_rate": 5e-05, "loss": 0.9026, "mean_token_accuracy": 0.8636918067932129, "num_tokens": 360717553.0, "step": 9500 }, { "epoch": 1.2086248568884366, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.835986137390137, "learning_rate": 5e-05, "loss": 1.0407, "mean_token_accuracy": 0.8760569095611572, "num_tokens": 360761013.0, "step": 9501 }, { "epoch": 1.208752067167027, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.952322483062744, "learning_rate": 5e-05, "loss": 0.8725, "mean_token_accuracy": 0.873388409614563, "num_tokens": 360804302.0, "step": 9502 }, { "epoch": 1.2088792774456176, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.9779634475708, "learning_rate": 5e-05, "loss": 0.9687, "mean_token_accuracy": 0.8942806124687195, "num_tokens": 360837045.0, "step": 9503 }, { "epoch": 1.2090064877242082, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.324956893920898, "learning_rate": 5e-05, "loss": 0.9135, "mean_token_accuracy": 0.8720470070838928, "num_tokens": 360876243.0, "step": 9504 }, { "epoch": 1.2091336980027987, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.621834754943848, "learning_rate": 5e-05, "loss": 1.0021, "mean_token_accuracy": 0.8712475299835205, "num_tokens": 360915112.0, "step": 9505 }, { "epoch": 1.2092609082813892, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.50568962097168, "learning_rate": 5e-05, "loss": 0.9117, "mean_token_accuracy": 0.8762565851211548, "num_tokens": 360951856.0, "step": 9506 }, { "epoch": 1.2093881185599797, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.981529235839844, "learning_rate": 5e-05, "loss": 0.9306, "mean_token_accuracy": 0.8808959722518921, "num_tokens": 360987243.0, "step": 9507 }, { "epoch": 1.2095153288385703, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.263008117675781, "learning_rate": 5e-05, "loss": 0.857, "mean_token_accuracy": 0.8795270919799805, "num_tokens": 361025099.0, "step": 9508 }, { "epoch": 1.2096425391171606, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.615384101867676, "learning_rate": 5e-05, "loss": 0.9544, "mean_token_accuracy": 0.8568308353424072, "num_tokens": 361055232.0, "step": 9509 }, { "epoch": 1.209769749395751, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.757969856262207, "learning_rate": 5e-05, "loss": 0.8108, "mean_token_accuracy": 0.8737884163856506, "num_tokens": 361092500.0, "step": 9510 }, { "epoch": 1.2098969596743416, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.76504135131836, "learning_rate": 5e-05, "loss": 0.8999, "mean_token_accuracy": 0.8774113655090332, "num_tokens": 361127044.0, "step": 9511 }, { "epoch": 1.2100241699529322, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.312468528747559, "learning_rate": 5e-05, "loss": 0.7288, "mean_token_accuracy": 0.8929756879806519, "num_tokens": 361163065.0, "step": 9512 }, { "epoch": 1.2101513802315227, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.151957511901855, "learning_rate": 5e-05, "loss": 0.9742, "mean_token_accuracy": 0.8658401966094971, "num_tokens": 361204465.0, "step": 9513 }, { "epoch": 1.2102785905101132, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.345832347869873, "learning_rate": 5e-05, "loss": 0.8478, "mean_token_accuracy": 0.8512002229690552, "num_tokens": 361243788.0, "step": 9514 }, { "epoch": 1.2104058007887037, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.267254829406738, "learning_rate": 5e-05, "loss": 0.9763, "mean_token_accuracy": 0.8782559633255005, "num_tokens": 361281427.0, "step": 9515 }, { "epoch": 1.2105330110672943, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.801745414733887, "learning_rate": 5e-05, "loss": 0.8911, "mean_token_accuracy": 0.8600373864173889, "num_tokens": 361324695.0, "step": 9516 }, { "epoch": 1.2106602213458848, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.768444061279297, "learning_rate": 5e-05, "loss": 0.9455, "mean_token_accuracy": 0.8704097867012024, "num_tokens": 361370446.0, "step": 9517 }, { "epoch": 1.2107874316244753, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.069290161132812, "learning_rate": 5e-05, "loss": 0.8608, "mean_token_accuracy": 0.8710801005363464, "num_tokens": 361414622.0, "step": 9518 }, { "epoch": 1.2109146419030659, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.080781936645508, "learning_rate": 5e-05, "loss": 0.8319, "mean_token_accuracy": 0.8845434188842773, "num_tokens": 361452363.0, "step": 9519 }, { "epoch": 1.2110418521816562, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.914089679718018, "learning_rate": 5e-05, "loss": 0.8683, "mean_token_accuracy": 0.8581133484840393, "num_tokens": 361484195.0, "step": 9520 }, { "epoch": 1.2111690624602467, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.849266052246094, "learning_rate": 5e-05, "loss": 0.8154, "mean_token_accuracy": 0.8730669617652893, "num_tokens": 361523604.0, "step": 9521 }, { "epoch": 1.2112962727388372, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.546331405639648, "learning_rate": 5e-05, "loss": 0.8091, "mean_token_accuracy": 0.8578608632087708, "num_tokens": 361563723.0, "step": 9522 }, { "epoch": 1.2114234830174277, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.991483211517334, "learning_rate": 5e-05, "loss": 0.7983, "mean_token_accuracy": 0.8740695118904114, "num_tokens": 361602578.0, "step": 9523 }, { "epoch": 1.2115506932960183, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.221081256866455, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.8704403638839722, "num_tokens": 361644255.0, "step": 9524 }, { "epoch": 1.2116779035746088, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.186787605285645, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.8729829788208008, "num_tokens": 361683424.0, "step": 9525 }, { "epoch": 1.2118051138531993, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.045799255371094, "learning_rate": 5e-05, "loss": 0.6774, "mean_token_accuracy": 0.8835989236831665, "num_tokens": 361720965.0, "step": 9526 }, { "epoch": 1.2119323241317899, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.31261920928955, "learning_rate": 5e-05, "loss": 0.8749, "mean_token_accuracy": 0.8598760366439819, "num_tokens": 361753900.0, "step": 9527 }, { "epoch": 1.2120595344103804, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.970864772796631, "learning_rate": 5e-05, "loss": 0.7124, "mean_token_accuracy": 0.8742642402648926, "num_tokens": 361793072.0, "step": 9528 }, { "epoch": 1.212186744688971, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.405176162719727, "learning_rate": 5e-05, "loss": 0.8911, "mean_token_accuracy": 0.8560091853141785, "num_tokens": 361831082.0, "step": 9529 }, { "epoch": 1.2123139549675614, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.014432430267334, "learning_rate": 5e-05, "loss": 0.6903, "mean_token_accuracy": 0.8776916861534119, "num_tokens": 361864834.0, "step": 9530 }, { "epoch": 1.212441165246152, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.314346313476562, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.8728373050689697, "num_tokens": 361905567.0, "step": 9531 }, { "epoch": 1.2125683755247425, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.255328178405762, "learning_rate": 5e-05, "loss": 0.7168, "mean_token_accuracy": 0.8829663991928101, "num_tokens": 361945718.0, "step": 9532 }, { "epoch": 1.212695585803333, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.944845676422119, "learning_rate": 5e-05, "loss": 0.8051, "mean_token_accuracy": 0.8666715621948242, "num_tokens": 361985317.0, "step": 9533 }, { "epoch": 1.2128227960819233, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.567570209503174, "learning_rate": 5e-05, "loss": 0.75, "mean_token_accuracy": 0.8754730224609375, "num_tokens": 362020725.0, "step": 9534 }, { "epoch": 1.2129500063605139, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.453268527984619, "learning_rate": 5e-05, "loss": 0.7683, "mean_token_accuracy": 0.861875593662262, "num_tokens": 362054398.0, "step": 9535 }, { "epoch": 1.2130772166391044, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.877849102020264, "learning_rate": 5e-05, "loss": 0.8008, "mean_token_accuracy": 0.8586554527282715, "num_tokens": 362092075.0, "step": 9536 }, { "epoch": 1.213204426917695, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.053325176239014, "learning_rate": 5e-05, "loss": 0.6542, "mean_token_accuracy": 0.8926137685775757, "num_tokens": 362128286.0, "step": 9537 }, { "epoch": 1.2133316371962855, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.152572631835938, "learning_rate": 5e-05, "loss": 0.8201, "mean_token_accuracy": 0.8652540445327759, "num_tokens": 362169965.0, "step": 9538 }, { "epoch": 1.213458847474876, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.791722297668457, "learning_rate": 5e-05, "loss": 0.6985, "mean_token_accuracy": 0.8708685636520386, "num_tokens": 362211638.0, "step": 9539 }, { "epoch": 1.2135860577534665, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.399320602416992, "learning_rate": 5e-05, "loss": 0.7555, "mean_token_accuracy": 0.8888611793518066, "num_tokens": 362241597.0, "step": 9540 }, { "epoch": 1.213713268032057, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.64179801940918, "learning_rate": 5e-05, "loss": 0.6964, "mean_token_accuracy": 0.8706656694412231, "num_tokens": 362280028.0, "step": 9541 }, { "epoch": 1.2138404783106476, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.5450439453125, "learning_rate": 5e-05, "loss": 0.8352, "mean_token_accuracy": 0.8726300001144409, "num_tokens": 362323611.0, "step": 9542 }, { "epoch": 1.213967688589238, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.740029811859131, "learning_rate": 5e-05, "loss": 0.6914, "mean_token_accuracy": 0.8741017580032349, "num_tokens": 362359097.0, "step": 9543 }, { "epoch": 1.2140948988678284, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.367931365966797, "learning_rate": 5e-05, "loss": 0.8291, "mean_token_accuracy": 0.870521068572998, "num_tokens": 362402175.0, "step": 9544 }, { "epoch": 1.214222109146419, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.12130880355835, "learning_rate": 5e-05, "loss": 0.734, "mean_token_accuracy": 0.8727818727493286, "num_tokens": 362442581.0, "step": 9545 }, { "epoch": 1.2143493194250095, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.839230060577393, "learning_rate": 5e-05, "loss": 0.7515, "mean_token_accuracy": 0.8832826614379883, "num_tokens": 362480691.0, "step": 9546 }, { "epoch": 1.2144765297036, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.634574890136719, "learning_rate": 5e-05, "loss": 0.7521, "mean_token_accuracy": 0.8747860193252563, "num_tokens": 362522864.0, "step": 9547 }, { "epoch": 1.2146037399821905, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.227515697479248, "learning_rate": 5e-05, "loss": 0.7139, "mean_token_accuracy": 0.8749383091926575, "num_tokens": 362564431.0, "step": 9548 }, { "epoch": 1.214730950260781, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.06614875793457, "learning_rate": 5e-05, "loss": 0.7468, "mean_token_accuracy": 0.8854952454566956, "num_tokens": 362599670.0, "step": 9549 }, { "epoch": 1.2148581605393716, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.8645710945129395, "learning_rate": 5e-05, "loss": 0.7176, "mean_token_accuracy": 0.8702589273452759, "num_tokens": 362637588.0, "step": 9550 }, { "epoch": 1.214985370817962, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.243351936340332, "learning_rate": 5e-05, "loss": 0.7788, "mean_token_accuracy": 0.8752409219741821, "num_tokens": 362674679.0, "step": 9551 }, { "epoch": 1.2151125810965526, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.8235673904418945, "learning_rate": 5e-05, "loss": 0.699, "mean_token_accuracy": 0.8677072525024414, "num_tokens": 362705840.0, "step": 9552 }, { "epoch": 1.2152397913751432, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.163724899291992, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8707586526870728, "num_tokens": 362739880.0, "step": 9553 }, { "epoch": 1.2153670016537337, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.010584354400635, "learning_rate": 5e-05, "loss": 0.7019, "mean_token_accuracy": 0.8800079822540283, "num_tokens": 362779101.0, "step": 9554 }, { "epoch": 1.2154942119323242, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.905527591705322, "learning_rate": 5e-05, "loss": 0.745, "mean_token_accuracy": 0.878400444984436, "num_tokens": 362821034.0, "step": 9555 }, { "epoch": 1.2156214222109147, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.279540061950684, "learning_rate": 5e-05, "loss": 0.6874, "mean_token_accuracy": 0.8829014301300049, "num_tokens": 362856895.0, "step": 9556 }, { "epoch": 1.2157486324895053, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.592299938201904, "learning_rate": 5e-05, "loss": 0.7429, "mean_token_accuracy": 0.8717498779296875, "num_tokens": 362894955.0, "step": 9557 }, { "epoch": 1.2158758427680956, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.530935287475586, "learning_rate": 5e-05, "loss": 0.7475, "mean_token_accuracy": 0.8673374056816101, "num_tokens": 362930397.0, "step": 9558 }, { "epoch": 1.216003053046686, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.315331935882568, "learning_rate": 5e-05, "loss": 0.6639, "mean_token_accuracy": 0.8898406028747559, "num_tokens": 362970728.0, "step": 9559 }, { "epoch": 1.2161302633252766, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.705015659332275, "learning_rate": 5e-05, "loss": 0.6938, "mean_token_accuracy": 0.8870501518249512, "num_tokens": 363007803.0, "step": 9560 }, { "epoch": 1.2162574736038672, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.188049793243408, "learning_rate": 5e-05, "loss": 0.7047, "mean_token_accuracy": 0.8723418712615967, "num_tokens": 363047815.0, "step": 9561 }, { "epoch": 1.2163846838824577, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.725253105163574, "learning_rate": 5e-05, "loss": 0.7087, "mean_token_accuracy": 0.8811671733856201, "num_tokens": 363084156.0, "step": 9562 }, { "epoch": 1.2165118941610482, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.225067615509033, "learning_rate": 5e-05, "loss": 0.7166, "mean_token_accuracy": 0.8689879775047302, "num_tokens": 363125390.0, "step": 9563 }, { "epoch": 1.2166391044396387, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.619912624359131, "learning_rate": 5e-05, "loss": 0.7194, "mean_token_accuracy": 0.8744491338729858, "num_tokens": 363164210.0, "step": 9564 }, { "epoch": 1.2167663147182293, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.348275661468506, "learning_rate": 5e-05, "loss": 0.7531, "mean_token_accuracy": 0.8540970087051392, "num_tokens": 363200356.0, "step": 9565 }, { "epoch": 1.2168935249968198, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.457630157470703, "learning_rate": 5e-05, "loss": 0.7644, "mean_token_accuracy": 0.8550333380699158, "num_tokens": 363239622.0, "step": 9566 }, { "epoch": 1.2170207352754103, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.474762439727783, "learning_rate": 5e-05, "loss": 0.7177, "mean_token_accuracy": 0.8754917979240417, "num_tokens": 363277348.0, "step": 9567 }, { "epoch": 1.2171479455540009, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.319937229156494, "learning_rate": 5e-05, "loss": 0.6979, "mean_token_accuracy": 0.874053418636322, "num_tokens": 363310077.0, "step": 9568 }, { "epoch": 1.2172751558325912, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.589312553405762, "learning_rate": 5e-05, "loss": 0.8089, "mean_token_accuracy": 0.8407713174819946, "num_tokens": 363347919.0, "step": 9569 }, { "epoch": 1.2174023661111817, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 7.183781147003174, "learning_rate": 5e-05, "loss": 0.681, "mean_token_accuracy": 0.8744158744812012, "num_tokens": 363390818.0, "step": 9570 }, { "epoch": 1.2175295763897722, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.779620170593262, "learning_rate": 5e-05, "loss": 0.8143, "mean_token_accuracy": 0.8479738831520081, "num_tokens": 363424028.0, "step": 9571 }, { "epoch": 1.2176567866683627, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.934235572814941, "learning_rate": 5e-05, "loss": 0.671, "mean_token_accuracy": 0.8709636926651001, "num_tokens": 363456862.0, "step": 9572 }, { "epoch": 1.2177839969469533, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 8.09903621673584, "learning_rate": 5e-05, "loss": 0.7477, "mean_token_accuracy": 0.8749520778656006, "num_tokens": 363498538.0, "step": 9573 }, { "epoch": 1.2179112072255438, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 6.525078773498535, "learning_rate": 5e-05, "loss": 0.6327, "mean_token_accuracy": 0.8782694339752197, "num_tokens": 363533050.0, "step": 9574 }, { "epoch": 1.2180384175041343, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.700501441955566, "learning_rate": 5e-05, "loss": 0.8226, "mean_token_accuracy": 0.8703805208206177, "num_tokens": 363574151.0, "step": 9575 }, { "epoch": 1.2181656277827249, "ewc_loss": 0.259765625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00020885467529296875, "grad_norm": 5.8631415367126465, "learning_rate": 5e-05, "loss": 0.6337, "mean_token_accuracy": 0.8689274787902832, "num_tokens": 363615829.0, "step": 9576 }, { "epoch": 1.2182928380613154, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.88090991973877, "learning_rate": 5e-05, "loss": 0.9346, "mean_token_accuracy": 0.8720229268074036, "num_tokens": 363649974.0, "step": 9577 }, { "epoch": 1.218420048339906, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00021648406982421875, "grad_norm": 5.2143378257751465, "learning_rate": 5e-05, "loss": 0.6339, "mean_token_accuracy": 0.8753300905227661, "num_tokens": 363691196.0, "step": 9578 }, { "epoch": 1.2185472586184964, "ewc_loss": 0.8515625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 11.59531307220459, "learning_rate": 5e-05, "loss": 1.1795, "mean_token_accuracy": 0.8845500946044922, "num_tokens": 363726496.0, "step": 9579 }, { "epoch": 1.218674468897087, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.032824993133545, "learning_rate": 5e-05, "loss": 0.9675, "mean_token_accuracy": 0.8618611693382263, "num_tokens": 363769119.0, "step": 9580 }, { "epoch": 1.2188016791756775, "ewc_loss": 0.9609375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000911712646484375, "grad_norm": 12.257874488830566, "learning_rate": 5e-05, "loss": 1.389, "mean_token_accuracy": 0.8566322922706604, "num_tokens": 363803363.0, "step": 9581 }, { "epoch": 1.218928889454268, "ewc_loss": 1.0390625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00098419189453125, "grad_norm": 9.900827407836914, "learning_rate": 5e-05, "loss": 1.3907, "mean_token_accuracy": 0.8755266666412354, "num_tokens": 363842602.0, "step": 9582 }, { "epoch": 1.2190560997328583, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.677175521850586, "learning_rate": 5e-05, "loss": 1.0334, "mean_token_accuracy": 0.871263861656189, "num_tokens": 363882758.0, "step": 9583 }, { "epoch": 1.2191833100114489, "ewc_loss": 0.90234375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 9.680294036865234, "learning_rate": 5e-05, "loss": 1.2294, "mean_token_accuracy": 0.8857792019844055, "num_tokens": 363923738.0, "step": 9584 }, { "epoch": 1.2193105202900394, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.399115562438965, "learning_rate": 5e-05, "loss": 0.9742, "mean_token_accuracy": 0.8791366219520569, "num_tokens": 363957999.0, "step": 9585 }, { "epoch": 1.21943773056863, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 9.700220108032227, "learning_rate": 5e-05, "loss": 1.0678, "mean_token_accuracy": 0.8742412328720093, "num_tokens": 363992653.0, "step": 9586 }, { "epoch": 1.2195649408472204, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.26025676727295, "learning_rate": 5e-05, "loss": 0.9062, "mean_token_accuracy": 0.8863742351531982, "num_tokens": 364027376.0, "step": 9587 }, { "epoch": 1.219692151125811, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.994744300842285, "learning_rate": 5e-05, "loss": 1.0226, "mean_token_accuracy": 0.8729335069656372, "num_tokens": 364063299.0, "step": 9588 }, { "epoch": 1.2198193614044015, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.01362419128418, "learning_rate": 5e-05, "loss": 1.033, "mean_token_accuracy": 0.8468692898750305, "num_tokens": 364105724.0, "step": 9589 }, { "epoch": 1.219946571682992, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.805118560791016, "learning_rate": 5e-05, "loss": 1.0381, "mean_token_accuracy": 0.8643743395805359, "num_tokens": 364140427.0, "step": 9590 }, { "epoch": 1.2200737819615826, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.18210220336914, "learning_rate": 5e-05, "loss": 0.9957, "mean_token_accuracy": 0.8620591163635254, "num_tokens": 364185262.0, "step": 9591 }, { "epoch": 1.220200992240173, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.933252334594727, "learning_rate": 5e-05, "loss": 0.9171, "mean_token_accuracy": 0.8735673427581787, "num_tokens": 364219006.0, "step": 9592 }, { "epoch": 1.2203282025187634, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.53879165649414, "learning_rate": 5e-05, "loss": 0.9634, "mean_token_accuracy": 0.8781850337982178, "num_tokens": 364259699.0, "step": 9593 }, { "epoch": 1.220455412797354, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.5190534591674805, "learning_rate": 5e-05, "loss": 0.7647, "mean_token_accuracy": 0.8783489465713501, "num_tokens": 364297759.0, "step": 9594 }, { "epoch": 1.2205826230759445, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.56447982788086, "learning_rate": 5e-05, "loss": 1.0255, "mean_token_accuracy": 0.8683866858482361, "num_tokens": 364338683.0, "step": 9595 }, { "epoch": 1.220709833354535, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.510257720947266, "learning_rate": 5e-05, "loss": 0.7346, "mean_token_accuracy": 0.872920036315918, "num_tokens": 364377587.0, "step": 9596 }, { "epoch": 1.2208370436331255, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 11.3226900100708, "learning_rate": 5e-05, "loss": 1.1949, "mean_token_accuracy": 0.8671749234199524, "num_tokens": 364416740.0, "step": 9597 }, { "epoch": 1.220964253911716, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.008524894714355, "learning_rate": 5e-05, "loss": 1.0158, "mean_token_accuracy": 0.8623931407928467, "num_tokens": 364458695.0, "step": 9598 }, { "epoch": 1.2210914641903066, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 11.609587669372559, "learning_rate": 5e-05, "loss": 1.1962, "mean_token_accuracy": 0.8639105558395386, "num_tokens": 364501395.0, "step": 9599 }, { "epoch": 1.221218674468897, "ewc_loss": 0.87109375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 9.726020812988281, "learning_rate": 5e-05, "loss": 1.224, "mean_token_accuracy": 0.878750205039978, "num_tokens": 364534126.0, "step": 9600 }, { "epoch": 1.2213458847474876, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.450488090515137, "learning_rate": 5e-05, "loss": 0.9816, "mean_token_accuracy": 0.8843615651130676, "num_tokens": 364573776.0, "step": 9601 }, { "epoch": 1.2214730950260781, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 9.222153663635254, "learning_rate": 5e-05, "loss": 1.0885, "mean_token_accuracy": 0.8689910769462585, "num_tokens": 364617900.0, "step": 9602 }, { "epoch": 1.2216003053046687, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.82500171661377, "learning_rate": 5e-05, "loss": 0.9942, "mean_token_accuracy": 0.8817283511161804, "num_tokens": 364658992.0, "step": 9603 }, { "epoch": 1.2217275155832592, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.378693580627441, "learning_rate": 5e-05, "loss": 0.8843, "mean_token_accuracy": 0.8844475746154785, "num_tokens": 364695224.0, "step": 9604 }, { "epoch": 1.2218547258618497, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.713850975036621, "learning_rate": 5e-05, "loss": 1.0181, "mean_token_accuracy": 0.8749801516532898, "num_tokens": 364731476.0, "step": 9605 }, { "epoch": 1.2219819361404403, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.595818996429443, "learning_rate": 5e-05, "loss": 0.8122, "mean_token_accuracy": 0.8740369081497192, "num_tokens": 364774129.0, "step": 9606 }, { "epoch": 1.2221091464190306, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.442394256591797, "learning_rate": 5e-05, "loss": 1.0718, "mean_token_accuracy": 0.8743866682052612, "num_tokens": 364808097.0, "step": 9607 }, { "epoch": 1.222236356697621, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.911486625671387, "learning_rate": 5e-05, "loss": 0.852, "mean_token_accuracy": 0.8787589073181152, "num_tokens": 364847592.0, "step": 9608 }, { "epoch": 1.2223635669762116, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.659088134765625, "learning_rate": 5e-05, "loss": 1.0741, "mean_token_accuracy": 0.8777428269386292, "num_tokens": 364885846.0, "step": 9609 }, { "epoch": 1.2224907772548022, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.029326438903809, "learning_rate": 5e-05, "loss": 1.0612, "mean_token_accuracy": 0.8605809211730957, "num_tokens": 364927536.0, "step": 9610 }, { "epoch": 1.2226179875333927, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.44409465789795, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.8736155033111572, "num_tokens": 364961050.0, "step": 9611 }, { "epoch": 1.2227451978119832, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.084833145141602, "learning_rate": 5e-05, "loss": 0.9945, "mean_token_accuracy": 0.8665108680725098, "num_tokens": 365001058.0, "step": 9612 }, { "epoch": 1.2228724080905737, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.222602844238281, "learning_rate": 5e-05, "loss": 0.8283, "mean_token_accuracy": 0.880761981010437, "num_tokens": 365041345.0, "step": 9613 }, { "epoch": 1.2229996183691643, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.8175630569458, "learning_rate": 5e-05, "loss": 0.914, "mean_token_accuracy": 0.8713786602020264, "num_tokens": 365085669.0, "step": 9614 }, { "epoch": 1.2231268286477548, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.705966472625732, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8884539604187012, "num_tokens": 365125630.0, "step": 9615 }, { "epoch": 1.2232540389263453, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.346081733703613, "learning_rate": 5e-05, "loss": 0.8035, "mean_token_accuracy": 0.8771374225616455, "num_tokens": 365162904.0, "step": 9616 }, { "epoch": 1.2233812492049359, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.746902942657471, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8781766891479492, "num_tokens": 365196509.0, "step": 9617 }, { "epoch": 1.2235084594835262, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.834316253662109, "learning_rate": 5e-05, "loss": 0.8171, "mean_token_accuracy": 0.8616476058959961, "num_tokens": 365232947.0, "step": 9618 }, { "epoch": 1.2236356697621167, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.649108409881592, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8738092184066772, "num_tokens": 365273434.0, "step": 9619 }, { "epoch": 1.2237628800407072, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.869537353515625, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.8667138814926147, "num_tokens": 365316638.0, "step": 9620 }, { "epoch": 1.2238900903192977, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.289469242095947, "learning_rate": 5e-05, "loss": 0.7445, "mean_token_accuracy": 0.8683098554611206, "num_tokens": 365348519.0, "step": 9621 }, { "epoch": 1.2240173005978883, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.05757999420166, "learning_rate": 5e-05, "loss": 0.7341, "mean_token_accuracy": 0.8859584331512451, "num_tokens": 365383782.0, "step": 9622 }, { "epoch": 1.2241445108764788, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.144484043121338, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.864574134349823, "num_tokens": 365421363.0, "step": 9623 }, { "epoch": 1.2242717211550693, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.989786148071289, "learning_rate": 5e-05, "loss": 0.7609, "mean_token_accuracy": 0.8712673187255859, "num_tokens": 365468207.0, "step": 9624 }, { "epoch": 1.2243989314336599, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.188232421875, "learning_rate": 5e-05, "loss": 0.6744, "mean_token_accuracy": 0.8786469101905823, "num_tokens": 365498318.0, "step": 9625 }, { "epoch": 1.2245261417122504, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.9766645431518555, "learning_rate": 5e-05, "loss": 0.7511, "mean_token_accuracy": 0.8739681243896484, "num_tokens": 365536140.0, "step": 9626 }, { "epoch": 1.224653351990841, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.000092506408691, "learning_rate": 5e-05, "loss": 0.7403, "mean_token_accuracy": 0.8599435091018677, "num_tokens": 365576490.0, "step": 9627 }, { "epoch": 1.2247805622694314, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.29489803314209, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.8701676726341248, "num_tokens": 365618043.0, "step": 9628 }, { "epoch": 1.224907772548022, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.679742813110352, "learning_rate": 5e-05, "loss": 0.7029, "mean_token_accuracy": 0.8671826124191284, "num_tokens": 365648303.0, "step": 9629 }, { "epoch": 1.2250349828266125, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.611924171447754, "learning_rate": 5e-05, "loss": 0.8353, "mean_token_accuracy": 0.8721334934234619, "num_tokens": 365684739.0, "step": 9630 }, { "epoch": 1.225162193105203, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.538414478302002, "learning_rate": 5e-05, "loss": 0.7199, "mean_token_accuracy": 0.8595331907272339, "num_tokens": 365726542.0, "step": 9631 }, { "epoch": 1.2252894033837933, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.844059944152832, "learning_rate": 5e-05, "loss": 0.8239, "mean_token_accuracy": 0.8846325874328613, "num_tokens": 365769949.0, "step": 9632 }, { "epoch": 1.2254166136623839, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.58266544342041, "learning_rate": 5e-05, "loss": 0.6887, "mean_token_accuracy": 0.8743802905082703, "num_tokens": 365805891.0, "step": 9633 }, { "epoch": 1.2255438239409744, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.911157608032227, "learning_rate": 5e-05, "loss": 0.9522, "mean_token_accuracy": 0.8531869649887085, "num_tokens": 365847011.0, "step": 9634 }, { "epoch": 1.225671034219565, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.905846118927002, "learning_rate": 5e-05, "loss": 0.7143, "mean_token_accuracy": 0.8768053650856018, "num_tokens": 365884546.0, "step": 9635 }, { "epoch": 1.2257982444981554, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.502799987792969, "learning_rate": 5e-05, "loss": 0.8202, "mean_token_accuracy": 0.8817551732063293, "num_tokens": 365919619.0, "step": 9636 }, { "epoch": 1.225925454776746, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.485091209411621, "learning_rate": 5e-05, "loss": 0.7723, "mean_token_accuracy": 0.8809174299240112, "num_tokens": 365956020.0, "step": 9637 }, { "epoch": 1.2260526650553365, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.714035987854004, "learning_rate": 5e-05, "loss": 0.8368, "mean_token_accuracy": 0.8532220125198364, "num_tokens": 365994945.0, "step": 9638 }, { "epoch": 1.226179875333927, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.975432872772217, "learning_rate": 5e-05, "loss": 0.8224, "mean_token_accuracy": 0.8656840324401855, "num_tokens": 366036476.0, "step": 9639 }, { "epoch": 1.2263070856125176, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.141258239746094, "learning_rate": 5e-05, "loss": 0.6968, "mean_token_accuracy": 0.8812931776046753, "num_tokens": 366072966.0, "step": 9640 }, { "epoch": 1.226434295891108, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.18847370147705, "learning_rate": 5e-05, "loss": 0.7867, "mean_token_accuracy": 0.8736851215362549, "num_tokens": 366111535.0, "step": 9641 }, { "epoch": 1.2265615061696984, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.0290117263793945, "learning_rate": 5e-05, "loss": 0.7261, "mean_token_accuracy": 0.8692501783370972, "num_tokens": 366151104.0, "step": 9642 }, { "epoch": 1.226688716448289, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.073431968688965, "learning_rate": 5e-05, "loss": 0.8264, "mean_token_accuracy": 0.864031195640564, "num_tokens": 366184062.0, "step": 9643 }, { "epoch": 1.2268159267268794, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.186758041381836, "learning_rate": 5e-05, "loss": 0.7007, "mean_token_accuracy": 0.8765791654586792, "num_tokens": 366217774.0, "step": 9644 }, { "epoch": 1.22694313700547, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.814272880554199, "learning_rate": 5e-05, "loss": 0.7289, "mean_token_accuracy": 0.8810806274414062, "num_tokens": 366252776.0, "step": 9645 }, { "epoch": 1.2270703472840605, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.411797523498535, "learning_rate": 5e-05, "loss": 0.7281, "mean_token_accuracy": 0.8705103397369385, "num_tokens": 366289031.0, "step": 9646 }, { "epoch": 1.227197557562651, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.498739719390869, "learning_rate": 5e-05, "loss": 0.6992, "mean_token_accuracy": 0.8825005292892456, "num_tokens": 366327919.0, "step": 9647 }, { "epoch": 1.2273247678412416, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.648876667022705, "learning_rate": 5e-05, "loss": 0.7637, "mean_token_accuracy": 0.8594180345535278, "num_tokens": 366368655.0, "step": 9648 }, { "epoch": 1.227451978119832, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.289885997772217, "learning_rate": 5e-05, "loss": 0.7151, "mean_token_accuracy": 0.8731076717376709, "num_tokens": 366406987.0, "step": 9649 }, { "epoch": 1.2275791883984226, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.756845474243164, "learning_rate": 5e-05, "loss": 0.7786, "mean_token_accuracy": 0.8588531017303467, "num_tokens": 366447360.0, "step": 9650 }, { "epoch": 1.2277063986770131, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.190827369689941, "learning_rate": 5e-05, "loss": 0.6784, "mean_token_accuracy": 0.8819907903671265, "num_tokens": 366487640.0, "step": 9651 }, { "epoch": 1.2278336089556037, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.805635929107666, "learning_rate": 5e-05, "loss": 0.7778, "mean_token_accuracy": 0.8637503385543823, "num_tokens": 366526569.0, "step": 9652 }, { "epoch": 1.2279608192341942, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.149515151977539, "learning_rate": 5e-05, "loss": 0.7134, "mean_token_accuracy": 0.8667051196098328, "num_tokens": 366567503.0, "step": 9653 }, { "epoch": 1.2280880295127847, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.837815761566162, "learning_rate": 5e-05, "loss": 0.7407, "mean_token_accuracy": 0.8706640601158142, "num_tokens": 366609230.0, "step": 9654 }, { "epoch": 1.2282152397913753, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 7.070549964904785, "learning_rate": 5e-05, "loss": 0.6966, "mean_token_accuracy": 0.8736376762390137, "num_tokens": 366646982.0, "step": 9655 }, { "epoch": 1.2283424500699656, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.943598747253418, "learning_rate": 5e-05, "loss": 0.8032, "mean_token_accuracy": 0.8492992520332336, "num_tokens": 366684051.0, "step": 9656 }, { "epoch": 1.228469660348556, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.882241725921631, "learning_rate": 5e-05, "loss": 0.7236, "mean_token_accuracy": 0.859223484992981, "num_tokens": 366724617.0, "step": 9657 }, { "epoch": 1.2285968706271466, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.246097564697266, "learning_rate": 5e-05, "loss": 0.7357, "mean_token_accuracy": 0.8872975707054138, "num_tokens": 366756642.0, "step": 9658 }, { "epoch": 1.2287240809057371, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.4830851554870605, "learning_rate": 5e-05, "loss": 0.6779, "mean_token_accuracy": 0.8710550665855408, "num_tokens": 366789197.0, "step": 9659 }, { "epoch": 1.2288512911843277, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.864057540893555, "learning_rate": 5e-05, "loss": 0.8467, "mean_token_accuracy": 0.8697975873947144, "num_tokens": 366825684.0, "step": 9660 }, { "epoch": 1.2289785014629182, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 5.890459060668945, "learning_rate": 5e-05, "loss": 0.6779, "mean_token_accuracy": 0.8610358238220215, "num_tokens": 366871597.0, "step": 9661 }, { "epoch": 1.2291057117415087, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.890484809875488, "learning_rate": 5e-05, "loss": 0.9173, "mean_token_accuracy": 0.8865548372268677, "num_tokens": 366910961.0, "step": 9662 }, { "epoch": 1.2292329220200993, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 5.68613338470459, "learning_rate": 5e-05, "loss": 0.6412, "mean_token_accuracy": 0.8852307796478271, "num_tokens": 366951701.0, "step": 9663 }, { "epoch": 1.2293601322986898, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 10.799580574035645, "learning_rate": 5e-05, "loss": 1.1651, "mean_token_accuracy": 0.8618659973144531, "num_tokens": 366992472.0, "step": 9664 }, { "epoch": 1.2294873425772803, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.40388822555542, "learning_rate": 5e-05, "loss": 0.9805, "mean_token_accuracy": 0.8640753626823425, "num_tokens": 367031220.0, "step": 9665 }, { "epoch": 1.2296145528558708, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.047049522399902, "learning_rate": 5e-05, "loss": 1.0714, "mean_token_accuracy": 0.8642005324363708, "num_tokens": 367067275.0, "step": 9666 }, { "epoch": 1.2297417631344612, "ewc_loss": 0.796875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 9.114812850952148, "learning_rate": 5e-05, "loss": 1.2131, "mean_token_accuracy": 0.8566364049911499, "num_tokens": 367111851.0, "step": 9667 }, { "epoch": 1.2298689734130517, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.637781143188477, "learning_rate": 5e-05, "loss": 0.8041, "mean_token_accuracy": 0.8679133653640747, "num_tokens": 367150895.0, "step": 9668 }, { "epoch": 1.2299961836916422, "ewc_loss": 0.86328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 10.90380859375, "learning_rate": 5e-05, "loss": 1.2692, "mean_token_accuracy": 0.8591958284378052, "num_tokens": 367187874.0, "step": 9669 }, { "epoch": 1.2301233939702327, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.11677360534668, "learning_rate": 5e-05, "loss": 0.8287, "mean_token_accuracy": 0.8771671056747437, "num_tokens": 367228859.0, "step": 9670 }, { "epoch": 1.2302506042488233, "ewc_loss": 0.984375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000934600830078125, "grad_norm": 12.620532989501953, "learning_rate": 5e-05, "loss": 1.3101, "mean_token_accuracy": 0.8852280378341675, "num_tokens": 367264686.0, "step": 9671 }, { "epoch": 1.2303778145274138, "ewc_loss": 0.984375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000934600830078125, "grad_norm": 10.040955543518066, "learning_rate": 5e-05, "loss": 1.341, "mean_token_accuracy": 0.8759020566940308, "num_tokens": 367309909.0, "step": 9672 }, { "epoch": 1.2305050248060043, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 11.525598526000977, "learning_rate": 5e-05, "loss": 1.072, "mean_token_accuracy": 0.8791276216506958, "num_tokens": 367343996.0, "step": 9673 }, { "epoch": 1.2306322350845948, "ewc_loss": 0.796875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 9.500104904174805, "learning_rate": 5e-05, "loss": 1.1801, "mean_token_accuracy": 0.8686909675598145, "num_tokens": 367386267.0, "step": 9674 }, { "epoch": 1.2307594453631854, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00075531005859375, "grad_norm": 11.555839538574219, "learning_rate": 5e-05, "loss": 1.1945, "mean_token_accuracy": 0.8650518655776978, "num_tokens": 367430256.0, "step": 9675 }, { "epoch": 1.230886655641776, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.5980224609375, "learning_rate": 5e-05, "loss": 1.0246, "mean_token_accuracy": 0.8576316833496094, "num_tokens": 367469189.0, "step": 9676 }, { "epoch": 1.2310138659203664, "ewc_loss": 0.88671875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000835418701171875, "grad_norm": 11.910599708557129, "learning_rate": 5e-05, "loss": 1.3014, "mean_token_accuracy": 0.8592971563339233, "num_tokens": 367501323.0, "step": 9677 }, { "epoch": 1.231141076198957, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.116490364074707, "learning_rate": 5e-05, "loss": 1.056, "mean_token_accuracy": 0.8664689660072327, "num_tokens": 367542361.0, "step": 9678 }, { "epoch": 1.2312682864775475, "ewc_loss": 0.85546875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 12.0420503616333, "learning_rate": 5e-05, "loss": 1.2742, "mean_token_accuracy": 0.8557072877883911, "num_tokens": 367575859.0, "step": 9679 }, { "epoch": 1.231395496756138, "ewc_loss": 0.78125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 10.096453666687012, "learning_rate": 5e-05, "loss": 1.1495, "mean_token_accuracy": 0.8729636669158936, "num_tokens": 367611330.0, "step": 9680 }, { "epoch": 1.2315227070347283, "ewc_loss": 0.75, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 11.25321102142334, "learning_rate": 5e-05, "loss": 1.0842, "mean_token_accuracy": 0.882626473903656, "num_tokens": 367653110.0, "step": 9681 }, { "epoch": 1.2316499173133189, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 9.852788925170898, "learning_rate": 5e-05, "loss": 1.0451, "mean_token_accuracy": 0.8781607151031494, "num_tokens": 367687799.0, "step": 9682 }, { "epoch": 1.2317771275919094, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.686241149902344, "learning_rate": 5e-05, "loss": 1.0503, "mean_token_accuracy": 0.8742550611495972, "num_tokens": 367728253.0, "step": 9683 }, { "epoch": 1.2319043378705, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.321075439453125, "learning_rate": 5e-05, "loss": 0.8998, "mean_token_accuracy": 0.9007375240325928, "num_tokens": 367770397.0, "step": 9684 }, { "epoch": 1.2320315481490904, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.29361343383789, "learning_rate": 5e-05, "loss": 1.0237, "mean_token_accuracy": 0.8728041648864746, "num_tokens": 367810630.0, "step": 9685 }, { "epoch": 1.232158758427681, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.792732238769531, "learning_rate": 5e-05, "loss": 0.9374, "mean_token_accuracy": 0.8657166957855225, "num_tokens": 367853563.0, "step": 9686 }, { "epoch": 1.2322859687062715, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.239184379577637, "learning_rate": 5e-05, "loss": 1.0641, "mean_token_accuracy": 0.8613899946212769, "num_tokens": 367888113.0, "step": 9687 }, { "epoch": 1.232413178984862, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.435606002807617, "learning_rate": 5e-05, "loss": 0.8951, "mean_token_accuracy": 0.8691594004631042, "num_tokens": 367925839.0, "step": 9688 }, { "epoch": 1.2325403892634526, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.449249267578125, "learning_rate": 5e-05, "loss": 1.0696, "mean_token_accuracy": 0.8622472882270813, "num_tokens": 367966944.0, "step": 9689 }, { "epoch": 1.232667599542043, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.492880821228027, "learning_rate": 5e-05, "loss": 0.8637, "mean_token_accuracy": 0.8745461702346802, "num_tokens": 367999470.0, "step": 9690 }, { "epoch": 1.2327948098206334, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.558629989624023, "learning_rate": 5e-05, "loss": 1.0035, "mean_token_accuracy": 0.8885335922241211, "num_tokens": 368042780.0, "step": 9691 }, { "epoch": 1.232922020099224, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.879779815673828, "learning_rate": 5e-05, "loss": 0.8987, "mean_token_accuracy": 0.8773964047431946, "num_tokens": 368078062.0, "step": 9692 }, { "epoch": 1.2330492303778144, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.16150951385498, "learning_rate": 5e-05, "loss": 1.0254, "mean_token_accuracy": 0.8718403577804565, "num_tokens": 368121570.0, "step": 9693 }, { "epoch": 1.233176440656405, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.984456062316895, "learning_rate": 5e-05, "loss": 0.896, "mean_token_accuracy": 0.8795455098152161, "num_tokens": 368158576.0, "step": 9694 }, { "epoch": 1.2333036509349955, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.490949630737305, "learning_rate": 5e-05, "loss": 0.8995, "mean_token_accuracy": 0.8901457786560059, "num_tokens": 368199939.0, "step": 9695 }, { "epoch": 1.233430861213586, "ewc_loss": 0.5, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.604761123657227, "learning_rate": 5e-05, "loss": 0.8791, "mean_token_accuracy": 0.8670929670333862, "num_tokens": 368240156.0, "step": 9696 }, { "epoch": 1.2335580714921766, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.127937316894531, "learning_rate": 5e-05, "loss": 0.9663, "mean_token_accuracy": 0.8558397889137268, "num_tokens": 368270852.0, "step": 9697 }, { "epoch": 1.233685281770767, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.040839195251465, "learning_rate": 5e-05, "loss": 0.8296, "mean_token_accuracy": 0.865466833114624, "num_tokens": 368311986.0, "step": 9698 }, { "epoch": 1.2338124920493576, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.183533668518066, "learning_rate": 5e-05, "loss": 0.9434, "mean_token_accuracy": 0.8652113676071167, "num_tokens": 368354330.0, "step": 9699 }, { "epoch": 1.2339397023279481, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.641673564910889, "learning_rate": 5e-05, "loss": 0.7359, "mean_token_accuracy": 0.8812932968139648, "num_tokens": 368392712.0, "step": 9700 }, { "epoch": 1.2340669126065387, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.477226257324219, "learning_rate": 5e-05, "loss": 0.9818, "mean_token_accuracy": 0.8661277294158936, "num_tokens": 368430913.0, "step": 9701 }, { "epoch": 1.2341941228851292, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.693072319030762, "learning_rate": 5e-05, "loss": 0.74, "mean_token_accuracy": 0.8867697715759277, "num_tokens": 368462353.0, "step": 9702 }, { "epoch": 1.2343213331637197, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.561921119689941, "learning_rate": 5e-05, "loss": 0.9948, "mean_token_accuracy": 0.8661261796951294, "num_tokens": 368506061.0, "step": 9703 }, { "epoch": 1.2344485434423103, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.102112770080566, "learning_rate": 5e-05, "loss": 0.7862, "mean_token_accuracy": 0.8808172941207886, "num_tokens": 368539435.0, "step": 9704 }, { "epoch": 1.2345757537209006, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.089244842529297, "learning_rate": 5e-05, "loss": 0.9248, "mean_token_accuracy": 0.869896650314331, "num_tokens": 368578249.0, "step": 9705 }, { "epoch": 1.234702963999491, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.262096405029297, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.8870409727096558, "num_tokens": 368617766.0, "step": 9706 }, { "epoch": 1.2348301742780816, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.405674934387207, "learning_rate": 5e-05, "loss": 0.8256, "mean_token_accuracy": 0.8791451454162598, "num_tokens": 368651458.0, "step": 9707 }, { "epoch": 1.2349573845566721, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.027667045593262, "learning_rate": 5e-05, "loss": 0.8504, "mean_token_accuracy": 0.854845404624939, "num_tokens": 368688689.0, "step": 9708 }, { "epoch": 1.2350845948352627, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.07669448852539, "learning_rate": 5e-05, "loss": 0.8683, "mean_token_accuracy": 0.8534217476844788, "num_tokens": 368731996.0, "step": 9709 }, { "epoch": 1.2352118051138532, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.6785569190979, "learning_rate": 5e-05, "loss": 0.7268, "mean_token_accuracy": 0.8821920156478882, "num_tokens": 368772712.0, "step": 9710 }, { "epoch": 1.2353390153924437, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.044479370117188, "learning_rate": 5e-05, "loss": 0.7833, "mean_token_accuracy": 0.876032829284668, "num_tokens": 368816582.0, "step": 9711 }, { "epoch": 1.2354662256710343, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.45161771774292, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.8754795789718628, "num_tokens": 368854284.0, "step": 9712 }, { "epoch": 1.2355934359496248, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.996111869812012, "learning_rate": 5e-05, "loss": 0.7852, "mean_token_accuracy": 0.87526535987854, "num_tokens": 368891564.0, "step": 9713 }, { "epoch": 1.2357206462282153, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.391783237457275, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8559448719024658, "num_tokens": 368931176.0, "step": 9714 }, { "epoch": 1.2358478565068058, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.92390251159668, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.880064070224762, "num_tokens": 368967730.0, "step": 9715 }, { "epoch": 1.2359750667853961, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.318100452423096, "learning_rate": 5e-05, "loss": 0.6977, "mean_token_accuracy": 0.8760982751846313, "num_tokens": 369009378.0, "step": 9716 }, { "epoch": 1.2361022770639867, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00051116943359375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.962957382202148, "learning_rate": 5e-05, "loss": 0.7733, "mean_token_accuracy": 0.8656376004219055, "num_tokens": 369045716.0, "step": 9717 }, { "epoch": 1.2362294873425772, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.206455230712891, "learning_rate": 5e-05, "loss": 0.6716, "mean_token_accuracy": 0.8818514347076416, "num_tokens": 369082022.0, "step": 9718 }, { "epoch": 1.2363566976211677, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.037758827209473, "learning_rate": 5e-05, "loss": 0.7911, "mean_token_accuracy": 0.8650988936424255, "num_tokens": 369122456.0, "step": 9719 }, { "epoch": 1.2364839078997583, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.132681846618652, "learning_rate": 5e-05, "loss": 0.7416, "mean_token_accuracy": 0.8593347668647766, "num_tokens": 369156977.0, "step": 9720 }, { "epoch": 1.2366111181783488, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.063753128051758, "learning_rate": 5e-05, "loss": 0.7919, "mean_token_accuracy": 0.8636300563812256, "num_tokens": 369195113.0, "step": 9721 }, { "epoch": 1.2367383284569393, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.13224983215332, "learning_rate": 5e-05, "loss": 0.6689, "mean_token_accuracy": 0.8835171461105347, "num_tokens": 369238650.0, "step": 9722 }, { "epoch": 1.2368655387355298, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.068892478942871, "learning_rate": 5e-05, "loss": 0.8226, "mean_token_accuracy": 0.852424144744873, "num_tokens": 369275753.0, "step": 9723 }, { "epoch": 1.2369927490141204, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.148671627044678, "learning_rate": 5e-05, "loss": 0.769, "mean_token_accuracy": 0.8474538922309875, "num_tokens": 369317364.0, "step": 9724 }, { "epoch": 1.237119959292711, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.965206146240234, "learning_rate": 5e-05, "loss": 0.7644, "mean_token_accuracy": 0.8755495548248291, "num_tokens": 369355889.0, "step": 9725 }, { "epoch": 1.2372471695713014, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.296277046203613, "learning_rate": 5e-05, "loss": 0.6578, "mean_token_accuracy": 0.8848819136619568, "num_tokens": 369393749.0, "step": 9726 }, { "epoch": 1.237374379849892, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.706106662750244, "learning_rate": 5e-05, "loss": 0.7464, "mean_token_accuracy": 0.8709222078323364, "num_tokens": 369432442.0, "step": 9727 }, { "epoch": 1.2375015901284825, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.548248291015625, "learning_rate": 5e-05, "loss": 0.7439, "mean_token_accuracy": 0.8700138926506042, "num_tokens": 369469513.0, "step": 9728 }, { "epoch": 1.237628800407073, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.350790023803711, "learning_rate": 5e-05, "loss": 0.6997, "mean_token_accuracy": 0.8729286789894104, "num_tokens": 369505720.0, "step": 9729 }, { "epoch": 1.2377560106856633, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.9133429527282715, "learning_rate": 5e-05, "loss": 0.7039, "mean_token_accuracy": 0.8791327476501465, "num_tokens": 369535225.0, "step": 9730 }, { "epoch": 1.2378832209642538, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.90914249420166, "learning_rate": 5e-05, "loss": 0.6696, "mean_token_accuracy": 0.874845027923584, "num_tokens": 369572580.0, "step": 9731 }, { "epoch": 1.2380104312428444, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.409857749938965, "learning_rate": 5e-05, "loss": 0.7583, "mean_token_accuracy": 0.8789379000663757, "num_tokens": 369611566.0, "step": 9732 }, { "epoch": 1.238137641521435, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00023555755615234375, "grad_norm": 6.416872501373291, "learning_rate": 5e-05, "loss": 0.6695, "mean_token_accuracy": 0.8722584247589111, "num_tokens": 369652541.0, "step": 9733 }, { "epoch": 1.2382648518000254, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 9.065788269042969, "learning_rate": 5e-05, "loss": 0.8952, "mean_token_accuracy": 0.8605602979660034, "num_tokens": 369686744.0, "step": 9734 }, { "epoch": 1.238392062078616, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002269744873046875, "grad_norm": 6.035079479217529, "learning_rate": 5e-05, "loss": 0.607, "mean_token_accuracy": 0.8834747076034546, "num_tokens": 369725090.0, "step": 9735 }, { "epoch": 1.2385192723572065, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.605039596557617, "learning_rate": 5e-05, "loss": 0.8845, "mean_token_accuracy": 0.8909314870834351, "num_tokens": 369764255.0, "step": 9736 }, { "epoch": 1.238646482635797, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.418276786804199, "learning_rate": 5e-05, "loss": 0.6832, "mean_token_accuracy": 0.8769797086715698, "num_tokens": 369803589.0, "step": 9737 }, { "epoch": 1.2387736929143875, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.277892112731934, "learning_rate": 5e-05, "loss": 0.9168, "mean_token_accuracy": 0.8832790851593018, "num_tokens": 369837984.0, "step": 9738 }, { "epoch": 1.238900903192978, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.667603015899658, "learning_rate": 5e-05, "loss": 0.8233, "mean_token_accuracy": 0.8719364404678345, "num_tokens": 369881398.0, "step": 9739 }, { "epoch": 1.2390281134715684, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.718898296356201, "learning_rate": 5e-05, "loss": 0.8081, "mean_token_accuracy": 0.8699476718902588, "num_tokens": 369915230.0, "step": 9740 }, { "epoch": 1.239155323750159, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.886398315429688, "learning_rate": 5e-05, "loss": 0.8958, "mean_token_accuracy": 0.8717116713523865, "num_tokens": 369955237.0, "step": 9741 }, { "epoch": 1.2392825340287494, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.418412685394287, "learning_rate": 5e-05, "loss": 0.6989, "mean_token_accuracy": 0.8699514269828796, "num_tokens": 369995938.0, "step": 9742 }, { "epoch": 1.23940974430734, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.75500774383545, "learning_rate": 5e-05, "loss": 0.9829, "mean_token_accuracy": 0.8640173077583313, "num_tokens": 370037346.0, "step": 9743 }, { "epoch": 1.2395369545859305, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.625122547149658, "learning_rate": 5e-05, "loss": 0.7281, "mean_token_accuracy": 0.8710967898368835, "num_tokens": 370067222.0, "step": 9744 }, { "epoch": 1.239664164864521, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.504802703857422, "learning_rate": 5e-05, "loss": 0.9395, "mean_token_accuracy": 0.8822531700134277, "num_tokens": 370101438.0, "step": 9745 }, { "epoch": 1.2397913751431116, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.889594078063965, "learning_rate": 5e-05, "loss": 0.8858, "mean_token_accuracy": 0.871094822883606, "num_tokens": 370136887.0, "step": 9746 }, { "epoch": 1.239918585421702, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.067004203796387, "learning_rate": 5e-05, "loss": 0.8096, "mean_token_accuracy": 0.8808664083480835, "num_tokens": 370176797.0, "step": 9747 }, { "epoch": 1.2400457957002926, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.68698787689209, "learning_rate": 5e-05, "loss": 0.9204, "mean_token_accuracy": 0.8661746382713318, "num_tokens": 370206246.0, "step": 9748 }, { "epoch": 1.2401730059788831, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.22853422164917, "learning_rate": 5e-05, "loss": 0.7755, "mean_token_accuracy": 0.8721069693565369, "num_tokens": 370246132.0, "step": 9749 }, { "epoch": 1.2403002162574737, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.555822372436523, "learning_rate": 5e-05, "loss": 0.8439, "mean_token_accuracy": 0.8763911724090576, "num_tokens": 370280039.0, "step": 9750 }, { "epoch": 1.2404274265360642, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.6005940437316895, "learning_rate": 5e-05, "loss": 0.7757, "mean_token_accuracy": 0.8776844143867493, "num_tokens": 370317522.0, "step": 9751 }, { "epoch": 1.2405546368146547, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.572305202484131, "learning_rate": 5e-05, "loss": 0.7512, "mean_token_accuracy": 0.8742725253105164, "num_tokens": 370352350.0, "step": 9752 }, { "epoch": 1.2406818470932452, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.32574462890625, "learning_rate": 5e-05, "loss": 0.8106, "mean_token_accuracy": 0.8814194798469543, "num_tokens": 370382645.0, "step": 9753 }, { "epoch": 1.2408090573718356, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.8199310302734375, "learning_rate": 5e-05, "loss": 0.709, "mean_token_accuracy": 0.8762059211730957, "num_tokens": 370422679.0, "step": 9754 }, { "epoch": 1.240936267650426, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.890921592712402, "learning_rate": 5e-05, "loss": 0.8755, "mean_token_accuracy": 0.8691674470901489, "num_tokens": 370458475.0, "step": 9755 }, { "epoch": 1.2410634779290166, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.6227946281433105, "learning_rate": 5e-05, "loss": 0.7173, "mean_token_accuracy": 0.8667862415313721, "num_tokens": 370489641.0, "step": 9756 }, { "epoch": 1.2411906882076071, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.096834182739258, "learning_rate": 5e-05, "loss": 0.9026, "mean_token_accuracy": 0.8737534880638123, "num_tokens": 370526649.0, "step": 9757 }, { "epoch": 1.2413178984861977, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.825669288635254, "learning_rate": 5e-05, "loss": 0.7377, "mean_token_accuracy": 0.8705681562423706, "num_tokens": 370564664.0, "step": 9758 }, { "epoch": 1.2414451087647882, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.036704063415527, "learning_rate": 5e-05, "loss": 0.9029, "mean_token_accuracy": 0.8801354169845581, "num_tokens": 370598000.0, "step": 9759 }, { "epoch": 1.2415723190433787, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.027139663696289, "learning_rate": 5e-05, "loss": 0.7933, "mean_token_accuracy": 0.8623704314231873, "num_tokens": 370639318.0, "step": 9760 }, { "epoch": 1.2416995293219693, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.064042091369629, "learning_rate": 5e-05, "loss": 0.9829, "mean_token_accuracy": 0.8639750480651855, "num_tokens": 370676064.0, "step": 9761 }, { "epoch": 1.2418267396005598, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.00376558303833, "learning_rate": 5e-05, "loss": 0.7501, "mean_token_accuracy": 0.8755072355270386, "num_tokens": 370713190.0, "step": 9762 }, { "epoch": 1.2419539498791503, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.297385215759277, "learning_rate": 5e-05, "loss": 0.9355, "mean_token_accuracy": 0.8860578536987305, "num_tokens": 370748433.0, "step": 9763 }, { "epoch": 1.2420811601577408, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.074553966522217, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8797463178634644, "num_tokens": 370781522.0, "step": 9764 }, { "epoch": 1.2422083704363311, "ewc_loss": 0.625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.420499801635742, "learning_rate": 5e-05, "loss": 0.9812, "mean_token_accuracy": 0.8756169676780701, "num_tokens": 370821878.0, "step": 9765 }, { "epoch": 1.2423355807149217, "ewc_loss": 0.5, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.609600067138672, "learning_rate": 5e-05, "loss": 0.8834, "mean_token_accuracy": 0.8681402802467346, "num_tokens": 370857754.0, "step": 9766 }, { "epoch": 1.2424627909935122, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.824554443359375, "learning_rate": 5e-05, "loss": 0.9606, "mean_token_accuracy": 0.8674819469451904, "num_tokens": 370893672.0, "step": 9767 }, { "epoch": 1.2425900012721027, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.133854866027832, "learning_rate": 5e-05, "loss": 0.9164, "mean_token_accuracy": 0.868334174156189, "num_tokens": 370928118.0, "step": 9768 }, { "epoch": 1.2427172115506933, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.905519485473633, "learning_rate": 5e-05, "loss": 0.9014, "mean_token_accuracy": 0.8558729290962219, "num_tokens": 370960748.0, "step": 9769 }, { "epoch": 1.2428444218292838, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.17270278930664, "learning_rate": 5e-05, "loss": 0.8385, "mean_token_accuracy": 0.8849590420722961, "num_tokens": 370992408.0, "step": 9770 }, { "epoch": 1.2429716321078743, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.609061241149902, "learning_rate": 5e-05, "loss": 0.8471, "mean_token_accuracy": 0.8553833961486816, "num_tokens": 371035020.0, "step": 9771 }, { "epoch": 1.2430988423864648, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.7996602058410645, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.8798162341117859, "num_tokens": 371072738.0, "step": 9772 }, { "epoch": 1.2432260526650554, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.801461696624756, "learning_rate": 5e-05, "loss": 0.761, "mean_token_accuracy": 0.8785931468009949, "num_tokens": 371109246.0, "step": 9773 }, { "epoch": 1.243353262943646, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.468411922454834, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.8841017484664917, "num_tokens": 371148322.0, "step": 9774 }, { "epoch": 1.2434804732222364, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.823117256164551, "learning_rate": 5e-05, "loss": 0.8161, "mean_token_accuracy": 0.8617686033248901, "num_tokens": 371186534.0, "step": 9775 }, { "epoch": 1.243607683500827, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.497513771057129, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.8735209703445435, "num_tokens": 371224973.0, "step": 9776 }, { "epoch": 1.2437348937794175, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.593716144561768, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8759669065475464, "num_tokens": 371262629.0, "step": 9777 }, { "epoch": 1.243862104058008, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.646918296813965, "learning_rate": 5e-05, "loss": 0.7619, "mean_token_accuracy": 0.8710633516311646, "num_tokens": 371298421.0, "step": 9778 }, { "epoch": 1.2439893143365983, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.458324432373047, "learning_rate": 5e-05, "loss": 0.7985, "mean_token_accuracy": 0.8487939834594727, "num_tokens": 371337393.0, "step": 9779 }, { "epoch": 1.2441165246151888, "ewc_loss": 0.375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.662327766418457, "learning_rate": 5e-05, "loss": 0.7236, "mean_token_accuracy": 0.8765796422958374, "num_tokens": 371372436.0, "step": 9780 }, { "epoch": 1.2442437348937794, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.483166217803955, "learning_rate": 5e-05, "loss": 0.7184, "mean_token_accuracy": 0.8750529885292053, "num_tokens": 371404269.0, "step": 9781 }, { "epoch": 1.24437094517237, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.543146133422852, "learning_rate": 5e-05, "loss": 0.7335, "mean_token_accuracy": 0.8743653893470764, "num_tokens": 371446494.0, "step": 9782 }, { "epoch": 1.2444981554509604, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.546843528747559, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.863287091255188, "num_tokens": 371482524.0, "step": 9783 }, { "epoch": 1.244625365729551, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.502248287200928, "learning_rate": 5e-05, "loss": 0.8076, "mean_token_accuracy": 0.8506438136100769, "num_tokens": 371519745.0, "step": 9784 }, { "epoch": 1.2447525760081415, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.511831283569336, "learning_rate": 5e-05, "loss": 0.7257, "mean_token_accuracy": 0.8770589828491211, "num_tokens": 371563593.0, "step": 9785 }, { "epoch": 1.244879786286732, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.52852201461792, "learning_rate": 5e-05, "loss": 0.7034, "mean_token_accuracy": 0.8773688077926636, "num_tokens": 371600646.0, "step": 9786 }, { "epoch": 1.2450069965653225, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.463888645172119, "learning_rate": 5e-05, "loss": 0.6897, "mean_token_accuracy": 0.8821293115615845, "num_tokens": 371635071.0, "step": 9787 }, { "epoch": 1.245134206843913, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.55866813659668, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8727655410766602, "num_tokens": 371672441.0, "step": 9788 }, { "epoch": 1.2452614171225034, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.4188385009765625, "learning_rate": 5e-05, "loss": 0.6781, "mean_token_accuracy": 0.8792884349822998, "num_tokens": 371712751.0, "step": 9789 }, { "epoch": 1.245388627401094, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.603944778442383, "learning_rate": 5e-05, "loss": 0.6973, "mean_token_accuracy": 0.8747851848602295, "num_tokens": 371748863.0, "step": 9790 }, { "epoch": 1.2455158376796844, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.337792873382568, "learning_rate": 5e-05, "loss": 0.7344, "mean_token_accuracy": 0.863746702671051, "num_tokens": 371784763.0, "step": 9791 }, { "epoch": 1.245643047958275, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.671960353851318, "learning_rate": 5e-05, "loss": 0.7808, "mean_token_accuracy": 0.853005588054657, "num_tokens": 371822722.0, "step": 9792 }, { "epoch": 1.2457702582368655, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.230599880218506, "learning_rate": 5e-05, "loss": 0.634, "mean_token_accuracy": 0.891838788986206, "num_tokens": 371855423.0, "step": 9793 }, { "epoch": 1.245897468515456, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.8251214027404785, "learning_rate": 5e-05, "loss": 0.7566, "mean_token_accuracy": 0.8660211563110352, "num_tokens": 371897900.0, "step": 9794 }, { "epoch": 1.2460246787940465, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 7.01867151260376, "learning_rate": 5e-05, "loss": 0.6711, "mean_token_accuracy": 0.875830888748169, "num_tokens": 371942394.0, "step": 9795 }, { "epoch": 1.246151889072637, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 8.144804954528809, "learning_rate": 5e-05, "loss": 0.8128, "mean_token_accuracy": 0.8527700901031494, "num_tokens": 371982391.0, "step": 9796 }, { "epoch": 1.2462790993512276, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00023746490478515625, "grad_norm": 6.590758800506592, "learning_rate": 5e-05, "loss": 0.654, "mean_token_accuracy": 0.8722954988479614, "num_tokens": 372018599.0, "step": 9797 }, { "epoch": 1.2464063096298181, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.786853790283203, "learning_rate": 5e-05, "loss": 0.8352, "mean_token_accuracy": 0.8682372570037842, "num_tokens": 372058650.0, "step": 9798 }, { "epoch": 1.2465335199084087, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000209808349609375, "grad_norm": 5.876114368438721, "learning_rate": 5e-05, "loss": 0.6428, "mean_token_accuracy": 0.8661966323852539, "num_tokens": 372102272.0, "step": 9799 }, { "epoch": 1.2466607301869992, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 10.030377388000488, "learning_rate": 5e-05, "loss": 0.8914, "mean_token_accuracy": 0.8896528482437134, "num_tokens": 372139086.0, "step": 9800 }, { "epoch": 1.2467879404655897, "ewc_loss": 0.2734375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00022125244140625, "grad_norm": 5.254127025604248, "learning_rate": 5e-05, "loss": 0.6293, "mean_token_accuracy": 0.8745230436325073, "num_tokens": 372178418.0, "step": 9801 }, { "epoch": 1.2469151507441802, "ewc_loss": 0.875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000823974609375, "grad_norm": 11.766725540161133, "learning_rate": 5e-05, "loss": 1.2041, "mean_token_accuracy": 0.8846969604492188, "num_tokens": 372218837.0, "step": 9802 }, { "epoch": 1.2470423610227706, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 7.365966320037842, "learning_rate": 5e-05, "loss": 0.996, "mean_token_accuracy": 0.8570207357406616, "num_tokens": 372256620.0, "step": 9803 }, { "epoch": 1.247169571301361, "ewc_loss": 0.90625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 12.091886520385742, "learning_rate": 5e-05, "loss": 1.289, "mean_token_accuracy": 0.8679032325744629, "num_tokens": 372292379.0, "step": 9804 }, { "epoch": 1.2472967815799516, "ewc_loss": 1.046875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0009918212890625, "grad_norm": 9.986926078796387, "learning_rate": 5e-05, "loss": 1.3784, "mean_token_accuracy": 0.881604790687561, "num_tokens": 372336319.0, "step": 9805 }, { "epoch": 1.2474239918585421, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 10.33922290802002, "learning_rate": 5e-05, "loss": 1.0318, "mean_token_accuracy": 0.8668608665466309, "num_tokens": 372372965.0, "step": 9806 }, { "epoch": 1.2475512021371327, "ewc_loss": 0.90234375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 9.951900482177734, "learning_rate": 5e-05, "loss": 1.2596, "mean_token_accuracy": 0.8743447065353394, "num_tokens": 372415453.0, "step": 9807 }, { "epoch": 1.2476784124157232, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.88914966583252, "learning_rate": 5e-05, "loss": 0.9851, "mean_token_accuracy": 0.8674702644348145, "num_tokens": 372452822.0, "step": 9808 }, { "epoch": 1.2478056226943137, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.459010124206543, "learning_rate": 5e-05, "loss": 1.1487, "mean_token_accuracy": 0.8697601556777954, "num_tokens": 372494316.0, "step": 9809 }, { "epoch": 1.2479328329729042, "ewc_loss": 0.625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.423145294189453, "learning_rate": 5e-05, "loss": 0.9548, "mean_token_accuracy": 0.8816959857940674, "num_tokens": 372531334.0, "step": 9810 }, { "epoch": 1.2480600432514948, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.52216625213623, "learning_rate": 5e-05, "loss": 1.0464, "mean_token_accuracy": 0.8791794180870056, "num_tokens": 372568303.0, "step": 9811 }, { "epoch": 1.2481872535300853, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 8.646432876586914, "learning_rate": 5e-05, "loss": 1.0022, "mean_token_accuracy": 0.882442831993103, "num_tokens": 372602783.0, "step": 9812 }, { "epoch": 1.2483144638086758, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.711188316345215, "learning_rate": 5e-05, "loss": 0.9732, "mean_token_accuracy": 0.877575695514679, "num_tokens": 372643883.0, "step": 9813 }, { "epoch": 1.2484416740872661, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.701234817504883, "learning_rate": 5e-05, "loss": 1.0119, "mean_token_accuracy": 0.8697406649589539, "num_tokens": 372681864.0, "step": 9814 }, { "epoch": 1.2485688843658567, "ewc_loss": 0.5, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.442861557006836, "learning_rate": 5e-05, "loss": 0.8839, "mean_token_accuracy": 0.8694761395454407, "num_tokens": 372717657.0, "step": 9815 }, { "epoch": 1.2486960946444472, "ewc_loss": 0.625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.226277351379395, "learning_rate": 5e-05, "loss": 0.9417, "mean_token_accuracy": 0.8890600800514221, "num_tokens": 372756702.0, "step": 9816 }, { "epoch": 1.2488233049230377, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.162552356719971, "learning_rate": 5e-05, "loss": 0.7743, "mean_token_accuracy": 0.8694558143615723, "num_tokens": 372795620.0, "step": 9817 }, { "epoch": 1.2489505152016283, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.532346725463867, "learning_rate": 5e-05, "loss": 1.0845, "mean_token_accuracy": 0.866916298866272, "num_tokens": 372833803.0, "step": 9818 }, { "epoch": 1.2490777254802188, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.338413238525391, "learning_rate": 5e-05, "loss": 0.8616, "mean_token_accuracy": 0.8786689043045044, "num_tokens": 372864229.0, "step": 9819 }, { "epoch": 1.2492049357588093, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 11.363232612609863, "learning_rate": 5e-05, "loss": 1.1339, "mean_token_accuracy": 0.8743232488632202, "num_tokens": 372900321.0, "step": 9820 }, { "epoch": 1.2493321460373998, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 8.898449897766113, "learning_rate": 5e-05, "loss": 1.127, "mean_token_accuracy": 0.8704707622528076, "num_tokens": 372941237.0, "step": 9821 }, { "epoch": 1.2494593563159904, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.66135025024414, "learning_rate": 5e-05, "loss": 1.0261, "mean_token_accuracy": 0.8680230379104614, "num_tokens": 372980057.0, "step": 9822 }, { "epoch": 1.249586566594581, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 9.024239540100098, "learning_rate": 5e-05, "loss": 1.1113, "mean_token_accuracy": 0.8750429153442383, "num_tokens": 373014746.0, "step": 9823 }, { "epoch": 1.2497137768731714, "ewc_loss": 0.625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.968474388122559, "learning_rate": 5e-05, "loss": 1.017, "mean_token_accuracy": 0.8626188635826111, "num_tokens": 373053298.0, "step": 9824 }, { "epoch": 1.249840987151762, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.34933853149414, "learning_rate": 5e-05, "loss": 0.913, "mean_token_accuracy": 0.8867265582084656, "num_tokens": 373100155.0, "step": 9825 }, { "epoch": 1.2499681974303525, "ewc_loss": 0.625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.712254524230957, "learning_rate": 5e-05, "loss": 1.0153, "mean_token_accuracy": 0.864923894405365, "num_tokens": 373138554.0, "step": 9826 }, { "epoch": 1.250095407708943, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.580721855163574, "learning_rate": 5e-05, "loss": 0.878, "mean_token_accuracy": 0.8597652316093445, "num_tokens": 373173023.0, "step": 9827 }, { "epoch": 1.2502226179875333, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.165325164794922, "learning_rate": 5e-05, "loss": 1.0393, "mean_token_accuracy": 0.8671127557754517, "num_tokens": 373210630.0, "step": 9828 }, { "epoch": 1.2503498282661238, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.469245910644531, "learning_rate": 5e-05, "loss": 0.8673, "mean_token_accuracy": 0.8659216165542603, "num_tokens": 373247859.0, "step": 9829 }, { "epoch": 1.2504770385447144, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.718841552734375, "learning_rate": 5e-05, "loss": 1.074, "mean_token_accuracy": 0.8740441799163818, "num_tokens": 373285663.0, "step": 9830 }, { "epoch": 1.250604248823305, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.63146686553955, "learning_rate": 5e-05, "loss": 1.009, "mean_token_accuracy": 0.8637834787368774, "num_tokens": 373321743.0, "step": 9831 }, { "epoch": 1.2507314591018954, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.842597961425781, "learning_rate": 5e-05, "loss": 1.0702, "mean_token_accuracy": 0.8491876125335693, "num_tokens": 373355768.0, "step": 9832 }, { "epoch": 1.250858669380486, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.137224197387695, "learning_rate": 5e-05, "loss": 1.0214, "mean_token_accuracy": 0.8685986995697021, "num_tokens": 373399335.0, "step": 9833 }, { "epoch": 1.2509858796590765, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.412474632263184, "learning_rate": 5e-05, "loss": 0.8689, "mean_token_accuracy": 0.8667100667953491, "num_tokens": 373433708.0, "step": 9834 }, { "epoch": 1.251113089937667, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.003666877746582, "learning_rate": 5e-05, "loss": 0.9285, "mean_token_accuracy": 0.8767808675765991, "num_tokens": 373472145.0, "step": 9835 }, { "epoch": 1.2512403002162575, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.659607887268066, "learning_rate": 5e-05, "loss": 0.7795, "mean_token_accuracy": 0.8788560628890991, "num_tokens": 373506578.0, "step": 9836 }, { "epoch": 1.2513675104948478, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.704197883605957, "learning_rate": 5e-05, "loss": 0.8683, "mean_token_accuracy": 0.8806731700897217, "num_tokens": 373545988.0, "step": 9837 }, { "epoch": 1.2514947207734384, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.596245288848877, "learning_rate": 5e-05, "loss": 0.7547, "mean_token_accuracy": 0.8766926527023315, "num_tokens": 373581001.0, "step": 9838 }, { "epoch": 1.251621931052029, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.266816139221191, "learning_rate": 5e-05, "loss": 0.8507, "mean_token_accuracy": 0.8675941824913025, "num_tokens": 373614961.0, "step": 9839 }, { "epoch": 1.2517491413306194, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.583930015563965, "learning_rate": 5e-05, "loss": 0.7183, "mean_token_accuracy": 0.8857955932617188, "num_tokens": 373652385.0, "step": 9840 }, { "epoch": 1.25187635160921, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.106500625610352, "learning_rate": 5e-05, "loss": 0.8089, "mean_token_accuracy": 0.8680482506752014, "num_tokens": 373692468.0, "step": 9841 }, { "epoch": 1.2520035618878005, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.3750529289245605, "learning_rate": 5e-05, "loss": 0.7381, "mean_token_accuracy": 0.8762312531471252, "num_tokens": 373736885.0, "step": 9842 }, { "epoch": 1.252130772166391, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.244169235229492, "learning_rate": 5e-05, "loss": 0.8291, "mean_token_accuracy": 0.866492748260498, "num_tokens": 373777354.0, "step": 9843 }, { "epoch": 1.2522579824449815, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.09798526763916, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.8640932440757751, "num_tokens": 373812400.0, "step": 9844 }, { "epoch": 1.252385192723572, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.431827545166016, "learning_rate": 5e-05, "loss": 0.7889, "mean_token_accuracy": 0.8854217529296875, "num_tokens": 373846033.0, "step": 9845 }, { "epoch": 1.2525124030021626, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.929378509521484, "learning_rate": 5e-05, "loss": 0.6905, "mean_token_accuracy": 0.8803406953811646, "num_tokens": 373887888.0, "step": 9846 }, { "epoch": 1.2526396132807531, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.630377769470215, "learning_rate": 5e-05, "loss": 0.8153, "mean_token_accuracy": 0.8860546350479126, "num_tokens": 373928082.0, "step": 9847 }, { "epoch": 1.2527668235593437, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.786258220672607, "learning_rate": 5e-05, "loss": 0.7351, "mean_token_accuracy": 0.8615996837615967, "num_tokens": 373964078.0, "step": 9848 }, { "epoch": 1.2528940338379342, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.9107666015625, "learning_rate": 5e-05, "loss": 0.9341, "mean_token_accuracy": 0.8552305698394775, "num_tokens": 373998163.0, "step": 9849 }, { "epoch": 1.2530212441165247, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.828840732574463, "learning_rate": 5e-05, "loss": 0.723, "mean_token_accuracy": 0.8708688020706177, "num_tokens": 374036322.0, "step": 9850 }, { "epoch": 1.2531484543951152, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.914258003234863, "learning_rate": 5e-05, "loss": 0.9123, "mean_token_accuracy": 0.8691688179969788, "num_tokens": 374079144.0, "step": 9851 }, { "epoch": 1.2532756646737058, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.255100727081299, "learning_rate": 5e-05, "loss": 0.7826, "mean_token_accuracy": 0.8697168827056885, "num_tokens": 374116200.0, "step": 9852 }, { "epoch": 1.253402874952296, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.423462867736816, "learning_rate": 5e-05, "loss": 0.8917, "mean_token_accuracy": 0.8655835390090942, "num_tokens": 374151009.0, "step": 9853 }, { "epoch": 1.2535300852308866, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.697299003601074, "learning_rate": 5e-05, "loss": 0.8947, "mean_token_accuracy": 0.8492024540901184, "num_tokens": 374187974.0, "step": 9854 }, { "epoch": 1.2536572955094771, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.808746814727783, "learning_rate": 5e-05, "loss": 0.8365, "mean_token_accuracy": 0.8601353168487549, "num_tokens": 374229203.0, "step": 9855 }, { "epoch": 1.2537845057880677, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.906769275665283, "learning_rate": 5e-05, "loss": 0.7772, "mean_token_accuracy": 0.8812187910079956, "num_tokens": 374261606.0, "step": 9856 }, { "epoch": 1.2539117160666582, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.459644794464111, "learning_rate": 5e-05, "loss": 0.773, "mean_token_accuracy": 0.8645375967025757, "num_tokens": 374300274.0, "step": 9857 }, { "epoch": 1.2540389263452487, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.962368965148926, "learning_rate": 5e-05, "loss": 0.866, "mean_token_accuracy": 0.8505619764328003, "num_tokens": 374336942.0, "step": 9858 }, { "epoch": 1.2541661366238392, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.333467483520508, "learning_rate": 5e-05, "loss": 0.7632, "mean_token_accuracy": 0.8629584312438965, "num_tokens": 374370213.0, "step": 9859 }, { "epoch": 1.2542933469024298, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.89621639251709, "learning_rate": 5e-05, "loss": 0.8001, "mean_token_accuracy": 0.8645555973052979, "num_tokens": 374410653.0, "step": 9860 }, { "epoch": 1.2544205571810203, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.3680806159973145, "learning_rate": 5e-05, "loss": 0.7562, "mean_token_accuracy": 0.8596969842910767, "num_tokens": 374454449.0, "step": 9861 }, { "epoch": 1.2545477674596106, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.78788948059082, "learning_rate": 5e-05, "loss": 0.7408, "mean_token_accuracy": 0.8760545253753662, "num_tokens": 374497494.0, "step": 9862 }, { "epoch": 1.2546749777382011, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.35788106918335, "learning_rate": 5e-05, "loss": 0.7099, "mean_token_accuracy": 0.8765801191329956, "num_tokens": 374534739.0, "step": 9863 }, { "epoch": 1.2548021880167917, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.838672637939453, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8657147288322449, "num_tokens": 374573609.0, "step": 9864 }, { "epoch": 1.2549293982953822, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.225237846374512, "learning_rate": 5e-05, "loss": 0.6985, "mean_token_accuracy": 0.876961350440979, "num_tokens": 374610269.0, "step": 9865 }, { "epoch": 1.2550566085739727, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.960448741912842, "learning_rate": 5e-05, "loss": 0.7853, "mean_token_accuracy": 0.8682605028152466, "num_tokens": 374644888.0, "step": 9866 }, { "epoch": 1.2551838188525632, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000514984130859375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.076176166534424, "learning_rate": 5e-05, "loss": 0.6843, "mean_token_accuracy": 0.8749064207077026, "num_tokens": 374685845.0, "step": 9867 }, { "epoch": 1.2553110291311538, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.12633991241455, "learning_rate": 5e-05, "loss": 0.7664, "mean_token_accuracy": 0.8739739060401917, "num_tokens": 374723501.0, "step": 9868 }, { "epoch": 1.2554382394097443, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.862536430358887, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.8645068407058716, "num_tokens": 374763082.0, "step": 9869 }, { "epoch": 1.2555654496883348, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.440776824951172, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8783366084098816, "num_tokens": 374802829.0, "step": 9870 }, { "epoch": 1.2556926599669254, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.562706470489502, "learning_rate": 5e-05, "loss": 0.6672, "mean_token_accuracy": 0.8777024745941162, "num_tokens": 374838802.0, "step": 9871 }, { "epoch": 1.255819870245516, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.868669509887695, "learning_rate": 5e-05, "loss": 0.797, "mean_token_accuracy": 0.890965461730957, "num_tokens": 374876831.0, "step": 9872 }, { "epoch": 1.2559470805241064, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.328327655792236, "learning_rate": 5e-05, "loss": 0.6331, "mean_token_accuracy": 0.8885972499847412, "num_tokens": 374918120.0, "step": 9873 }, { "epoch": 1.256074290802697, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.227705001831055, "learning_rate": 5e-05, "loss": 0.9156, "mean_token_accuracy": 0.8661671876907349, "num_tokens": 374958247.0, "step": 9874 }, { "epoch": 1.2562015010812875, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.554342269897461, "learning_rate": 5e-05, "loss": 0.7371, "mean_token_accuracy": 0.8664589524269104, "num_tokens": 374998461.0, "step": 9875 }, { "epoch": 1.256328711359878, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.969009399414062, "learning_rate": 5e-05, "loss": 0.9009, "mean_token_accuracy": 0.8779173493385315, "num_tokens": 375029813.0, "step": 9876 }, { "epoch": 1.2564559216384683, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.347646713256836, "learning_rate": 5e-05, "loss": 0.8076, "mean_token_accuracy": 0.8762240409851074, "num_tokens": 375070724.0, "step": 9877 }, { "epoch": 1.2565831319170588, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.922330856323242, "learning_rate": 5e-05, "loss": 0.83, "mean_token_accuracy": 0.8697550296783447, "num_tokens": 375110634.0, "step": 9878 }, { "epoch": 1.2567103421956494, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.268548965454102, "learning_rate": 5e-05, "loss": 0.8436, "mean_token_accuracy": 0.8794794678688049, "num_tokens": 375150054.0, "step": 9879 }, { "epoch": 1.25683755247424, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.788022041320801, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.8608236908912659, "num_tokens": 375185406.0, "step": 9880 }, { "epoch": 1.2569647627528304, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.048327445983887, "learning_rate": 5e-05, "loss": 0.8947, "mean_token_accuracy": 0.8766293525695801, "num_tokens": 375227544.0, "step": 9881 }, { "epoch": 1.257091973031421, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.28022575378418, "learning_rate": 5e-05, "loss": 0.6708, "mean_token_accuracy": 0.8754018545150757, "num_tokens": 375263523.0, "step": 9882 }, { "epoch": 1.2572191833100115, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.517951965332031, "learning_rate": 5e-05, "loss": 0.9476, "mean_token_accuracy": 0.8783652186393738, "num_tokens": 375304764.0, "step": 9883 }, { "epoch": 1.257346393588602, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.606910228729248, "learning_rate": 5e-05, "loss": 0.7291, "mean_token_accuracy": 0.8773815631866455, "num_tokens": 375340989.0, "step": 9884 }, { "epoch": 1.2574736038671925, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.312926292419434, "learning_rate": 5e-05, "loss": 0.9752, "mean_token_accuracy": 0.8711643218994141, "num_tokens": 375374987.0, "step": 9885 }, { "epoch": 1.2576008141457828, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.549539566040039, "learning_rate": 5e-05, "loss": 0.9132, "mean_token_accuracy": 0.8637798428535461, "num_tokens": 375415847.0, "step": 9886 }, { "epoch": 1.2577280244243734, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.32927131652832, "learning_rate": 5e-05, "loss": 0.8582, "mean_token_accuracy": 0.8858321905136108, "num_tokens": 375455761.0, "step": 9887 }, { "epoch": 1.257855234702964, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.22774600982666, "learning_rate": 5e-05, "loss": 0.8676, "mean_token_accuracy": 0.8869196772575378, "num_tokens": 375493774.0, "step": 9888 }, { "epoch": 1.2579824449815544, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.472630023956299, "learning_rate": 5e-05, "loss": 0.8285, "mean_token_accuracy": 0.8645970821380615, "num_tokens": 375531947.0, "step": 9889 }, { "epoch": 1.258109655260145, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.253588676452637, "learning_rate": 5e-05, "loss": 0.8458, "mean_token_accuracy": 0.8833715915679932, "num_tokens": 375573546.0, "step": 9890 }, { "epoch": 1.2582368655387355, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.541362285614014, "learning_rate": 5e-05, "loss": 0.8555, "mean_token_accuracy": 0.8516077399253845, "num_tokens": 375617351.0, "step": 9891 }, { "epoch": 1.258364075817326, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.583579063415527, "learning_rate": 5e-05, "loss": 0.7927, "mean_token_accuracy": 0.8715648651123047, "num_tokens": 375653939.0, "step": 9892 }, { "epoch": 1.2584912860959165, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.104982376098633, "learning_rate": 5e-05, "loss": 0.8536, "mean_token_accuracy": 0.8625465631484985, "num_tokens": 375695345.0, "step": 9893 }, { "epoch": 1.258618496374507, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.925825595855713, "learning_rate": 5e-05, "loss": 0.6665, "mean_token_accuracy": 0.8894318342208862, "num_tokens": 375731786.0, "step": 9894 }, { "epoch": 1.2587457066530976, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.579351425170898, "learning_rate": 5e-05, "loss": 0.8547, "mean_token_accuracy": 0.871131420135498, "num_tokens": 375773258.0, "step": 9895 }, { "epoch": 1.2588729169316881, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.604788303375244, "learning_rate": 5e-05, "loss": 0.6622, "mean_token_accuracy": 0.8850147128105164, "num_tokens": 375805712.0, "step": 9896 }, { "epoch": 1.2590001272102787, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.932513236999512, "learning_rate": 5e-05, "loss": 0.8965, "mean_token_accuracy": 0.8699930906295776, "num_tokens": 375835927.0, "step": 9897 }, { "epoch": 1.2591273374888692, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.518395900726318, "learning_rate": 5e-05, "loss": 0.6911, "mean_token_accuracy": 0.8815130591392517, "num_tokens": 375870568.0, "step": 9898 }, { "epoch": 1.2592545477674597, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.228826522827148, "learning_rate": 5e-05, "loss": 0.9578, "mean_token_accuracy": 0.8719567060470581, "num_tokens": 375904941.0, "step": 9899 }, { "epoch": 1.2593817580460502, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.785369396209717, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.856934666633606, "num_tokens": 375941193.0, "step": 9900 }, { "epoch": 1.2595089683246408, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.106595039367676, "learning_rate": 5e-05, "loss": 0.9422, "mean_token_accuracy": 0.8783091902732849, "num_tokens": 375977952.0, "step": 9901 }, { "epoch": 1.259636178603231, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.390427112579346, "learning_rate": 5e-05, "loss": 0.8779, "mean_token_accuracy": 0.86339271068573, "num_tokens": 376016711.0, "step": 9902 }, { "epoch": 1.2597633888818216, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.489350318908691, "learning_rate": 5e-05, "loss": 0.9582, "mean_token_accuracy": 0.8602988719940186, "num_tokens": 376061753.0, "step": 9903 }, { "epoch": 1.2598905991604121, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.808081150054932, "learning_rate": 5e-05, "loss": 0.9077, "mean_token_accuracy": 0.8649318218231201, "num_tokens": 376100148.0, "step": 9904 }, { "epoch": 1.2600178094390027, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.940168380737305, "learning_rate": 5e-05, "loss": 0.8392, "mean_token_accuracy": 0.876534104347229, "num_tokens": 376140663.0, "step": 9905 }, { "epoch": 1.2601450197175932, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.754483699798584, "learning_rate": 5e-05, "loss": 0.9153, "mean_token_accuracy": 0.8539596796035767, "num_tokens": 376186702.0, "step": 9906 }, { "epoch": 1.2602722299961837, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.894284725189209, "learning_rate": 5e-05, "loss": 0.8492, "mean_token_accuracy": 0.8714923858642578, "num_tokens": 376222736.0, "step": 9907 }, { "epoch": 1.2603994402747742, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.4533796310424805, "learning_rate": 5e-05, "loss": 0.8318, "mean_token_accuracy": 0.858523964881897, "num_tokens": 376258871.0, "step": 9908 }, { "epoch": 1.2605266505533648, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.979379177093506, "learning_rate": 5e-05, "loss": 0.8421, "mean_token_accuracy": 0.8689930438995361, "num_tokens": 376295486.0, "step": 9909 }, { "epoch": 1.2606538608319553, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.357857704162598, "learning_rate": 5e-05, "loss": 0.752, "mean_token_accuracy": 0.8755578398704529, "num_tokens": 376333138.0, "step": 9910 }, { "epoch": 1.2607810711105456, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.846562385559082, "learning_rate": 5e-05, "loss": 0.7557, "mean_token_accuracy": 0.8807944655418396, "num_tokens": 376367284.0, "step": 9911 }, { "epoch": 1.2609082813891361, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.507862091064453, "learning_rate": 5e-05, "loss": 0.8002, "mean_token_accuracy": 0.8593567609786987, "num_tokens": 376398318.0, "step": 9912 }, { "epoch": 1.2610354916677267, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.5380024909973145, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.8803800344467163, "num_tokens": 376430599.0, "step": 9913 }, { "epoch": 1.2611627019463172, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.752614974975586, "learning_rate": 5e-05, "loss": 0.7587, "mean_token_accuracy": 0.8753636479377747, "num_tokens": 376467238.0, "step": 9914 }, { "epoch": 1.2612899122249077, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.262068271636963, "learning_rate": 5e-05, "loss": 0.7347, "mean_token_accuracy": 0.8670902252197266, "num_tokens": 376505728.0, "step": 9915 }, { "epoch": 1.2614171225034982, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.931063175201416, "learning_rate": 5e-05, "loss": 0.8431, "mean_token_accuracy": 0.8535804152488708, "num_tokens": 376550387.0, "step": 9916 }, { "epoch": 1.2615443327820888, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.0971150398254395, "learning_rate": 5e-05, "loss": 0.6624, "mean_token_accuracy": 0.8886516690254211, "num_tokens": 376583365.0, "step": 9917 }, { "epoch": 1.2616715430606793, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.064352035522461, "learning_rate": 5e-05, "loss": 0.7602, "mean_token_accuracy": 0.8801407814025879, "num_tokens": 376620625.0, "step": 9918 }, { "epoch": 1.2617987533392698, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.9240312576293945, "learning_rate": 5e-05, "loss": 0.6897, "mean_token_accuracy": 0.8741564750671387, "num_tokens": 376657048.0, "step": 9919 }, { "epoch": 1.2619259636178604, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.307621002197266, "learning_rate": 5e-05, "loss": 0.8146, "mean_token_accuracy": 0.8661742806434631, "num_tokens": 376697000.0, "step": 9920 }, { "epoch": 1.2620531738964509, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.63450813293457, "learning_rate": 5e-05, "loss": 0.6978, "mean_token_accuracy": 0.8708341121673584, "num_tokens": 376739306.0, "step": 9921 }, { "epoch": 1.2621803841750414, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.752668380737305, "learning_rate": 5e-05, "loss": 0.832, "mean_token_accuracy": 0.8772099018096924, "num_tokens": 376777136.0, "step": 9922 }, { "epoch": 1.262307594453632, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.296100616455078, "learning_rate": 5e-05, "loss": 0.6546, "mean_token_accuracy": 0.8793978691101074, "num_tokens": 376823875.0, "step": 9923 }, { "epoch": 1.2624348047322225, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.303281784057617, "learning_rate": 5e-05, "loss": 0.9321, "mean_token_accuracy": 0.8685553073883057, "num_tokens": 376861335.0, "step": 9924 }, { "epoch": 1.262562015010813, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.287527561187744, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8586612939834595, "num_tokens": 376896076.0, "step": 9925 }, { "epoch": 1.2626892252894033, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.488685607910156, "learning_rate": 5e-05, "loss": 1.0073, "mean_token_accuracy": 0.8623501658439636, "num_tokens": 376943761.0, "step": 9926 }, { "epoch": 1.2628164355679938, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.0820159912109375, "learning_rate": 5e-05, "loss": 0.8555, "mean_token_accuracy": 0.8672348856925964, "num_tokens": 376983526.0, "step": 9927 }, { "epoch": 1.2629436458465844, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.65682601928711, "learning_rate": 5e-05, "loss": 0.8943, "mean_token_accuracy": 0.884645938873291, "num_tokens": 377025293.0, "step": 9928 }, { "epoch": 1.263070856125175, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.184895515441895, "learning_rate": 5e-05, "loss": 0.9304, "mean_token_accuracy": 0.8712302446365356, "num_tokens": 377071380.0, "step": 9929 }, { "epoch": 1.2631980664037654, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.13722562789917, "learning_rate": 5e-05, "loss": 0.7975, "mean_token_accuracy": 0.861919105052948, "num_tokens": 377110215.0, "step": 9930 }, { "epoch": 1.263325276682356, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.096848487854004, "learning_rate": 5e-05, "loss": 0.9245, "mean_token_accuracy": 0.8824424147605896, "num_tokens": 377150191.0, "step": 9931 }, { "epoch": 1.2634524869609465, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.337579727172852, "learning_rate": 5e-05, "loss": 0.769, "mean_token_accuracy": 0.8541231155395508, "num_tokens": 377190636.0, "step": 9932 }, { "epoch": 1.263579697239537, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.595433235168457, "learning_rate": 5e-05, "loss": 0.9851, "mean_token_accuracy": 0.8724343180656433, "num_tokens": 377230314.0, "step": 9933 }, { "epoch": 1.2637069075181275, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 6.717499256134033, "learning_rate": 5e-05, "loss": 0.8102, "mean_token_accuracy": 0.8659144043922424, "num_tokens": 377270830.0, "step": 9934 }, { "epoch": 1.2638341177967178, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.235101699829102, "learning_rate": 5e-05, "loss": 1.0406, "mean_token_accuracy": 0.8543561100959778, "num_tokens": 377313862.0, "step": 9935 }, { "epoch": 1.2639613280753084, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.716870307922363, "learning_rate": 5e-05, "loss": 0.937, "mean_token_accuracy": 0.8674089908599854, "num_tokens": 377355451.0, "step": 9936 }, { "epoch": 1.264088538353899, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.116848945617676, "learning_rate": 5e-05, "loss": 0.8967, "mean_token_accuracy": 0.8669967651367188, "num_tokens": 377397608.0, "step": 9937 }, { "epoch": 1.2642157486324894, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.309160232543945, "learning_rate": 5e-05, "loss": 0.9285, "mean_token_accuracy": 0.8724745512008667, "num_tokens": 377436268.0, "step": 9938 }, { "epoch": 1.26434295891108, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.359870910644531, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.8706912994384766, "num_tokens": 377470896.0, "step": 9939 }, { "epoch": 1.2644701691896705, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.298019409179688, "learning_rate": 5e-05, "loss": 0.8721, "mean_token_accuracy": 0.8770509958267212, "num_tokens": 377508828.0, "step": 9940 }, { "epoch": 1.264597379468261, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.439930438995361, "learning_rate": 5e-05, "loss": 0.8515, "mean_token_accuracy": 0.856303334236145, "num_tokens": 377542007.0, "step": 9941 }, { "epoch": 1.2647245897468515, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.76474666595459, "learning_rate": 5e-05, "loss": 0.843, "mean_token_accuracy": 0.8671495318412781, "num_tokens": 377582671.0, "step": 9942 }, { "epoch": 1.264851800025442, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.790849208831787, "learning_rate": 5e-05, "loss": 0.8094, "mean_token_accuracy": 0.8749814033508301, "num_tokens": 377624840.0, "step": 9943 }, { "epoch": 1.2649790103040326, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.34574556350708, "learning_rate": 5e-05, "loss": 0.7552, "mean_token_accuracy": 0.8713192939758301, "num_tokens": 377661955.0, "step": 9944 }, { "epoch": 1.2651062205826231, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.959601402282715, "learning_rate": 5e-05, "loss": 0.7969, "mean_token_accuracy": 0.8765724897384644, "num_tokens": 377698574.0, "step": 9945 }, { "epoch": 1.2652334308612136, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.143411636352539, "learning_rate": 5e-05, "loss": 0.757, "mean_token_accuracy": 0.862364649772644, "num_tokens": 377739874.0, "step": 9946 }, { "epoch": 1.2653606411398042, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.14767837524414, "learning_rate": 5e-05, "loss": 0.8697, "mean_token_accuracy": 0.8550534844398499, "num_tokens": 377775644.0, "step": 9947 }, { "epoch": 1.2654878514183947, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.871685028076172, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8635405898094177, "num_tokens": 377812648.0, "step": 9948 }, { "epoch": 1.2656150616969852, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.526744842529297, "learning_rate": 5e-05, "loss": 0.8889, "mean_token_accuracy": 0.8582931756973267, "num_tokens": 377849928.0, "step": 9949 }, { "epoch": 1.2657422719755758, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.545838832855225, "learning_rate": 5e-05, "loss": 0.6889, "mean_token_accuracy": 0.876381516456604, "num_tokens": 377891154.0, "step": 9950 }, { "epoch": 1.265869482254166, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.948888778686523, "learning_rate": 5e-05, "loss": 0.904, "mean_token_accuracy": 0.8678871393203735, "num_tokens": 377926858.0, "step": 9951 }, { "epoch": 1.2659966925327566, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.408807277679443, "learning_rate": 5e-05, "loss": 0.693, "mean_token_accuracy": 0.8786389827728271, "num_tokens": 377976895.0, "step": 9952 }, { "epoch": 1.2661239028113471, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.257660865783691, "learning_rate": 5e-05, "loss": 0.9488, "mean_token_accuracy": 0.870978593826294, "num_tokens": 378010300.0, "step": 9953 }, { "epoch": 1.2662511130899377, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.616356372833252, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8700985312461853, "num_tokens": 378051764.0, "step": 9954 }, { "epoch": 1.2663783233685282, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.246012687683105, "learning_rate": 5e-05, "loss": 0.9776, "mean_token_accuracy": 0.8668580055236816, "num_tokens": 378087976.0, "step": 9955 }, { "epoch": 1.2665055336471187, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.193749904632568, "learning_rate": 5e-05, "loss": 0.8599, "mean_token_accuracy": 0.8701361417770386, "num_tokens": 378124062.0, "step": 9956 }, { "epoch": 1.2666327439257092, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.739066123962402, "learning_rate": 5e-05, "loss": 0.9348, "mean_token_accuracy": 0.8703010082244873, "num_tokens": 378161248.0, "step": 9957 }, { "epoch": 1.2667599542042998, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 7.7270708084106445, "learning_rate": 5e-05, "loss": 0.9183, "mean_token_accuracy": 0.8636173009872437, "num_tokens": 378197959.0, "step": 9958 }, { "epoch": 1.2668871644828903, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.061649322509766, "learning_rate": 5e-05, "loss": 0.8822, "mean_token_accuracy": 0.8665633797645569, "num_tokens": 378233030.0, "step": 9959 }, { "epoch": 1.2670143747614806, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.841951847076416, "learning_rate": 5e-05, "loss": 0.9133, "mean_token_accuracy": 0.8639819622039795, "num_tokens": 378265147.0, "step": 9960 }, { "epoch": 1.2671415850400711, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.8404669761657715, "learning_rate": 5e-05, "loss": 0.8422, "mean_token_accuracy": 0.8657068610191345, "num_tokens": 378298759.0, "step": 9961 }, { "epoch": 1.2672687953186617, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.497162818908691, "learning_rate": 5e-05, "loss": 0.8591, "mean_token_accuracy": 0.8539174199104309, "num_tokens": 378338456.0, "step": 9962 }, { "epoch": 1.2673960055972522, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.096324920654297, "learning_rate": 5e-05, "loss": 0.8596, "mean_token_accuracy": 0.8726235032081604, "num_tokens": 378374755.0, "step": 9963 }, { "epoch": 1.2675232158758427, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.090392112731934, "learning_rate": 5e-05, "loss": 0.7092, "mean_token_accuracy": 0.8870183229446411, "num_tokens": 378415302.0, "step": 9964 }, { "epoch": 1.2676504261544332, "ewc_loss": 0.5, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.36453628540039, "learning_rate": 5e-05, "loss": 0.8727, "mean_token_accuracy": 0.8673803210258484, "num_tokens": 378454400.0, "step": 9965 }, { "epoch": 1.2677776364330238, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.9554548263549805, "learning_rate": 5e-05, "loss": 0.7343, "mean_token_accuracy": 0.8738903999328613, "num_tokens": 378496986.0, "step": 9966 }, { "epoch": 1.2679048467116143, "ewc_loss": 0.5, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.371246337890625, "learning_rate": 5e-05, "loss": 0.8451, "mean_token_accuracy": 0.8798978328704834, "num_tokens": 378533937.0, "step": 9967 }, { "epoch": 1.2680320569902048, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.062281608581543, "learning_rate": 5e-05, "loss": 0.7349, "mean_token_accuracy": 0.8773566484451294, "num_tokens": 378567420.0, "step": 9968 }, { "epoch": 1.2681592672687954, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.214015007019043, "learning_rate": 5e-05, "loss": 0.8382, "mean_token_accuracy": 0.8775951862335205, "num_tokens": 378600939.0, "step": 9969 }, { "epoch": 1.2682864775473859, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.20556116104126, "learning_rate": 5e-05, "loss": 0.7978, "mean_token_accuracy": 0.8643043041229248, "num_tokens": 378647858.0, "step": 9970 }, { "epoch": 1.2684136878259764, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.04347038269043, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.8703778982162476, "num_tokens": 378686695.0, "step": 9971 }, { "epoch": 1.268540898104567, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.317674160003662, "learning_rate": 5e-05, "loss": 0.7614, "mean_token_accuracy": 0.8760052919387817, "num_tokens": 378732583.0, "step": 9972 }, { "epoch": 1.2686681083831575, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.895742416381836, "learning_rate": 5e-05, "loss": 0.78, "mean_token_accuracy": 0.8806043863296509, "num_tokens": 378768134.0, "step": 9973 }, { "epoch": 1.268795318661748, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.37897253036499, "learning_rate": 5e-05, "loss": 0.7296, "mean_token_accuracy": 0.8843667507171631, "num_tokens": 378799684.0, "step": 9974 }, { "epoch": 1.2689225289403383, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.771804332733154, "learning_rate": 5e-05, "loss": 0.7994, "mean_token_accuracy": 0.8654265403747559, "num_tokens": 378837831.0, "step": 9975 }, { "epoch": 1.2690497392189288, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.42474365234375, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.8707000613212585, "num_tokens": 378876697.0, "step": 9976 }, { "epoch": 1.2691769494975194, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.668411731719971, "learning_rate": 5e-05, "loss": 0.7407, "mean_token_accuracy": 0.8842216730117798, "num_tokens": 378916301.0, "step": 9977 }, { "epoch": 1.2693041597761099, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.435474872589111, "learning_rate": 5e-05, "loss": 0.765, "mean_token_accuracy": 0.8664806485176086, "num_tokens": 378956294.0, "step": 9978 }, { "epoch": 1.2694313700547004, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.635142803192139, "learning_rate": 5e-05, "loss": 0.7869, "mean_token_accuracy": 0.8628345131874084, "num_tokens": 378992704.0, "step": 9979 }, { "epoch": 1.269558580333291, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.408388614654541, "learning_rate": 5e-05, "loss": 0.7187, "mean_token_accuracy": 0.8786939978599548, "num_tokens": 379034955.0, "step": 9980 }, { "epoch": 1.2696857906118815, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.624740123748779, "learning_rate": 5e-05, "loss": 0.7172, "mean_token_accuracy": 0.8824314475059509, "num_tokens": 379075026.0, "step": 9981 }, { "epoch": 1.269813000890472, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.394464015960693, "learning_rate": 5e-05, "loss": 0.7358, "mean_token_accuracy": 0.8682820796966553, "num_tokens": 379113452.0, "step": 9982 }, { "epoch": 1.2699402111690625, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.625813961029053, "learning_rate": 5e-05, "loss": 0.8099, "mean_token_accuracy": 0.8467158079147339, "num_tokens": 379155632.0, "step": 9983 }, { "epoch": 1.2700674214476528, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.362003803253174, "learning_rate": 5e-05, "loss": 0.7051, "mean_token_accuracy": 0.8799213171005249, "num_tokens": 379188987.0, "step": 9984 }, { "epoch": 1.2701946317262434, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.66132116317749, "learning_rate": 5e-05, "loss": 0.7165, "mean_token_accuracy": 0.8749972581863403, "num_tokens": 379229099.0, "step": 9985 }, { "epoch": 1.270321842004834, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.274526596069336, "learning_rate": 5e-05, "loss": 0.6872, "mean_token_accuracy": 0.8814602494239807, "num_tokens": 379265365.0, "step": 9986 }, { "epoch": 1.2704490522834244, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.737818241119385, "learning_rate": 5e-05, "loss": 0.7277, "mean_token_accuracy": 0.8784692883491516, "num_tokens": 379301170.0, "step": 9987 }, { "epoch": 1.270576262562015, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.160465717315674, "learning_rate": 5e-05, "loss": 0.7012, "mean_token_accuracy": 0.8692072033882141, "num_tokens": 379334983.0, "step": 9988 }, { "epoch": 1.2707034728406055, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.885164260864258, "learning_rate": 5e-05, "loss": 0.7728, "mean_token_accuracy": 0.8612387180328369, "num_tokens": 379371584.0, "step": 9989 }, { "epoch": 1.270830683119196, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.9550461769104, "learning_rate": 5e-05, "loss": 0.694, "mean_token_accuracy": 0.8719006180763245, "num_tokens": 379415363.0, "step": 9990 }, { "epoch": 1.2709578933977865, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.15892219543457, "learning_rate": 5e-05, "loss": 0.7345, "mean_token_accuracy": 0.8826326727867126, "num_tokens": 379447019.0, "step": 9991 }, { "epoch": 1.271085103676377, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.610631942749023, "learning_rate": 5e-05, "loss": 0.6507, "mean_token_accuracy": 0.8799394965171814, "num_tokens": 379490229.0, "step": 9992 }, { "epoch": 1.2712123139549676, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.666181564331055, "learning_rate": 5e-05, "loss": 0.8197, "mean_token_accuracy": 0.8726999759674072, "num_tokens": 379528608.0, "step": 9993 }, { "epoch": 1.2713395242335581, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.14880895614624, "learning_rate": 5e-05, "loss": 0.6967, "mean_token_accuracy": 0.8639963865280151, "num_tokens": 379566916.0, "step": 9994 }, { "epoch": 1.2714667345121486, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.39372444152832, "learning_rate": 5e-05, "loss": 0.8717, "mean_token_accuracy": 0.8857579231262207, "num_tokens": 379600648.0, "step": 9995 }, { "epoch": 1.2715939447907392, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.9060845375061035, "learning_rate": 5e-05, "loss": 0.6952, "mean_token_accuracy": 0.8666234016418457, "num_tokens": 379639519.0, "step": 9996 }, { "epoch": 1.2717211550693297, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.950438499450684, "learning_rate": 5e-05, "loss": 0.9774, "mean_token_accuracy": 0.8791886568069458, "num_tokens": 379674426.0, "step": 9997 }, { "epoch": 1.2718483653479202, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 6.769602298736572, "learning_rate": 5e-05, "loss": 0.8012, "mean_token_accuracy": 0.8790401220321655, "num_tokens": 379713121.0, "step": 9998 }, { "epoch": 1.2719755756265108, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.255474090576172, "learning_rate": 5e-05, "loss": 0.9514, "mean_token_accuracy": 0.87758469581604, "num_tokens": 379749952.0, "step": 9999 }, { "epoch": 1.272102785905101, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.366775512695312, "learning_rate": 5e-05, "loss": 1.0355, "mean_token_accuracy": 0.8622448444366455, "num_tokens": 379786057.0, "step": 10000 }, { "epoch": 1.2722299961836916, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.029614448547363, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8622470498085022, "num_tokens": 379820034.0, "step": 10001 }, { "epoch": 1.2723572064622821, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.319669723510742, "learning_rate": 5e-05, "loss": 1.06, "mean_token_accuracy": 0.8771086931228638, "num_tokens": 379863969.0, "step": 10002 }, { "epoch": 1.2724844167408726, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 5.836548328399658, "learning_rate": 5e-05, "loss": 0.6814, "mean_token_accuracy": 0.8794987201690674, "num_tokens": 379903360.0, "step": 10003 }, { "epoch": 1.2726116270194632, "ewc_loss": 1.0703125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00101470947265625, "grad_norm": 12.781448364257812, "learning_rate": 5e-05, "loss": 1.4332, "mean_token_accuracy": 0.8742527961730957, "num_tokens": 379935874.0, "step": 10004 }, { "epoch": 1.2727388372980537, "ewc_loss": 0.921875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0008697509765625, "grad_norm": 9.539290428161621, "learning_rate": 5e-05, "loss": 1.3002, "mean_token_accuracy": 0.8637210726737976, "num_tokens": 379972525.0, "step": 10005 }, { "epoch": 1.2728660475766442, "ewc_loss": 0.78125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 11.77199935913086, "learning_rate": 5e-05, "loss": 1.142, "mean_token_accuracy": 0.8731774091720581, "num_tokens": 380012925.0, "step": 10006 }, { "epoch": 1.2729932578552348, "ewc_loss": 0.9453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000896453857421875, "grad_norm": 10.018195152282715, "learning_rate": 5e-05, "loss": 1.3501, "mean_token_accuracy": 0.8613492250442505, "num_tokens": 380048235.0, "step": 10007 }, { "epoch": 1.2731204681338253, "ewc_loss": 0.765625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 11.232609748840332, "learning_rate": 5e-05, "loss": 1.207, "mean_token_accuracy": 0.8510690927505493, "num_tokens": 380084607.0, "step": 10008 }, { "epoch": 1.2732476784124156, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.022613525390625, "learning_rate": 5e-05, "loss": 1.0669, "mean_token_accuracy": 0.86097651720047, "num_tokens": 380121514.0, "step": 10009 }, { "epoch": 1.2733748886910061, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 10.754881858825684, "learning_rate": 5e-05, "loss": 1.136, "mean_token_accuracy": 0.8640455603599548, "num_tokens": 380158640.0, "step": 10010 }, { "epoch": 1.2735020989695967, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.731375694274902, "learning_rate": 5e-05, "loss": 0.9929, "mean_token_accuracy": 0.8680927157402039, "num_tokens": 380196195.0, "step": 10011 }, { "epoch": 1.2736293092481872, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.330408096313477, "learning_rate": 5e-05, "loss": 1.0828, "mean_token_accuracy": 0.8687238097190857, "num_tokens": 380234777.0, "step": 10012 }, { "epoch": 1.2737565195267777, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.68549919128418, "learning_rate": 5e-05, "loss": 1.0154, "mean_token_accuracy": 0.8654645681381226, "num_tokens": 380276494.0, "step": 10013 }, { "epoch": 1.2738837298053682, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.892291069030762, "learning_rate": 5e-05, "loss": 1.0313, "mean_token_accuracy": 0.8696837425231934, "num_tokens": 380314791.0, "step": 10014 }, { "epoch": 1.2740109400839588, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.503764152526855, "learning_rate": 5e-05, "loss": 0.9895, "mean_token_accuracy": 0.8689671158790588, "num_tokens": 380354372.0, "step": 10015 }, { "epoch": 1.2741381503625493, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.486387252807617, "learning_rate": 5e-05, "loss": 0.9546, "mean_token_accuracy": 0.8822368383407593, "num_tokens": 380388022.0, "step": 10016 }, { "epoch": 1.2742653606411398, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.083024978637695, "learning_rate": 5e-05, "loss": 0.908, "mean_token_accuracy": 0.8632645606994629, "num_tokens": 380426665.0, "step": 10017 }, { "epoch": 1.2743925709197303, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.548869132995605, "learning_rate": 5e-05, "loss": 0.9686, "mean_token_accuracy": 0.8808959722518921, "num_tokens": 380466068.0, "step": 10018 }, { "epoch": 1.2745197811983209, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.608908653259277, "learning_rate": 5e-05, "loss": 0.7945, "mean_token_accuracy": 0.8852030634880066, "num_tokens": 380500370.0, "step": 10019 }, { "epoch": 1.2746469914769114, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.102615356445312, "learning_rate": 5e-05, "loss": 1.0608, "mean_token_accuracy": 0.8720720410346985, "num_tokens": 380538886.0, "step": 10020 }, { "epoch": 1.274774201755502, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.857629776000977, "learning_rate": 5e-05, "loss": 0.8924, "mean_token_accuracy": 0.8728851079940796, "num_tokens": 380577425.0, "step": 10021 }, { "epoch": 1.2749014120340925, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 10.308258056640625, "learning_rate": 5e-05, "loss": 1.137, "mean_token_accuracy": 0.8586097359657288, "num_tokens": 380616999.0, "step": 10022 }, { "epoch": 1.275028622312683, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.666476249694824, "learning_rate": 5e-05, "loss": 1.0034, "mean_token_accuracy": 0.8696389198303223, "num_tokens": 380654065.0, "step": 10023 }, { "epoch": 1.2751558325912733, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.507987022399902, "learning_rate": 5e-05, "loss": 1.0112, "mean_token_accuracy": 0.8674715757369995, "num_tokens": 380687947.0, "step": 10024 }, { "epoch": 1.2752830428698638, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.829713821411133, "learning_rate": 5e-05, "loss": 0.952, "mean_token_accuracy": 0.8793458342552185, "num_tokens": 380728387.0, "step": 10025 }, { "epoch": 1.2754102531484544, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.508088111877441, "learning_rate": 5e-05, "loss": 0.9321, "mean_token_accuracy": 0.8649478554725647, "num_tokens": 380763906.0, "step": 10026 }, { "epoch": 1.2755374634270449, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.444265365600586, "learning_rate": 5e-05, "loss": 0.9211, "mean_token_accuracy": 0.860923171043396, "num_tokens": 380806904.0, "step": 10027 }, { "epoch": 1.2756646737056354, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.114754676818848, "learning_rate": 5e-05, "loss": 0.8385, "mean_token_accuracy": 0.8796387910842896, "num_tokens": 380842694.0, "step": 10028 }, { "epoch": 1.275791883984226, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.7378058433532715, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8808076977729797, "num_tokens": 380879624.0, "step": 10029 }, { "epoch": 1.2759190942628165, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.461461067199707, "learning_rate": 5e-05, "loss": 0.8712, "mean_token_accuracy": 0.8740642070770264, "num_tokens": 380916358.0, "step": 10030 }, { "epoch": 1.276046304541407, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.0140252113342285, "learning_rate": 5e-05, "loss": 0.7107, "mean_token_accuracy": 0.8824827075004578, "num_tokens": 380957858.0, "step": 10031 }, { "epoch": 1.2761735148199975, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.127541542053223, "learning_rate": 5e-05, "loss": 0.9103, "mean_token_accuracy": 0.8797527551651001, "num_tokens": 380997301.0, "step": 10032 }, { "epoch": 1.2763007250985878, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.795457363128662, "learning_rate": 5e-05, "loss": 0.7022, "mean_token_accuracy": 0.8792175054550171, "num_tokens": 381034023.0, "step": 10033 }, { "epoch": 1.2764279353771784, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.698588371276855, "learning_rate": 5e-05, "loss": 1.049, "mean_token_accuracy": 0.8586947917938232, "num_tokens": 381073402.0, "step": 10034 }, { "epoch": 1.2765551456557689, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.412418365478516, "learning_rate": 5e-05, "loss": 0.8098, "mean_token_accuracy": 0.880326509475708, "num_tokens": 381111182.0, "step": 10035 }, { "epoch": 1.2766823559343594, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.550857543945312, "learning_rate": 5e-05, "loss": 1.0027, "mean_token_accuracy": 0.873056173324585, "num_tokens": 381153527.0, "step": 10036 }, { "epoch": 1.27680956621295, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.422636032104492, "learning_rate": 5e-05, "loss": 0.9484, "mean_token_accuracy": 0.8706679344177246, "num_tokens": 381196337.0, "step": 10037 }, { "epoch": 1.2769367764915405, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.251465797424316, "learning_rate": 5e-05, "loss": 0.8824, "mean_token_accuracy": 0.8695551753044128, "num_tokens": 381239065.0, "step": 10038 }, { "epoch": 1.277063986770131, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.695833206176758, "learning_rate": 5e-05, "loss": 0.8705, "mean_token_accuracy": 0.8880600333213806, "num_tokens": 381274826.0, "step": 10039 }, { "epoch": 1.2771911970487215, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.2008585929870605, "learning_rate": 5e-05, "loss": 0.7698, "mean_token_accuracy": 0.8732198476791382, "num_tokens": 381318043.0, "step": 10040 }, { "epoch": 1.277318407327312, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.740367889404297, "learning_rate": 5e-05, "loss": 0.8968, "mean_token_accuracy": 0.8695307970046997, "num_tokens": 381352583.0, "step": 10041 }, { "epoch": 1.2774456176059026, "ewc_loss": 0.375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.035094738006592, "learning_rate": 5e-05, "loss": 0.7658, "mean_token_accuracy": 0.8652068376541138, "num_tokens": 381387271.0, "step": 10042 }, { "epoch": 1.2775728278844931, "ewc_loss": 0.5, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.60589599609375, "learning_rate": 5e-05, "loss": 0.8965, "mean_token_accuracy": 0.8636452555656433, "num_tokens": 381429749.0, "step": 10043 }, { "epoch": 1.2777000381630836, "ewc_loss": 0.375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.1534013748168945, "learning_rate": 5e-05, "loss": 0.7393, "mean_token_accuracy": 0.882047176361084, "num_tokens": 381468763.0, "step": 10044 }, { "epoch": 1.2778272484416742, "ewc_loss": 0.5, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.506345748901367, "learning_rate": 5e-05, "loss": 0.8546, "mean_token_accuracy": 0.8762408494949341, "num_tokens": 381505789.0, "step": 10045 }, { "epoch": 1.2779544587202647, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.204195976257324, "learning_rate": 5e-05, "loss": 0.6758, "mean_token_accuracy": 0.8932594656944275, "num_tokens": 381545861.0, "step": 10046 }, { "epoch": 1.2780816689988552, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.359359741210938, "learning_rate": 5e-05, "loss": 0.9259, "mean_token_accuracy": 0.8541948199272156, "num_tokens": 381585373.0, "step": 10047 }, { "epoch": 1.2782088792774458, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.3666672706604, "learning_rate": 5e-05, "loss": 0.7504, "mean_token_accuracy": 0.8760679364204407, "num_tokens": 381623029.0, "step": 10048 }, { "epoch": 1.278336089556036, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.979538440704346, "learning_rate": 5e-05, "loss": 0.8525, "mean_token_accuracy": 0.8631541132926941, "num_tokens": 381663481.0, "step": 10049 }, { "epoch": 1.2784632998346266, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.578222274780273, "learning_rate": 5e-05, "loss": 0.7043, "mean_token_accuracy": 0.8915751576423645, "num_tokens": 381703977.0, "step": 10050 }, { "epoch": 1.2785905101132171, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.7009735107421875, "learning_rate": 5e-05, "loss": 0.7943, "mean_token_accuracy": 0.8672727346420288, "num_tokens": 381741409.0, "step": 10051 }, { "epoch": 1.2787177203918076, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.571303844451904, "learning_rate": 5e-05, "loss": 0.7691, "mean_token_accuracy": 0.8689422607421875, "num_tokens": 381779013.0, "step": 10052 }, { "epoch": 1.2788449306703982, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.704699516296387, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8710392117500305, "num_tokens": 381813460.0, "step": 10053 }, { "epoch": 1.2789721409489887, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.390616416931152, "learning_rate": 5e-05, "loss": 0.7465, "mean_token_accuracy": 0.8661964535713196, "num_tokens": 381853821.0, "step": 10054 }, { "epoch": 1.2790993512275792, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.858809471130371, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.8622497320175171, "num_tokens": 381895411.0, "step": 10055 }, { "epoch": 1.2792265615061698, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.167073726654053, "learning_rate": 5e-05, "loss": 0.7329, "mean_token_accuracy": 0.8660025000572205, "num_tokens": 381929067.0, "step": 10056 }, { "epoch": 1.2793537717847603, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.08049488067627, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8745230436325073, "num_tokens": 381964002.0, "step": 10057 }, { "epoch": 1.2794809820633506, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.86519718170166, "learning_rate": 5e-05, "loss": 0.6459, "mean_token_accuracy": 0.8872649669647217, "num_tokens": 382002913.0, "step": 10058 }, { "epoch": 1.2796081923419411, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.476300239562988, "learning_rate": 5e-05, "loss": 0.8003, "mean_token_accuracy": 0.8809700012207031, "num_tokens": 382039026.0, "step": 10059 }, { "epoch": 1.2797354026205316, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.516839504241943, "learning_rate": 5e-05, "loss": 0.6839, "mean_token_accuracy": 0.87211012840271, "num_tokens": 382077413.0, "step": 10060 }, { "epoch": 1.2798626128991222, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.916887283325195, "learning_rate": 5e-05, "loss": 0.8828, "mean_token_accuracy": 0.8660058379173279, "num_tokens": 382112686.0, "step": 10061 }, { "epoch": 1.2799898231777127, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.350656032562256, "learning_rate": 5e-05, "loss": 0.7002, "mean_token_accuracy": 0.8704131245613098, "num_tokens": 382152184.0, "step": 10062 }, { "epoch": 1.2801170334563032, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.189964294433594, "learning_rate": 5e-05, "loss": 0.9667, "mean_token_accuracy": 0.8641979694366455, "num_tokens": 382190916.0, "step": 10063 }, { "epoch": 1.2802442437348938, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.660533428192139, "learning_rate": 5e-05, "loss": 0.759, "mean_token_accuracy": 0.8598598837852478, "num_tokens": 382222986.0, "step": 10064 }, { "epoch": 1.2803714540134843, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.88651180267334, "learning_rate": 5e-05, "loss": 0.9095, "mean_token_accuracy": 0.8802883625030518, "num_tokens": 382266039.0, "step": 10065 }, { "epoch": 1.2804986642920748, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.319931983947754, "learning_rate": 5e-05, "loss": 0.7844, "mean_token_accuracy": 0.8743810653686523, "num_tokens": 382302071.0, "step": 10066 }, { "epoch": 1.2806258745706653, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.145013809204102, "learning_rate": 5e-05, "loss": 0.8622, "mean_token_accuracy": 0.869816780090332, "num_tokens": 382340240.0, "step": 10067 }, { "epoch": 1.2807530848492559, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.847927570343018, "learning_rate": 5e-05, "loss": 0.8188, "mean_token_accuracy": 0.8743890523910522, "num_tokens": 382368308.0, "step": 10068 }, { "epoch": 1.2808802951278464, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.515149116516113, "learning_rate": 5e-05, "loss": 0.7687, "mean_token_accuracy": 0.8748183250427246, "num_tokens": 382406365.0, "step": 10069 }, { "epoch": 1.281007505406437, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.980523586273193, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8619749546051025, "num_tokens": 382449005.0, "step": 10070 }, { "epoch": 1.2811347156850275, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.380884647369385, "learning_rate": 5e-05, "loss": 0.7691, "mean_token_accuracy": 0.8674612045288086, "num_tokens": 382487744.0, "step": 10071 }, { "epoch": 1.281261925963618, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.8074259757995605, "learning_rate": 5e-05, "loss": 0.7892, "mean_token_accuracy": 0.8704978823661804, "num_tokens": 382526954.0, "step": 10072 }, { "epoch": 1.2813891362422083, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.466552734375, "learning_rate": 5e-05, "loss": 0.7203, "mean_token_accuracy": 0.8806131482124329, "num_tokens": 382565257.0, "step": 10073 }, { "epoch": 1.2815163465207988, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.671608924865723, "learning_rate": 5e-05, "loss": 0.7837, "mean_token_accuracy": 0.8678271174430847, "num_tokens": 382601992.0, "step": 10074 }, { "epoch": 1.2816435567993893, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.425411224365234, "learning_rate": 5e-05, "loss": 0.743, "mean_token_accuracy": 0.8686560988426208, "num_tokens": 382637181.0, "step": 10075 }, { "epoch": 1.2817707670779799, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.756782054901123, "learning_rate": 5e-05, "loss": 0.797, "mean_token_accuracy": 0.8616517782211304, "num_tokens": 382679886.0, "step": 10076 }, { "epoch": 1.2818979773565704, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.236576557159424, "learning_rate": 5e-05, "loss": 0.7498, "mean_token_accuracy": 0.8662166595458984, "num_tokens": 382717062.0, "step": 10077 }, { "epoch": 1.282025187635161, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.958381652832031, "learning_rate": 5e-05, "loss": 0.7828, "mean_token_accuracy": 0.867290735244751, "num_tokens": 382752188.0, "step": 10078 }, { "epoch": 1.2821523979137515, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.997478485107422, "learning_rate": 5e-05, "loss": 0.7005, "mean_token_accuracy": 0.8693286180496216, "num_tokens": 382785229.0, "step": 10079 }, { "epoch": 1.282279608192342, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.180709838867188, "learning_rate": 5e-05, "loss": 0.7981, "mean_token_accuracy": 0.8684073686599731, "num_tokens": 382823432.0, "step": 10080 }, { "epoch": 1.2824068184709325, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.779860973358154, "learning_rate": 5e-05, "loss": 0.7248, "mean_token_accuracy": 0.8624378442764282, "num_tokens": 382862763.0, "step": 10081 }, { "epoch": 1.2825340287495228, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.46455192565918, "learning_rate": 5e-05, "loss": 0.8155, "mean_token_accuracy": 0.8700635433197021, "num_tokens": 382898312.0, "step": 10082 }, { "epoch": 1.2826612390281134, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.5651984214782715, "learning_rate": 5e-05, "loss": 0.6834, "mean_token_accuracy": 0.8704797029495239, "num_tokens": 382933494.0, "step": 10083 }, { "epoch": 1.2827884493067039, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.753931999206543, "learning_rate": 5e-05, "loss": 0.8208, "mean_token_accuracy": 0.8844640254974365, "num_tokens": 382971029.0, "step": 10084 }, { "epoch": 1.2829156595852944, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.445668697357178, "learning_rate": 5e-05, "loss": 0.7209, "mean_token_accuracy": 0.8639629483222961, "num_tokens": 383003911.0, "step": 10085 }, { "epoch": 1.283042869863885, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.959973335266113, "learning_rate": 5e-05, "loss": 0.9201, "mean_token_accuracy": 0.8646740317344666, "num_tokens": 383042945.0, "step": 10086 }, { "epoch": 1.2831700801424755, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.5979461669921875, "learning_rate": 5e-05, "loss": 0.7121, "mean_token_accuracy": 0.8741593956947327, "num_tokens": 383085944.0, "step": 10087 }, { "epoch": 1.283297290421066, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.810203552246094, "learning_rate": 5e-05, "loss": 0.9334, "mean_token_accuracy": 0.8668918609619141, "num_tokens": 383124854.0, "step": 10088 }, { "epoch": 1.2834245006996565, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.11344575881958, "learning_rate": 5e-05, "loss": 0.7713, "mean_token_accuracy": 0.8759124279022217, "num_tokens": 383159427.0, "step": 10089 }, { "epoch": 1.283551710978247, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.156026840209961, "learning_rate": 5e-05, "loss": 0.8393, "mean_token_accuracy": 0.8734108209609985, "num_tokens": 383199320.0, "step": 10090 }, { "epoch": 1.2836789212568376, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.732923984527588, "learning_rate": 5e-05, "loss": 0.7888, "mean_token_accuracy": 0.8836625814437866, "num_tokens": 383231367.0, "step": 10091 }, { "epoch": 1.283806131535428, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.369401931762695, "learning_rate": 5e-05, "loss": 0.748, "mean_token_accuracy": 0.8743276596069336, "num_tokens": 383268238.0, "step": 10092 }, { "epoch": 1.2839333418140186, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.160758018493652, "learning_rate": 5e-05, "loss": 0.8288, "mean_token_accuracy": 0.8771299719810486, "num_tokens": 383303445.0, "step": 10093 }, { "epoch": 1.2840605520926092, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.998091220855713, "learning_rate": 5e-05, "loss": 0.7729, "mean_token_accuracy": 0.8541164398193359, "num_tokens": 383341106.0, "step": 10094 }, { "epoch": 1.2841877623711997, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.214016914367676, "learning_rate": 5e-05, "loss": 0.8621, "mean_token_accuracy": 0.8647094368934631, "num_tokens": 383379225.0, "step": 10095 }, { "epoch": 1.2843149726497902, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.0260090827941895, "learning_rate": 5e-05, "loss": 0.7222, "mean_token_accuracy": 0.8735561966896057, "num_tokens": 383420798.0, "step": 10096 }, { "epoch": 1.2844421829283807, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.044743537902832, "learning_rate": 5e-05, "loss": 0.8476, "mean_token_accuracy": 0.8583385944366455, "num_tokens": 383467145.0, "step": 10097 }, { "epoch": 1.284569393206971, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000518798828125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.191317558288574, "learning_rate": 5e-05, "loss": 0.7704, "mean_token_accuracy": 0.8589469790458679, "num_tokens": 383505143.0, "step": 10098 }, { "epoch": 1.2846966034855616, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.848390579223633, "learning_rate": 5e-05, "loss": 0.8078, "mean_token_accuracy": 0.8626360893249512, "num_tokens": 383545987.0, "step": 10099 }, { "epoch": 1.2848238137641521, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.310424327850342, "learning_rate": 5e-05, "loss": 0.7307, "mean_token_accuracy": 0.8748002052307129, "num_tokens": 383578546.0, "step": 10100 }, { "epoch": 1.2849510240427426, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.701183795928955, "learning_rate": 5e-05, "loss": 0.7569, "mean_token_accuracy": 0.8752239346504211, "num_tokens": 383619603.0, "step": 10101 }, { "epoch": 1.2850782343213332, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.380333423614502, "learning_rate": 5e-05, "loss": 0.7723, "mean_token_accuracy": 0.859758198261261, "num_tokens": 383660716.0, "step": 10102 }, { "epoch": 1.2852054445999237, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.617268085479736, "learning_rate": 5e-05, "loss": 0.7385, "mean_token_accuracy": 0.8751558065414429, "num_tokens": 383704344.0, "step": 10103 }, { "epoch": 1.2853326548785142, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.434881210327148, "learning_rate": 5e-05, "loss": 0.7704, "mean_token_accuracy": 0.8607886433601379, "num_tokens": 383738141.0, "step": 10104 }, { "epoch": 1.2854598651571048, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.530031204223633, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8612716794013977, "num_tokens": 383774682.0, "step": 10105 }, { "epoch": 1.2855870754356953, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.460627555847168, "learning_rate": 5e-05, "loss": 0.7123, "mean_token_accuracy": 0.8799259662628174, "num_tokens": 383814056.0, "step": 10106 }, { "epoch": 1.2857142857142856, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.500843048095703, "learning_rate": 5e-05, "loss": 0.7649, "mean_token_accuracy": 0.8584328889846802, "num_tokens": 383858151.0, "step": 10107 }, { "epoch": 1.2858414959928761, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.460689067840576, "learning_rate": 5e-05, "loss": 0.7292, "mean_token_accuracy": 0.8677514791488647, "num_tokens": 383895121.0, "step": 10108 }, { "epoch": 1.2859687062714666, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.497284889221191, "learning_rate": 5e-05, "loss": 0.773, "mean_token_accuracy": 0.8566027879714966, "num_tokens": 383929273.0, "step": 10109 }, { "epoch": 1.2860959165500572, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.403566837310791, "learning_rate": 5e-05, "loss": 0.711, "mean_token_accuracy": 0.870681643486023, "num_tokens": 383970374.0, "step": 10110 }, { "epoch": 1.2862231268286477, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.5567193031311035, "learning_rate": 5e-05, "loss": 0.7255, "mean_token_accuracy": 0.870148777961731, "num_tokens": 384010375.0, "step": 10111 }, { "epoch": 1.2863503371072382, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.299234867095947, "learning_rate": 5e-05, "loss": 0.7346, "mean_token_accuracy": 0.8639211058616638, "num_tokens": 384040547.0, "step": 10112 }, { "epoch": 1.2864775473858288, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.6766357421875, "learning_rate": 5e-05, "loss": 0.743, "mean_token_accuracy": 0.866231381893158, "num_tokens": 384076005.0, "step": 10113 }, { "epoch": 1.2866047576644193, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 7.1367034912109375, "learning_rate": 5e-05, "loss": 0.6647, "mean_token_accuracy": 0.8802021145820618, "num_tokens": 384112568.0, "step": 10114 }, { "epoch": 1.2867319679430098, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.897805690765381, "learning_rate": 5e-05, "loss": 0.733, "mean_token_accuracy": 0.8794500231742859, "num_tokens": 384151942.0, "step": 10115 }, { "epoch": 1.2868591782216003, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.822713375091553, "learning_rate": 5e-05, "loss": 0.6758, "mean_token_accuracy": 0.8697028160095215, "num_tokens": 384192534.0, "step": 10116 }, { "epoch": 1.2869863885001909, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.361434936523438, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8776616454124451, "num_tokens": 384230600.0, "step": 10117 }, { "epoch": 1.2871135987787814, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00022602081298828125, "grad_norm": 6.257691860198975, "learning_rate": 5e-05, "loss": 0.592, "mean_token_accuracy": 0.8888726830482483, "num_tokens": 384266566.0, "step": 10118 }, { "epoch": 1.287240809057372, "ewc_loss": 0.5, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 9.277235984802246, "learning_rate": 5e-05, "loss": 0.892, "mean_token_accuracy": 0.8641164302825928, "num_tokens": 384304536.0, "step": 10119 }, { "epoch": 1.2873680193359625, "ewc_loss": 0.251953125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00019931793212890625, "grad_norm": 5.425739765167236, "learning_rate": 5e-05, "loss": 0.6192, "mean_token_accuracy": 0.8699853420257568, "num_tokens": 384347865.0, "step": 10120 }, { "epoch": 1.287495229614553, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.907533645629883, "learning_rate": 5e-05, "loss": 1.0668, "mean_token_accuracy": 0.8662031888961792, "num_tokens": 384380060.0, "step": 10121 }, { "epoch": 1.2876224398931433, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 5.715022563934326, "learning_rate": 5e-05, "loss": 0.7689, "mean_token_accuracy": 0.8552765846252441, "num_tokens": 384417574.0, "step": 10122 }, { "epoch": 1.2877496501717338, "ewc_loss": 1.0703125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00101470947265625, "grad_norm": 12.646647453308105, "learning_rate": 5e-05, "loss": 1.4383, "mean_token_accuracy": 0.875611424446106, "num_tokens": 384454591.0, "step": 10123 }, { "epoch": 1.2878768604503243, "ewc_loss": 0.9921875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00093841552734375, "grad_norm": 9.448897361755371, "learning_rate": 5e-05, "loss": 1.3594, "mean_token_accuracy": 0.8711817860603333, "num_tokens": 384489642.0, "step": 10124 }, { "epoch": 1.2880040707289149, "ewc_loss": 0.734375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 11.382055282592773, "learning_rate": 5e-05, "loss": 1.1182, "mean_token_accuracy": 0.8678576350212097, "num_tokens": 384528993.0, "step": 10125 }, { "epoch": 1.2881312810075054, "ewc_loss": 1.0078125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000957489013671875, "grad_norm": 9.896306037902832, "learning_rate": 5e-05, "loss": 1.3875, "mean_token_accuracy": 0.8673557639122009, "num_tokens": 384569321.0, "step": 10126 }, { "epoch": 1.288258491286096, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.28630256652832, "learning_rate": 5e-05, "loss": 1.0654, "mean_token_accuracy": 0.8655402064323425, "num_tokens": 384614821.0, "step": 10127 }, { "epoch": 1.2883857015646865, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 9.37732982635498, "learning_rate": 5e-05, "loss": 1.0918, "mean_token_accuracy": 0.8816392421722412, "num_tokens": 384658557.0, "step": 10128 }, { "epoch": 1.288512911843277, "ewc_loss": 0.625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.909914016723633, "learning_rate": 5e-05, "loss": 1.0362, "mean_token_accuracy": 0.8597205877304077, "num_tokens": 384697121.0, "step": 10129 }, { "epoch": 1.2886401221218675, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.518996238708496, "learning_rate": 5e-05, "loss": 1.0825, "mean_token_accuracy": 0.8546216487884521, "num_tokens": 384735410.0, "step": 10130 }, { "epoch": 1.2887673324004578, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.003274917602539, "learning_rate": 5e-05, "loss": 0.9456, "mean_token_accuracy": 0.8683151006698608, "num_tokens": 384769394.0, "step": 10131 }, { "epoch": 1.2888945426790483, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.596786499023438, "learning_rate": 5e-05, "loss": 1.0137, "mean_token_accuracy": 0.8700424432754517, "num_tokens": 384808845.0, "step": 10132 }, { "epoch": 1.2890217529576389, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.701014518737793, "learning_rate": 5e-05, "loss": 0.9202, "mean_token_accuracy": 0.871575117111206, "num_tokens": 384848129.0, "step": 10133 }, { "epoch": 1.2891489632362294, "ewc_loss": 0.625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.394840240478516, "learning_rate": 5e-05, "loss": 0.9554, "mean_token_accuracy": 0.8870687484741211, "num_tokens": 384885181.0, "step": 10134 }, { "epoch": 1.28927617351482, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.82753849029541, "learning_rate": 5e-05, "loss": 0.9012, "mean_token_accuracy": 0.8780088424682617, "num_tokens": 384923499.0, "step": 10135 }, { "epoch": 1.2894033837934105, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.734698295593262, "learning_rate": 5e-05, "loss": 0.8945, "mean_token_accuracy": 0.8811349868774414, "num_tokens": 384957367.0, "step": 10136 }, { "epoch": 1.289530594072001, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.123929023742676, "learning_rate": 5e-05, "loss": 0.9237, "mean_token_accuracy": 0.8768948316574097, "num_tokens": 385000409.0, "step": 10137 }, { "epoch": 1.2896578043505915, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.676520347595215, "learning_rate": 5e-05, "loss": 0.8114, "mean_token_accuracy": 0.8665059804916382, "num_tokens": 385038439.0, "step": 10138 }, { "epoch": 1.289785014629182, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.728907585144043, "learning_rate": 5e-05, "loss": 0.9361, "mean_token_accuracy": 0.8755295276641846, "num_tokens": 385073950.0, "step": 10139 }, { "epoch": 1.2899122249077726, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.683485984802246, "learning_rate": 5e-05, "loss": 0.7226, "mean_token_accuracy": 0.8652881383895874, "num_tokens": 385112394.0, "step": 10140 }, { "epoch": 1.290039435186363, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.753989219665527, "learning_rate": 5e-05, "loss": 1.0037, "mean_token_accuracy": 0.8691462278366089, "num_tokens": 385149096.0, "step": 10141 }, { "epoch": 1.2901666454649536, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.3809733390808105, "learning_rate": 5e-05, "loss": 0.7565, "mean_token_accuracy": 0.8641738891601562, "num_tokens": 385181986.0, "step": 10142 }, { "epoch": 1.2902938557435442, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.936914443969727, "learning_rate": 5e-05, "loss": 1.2391, "mean_token_accuracy": 0.8429625034332275, "num_tokens": 385220730.0, "step": 10143 }, { "epoch": 1.2904210660221347, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 7.984492301940918, "learning_rate": 5e-05, "loss": 1.0575, "mean_token_accuracy": 0.8540948629379272, "num_tokens": 385262250.0, "step": 10144 }, { "epoch": 1.2905482763007252, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.655312538146973, "learning_rate": 5e-05, "loss": 1.1074, "mean_token_accuracy": 0.8652986884117126, "num_tokens": 385303518.0, "step": 10145 }, { "epoch": 1.2906754865793157, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00077056884765625, "grad_norm": 9.323185920715332, "learning_rate": 5e-05, "loss": 1.1943, "mean_token_accuracy": 0.8697043657302856, "num_tokens": 385342303.0, "step": 10146 }, { "epoch": 1.290802696857906, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.048140525817871, "learning_rate": 5e-05, "loss": 0.9093, "mean_token_accuracy": 0.8753840327262878, "num_tokens": 385381581.0, "step": 10147 }, { "epoch": 1.2909299071364966, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 9.386881828308105, "learning_rate": 5e-05, "loss": 1.0441, "mean_token_accuracy": 0.8778752088546753, "num_tokens": 385414835.0, "step": 10148 }, { "epoch": 1.291057117415087, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.9330573081970215, "learning_rate": 5e-05, "loss": 0.877, "mean_token_accuracy": 0.870361328125, "num_tokens": 385454022.0, "step": 10149 }, { "epoch": 1.2911843276936776, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.311851501464844, "learning_rate": 5e-05, "loss": 1.0362, "mean_token_accuracy": 0.8594328165054321, "num_tokens": 385486479.0, "step": 10150 }, { "epoch": 1.2913115379722682, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.83746862411499, "learning_rate": 5e-05, "loss": 0.9337, "mean_token_accuracy": 0.8615832328796387, "num_tokens": 385527942.0, "step": 10151 }, { "epoch": 1.2914387482508587, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.618735313415527, "learning_rate": 5e-05, "loss": 0.9308, "mean_token_accuracy": 0.8674409985542297, "num_tokens": 385567994.0, "step": 10152 }, { "epoch": 1.2915659585294492, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.11857795715332, "learning_rate": 5e-05, "loss": 0.9054, "mean_token_accuracy": 0.8764209747314453, "num_tokens": 385603908.0, "step": 10153 }, { "epoch": 1.2916931688080397, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.664265155792236, "learning_rate": 5e-05, "loss": 0.8052, "mean_token_accuracy": 0.8716404438018799, "num_tokens": 385641461.0, "step": 10154 }, { "epoch": 1.2918203790866303, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.451143264770508, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.877490758895874, "num_tokens": 385676677.0, "step": 10155 }, { "epoch": 1.2919475893652206, "ewc_loss": 0.375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.026381969451904, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.8706165552139282, "num_tokens": 385715674.0, "step": 10156 }, { "epoch": 1.2920747996438111, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.88709545135498, "learning_rate": 5e-05, "loss": 0.9395, "mean_token_accuracy": 0.8673663139343262, "num_tokens": 385747947.0, "step": 10157 }, { "epoch": 1.2922020099224016, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.733180999755859, "learning_rate": 5e-05, "loss": 0.7586, "mean_token_accuracy": 0.8647785186767578, "num_tokens": 385787395.0, "step": 10158 }, { "epoch": 1.2923292202009922, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.412373542785645, "learning_rate": 5e-05, "loss": 0.9489, "mean_token_accuracy": 0.881300687789917, "num_tokens": 385821724.0, "step": 10159 }, { "epoch": 1.2924564304795827, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 6.964312553405762, "learning_rate": 5e-05, "loss": 0.7726, "mean_token_accuracy": 0.8916483521461487, "num_tokens": 385856281.0, "step": 10160 }, { "epoch": 1.2925836407581732, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.468812942504883, "learning_rate": 5e-05, "loss": 1.008, "mean_token_accuracy": 0.8670568466186523, "num_tokens": 385895979.0, "step": 10161 }, { "epoch": 1.2927108510367638, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.650575160980225, "learning_rate": 5e-05, "loss": 0.9064, "mean_token_accuracy": 0.8790537714958191, "num_tokens": 385933134.0, "step": 10162 }, { "epoch": 1.2928380613153543, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.84821891784668, "learning_rate": 5e-05, "loss": 0.9461, "mean_token_accuracy": 0.8663602471351624, "num_tokens": 385974487.0, "step": 10163 }, { "epoch": 1.2929652715939448, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.986705303192139, "learning_rate": 5e-05, "loss": 0.9884, "mean_token_accuracy": 0.8601039052009583, "num_tokens": 386014594.0, "step": 10164 }, { "epoch": 1.2930924818725353, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.20981216430664, "learning_rate": 5e-05, "loss": 0.8916, "mean_token_accuracy": 0.8643472790718079, "num_tokens": 386044982.0, "step": 10165 }, { "epoch": 1.2932196921511259, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.82852029800415, "learning_rate": 5e-05, "loss": 0.8785, "mean_token_accuracy": 0.8767995834350586, "num_tokens": 386084517.0, "step": 10166 }, { "epoch": 1.2933469024297164, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.054627418518066, "learning_rate": 5e-05, "loss": 0.8398, "mean_token_accuracy": 0.8722303509712219, "num_tokens": 386121841.0, "step": 10167 }, { "epoch": 1.293474112708307, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.437349319458008, "learning_rate": 5e-05, "loss": 0.7884, "mean_token_accuracy": 0.8734849691390991, "num_tokens": 386157426.0, "step": 10168 }, { "epoch": 1.2936013229868975, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.221010208129883, "learning_rate": 5e-05, "loss": 0.8553, "mean_token_accuracy": 0.8670318722724915, "num_tokens": 386196739.0, "step": 10169 }, { "epoch": 1.293728533265488, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.049051284790039, "learning_rate": 5e-05, "loss": 0.694, "mean_token_accuracy": 0.8859541416168213, "num_tokens": 386234335.0, "step": 10170 }, { "epoch": 1.2938557435440783, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000522613525390625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.500899314880371, "learning_rate": 5e-05, "loss": 0.8797, "mean_token_accuracy": 0.8690733909606934, "num_tokens": 386274166.0, "step": 10171 }, { "epoch": 1.2939829538226688, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.830308437347412, "learning_rate": 5e-05, "loss": 0.753, "mean_token_accuracy": 0.8615472316741943, "num_tokens": 386313709.0, "step": 10172 }, { "epoch": 1.2941101641012593, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.748602867126465, "learning_rate": 5e-05, "loss": 0.9472, "mean_token_accuracy": 0.8535856604576111, "num_tokens": 386345516.0, "step": 10173 }, { "epoch": 1.2942373743798499, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.803938388824463, "learning_rate": 5e-05, "loss": 0.6968, "mean_token_accuracy": 0.8826974630355835, "num_tokens": 386383284.0, "step": 10174 }, { "epoch": 1.2943645846584404, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.909957885742188, "learning_rate": 5e-05, "loss": 0.9649, "mean_token_accuracy": 0.8654549717903137, "num_tokens": 386423141.0, "step": 10175 }, { "epoch": 1.294491794937031, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.0057454109191895, "learning_rate": 5e-05, "loss": 0.7668, "mean_token_accuracy": 0.8792040348052979, "num_tokens": 386461888.0, "step": 10176 }, { "epoch": 1.2946190052156215, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.776622772216797, "learning_rate": 5e-05, "loss": 0.948, "mean_token_accuracy": 0.8690112829208374, "num_tokens": 386500061.0, "step": 10177 }, { "epoch": 1.294746215494212, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.445459842681885, "learning_rate": 5e-05, "loss": 0.7888, "mean_token_accuracy": 0.8897461891174316, "num_tokens": 386538843.0, "step": 10178 }, { "epoch": 1.2948734257728025, "ewc_loss": 0.5, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.172371864318848, "learning_rate": 5e-05, "loss": 0.8737, "mean_token_accuracy": 0.8695562481880188, "num_tokens": 386584346.0, "step": 10179 }, { "epoch": 1.2950006360513928, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.851386547088623, "learning_rate": 5e-05, "loss": 0.8163, "mean_token_accuracy": 0.8834807276725769, "num_tokens": 386619574.0, "step": 10180 }, { "epoch": 1.2951278463299833, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.521114349365234, "learning_rate": 5e-05, "loss": 0.7828, "mean_token_accuracy": 0.873988926410675, "num_tokens": 386655111.0, "step": 10181 }, { "epoch": 1.2952550566085739, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.01900577545166, "learning_rate": 5e-05, "loss": 0.8395, "mean_token_accuracy": 0.8728393316268921, "num_tokens": 386701102.0, "step": 10182 }, { "epoch": 1.2953822668871644, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.259546279907227, "learning_rate": 5e-05, "loss": 0.751, "mean_token_accuracy": 0.8712457418441772, "num_tokens": 386740259.0, "step": 10183 }, { "epoch": 1.295509477165755, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.03991985321045, "learning_rate": 5e-05, "loss": 0.7865, "mean_token_accuracy": 0.8785428404808044, "num_tokens": 386780345.0, "step": 10184 }, { "epoch": 1.2956366874443455, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.158754348754883, "learning_rate": 5e-05, "loss": 0.7231, "mean_token_accuracy": 0.8745365142822266, "num_tokens": 386819401.0, "step": 10185 }, { "epoch": 1.295763897722936, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.090423583984375, "learning_rate": 5e-05, "loss": 0.7916, "mean_token_accuracy": 0.8777344226837158, "num_tokens": 386856617.0, "step": 10186 }, { "epoch": 1.2958911080015265, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.037380218505859, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.8599706292152405, "num_tokens": 386895426.0, "step": 10187 }, { "epoch": 1.296018318280117, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.204767227172852, "learning_rate": 5e-05, "loss": 0.8252, "mean_token_accuracy": 0.86594158411026, "num_tokens": 386931577.0, "step": 10188 }, { "epoch": 1.2961455285587076, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.900949954986572, "learning_rate": 5e-05, "loss": 0.6972, "mean_token_accuracy": 0.8773974180221558, "num_tokens": 386968658.0, "step": 10189 }, { "epoch": 1.296272738837298, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.337918281555176, "learning_rate": 5e-05, "loss": 0.8186, "mean_token_accuracy": 0.8752315044403076, "num_tokens": 387005364.0, "step": 10190 }, { "epoch": 1.2963999491158886, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.841856479644775, "learning_rate": 5e-05, "loss": 0.7069, "mean_token_accuracy": 0.8748860359191895, "num_tokens": 387048767.0, "step": 10191 }, { "epoch": 1.2965271593944792, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.407389640808105, "learning_rate": 5e-05, "loss": 0.8012, "mean_token_accuracy": 0.8816418647766113, "num_tokens": 387083456.0, "step": 10192 }, { "epoch": 1.2966543696730697, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.861067771911621, "learning_rate": 5e-05, "loss": 0.7067, "mean_token_accuracy": 0.8776001930236816, "num_tokens": 387124220.0, "step": 10193 }, { "epoch": 1.2967815799516602, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.405787467956543, "learning_rate": 5e-05, "loss": 0.8824, "mean_token_accuracy": 0.8595885038375854, "num_tokens": 387163825.0, "step": 10194 }, { "epoch": 1.2969087902302507, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.930569648742676, "learning_rate": 5e-05, "loss": 0.7335, "mean_token_accuracy": 0.8713335394859314, "num_tokens": 387201230.0, "step": 10195 }, { "epoch": 1.297036000508841, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.339783668518066, "learning_rate": 5e-05, "loss": 0.8499, "mean_token_accuracy": 0.8682746291160583, "num_tokens": 387238426.0, "step": 10196 }, { "epoch": 1.2971632107874316, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.047410011291504, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8745523691177368, "num_tokens": 387275883.0, "step": 10197 }, { "epoch": 1.297290421066022, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.200332641601562, "learning_rate": 5e-05, "loss": 0.8166, "mean_token_accuracy": 0.8804396986961365, "num_tokens": 387319053.0, "step": 10198 }, { "epoch": 1.2974176313446126, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.1771087646484375, "learning_rate": 5e-05, "loss": 0.7183, "mean_token_accuracy": 0.8795481324195862, "num_tokens": 387355013.0, "step": 10199 }, { "epoch": 1.2975448416232032, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.035649299621582, "learning_rate": 5e-05, "loss": 0.7831, "mean_token_accuracy": 0.881308913230896, "num_tokens": 387392736.0, "step": 10200 }, { "epoch": 1.2976720519017937, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.274742603302002, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.8572055101394653, "num_tokens": 387431938.0, "step": 10201 }, { "epoch": 1.2977992621803842, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.907268047332764, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8717004060745239, "num_tokens": 387472724.0, "step": 10202 }, { "epoch": 1.2979264724589747, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.273114204406738, "learning_rate": 5e-05, "loss": 0.751, "mean_token_accuracy": 0.8722121715545654, "num_tokens": 387513666.0, "step": 10203 }, { "epoch": 1.2980536827375653, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.920292377471924, "learning_rate": 5e-05, "loss": 0.76, "mean_token_accuracy": 0.8797364234924316, "num_tokens": 387544691.0, "step": 10204 }, { "epoch": 1.2981808930161556, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.182167053222656, "learning_rate": 5e-05, "loss": 0.7046, "mean_token_accuracy": 0.8815815448760986, "num_tokens": 387580839.0, "step": 10205 }, { "epoch": 1.298308103294746, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.996352195739746, "learning_rate": 5e-05, "loss": 0.8059, "mean_token_accuracy": 0.8672058582305908, "num_tokens": 387616258.0, "step": 10206 }, { "epoch": 1.2984353135733366, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.060225486755371, "learning_rate": 5e-05, "loss": 0.748, "mean_token_accuracy": 0.8637627363204956, "num_tokens": 387654058.0, "step": 10207 }, { "epoch": 1.2985625238519272, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.110801696777344, "learning_rate": 5e-05, "loss": 0.8209, "mean_token_accuracy": 0.865348219871521, "num_tokens": 387690036.0, "step": 10208 }, { "epoch": 1.2986897341305177, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.9624409675598145, "learning_rate": 5e-05, "loss": 0.7221, "mean_token_accuracy": 0.8704313635826111, "num_tokens": 387729114.0, "step": 10209 }, { "epoch": 1.2988169444091082, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.190415382385254, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.8745015263557434, "num_tokens": 387768225.0, "step": 10210 }, { "epoch": 1.2989441546876987, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.955691337585449, "learning_rate": 5e-05, "loss": 0.7502, "mean_token_accuracy": 0.8621929883956909, "num_tokens": 387802093.0, "step": 10211 }, { "epoch": 1.2990713649662893, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.148691177368164, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.8746684193611145, "num_tokens": 387838857.0, "step": 10212 }, { "epoch": 1.2991985752448798, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.064274787902832, "learning_rate": 5e-05, "loss": 0.7333, "mean_token_accuracy": 0.8720015287399292, "num_tokens": 387878974.0, "step": 10213 }, { "epoch": 1.2993257855234703, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.9681501388549805, "learning_rate": 5e-05, "loss": 0.8199, "mean_token_accuracy": 0.8654857873916626, "num_tokens": 387920157.0, "step": 10214 }, { "epoch": 1.2994529958020609, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.281763553619385, "learning_rate": 5e-05, "loss": 0.7213, "mean_token_accuracy": 0.8747949600219727, "num_tokens": 387953345.0, "step": 10215 }, { "epoch": 1.2995802060806514, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.660619258880615, "learning_rate": 5e-05, "loss": 0.7438, "mean_token_accuracy": 0.8802428245544434, "num_tokens": 387990696.0, "step": 10216 }, { "epoch": 1.299707416359242, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.540114402770996, "learning_rate": 5e-05, "loss": 0.7836, "mean_token_accuracy": 0.8618553876876831, "num_tokens": 388028142.0, "step": 10217 }, { "epoch": 1.2998346266378324, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.343216896057129, "learning_rate": 5e-05, "loss": 0.786, "mean_token_accuracy": 0.8492492437362671, "num_tokens": 388058095.0, "step": 10218 }, { "epoch": 1.299961836916423, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.8163909912109375, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.8709695339202881, "num_tokens": 388097141.0, "step": 10219 }, { "epoch": 1.3000890471950133, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.017197608947754, "learning_rate": 5e-05, "loss": 0.712, "mean_token_accuracy": 0.8719391226768494, "num_tokens": 388132138.0, "step": 10220 }, { "epoch": 1.3002162574736038, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.14163875579834, "learning_rate": 5e-05, "loss": 0.784, "mean_token_accuracy": 0.8738901019096375, "num_tokens": 388172837.0, "step": 10221 }, { "epoch": 1.3003434677521943, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.671139240264893, "learning_rate": 5e-05, "loss": 0.7118, "mean_token_accuracy": 0.8605232238769531, "num_tokens": 388214797.0, "step": 10222 }, { "epoch": 1.3004706780307849, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.572774887084961, "learning_rate": 5e-05, "loss": 0.8349, "mean_token_accuracy": 0.863067626953125, "num_tokens": 388256374.0, "step": 10223 }, { "epoch": 1.3005978883093754, "ewc_loss": 0.296875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.252950191497803, "learning_rate": 5e-05, "loss": 0.6533, "mean_token_accuracy": 0.8747540712356567, "num_tokens": 388291227.0, "step": 10224 }, { "epoch": 1.300725098587966, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 9.21164321899414, "learning_rate": 5e-05, "loss": 0.8674, "mean_token_accuracy": 0.8789963722229004, "num_tokens": 388332963.0, "step": 10225 }, { "epoch": 1.3008523088665565, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.950447082519531, "learning_rate": 5e-05, "loss": 0.6328, "mean_token_accuracy": 0.8867278099060059, "num_tokens": 388372825.0, "step": 10226 }, { "epoch": 1.300979519145147, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.800263404846191, "learning_rate": 5e-05, "loss": 0.9772, "mean_token_accuracy": 0.8722690343856812, "num_tokens": 388411850.0, "step": 10227 }, { "epoch": 1.3011067294237375, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.509130477905273, "learning_rate": 5e-05, "loss": 0.7783, "mean_token_accuracy": 0.8682867884635925, "num_tokens": 388449509.0, "step": 10228 }, { "epoch": 1.3012339397023278, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.442731857299805, "learning_rate": 5e-05, "loss": 1.0229, "mean_token_accuracy": 0.8649379014968872, "num_tokens": 388483698.0, "step": 10229 }, { "epoch": 1.3013611499809183, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 7.933804035186768, "learning_rate": 5e-05, "loss": 0.9692, "mean_token_accuracy": 0.8689786791801453, "num_tokens": 388516832.0, "step": 10230 }, { "epoch": 1.3014883602595089, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.628910541534424, "learning_rate": 5e-05, "loss": 0.8019, "mean_token_accuracy": 0.8818975687026978, "num_tokens": 388553272.0, "step": 10231 }, { "epoch": 1.3016155705380994, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.375711441040039, "learning_rate": 5e-05, "loss": 0.9531, "mean_token_accuracy": 0.8881515264511108, "num_tokens": 388589163.0, "step": 10232 }, { "epoch": 1.30174278081669, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.047229766845703, "learning_rate": 5e-05, "loss": 0.7278, "mean_token_accuracy": 0.8622526526451111, "num_tokens": 388628305.0, "step": 10233 }, { "epoch": 1.3018699910952805, "ewc_loss": 0.78125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.955456733703613, "learning_rate": 5e-05, "loss": 1.1801, "mean_token_accuracy": 0.8632364273071289, "num_tokens": 388673462.0, "step": 10234 }, { "epoch": 1.301997201373871, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.081185817718506, "learning_rate": 5e-05, "loss": 0.9323, "mean_token_accuracy": 0.8619586229324341, "num_tokens": 388709357.0, "step": 10235 }, { "epoch": 1.3021244116524615, "ewc_loss": 0.8125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 11.185206413269043, "learning_rate": 5e-05, "loss": 1.1374, "mean_token_accuracy": 0.8874563574790955, "num_tokens": 388746550.0, "step": 10236 }, { "epoch": 1.302251621931052, "ewc_loss": 0.875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000823974609375, "grad_norm": 9.156952857971191, "learning_rate": 5e-05, "loss": 1.2988, "mean_token_accuracy": 0.8563399910926819, "num_tokens": 388795139.0, "step": 10237 }, { "epoch": 1.3023788322096426, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.623981475830078, "learning_rate": 5e-05, "loss": 0.9736, "mean_token_accuracy": 0.8763842582702637, "num_tokens": 388832282.0, "step": 10238 }, { "epoch": 1.302506042488233, "ewc_loss": 0.75, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 9.283402442932129, "learning_rate": 5e-05, "loss": 1.0929, "mean_token_accuracy": 0.8799543976783752, "num_tokens": 388873974.0, "step": 10239 }, { "epoch": 1.3026332527668236, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.461570739746094, "learning_rate": 5e-05, "loss": 0.9105, "mean_token_accuracy": 0.8856462836265564, "num_tokens": 388914562.0, "step": 10240 }, { "epoch": 1.3027604630454142, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.058648109436035, "learning_rate": 5e-05, "loss": 1.0131, "mean_token_accuracy": 0.8628830313682556, "num_tokens": 388950293.0, "step": 10241 }, { "epoch": 1.3028876733240047, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.203060150146484, "learning_rate": 5e-05, "loss": 0.9619, "mean_token_accuracy": 0.8706066012382507, "num_tokens": 388986649.0, "step": 10242 }, { "epoch": 1.3030148836025952, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.139679908752441, "learning_rate": 5e-05, "loss": 0.8297, "mean_token_accuracy": 0.8835172653198242, "num_tokens": 389027779.0, "step": 10243 }, { "epoch": 1.3031420938811857, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.557390213012695, "learning_rate": 5e-05, "loss": 0.9858, "mean_token_accuracy": 0.8561774492263794, "num_tokens": 389062760.0, "step": 10244 }, { "epoch": 1.303269304159776, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.228323936462402, "learning_rate": 5e-05, "loss": 0.826, "mean_token_accuracy": 0.8601726293563843, "num_tokens": 389101892.0, "step": 10245 }, { "epoch": 1.3033965144383666, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.103364944458008, "learning_rate": 5e-05, "loss": 0.9611, "mean_token_accuracy": 0.8747277855873108, "num_tokens": 389140155.0, "step": 10246 }, { "epoch": 1.303523724716957, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.935020923614502, "learning_rate": 5e-05, "loss": 0.8087, "mean_token_accuracy": 0.8646109104156494, "num_tokens": 389181563.0, "step": 10247 }, { "epoch": 1.3036509349955476, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.319894790649414, "learning_rate": 5e-05, "loss": 0.9644, "mean_token_accuracy": 0.8775500059127808, "num_tokens": 389217415.0, "step": 10248 }, { "epoch": 1.3037781452741382, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.1932244300842285, "learning_rate": 5e-05, "loss": 0.823, "mean_token_accuracy": 0.885719358921051, "num_tokens": 389255889.0, "step": 10249 }, { "epoch": 1.3039053555527287, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.135778427124023, "learning_rate": 5e-05, "loss": 0.9339, "mean_token_accuracy": 0.8782965540885925, "num_tokens": 389297650.0, "step": 10250 }, { "epoch": 1.3040325658313192, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.482608795166016, "learning_rate": 5e-05, "loss": 0.9387, "mean_token_accuracy": 0.8567481637001038, "num_tokens": 389339647.0, "step": 10251 }, { "epoch": 1.3041597761099097, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.70714282989502, "learning_rate": 5e-05, "loss": 0.9227, "mean_token_accuracy": 0.8737029433250427, "num_tokens": 389376135.0, "step": 10252 }, { "epoch": 1.3042869863885003, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.597343921661377, "learning_rate": 5e-05, "loss": 0.9084, "mean_token_accuracy": 0.861769437789917, "num_tokens": 389410516.0, "step": 10253 }, { "epoch": 1.3044141966670906, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.335939407348633, "learning_rate": 5e-05, "loss": 0.8965, "mean_token_accuracy": 0.8742594718933105, "num_tokens": 389452712.0, "step": 10254 }, { "epoch": 1.304541406945681, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.416926860809326, "learning_rate": 5e-05, "loss": 0.828, "mean_token_accuracy": 0.8731684684753418, "num_tokens": 389489257.0, "step": 10255 }, { "epoch": 1.3046686172242716, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.603547096252441, "learning_rate": 5e-05, "loss": 0.9409, "mean_token_accuracy": 0.8582274913787842, "num_tokens": 389524344.0, "step": 10256 }, { "epoch": 1.3047958275028622, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.879331111907959, "learning_rate": 5e-05, "loss": 0.8087, "mean_token_accuracy": 0.8525701761245728, "num_tokens": 389564351.0, "step": 10257 }, { "epoch": 1.3049230377814527, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.168147087097168, "learning_rate": 5e-05, "loss": 0.9614, "mean_token_accuracy": 0.872317910194397, "num_tokens": 389604606.0, "step": 10258 }, { "epoch": 1.3050502480600432, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 6.8163909912109375, "learning_rate": 5e-05, "loss": 0.7802, "mean_token_accuracy": 0.8740917444229126, "num_tokens": 389639600.0, "step": 10259 }, { "epoch": 1.3051774583386337, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.403282165527344, "learning_rate": 5e-05, "loss": 1.0593, "mean_token_accuracy": 0.8616830110549927, "num_tokens": 389677770.0, "step": 10260 }, { "epoch": 1.3053046686172243, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.456961631774902, "learning_rate": 5e-05, "loss": 0.9355, "mean_token_accuracy": 0.8577276468276978, "num_tokens": 389715366.0, "step": 10261 }, { "epoch": 1.3054318788958148, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.878008842468262, "learning_rate": 5e-05, "loss": 0.9646, "mean_token_accuracy": 0.8768502473831177, "num_tokens": 389754745.0, "step": 10262 }, { "epoch": 1.3055590891744053, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.084717750549316, "learning_rate": 5e-05, "loss": 0.9828, "mean_token_accuracy": 0.8632699251174927, "num_tokens": 389798599.0, "step": 10263 }, { "epoch": 1.3056862994529959, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.840744972229004, "learning_rate": 5e-05, "loss": 0.8605, "mean_token_accuracy": 0.8686689138412476, "num_tokens": 389831861.0, "step": 10264 }, { "epoch": 1.3058135097315864, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.385486602783203, "learning_rate": 5e-05, "loss": 0.9738, "mean_token_accuracy": 0.858588695526123, "num_tokens": 389877785.0, "step": 10265 }, { "epoch": 1.305940720010177, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.116592884063721, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8853511214256287, "num_tokens": 389921969.0, "step": 10266 }, { "epoch": 1.3060679302887674, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.461277961730957, "learning_rate": 5e-05, "loss": 0.877, "mean_token_accuracy": 0.8752760291099548, "num_tokens": 389959281.0, "step": 10267 }, { "epoch": 1.306195140567358, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.998961448669434, "learning_rate": 5e-05, "loss": 0.7163, "mean_token_accuracy": 0.8844537734985352, "num_tokens": 389992580.0, "step": 10268 }, { "epoch": 1.3063223508459483, "ewc_loss": 0.5, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.373062133789062, "learning_rate": 5e-05, "loss": 0.9019, "mean_token_accuracy": 0.8629158735275269, "num_tokens": 390029377.0, "step": 10269 }, { "epoch": 1.3064495611245388, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.036511421203613, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8757776618003845, "num_tokens": 390059679.0, "step": 10270 }, { "epoch": 1.3065767714031293, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.336568832397461, "learning_rate": 5e-05, "loss": 0.8667, "mean_token_accuracy": 0.8701825141906738, "num_tokens": 390096751.0, "step": 10271 }, { "epoch": 1.3067039816817199, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.0262322425842285, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8614650368690491, "num_tokens": 390138648.0, "step": 10272 }, { "epoch": 1.3068311919603104, "ewc_loss": 0.5, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.59203815460205, "learning_rate": 5e-05, "loss": 0.8734, "mean_token_accuracy": 0.8707602024078369, "num_tokens": 390176008.0, "step": 10273 }, { "epoch": 1.306958402238901, "ewc_loss": 0.375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.825717926025391, "learning_rate": 5e-05, "loss": 0.7528, "mean_token_accuracy": 0.8676313161849976, "num_tokens": 390218120.0, "step": 10274 }, { "epoch": 1.3070856125174914, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.628315925598145, "learning_rate": 5e-05, "loss": 0.8959, "mean_token_accuracy": 0.8763923048973083, "num_tokens": 390264776.0, "step": 10275 }, { "epoch": 1.307212822796082, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.0442094802856445, "learning_rate": 5e-05, "loss": 0.7592, "mean_token_accuracy": 0.8814400434494019, "num_tokens": 390298798.0, "step": 10276 }, { "epoch": 1.3073400330746725, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.366520881652832, "learning_rate": 5e-05, "loss": 0.9092, "mean_token_accuracy": 0.8653632402420044, "num_tokens": 390333332.0, "step": 10277 }, { "epoch": 1.3074672433532628, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.279608726501465, "learning_rate": 5e-05, "loss": 0.8297, "mean_token_accuracy": 0.8638536930084229, "num_tokens": 390374128.0, "step": 10278 }, { "epoch": 1.3075944536318533, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.112288475036621, "learning_rate": 5e-05, "loss": 0.8651, "mean_token_accuracy": 0.8686274290084839, "num_tokens": 390416509.0, "step": 10279 }, { "epoch": 1.3077216639104439, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.377257823944092, "learning_rate": 5e-05, "loss": 0.7988, "mean_token_accuracy": 0.8730592131614685, "num_tokens": 390448129.0, "step": 10280 }, { "epoch": 1.3078488741890344, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.971301078796387, "learning_rate": 5e-05, "loss": 0.8508, "mean_token_accuracy": 0.8683223128318787, "num_tokens": 390489329.0, "step": 10281 }, { "epoch": 1.307976084467625, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.3119215965271, "learning_rate": 5e-05, "loss": 0.7799, "mean_token_accuracy": 0.8726643323898315, "num_tokens": 390529839.0, "step": 10282 }, { "epoch": 1.3081032947462155, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.040306091308594, "learning_rate": 5e-05, "loss": 0.8501, "mean_token_accuracy": 0.8701809644699097, "num_tokens": 390568888.0, "step": 10283 }, { "epoch": 1.308230505024806, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.122061252593994, "learning_rate": 5e-05, "loss": 0.7351, "mean_token_accuracy": 0.8774633407592773, "num_tokens": 390606709.0, "step": 10284 }, { "epoch": 1.3083577153033965, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.207587242126465, "learning_rate": 5e-05, "loss": 0.8607, "mean_token_accuracy": 0.86835777759552, "num_tokens": 390645027.0, "step": 10285 }, { "epoch": 1.308484925581987, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.952755928039551, "learning_rate": 5e-05, "loss": 0.7485, "mean_token_accuracy": 0.8698472380638123, "num_tokens": 390685923.0, "step": 10286 }, { "epoch": 1.3086121358605776, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.36598014831543, "learning_rate": 5e-05, "loss": 0.8106, "mean_token_accuracy": 0.8860483169555664, "num_tokens": 390728410.0, "step": 10287 }, { "epoch": 1.308739346139168, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.875798225402832, "learning_rate": 5e-05, "loss": 0.7589, "mean_token_accuracy": 0.8624591827392578, "num_tokens": 390764146.0, "step": 10288 }, { "epoch": 1.3088665564177586, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.444106101989746, "learning_rate": 5e-05, "loss": 0.8649, "mean_token_accuracy": 0.8697782754898071, "num_tokens": 390800363.0, "step": 10289 }, { "epoch": 1.3089937666963491, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.980985164642334, "learning_rate": 5e-05, "loss": 0.7502, "mean_token_accuracy": 0.8724673986434937, "num_tokens": 390840213.0, "step": 10290 }, { "epoch": 1.3091209769749397, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.252420425415039, "learning_rate": 5e-05, "loss": 0.8019, "mean_token_accuracy": 0.8869304656982422, "num_tokens": 390879087.0, "step": 10291 }, { "epoch": 1.3092481872535302, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.26016902923584, "learning_rate": 5e-05, "loss": 0.7088, "mean_token_accuracy": 0.8900124430656433, "num_tokens": 390914074.0, "step": 10292 }, { "epoch": 1.3093753975321207, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.83422327041626, "learning_rate": 5e-05, "loss": 0.788, "mean_token_accuracy": 0.8806354403495789, "num_tokens": 390956283.0, "step": 10293 }, { "epoch": 1.309502607810711, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.598611831665039, "learning_rate": 5e-05, "loss": 0.8049, "mean_token_accuracy": 0.8709207773208618, "num_tokens": 390996898.0, "step": 10294 }, { "epoch": 1.3096298180893016, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.415276050567627, "learning_rate": 5e-05, "loss": 0.7262, "mean_token_accuracy": 0.882576584815979, "num_tokens": 391033280.0, "step": 10295 }, { "epoch": 1.309757028367892, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.830243110656738, "learning_rate": 5e-05, "loss": 0.7381, "mean_token_accuracy": 0.8876774907112122, "num_tokens": 391074159.0, "step": 10296 }, { "epoch": 1.3098842386464826, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.143490314483643, "learning_rate": 5e-05, "loss": 0.7123, "mean_token_accuracy": 0.8745107054710388, "num_tokens": 391115649.0, "step": 10297 }, { "epoch": 1.3100114489250732, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.985100269317627, "learning_rate": 5e-05, "loss": 0.7347, "mean_token_accuracy": 0.892339825630188, "num_tokens": 391147601.0, "step": 10298 }, { "epoch": 1.3101386592036637, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.9806742668151855, "learning_rate": 5e-05, "loss": 0.7072, "mean_token_accuracy": 0.8733367323875427, "num_tokens": 391183483.0, "step": 10299 }, { "epoch": 1.3102658694822542, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.099588394165039, "learning_rate": 5e-05, "loss": 0.7493, "mean_token_accuracy": 0.8880366086959839, "num_tokens": 391217346.0, "step": 10300 }, { "epoch": 1.3103930797608447, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.863757133483887, "learning_rate": 5e-05, "loss": 0.6889, "mean_token_accuracy": 0.8787920475006104, "num_tokens": 391255575.0, "step": 10301 }, { "epoch": 1.3105202900394353, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.2623872756958, "learning_rate": 5e-05, "loss": 0.7816, "mean_token_accuracy": 0.8768419027328491, "num_tokens": 391291680.0, "step": 10302 }, { "epoch": 1.3106475003180256, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.7156596183776855, "learning_rate": 5e-05, "loss": 0.7085, "mean_token_accuracy": 0.87286376953125, "num_tokens": 391330034.0, "step": 10303 }, { "epoch": 1.310774710596616, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.458314895629883, "learning_rate": 5e-05, "loss": 0.8156, "mean_token_accuracy": 0.880696177482605, "num_tokens": 391370558.0, "step": 10304 }, { "epoch": 1.3109019208752066, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.614968299865723, "learning_rate": 5e-05, "loss": 0.7097, "mean_token_accuracy": 0.8687306046485901, "num_tokens": 391414091.0, "step": 10305 }, { "epoch": 1.3110291311537972, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.610207557678223, "learning_rate": 5e-05, "loss": 0.8855, "mean_token_accuracy": 0.8615314960479736, "num_tokens": 391458057.0, "step": 10306 }, { "epoch": 1.3111563414323877, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.651931285858154, "learning_rate": 5e-05, "loss": 0.6964, "mean_token_accuracy": 0.8762708306312561, "num_tokens": 391495419.0, "step": 10307 }, { "epoch": 1.3112835517109782, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.558114051818848, "learning_rate": 5e-05, "loss": 0.8724, "mean_token_accuracy": 0.8710657358169556, "num_tokens": 391529353.0, "step": 10308 }, { "epoch": 1.3114107619895687, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.950320243835449, "learning_rate": 5e-05, "loss": 0.8296, "mean_token_accuracy": 0.8477984666824341, "num_tokens": 391571111.0, "step": 10309 }, { "epoch": 1.3115379722681593, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.138652801513672, "learning_rate": 5e-05, "loss": 0.853, "mean_token_accuracy": 0.8685433864593506, "num_tokens": 391612284.0, "step": 10310 }, { "epoch": 1.3116651825467498, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.424910068511963, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.8652587532997131, "num_tokens": 391646934.0, "step": 10311 }, { "epoch": 1.3117923928253403, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.540599346160889, "learning_rate": 5e-05, "loss": 0.7478, "mean_token_accuracy": 0.8801231980323792, "num_tokens": 391686871.0, "step": 10312 }, { "epoch": 1.3119196031039309, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.85641622543335, "learning_rate": 5e-05, "loss": 0.7964, "mean_token_accuracy": 0.8720632791519165, "num_tokens": 391728188.0, "step": 10313 }, { "epoch": 1.3120468133825214, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.1026787757873535, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.8655384182929993, "num_tokens": 391767434.0, "step": 10314 }, { "epoch": 1.312174023661112, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.07351303100586, "learning_rate": 5e-05, "loss": 0.7883, "mean_token_accuracy": 0.8773269653320312, "num_tokens": 391809581.0, "step": 10315 }, { "epoch": 1.3123012339397024, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.949893474578857, "learning_rate": 5e-05, "loss": 0.7404, "mean_token_accuracy": 0.8665656447410583, "num_tokens": 391849877.0, "step": 10316 }, { "epoch": 1.312428444218293, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.085691452026367, "learning_rate": 5e-05, "loss": 0.8206, "mean_token_accuracy": 0.8613286018371582, "num_tokens": 391885995.0, "step": 10317 }, { "epoch": 1.3125556544968833, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.984882831573486, "learning_rate": 5e-05, "loss": 0.6687, "mean_token_accuracy": 0.8816376328468323, "num_tokens": 391919100.0, "step": 10318 }, { "epoch": 1.3126828647754738, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.979948997497559, "learning_rate": 5e-05, "loss": 0.7905, "mean_token_accuracy": 0.8702894449234009, "num_tokens": 391953246.0, "step": 10319 }, { "epoch": 1.3128100750540643, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.112362384796143, "learning_rate": 5e-05, "loss": 0.7052, "mean_token_accuracy": 0.8772048354148865, "num_tokens": 391986618.0, "step": 10320 }, { "epoch": 1.3129372853326549, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.8182597160339355, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8681179285049438, "num_tokens": 392029663.0, "step": 10321 }, { "epoch": 1.3130644956112454, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.23744010925293, "learning_rate": 5e-05, "loss": 0.7302, "mean_token_accuracy": 0.8713333606719971, "num_tokens": 392063329.0, "step": 10322 }, { "epoch": 1.313191705889836, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.678550720214844, "learning_rate": 5e-05, "loss": 0.7309, "mean_token_accuracy": 0.8764504194259644, "num_tokens": 392101400.0, "step": 10323 }, { "epoch": 1.3133189161684264, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.319257736206055, "learning_rate": 5e-05, "loss": 0.7309, "mean_token_accuracy": 0.8669129014015198, "num_tokens": 392140146.0, "step": 10324 }, { "epoch": 1.313446126447017, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.592820167541504, "learning_rate": 5e-05, "loss": 0.7203, "mean_token_accuracy": 0.8798462152481079, "num_tokens": 392176427.0, "step": 10325 }, { "epoch": 1.3135733367256075, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.344147205352783, "learning_rate": 5e-05, "loss": 0.7398, "mean_token_accuracy": 0.8676202893257141, "num_tokens": 392221105.0, "step": 10326 }, { "epoch": 1.3137005470041978, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.588525295257568, "learning_rate": 5e-05, "loss": 0.7875, "mean_token_accuracy": 0.8560455441474915, "num_tokens": 392257301.0, "step": 10327 }, { "epoch": 1.3138277572827883, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.264831066131592, "learning_rate": 5e-05, "loss": 0.7081, "mean_token_accuracy": 0.87483811378479, "num_tokens": 392295849.0, "step": 10328 }, { "epoch": 1.3139549675613789, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.705753326416016, "learning_rate": 5e-05, "loss": 0.7542, "mean_token_accuracy": 0.8624436855316162, "num_tokens": 392336480.0, "step": 10329 }, { "epoch": 1.3140821778399694, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.049246311187744, "learning_rate": 5e-05, "loss": 0.7184, "mean_token_accuracy": 0.8650720119476318, "num_tokens": 392372562.0, "step": 10330 }, { "epoch": 1.31420938811856, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.972897529602051, "learning_rate": 5e-05, "loss": 0.7319, "mean_token_accuracy": 0.8810924887657166, "num_tokens": 392408843.0, "step": 10331 }, { "epoch": 1.3143365983971504, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.72917366027832, "learning_rate": 5e-05, "loss": 0.6709, "mean_token_accuracy": 0.8738319873809814, "num_tokens": 392440725.0, "step": 10332 }, { "epoch": 1.314463808675741, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.436628341674805, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.8594746589660645, "num_tokens": 392484643.0, "step": 10333 }, { "epoch": 1.3145910189543315, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 6.175858974456787, "learning_rate": 5e-05, "loss": 0.6432, "mean_token_accuracy": 0.8755326271057129, "num_tokens": 392523849.0, "step": 10334 }, { "epoch": 1.314718229232922, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 9.347325325012207, "learning_rate": 5e-05, "loss": 0.9018, "mean_token_accuracy": 0.8722577095031738, "num_tokens": 392559598.0, "step": 10335 }, { "epoch": 1.3148454395115126, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 5.516273498535156, "learning_rate": 5e-05, "loss": 0.7037, "mean_token_accuracy": 0.8528569936752319, "num_tokens": 392603856.0, "step": 10336 }, { "epoch": 1.314972649790103, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.663686752319336, "learning_rate": 5e-05, "loss": 1.0385, "mean_token_accuracy": 0.8800537586212158, "num_tokens": 392645317.0, "step": 10337 }, { "epoch": 1.3150998600686936, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.1552252769470215, "learning_rate": 5e-05, "loss": 0.7897, "mean_token_accuracy": 0.8685712814331055, "num_tokens": 392681650.0, "step": 10338 }, { "epoch": 1.3152270703472841, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000843048095703125, "grad_norm": 11.374276161193848, "learning_rate": 5e-05, "loss": 1.2722, "mean_token_accuracy": 0.8632184863090515, "num_tokens": 392711588.0, "step": 10339 }, { "epoch": 1.3153542806258747, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 8.786981582641602, "learning_rate": 5e-05, "loss": 1.2219, "mean_token_accuracy": 0.8764013051986694, "num_tokens": 392758504.0, "step": 10340 }, { "epoch": 1.3154814909044652, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.665581703186035, "learning_rate": 5e-05, "loss": 0.9552, "mean_token_accuracy": 0.8762849569320679, "num_tokens": 392792153.0, "step": 10341 }, { "epoch": 1.3156087011830557, "ewc_loss": 0.90234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00084686279296875, "grad_norm": 9.746967315673828, "learning_rate": 5e-05, "loss": 1.2303, "mean_token_accuracy": 0.8852468729019165, "num_tokens": 392840971.0, "step": 10342 }, { "epoch": 1.315735911461646, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.686617374420166, "learning_rate": 5e-05, "loss": 0.8621, "mean_token_accuracy": 0.8672440052032471, "num_tokens": 392882760.0, "step": 10343 }, { "epoch": 1.3158631217402366, "ewc_loss": 0.94140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000885009765625, "grad_norm": 11.440498352050781, "learning_rate": 5e-05, "loss": 1.3441, "mean_token_accuracy": 0.8601711392402649, "num_tokens": 392922165.0, "step": 10344 }, { "epoch": 1.315990332018827, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 8.546065330505371, "learning_rate": 5e-05, "loss": 1.1086, "mean_token_accuracy": 0.8622427582740784, "num_tokens": 392957181.0, "step": 10345 }, { "epoch": 1.3161175422974176, "ewc_loss": 0.796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000743865966796875, "grad_norm": 11.653897285461426, "learning_rate": 5e-05, "loss": 1.1932, "mean_token_accuracy": 0.8638054728507996, "num_tokens": 392998405.0, "step": 10346 }, { "epoch": 1.3162447525760081, "ewc_loss": 0.93359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00087738037109375, "grad_norm": 9.570917129516602, "learning_rate": 5e-05, "loss": 1.2933, "mean_token_accuracy": 0.8781467080116272, "num_tokens": 393036437.0, "step": 10347 }, { "epoch": 1.3163719628545987, "ewc_loss": 0.703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.990330696105957, "learning_rate": 5e-05, "loss": 1.0509, "mean_token_accuracy": 0.8786928057670593, "num_tokens": 393070711.0, "step": 10348 }, { "epoch": 1.3164991731331892, "ewc_loss": 0.71875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 8.788106918334961, "learning_rate": 5e-05, "loss": 1.1325, "mean_token_accuracy": 0.8556890487670898, "num_tokens": 393109052.0, "step": 10349 }, { "epoch": 1.3166263834117797, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.837767601013184, "learning_rate": 5e-05, "loss": 1.1516, "mean_token_accuracy": 0.8639976978302002, "num_tokens": 393146244.0, "step": 10350 }, { "epoch": 1.3167535936903703, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.982402801513672, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.8749294877052307, "num_tokens": 393180630.0, "step": 10351 }, { "epoch": 1.3168808039689606, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 11.227508544921875, "learning_rate": 5e-05, "loss": 1.1427, "mean_token_accuracy": 0.8715369701385498, "num_tokens": 393222687.0, "step": 10352 }, { "epoch": 1.317008014247551, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.130936622619629, "learning_rate": 5e-05, "loss": 0.9811, "mean_token_accuracy": 0.8662651777267456, "num_tokens": 393267383.0, "step": 10353 }, { "epoch": 1.3171352245261416, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 11.890471458435059, "learning_rate": 5e-05, "loss": 1.192, "mean_token_accuracy": 0.877341628074646, "num_tokens": 393302055.0, "step": 10354 }, { "epoch": 1.3172624348047322, "ewc_loss": 0.75, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 9.480149269104004, "learning_rate": 5e-05, "loss": 1.0762, "mean_token_accuracy": 0.883111834526062, "num_tokens": 393336863.0, "step": 10355 }, { "epoch": 1.3173896450833227, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 11.31946086883545, "learning_rate": 5e-05, "loss": 1.1875, "mean_token_accuracy": 0.8513573408126831, "num_tokens": 393373602.0, "step": 10356 }, { "epoch": 1.3175168553619132, "ewc_loss": 0.734375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 9.810221672058105, "learning_rate": 5e-05, "loss": 1.128, "mean_token_accuracy": 0.8629121780395508, "num_tokens": 393409858.0, "step": 10357 }, { "epoch": 1.3176440656405037, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.455780982971191, "learning_rate": 5e-05, "loss": 1.0398, "mean_token_accuracy": 0.8778829574584961, "num_tokens": 393451198.0, "step": 10358 }, { "epoch": 1.3177712759190943, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.35240650177002, "learning_rate": 5e-05, "loss": 0.9763, "mean_token_accuracy": 0.8812406659126282, "num_tokens": 393486916.0, "step": 10359 }, { "epoch": 1.3178984861976848, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.750777244567871, "learning_rate": 5e-05, "loss": 0.973, "mean_token_accuracy": 0.8839917778968811, "num_tokens": 393520039.0, "step": 10360 }, { "epoch": 1.3180256964762753, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.740155220031738, "learning_rate": 5e-05, "loss": 0.9581, "mean_token_accuracy": 0.862139105796814, "num_tokens": 393557960.0, "step": 10361 }, { "epoch": 1.3181529067548658, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.427933692932129, "learning_rate": 5e-05, "loss": 1.0015, "mean_token_accuracy": 0.8610188961029053, "num_tokens": 393592697.0, "step": 10362 }, { "epoch": 1.3182801170334564, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.011245727539062, "learning_rate": 5e-05, "loss": 0.8477, "mean_token_accuracy": 0.8628374934196472, "num_tokens": 393635423.0, "step": 10363 }, { "epoch": 1.318407327312047, "ewc_loss": 0.625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.698724746704102, "learning_rate": 5e-05, "loss": 0.9683, "mean_token_accuracy": 0.8788384199142456, "num_tokens": 393673508.0, "step": 10364 }, { "epoch": 1.3185345375906374, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.6121673583984375, "learning_rate": 5e-05, "loss": 0.7521, "mean_token_accuracy": 0.8871102929115295, "num_tokens": 393711974.0, "step": 10365 }, { "epoch": 1.318661747869228, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.26594066619873, "learning_rate": 5e-05, "loss": 1.1218, "mean_token_accuracy": 0.8509175777435303, "num_tokens": 393752670.0, "step": 10366 }, { "epoch": 1.3187889581478183, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.058188438415527, "learning_rate": 5e-05, "loss": 0.8731, "mean_token_accuracy": 0.867347240447998, "num_tokens": 393794312.0, "step": 10367 }, { "epoch": 1.3189161684264088, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.31617259979248, "learning_rate": 5e-05, "loss": 1.1064, "mean_token_accuracy": 0.8603342771530151, "num_tokens": 393833830.0, "step": 10368 }, { "epoch": 1.3190433787049993, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.839797019958496, "learning_rate": 5e-05, "loss": 0.9462, "mean_token_accuracy": 0.876358151435852, "num_tokens": 393874844.0, "step": 10369 }, { "epoch": 1.3191705889835899, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.387338638305664, "learning_rate": 5e-05, "loss": 1.0197, "mean_token_accuracy": 0.8576215505599976, "num_tokens": 393913861.0, "step": 10370 }, { "epoch": 1.3192977992621804, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.821623802185059, "learning_rate": 5e-05, "loss": 0.9527, "mean_token_accuracy": 0.8658685088157654, "num_tokens": 393953014.0, "step": 10371 }, { "epoch": 1.319425009540771, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.462373733520508, "learning_rate": 5e-05, "loss": 0.8808, "mean_token_accuracy": 0.8720666766166687, "num_tokens": 393995603.0, "step": 10372 }, { "epoch": 1.3195522198193614, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.309809684753418, "learning_rate": 5e-05, "loss": 0.8027, "mean_token_accuracy": 0.8794138431549072, "num_tokens": 394032127.0, "step": 10373 }, { "epoch": 1.319679430097952, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.219484329223633, "learning_rate": 5e-05, "loss": 0.7855, "mean_token_accuracy": 0.8876842260360718, "num_tokens": 394067153.0, "step": 10374 }, { "epoch": 1.3198066403765425, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.6113433837890625, "learning_rate": 5e-05, "loss": 0.7556, "mean_token_accuracy": 0.8743395209312439, "num_tokens": 394103484.0, "step": 10375 }, { "epoch": 1.3199338506551328, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.396612167358398, "learning_rate": 5e-05, "loss": 0.8278, "mean_token_accuracy": 0.8781673908233643, "num_tokens": 394144302.0, "step": 10376 }, { "epoch": 1.3200610609337233, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.146983623504639, "learning_rate": 5e-05, "loss": 0.7251, "mean_token_accuracy": 0.8699362277984619, "num_tokens": 394179258.0, "step": 10377 }, { "epoch": 1.3201882712123139, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.660088539123535, "learning_rate": 5e-05, "loss": 0.9009, "mean_token_accuracy": 0.8617695569992065, "num_tokens": 394221498.0, "step": 10378 }, { "epoch": 1.3203154814909044, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.883324146270752, "learning_rate": 5e-05, "loss": 0.6879, "mean_token_accuracy": 0.8843384385108948, "num_tokens": 394255985.0, "step": 10379 }, { "epoch": 1.320442691769495, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.0496826171875, "learning_rate": 5e-05, "loss": 0.9457, "mean_token_accuracy": 0.8595269918441772, "num_tokens": 394291020.0, "step": 10380 }, { "epoch": 1.3205699020480854, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.879472255706787, "learning_rate": 5e-05, "loss": 0.7302, "mean_token_accuracy": 0.8700083494186401, "num_tokens": 394329415.0, "step": 10381 }, { "epoch": 1.320697112326676, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.33593463897705, "learning_rate": 5e-05, "loss": 0.966, "mean_token_accuracy": 0.8717516660690308, "num_tokens": 394365287.0, "step": 10382 }, { "epoch": 1.3208243226052665, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.496337890625, "learning_rate": 5e-05, "loss": 0.7743, "mean_token_accuracy": 0.8768051862716675, "num_tokens": 394403749.0, "step": 10383 }, { "epoch": 1.320951532883857, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.783256530761719, "learning_rate": 5e-05, "loss": 0.8544, "mean_token_accuracy": 0.888883113861084, "num_tokens": 394445443.0, "step": 10384 }, { "epoch": 1.3210787431624476, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.146285057067871, "learning_rate": 5e-05, "loss": 0.7818, "mean_token_accuracy": 0.8866133689880371, "num_tokens": 394487650.0, "step": 10385 }, { "epoch": 1.321205953441038, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.650419235229492, "learning_rate": 5e-05, "loss": 0.7637, "mean_token_accuracy": 0.8819541931152344, "num_tokens": 394524606.0, "step": 10386 }, { "epoch": 1.3213331637196286, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.410122871398926, "learning_rate": 5e-05, "loss": 0.8427, "mean_token_accuracy": 0.866162896156311, "num_tokens": 394561177.0, "step": 10387 }, { "epoch": 1.3214603739982191, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.0479512214660645, "learning_rate": 5e-05, "loss": 0.7058, "mean_token_accuracy": 0.8806142210960388, "num_tokens": 394605212.0, "step": 10388 }, { "epoch": 1.3215875842768097, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.605231285095215, "learning_rate": 5e-05, "loss": 0.8371, "mean_token_accuracy": 0.8741844296455383, "num_tokens": 394637662.0, "step": 10389 }, { "epoch": 1.3217147945554002, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.885271072387695, "learning_rate": 5e-05, "loss": 0.6886, "mean_token_accuracy": 0.8803915977478027, "num_tokens": 394675551.0, "step": 10390 }, { "epoch": 1.3218420048339907, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.808513641357422, "learning_rate": 5e-05, "loss": 0.9136, "mean_token_accuracy": 0.8560193181037903, "num_tokens": 394715225.0, "step": 10391 }, { "epoch": 1.321969215112581, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.048498630523682, "learning_rate": 5e-05, "loss": 0.692, "mean_token_accuracy": 0.883114218711853, "num_tokens": 394749872.0, "step": 10392 }, { "epoch": 1.3220964253911716, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.559683799743652, "learning_rate": 5e-05, "loss": 0.9019, "mean_token_accuracy": 0.8604997396469116, "num_tokens": 394791674.0, "step": 10393 }, { "epoch": 1.322223635669762, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.51432991027832, "learning_rate": 5e-05, "loss": 0.7908, "mean_token_accuracy": 0.8631183505058289, "num_tokens": 394829556.0, "step": 10394 }, { "epoch": 1.3223508459483526, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.828097820281982, "learning_rate": 5e-05, "loss": 0.781, "mean_token_accuracy": 0.8739197254180908, "num_tokens": 394864421.0, "step": 10395 }, { "epoch": 1.3224780562269431, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.928892612457275, "learning_rate": 5e-05, "loss": 0.7622, "mean_token_accuracy": 0.8763601183891296, "num_tokens": 394901008.0, "step": 10396 }, { "epoch": 1.3226052665055337, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.334667205810547, "learning_rate": 5e-05, "loss": 0.8351, "mean_token_accuracy": 0.8440514802932739, "num_tokens": 394947172.0, "step": 10397 }, { "epoch": 1.3227324767841242, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.057588577270508, "learning_rate": 5e-05, "loss": 0.805, "mean_token_accuracy": 0.8625690937042236, "num_tokens": 394984515.0, "step": 10398 }, { "epoch": 1.3228596870627147, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.261196136474609, "learning_rate": 5e-05, "loss": 0.762, "mean_token_accuracy": 0.8609458208084106, "num_tokens": 395019819.0, "step": 10399 }, { "epoch": 1.3229868973413053, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.9481892585754395, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8704168796539307, "num_tokens": 395058606.0, "step": 10400 }, { "epoch": 1.3231141076198956, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.349713325500488, "learning_rate": 5e-05, "loss": 0.6892, "mean_token_accuracy": 0.8809766173362732, "num_tokens": 395094354.0, "step": 10401 }, { "epoch": 1.323241317898486, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.73538064956665, "learning_rate": 5e-05, "loss": 0.7574, "mean_token_accuracy": 0.8705798387527466, "num_tokens": 395136034.0, "step": 10402 }, { "epoch": 1.3233685281770766, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.513920307159424, "learning_rate": 5e-05, "loss": 0.683, "mean_token_accuracy": 0.8873562812805176, "num_tokens": 395176335.0, "step": 10403 }, { "epoch": 1.3234957384556671, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.475649833679199, "learning_rate": 5e-05, "loss": 0.6912, "mean_token_accuracy": 0.8816838264465332, "num_tokens": 395216297.0, "step": 10404 }, { "epoch": 1.3236229487342577, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.731163501739502, "learning_rate": 5e-05, "loss": 0.7422, "mean_token_accuracy": 0.8728240728378296, "num_tokens": 395255685.0, "step": 10405 }, { "epoch": 1.3237501590128482, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.249396324157715, "learning_rate": 5e-05, "loss": 0.7314, "mean_token_accuracy": 0.8616676926612854, "num_tokens": 395300484.0, "step": 10406 }, { "epoch": 1.3238773692914387, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.875153064727783, "learning_rate": 5e-05, "loss": 0.7764, "mean_token_accuracy": 0.8661391735076904, "num_tokens": 395333760.0, "step": 10407 }, { "epoch": 1.3240045795700293, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.139318466186523, "learning_rate": 5e-05, "loss": 0.7208, "mean_token_accuracy": 0.8701900839805603, "num_tokens": 395376264.0, "step": 10408 }, { "epoch": 1.3241317898486198, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.928699016571045, "learning_rate": 5e-05, "loss": 0.7928, "mean_token_accuracy": 0.8591775298118591, "num_tokens": 395417758.0, "step": 10409 }, { "epoch": 1.3242590001272103, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.092164993286133, "learning_rate": 5e-05, "loss": 0.7165, "mean_token_accuracy": 0.8667620420455933, "num_tokens": 395458657.0, "step": 10410 }, { "epoch": 1.3243862104058008, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.964319229125977, "learning_rate": 5e-05, "loss": 0.7953, "mean_token_accuracy": 0.8581279516220093, "num_tokens": 395500838.0, "step": 10411 }, { "epoch": 1.3245134206843914, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.992039203643799, "learning_rate": 5e-05, "loss": 0.7081, "mean_token_accuracy": 0.8695278167724609, "num_tokens": 395544912.0, "step": 10412 }, { "epoch": 1.324640630962982, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.140251159667969, "learning_rate": 5e-05, "loss": 0.7744, "mean_token_accuracy": 0.87328040599823, "num_tokens": 395584661.0, "step": 10413 }, { "epoch": 1.3247678412415724, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.731194496154785, "learning_rate": 5e-05, "loss": 0.6652, "mean_token_accuracy": 0.870741605758667, "num_tokens": 395623374.0, "step": 10414 }, { "epoch": 1.324895051520163, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.566322326660156, "learning_rate": 5e-05, "loss": 0.7722, "mean_token_accuracy": 0.8822828531265259, "num_tokens": 395658854.0, "step": 10415 }, { "epoch": 1.3250222617987533, "ewc_loss": 0.283203125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0002307891845703125, "grad_norm": 6.250255584716797, "learning_rate": 5e-05, "loss": 0.6377, "mean_token_accuracy": 0.8780007362365723, "num_tokens": 395695298.0, "step": 10416 }, { "epoch": 1.3251494720773438, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 9.33032512664795, "learning_rate": 5e-05, "loss": 0.957, "mean_token_accuracy": 0.8505880832672119, "num_tokens": 395731800.0, "step": 10417 }, { "epoch": 1.3252766823559343, "ewc_loss": 0.28125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00022983551025390625, "grad_norm": 5.7770233154296875, "learning_rate": 5e-05, "loss": 0.668, "mean_token_accuracy": 0.8636149168014526, "num_tokens": 395773811.0, "step": 10418 }, { "epoch": 1.3254038926345248, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.262528419494629, "learning_rate": 5e-05, "loss": 1.0097, "mean_token_accuracy": 0.8743813037872314, "num_tokens": 395817280.0, "step": 10419 }, { "epoch": 1.3255311029131154, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.3864617347717285, "learning_rate": 5e-05, "loss": 0.7766, "mean_token_accuracy": 0.8657999038696289, "num_tokens": 395857950.0, "step": 10420 }, { "epoch": 1.325658313191706, "ewc_loss": 0.703125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.372842788696289, "learning_rate": 5e-05, "loss": 1.0935, "mean_token_accuracy": 0.8615283966064453, "num_tokens": 395893048.0, "step": 10421 }, { "epoch": 1.3257855234702964, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.300180435180664, "learning_rate": 5e-05, "loss": 0.9954, "mean_token_accuracy": 0.8766820430755615, "num_tokens": 395932678.0, "step": 10422 }, { "epoch": 1.325912733748887, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.590702056884766, "learning_rate": 5e-05, "loss": 0.8882, "mean_token_accuracy": 0.8753889203071594, "num_tokens": 395967253.0, "step": 10423 }, { "epoch": 1.3260399440274775, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.502368927001953, "learning_rate": 5e-05, "loss": 0.9957, "mean_token_accuracy": 0.8878254890441895, "num_tokens": 396000935.0, "step": 10424 }, { "epoch": 1.3261671543060678, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.673561096191406, "learning_rate": 5e-05, "loss": 0.7818, "mean_token_accuracy": 0.8544734120368958, "num_tokens": 396039148.0, "step": 10425 }, { "epoch": 1.3262943645846583, "ewc_loss": 0.78125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 10.998305320739746, "learning_rate": 5e-05, "loss": 1.1077, "mean_token_accuracy": 0.8841032981872559, "num_tokens": 396072035.0, "step": 10426 }, { "epoch": 1.3264215748632489, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.425875186920166, "learning_rate": 5e-05, "loss": 0.9371, "mean_token_accuracy": 0.8671296834945679, "num_tokens": 396112047.0, "step": 10427 }, { "epoch": 1.3265487851418394, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000732421875, "grad_norm": 11.235669136047363, "learning_rate": 5e-05, "loss": 1.158, "mean_token_accuracy": 0.8705118894577026, "num_tokens": 396149844.0, "step": 10428 }, { "epoch": 1.32667599542043, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 9.171637535095215, "learning_rate": 5e-05, "loss": 1.0928, "mean_token_accuracy": 0.892441987991333, "num_tokens": 396187493.0, "step": 10429 }, { "epoch": 1.3268032056990204, "ewc_loss": 0.625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.901529312133789, "learning_rate": 5e-05, "loss": 0.9966, "mean_token_accuracy": 0.872620701789856, "num_tokens": 396220590.0, "step": 10430 }, { "epoch": 1.326930415977611, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 9.12990951538086, "learning_rate": 5e-05, "loss": 1.0774, "mean_token_accuracy": 0.8674877285957336, "num_tokens": 396262590.0, "step": 10431 }, { "epoch": 1.3270576262562015, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.904629707336426, "learning_rate": 5e-05, "loss": 0.9468, "mean_token_accuracy": 0.8735363483428955, "num_tokens": 396296491.0, "step": 10432 }, { "epoch": 1.327184836534792, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.66480541229248, "learning_rate": 5e-05, "loss": 0.9411, "mean_token_accuracy": 0.8733266592025757, "num_tokens": 396334798.0, "step": 10433 }, { "epoch": 1.3273120468133826, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.534255027770996, "learning_rate": 5e-05, "loss": 0.9189, "mean_token_accuracy": 0.869774341583252, "num_tokens": 396368278.0, "step": 10434 }, { "epoch": 1.327439257091973, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.856092929840088, "learning_rate": 5e-05, "loss": 0.8149, "mean_token_accuracy": 0.874370813369751, "num_tokens": 396405034.0, "step": 10435 }, { "epoch": 1.3275664673705636, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.736688613891602, "learning_rate": 5e-05, "loss": 0.966, "mean_token_accuracy": 0.8565306663513184, "num_tokens": 396445503.0, "step": 10436 }, { "epoch": 1.3276936776491541, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.109128475189209, "learning_rate": 5e-05, "loss": 0.7391, "mean_token_accuracy": 0.8769102096557617, "num_tokens": 396482462.0, "step": 10437 }, { "epoch": 1.3278208879277447, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.270001411437988, "learning_rate": 5e-05, "loss": 0.8838, "mean_token_accuracy": 0.8895263671875, "num_tokens": 396519186.0, "step": 10438 }, { "epoch": 1.3279480982063352, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.812808513641357, "learning_rate": 5e-05, "loss": 0.7538, "mean_token_accuracy": 0.8699842691421509, "num_tokens": 396558996.0, "step": 10439 }, { "epoch": 1.3280753084849257, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.721517562866211, "learning_rate": 5e-05, "loss": 1.0545, "mean_token_accuracy": 0.8548271059989929, "num_tokens": 396593158.0, "step": 10440 }, { "epoch": 1.328202518763516, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.184184551239014, "learning_rate": 5e-05, "loss": 0.8567, "mean_token_accuracy": 0.8620707988739014, "num_tokens": 396629189.0, "step": 10441 }, { "epoch": 1.3283297290421066, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.713947296142578, "learning_rate": 5e-05, "loss": 1.0686, "mean_token_accuracy": 0.8543127775192261, "num_tokens": 396665932.0, "step": 10442 }, { "epoch": 1.328456939320697, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 7.801122665405273, "learning_rate": 5e-05, "loss": 0.9809, "mean_token_accuracy": 0.8582810163497925, "num_tokens": 396706638.0, "step": 10443 }, { "epoch": 1.3285841495992876, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.140336990356445, "learning_rate": 5e-05, "loss": 0.951, "mean_token_accuracy": 0.8750759959220886, "num_tokens": 396744354.0, "step": 10444 }, { "epoch": 1.3287113598778781, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.064257621765137, "learning_rate": 5e-05, "loss": 0.9324, "mean_token_accuracy": 0.8742972016334534, "num_tokens": 396783806.0, "step": 10445 }, { "epoch": 1.3288385701564687, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.443890571594238, "learning_rate": 5e-05, "loss": 0.8946, "mean_token_accuracy": 0.8723121881484985, "num_tokens": 396822337.0, "step": 10446 }, { "epoch": 1.3289657804350592, "ewc_loss": 0.5, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.898282051086426, "learning_rate": 5e-05, "loss": 0.9261, "mean_token_accuracy": 0.8557191491127014, "num_tokens": 396861674.0, "step": 10447 }, { "epoch": 1.3290929907136497, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.253790855407715, "learning_rate": 5e-05, "loss": 0.8939, "mean_token_accuracy": 0.8673235177993774, "num_tokens": 396897686.0, "step": 10448 }, { "epoch": 1.3292202009922403, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.447500228881836, "learning_rate": 5e-05, "loss": 0.7975, "mean_token_accuracy": 0.8648576140403748, "num_tokens": 396936261.0, "step": 10449 }, { "epoch": 1.3293474112708306, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.47540283203125, "learning_rate": 5e-05, "loss": 0.896, "mean_token_accuracy": 0.8675435781478882, "num_tokens": 396967452.0, "step": 10450 }, { "epoch": 1.329474621549421, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.073386192321777, "learning_rate": 5e-05, "loss": 0.7066, "mean_token_accuracy": 0.8862665891647339, "num_tokens": 397004672.0, "step": 10451 }, { "epoch": 1.3296018318280116, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.830767631530762, "learning_rate": 5e-05, "loss": 0.932, "mean_token_accuracy": 0.8689208626747131, "num_tokens": 397039310.0, "step": 10452 }, { "epoch": 1.3297290421066021, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.916540145874023, "learning_rate": 5e-05, "loss": 0.7286, "mean_token_accuracy": 0.8793978095054626, "num_tokens": 397077465.0, "step": 10453 }, { "epoch": 1.3298562523851927, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.15503215789795, "learning_rate": 5e-05, "loss": 0.9463, "mean_token_accuracy": 0.8781212568283081, "num_tokens": 397108606.0, "step": 10454 }, { "epoch": 1.3299834626637832, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.158548355102539, "learning_rate": 5e-05, "loss": 0.8178, "mean_token_accuracy": 0.8675597906112671, "num_tokens": 397146628.0, "step": 10455 }, { "epoch": 1.3301106729423737, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.001935005187988, "learning_rate": 5e-05, "loss": 0.9671, "mean_token_accuracy": 0.8726575970649719, "num_tokens": 397182345.0, "step": 10456 }, { "epoch": 1.3302378832209643, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.75615119934082, "learning_rate": 5e-05, "loss": 0.8867, "mean_token_accuracy": 0.8721460103988647, "num_tokens": 397218673.0, "step": 10457 }, { "epoch": 1.3303650934995548, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.1798095703125, "learning_rate": 5e-05, "loss": 0.8616, "mean_token_accuracy": 0.8810579776763916, "num_tokens": 397250186.0, "step": 10458 }, { "epoch": 1.3304923037781453, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.137411117553711, "learning_rate": 5e-05, "loss": 0.8906, "mean_token_accuracy": 0.8694822788238525, "num_tokens": 397287947.0, "step": 10459 }, { "epoch": 1.3306195140567358, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.46466588973999, "learning_rate": 5e-05, "loss": 0.7865, "mean_token_accuracy": 0.8685985207557678, "num_tokens": 397328523.0, "step": 10460 }, { "epoch": 1.3307467243353264, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.157679557800293, "learning_rate": 5e-05, "loss": 0.8323, "mean_token_accuracy": 0.8737170696258545, "num_tokens": 397366238.0, "step": 10461 }, { "epoch": 1.330873934613917, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.260504722595215, "learning_rate": 5e-05, "loss": 0.7057, "mean_token_accuracy": 0.8846333622932434, "num_tokens": 397404324.0, "step": 10462 }, { "epoch": 1.3310011448925074, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.089475631713867, "learning_rate": 5e-05, "loss": 0.8078, "mean_token_accuracy": 0.8717736005783081, "num_tokens": 397441074.0, "step": 10463 }, { "epoch": 1.331128355171098, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.236878871917725, "learning_rate": 5e-05, "loss": 0.7537, "mean_token_accuracy": 0.8608956933021545, "num_tokens": 397476913.0, "step": 10464 }, { "epoch": 1.3312555654496883, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.065450668334961, "learning_rate": 5e-05, "loss": 0.7643, "mean_token_accuracy": 0.8769710063934326, "num_tokens": 397507569.0, "step": 10465 }, { "epoch": 1.3313827757282788, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.163204193115234, "learning_rate": 5e-05, "loss": 0.7075, "mean_token_accuracy": 0.8766330480575562, "num_tokens": 397544131.0, "step": 10466 }, { "epoch": 1.3315099860068693, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.05705738067627, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8863648176193237, "num_tokens": 397585042.0, "step": 10467 }, { "epoch": 1.3316371962854598, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.153440952301025, "learning_rate": 5e-05, "loss": 0.7382, "mean_token_accuracy": 0.8656001091003418, "num_tokens": 397620169.0, "step": 10468 }, { "epoch": 1.3317644065640504, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.060271263122559, "learning_rate": 5e-05, "loss": 0.8049, "mean_token_accuracy": 0.8667581081390381, "num_tokens": 397657349.0, "step": 10469 }, { "epoch": 1.331891616842641, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.100433826446533, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8572853207588196, "num_tokens": 397696309.0, "step": 10470 }, { "epoch": 1.3320188271212314, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.136938095092773, "learning_rate": 5e-05, "loss": 0.7893, "mean_token_accuracy": 0.8763359785079956, "num_tokens": 397734520.0, "step": 10471 }, { "epoch": 1.332146037399822, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.954237937927246, "learning_rate": 5e-05, "loss": 0.7323, "mean_token_accuracy": 0.8629587888717651, "num_tokens": 397773078.0, "step": 10472 }, { "epoch": 1.3322732476784125, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.355010032653809, "learning_rate": 5e-05, "loss": 0.7883, "mean_token_accuracy": 0.8827422857284546, "num_tokens": 397814180.0, "step": 10473 }, { "epoch": 1.3324004579570028, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.737034320831299, "learning_rate": 5e-05, "loss": 0.6744, "mean_token_accuracy": 0.8814215660095215, "num_tokens": 397850618.0, "step": 10474 }, { "epoch": 1.3325276682355933, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.680419921875, "learning_rate": 5e-05, "loss": 0.9112, "mean_token_accuracy": 0.8545966148376465, "num_tokens": 397888538.0, "step": 10475 }, { "epoch": 1.3326548785141838, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.585930347442627, "learning_rate": 5e-05, "loss": 0.6832, "mean_token_accuracy": 0.8801605701446533, "num_tokens": 397917861.0, "step": 10476 }, { "epoch": 1.3327820887927744, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.924962043762207, "learning_rate": 5e-05, "loss": 0.8449, "mean_token_accuracy": 0.8881023526191711, "num_tokens": 397954052.0, "step": 10477 }, { "epoch": 1.332909299071365, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.714341640472412, "learning_rate": 5e-05, "loss": 0.7205, "mean_token_accuracy": 0.872840166091919, "num_tokens": 397996458.0, "step": 10478 }, { "epoch": 1.3330365093499554, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.782791137695312, "learning_rate": 5e-05, "loss": 0.8749, "mean_token_accuracy": 0.8826892971992493, "num_tokens": 398034164.0, "step": 10479 }, { "epoch": 1.333163719628546, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.2395524978637695, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.8801738023757935, "num_tokens": 398075828.0, "step": 10480 }, { "epoch": 1.3332909299071365, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.107617378234863, "learning_rate": 5e-05, "loss": 0.8545, "mean_token_accuracy": 0.8713955283164978, "num_tokens": 398112941.0, "step": 10481 }, { "epoch": 1.333418140185727, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.8281450271606445, "learning_rate": 5e-05, "loss": 0.8166, "mean_token_accuracy": 0.8828585743904114, "num_tokens": 398146255.0, "step": 10482 }, { "epoch": 1.3335453504643175, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.324268817901611, "learning_rate": 5e-05, "loss": 0.7312, "mean_token_accuracy": 0.8819640874862671, "num_tokens": 398178552.0, "step": 10483 }, { "epoch": 1.333672560742908, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.224431037902832, "learning_rate": 5e-05, "loss": 0.8101, "mean_token_accuracy": 0.8825148344039917, "num_tokens": 398216310.0, "step": 10484 }, { "epoch": 1.3337997710214986, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.962843894958496, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.8716237545013428, "num_tokens": 398249756.0, "step": 10485 }, { "epoch": 1.3339269813000891, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.310285568237305, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.8705819845199585, "num_tokens": 398285264.0, "step": 10486 }, { "epoch": 1.3340541915786797, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.9588623046875, "learning_rate": 5e-05, "loss": 0.7757, "mean_token_accuracy": 0.8546767830848694, "num_tokens": 398320964.0, "step": 10487 }, { "epoch": 1.3341814018572702, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.206295013427734, "learning_rate": 5e-05, "loss": 0.8763, "mean_token_accuracy": 0.8541377782821655, "num_tokens": 398355392.0, "step": 10488 }, { "epoch": 1.3343086121358605, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.089007377624512, "learning_rate": 5e-05, "loss": 0.7323, "mean_token_accuracy": 0.8694183826446533, "num_tokens": 398393551.0, "step": 10489 }, { "epoch": 1.334435822414451, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.007354736328125, "learning_rate": 5e-05, "loss": 0.8416, "mean_token_accuracy": 0.8553286790847778, "num_tokens": 398433812.0, "step": 10490 }, { "epoch": 1.3345630326930416, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.243057727813721, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.8704159259796143, "num_tokens": 398469790.0, "step": 10491 }, { "epoch": 1.334690242971632, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.835982322692871, "learning_rate": 5e-05, "loss": 0.7624, "mean_token_accuracy": 0.8778680562973022, "num_tokens": 398512253.0, "step": 10492 }, { "epoch": 1.3348174532502226, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.363725185394287, "learning_rate": 5e-05, "loss": 0.7942, "mean_token_accuracy": 0.8529618382453918, "num_tokens": 398549790.0, "step": 10493 }, { "epoch": 1.3349446635288131, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.68735933303833, "learning_rate": 5e-05, "loss": 0.7035, "mean_token_accuracy": 0.8865507245063782, "num_tokens": 398586844.0, "step": 10494 }, { "epoch": 1.3350718738074037, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.423816680908203, "learning_rate": 5e-05, "loss": 0.7503, "mean_token_accuracy": 0.8662657737731934, "num_tokens": 398624177.0, "step": 10495 }, { "epoch": 1.3351990840859942, "ewc_loss": 0.375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.615103244781494, "learning_rate": 5e-05, "loss": 0.7086, "mean_token_accuracy": 0.8821260929107666, "num_tokens": 398661840.0, "step": 10496 }, { "epoch": 1.3353262943645847, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.412285327911377, "learning_rate": 5e-05, "loss": 0.6997, "mean_token_accuracy": 0.8820422887802124, "num_tokens": 398699535.0, "step": 10497 }, { "epoch": 1.3354535046431752, "ewc_loss": 0.375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.654837608337402, "learning_rate": 5e-05, "loss": 0.7736, "mean_token_accuracy": 0.8588102459907532, "num_tokens": 398736294.0, "step": 10498 }, { "epoch": 1.3355807149217656, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.290724277496338, "learning_rate": 5e-05, "loss": 0.693, "mean_token_accuracy": 0.8769279718399048, "num_tokens": 398776372.0, "step": 10499 }, { "epoch": 1.335707925200356, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.831431865692139, "learning_rate": 5e-05, "loss": 0.7716, "mean_token_accuracy": 0.8671221137046814, "num_tokens": 398816000.0, "step": 10500 }, { "epoch": 1.3358351354789466, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.027548789978027, "learning_rate": 5e-05, "loss": 0.6784, "mean_token_accuracy": 0.879946231842041, "num_tokens": 398851960.0, "step": 10501 }, { "epoch": 1.3359623457575371, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.163695335388184, "learning_rate": 5e-05, "loss": 0.7961, "mean_token_accuracy": 0.870669960975647, "num_tokens": 398886485.0, "step": 10502 }, { "epoch": 1.3360895560361277, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.605898857116699, "learning_rate": 5e-05, "loss": 0.7041, "mean_token_accuracy": 0.8634544610977173, "num_tokens": 398923800.0, "step": 10503 }, { "epoch": 1.3362167663147182, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.7373046875, "learning_rate": 5e-05, "loss": 0.8204, "mean_token_accuracy": 0.8758476376533508, "num_tokens": 398963578.0, "step": 10504 }, { "epoch": 1.3363439765933087, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00023651123046875, "grad_norm": 6.070059776306152, "learning_rate": 5e-05, "loss": 0.653, "mean_token_accuracy": 0.8737384080886841, "num_tokens": 399009791.0, "step": 10505 }, { "epoch": 1.3364711868718993, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.632646560668945, "learning_rate": 5e-05, "loss": 0.9105, "mean_token_accuracy": 0.8791749477386475, "num_tokens": 399048701.0, "step": 10506 }, { "epoch": 1.3365983971504898, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.724017143249512, "learning_rate": 5e-05, "loss": 0.6444, "mean_token_accuracy": 0.8820462226867676, "num_tokens": 399083441.0, "step": 10507 }, { "epoch": 1.3367256074290803, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.491381645202637, "learning_rate": 5e-05, "loss": 1.0775, "mean_token_accuracy": 0.8763197660446167, "num_tokens": 399113762.0, "step": 10508 }, { "epoch": 1.3368528177076708, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 6.8923773765563965, "learning_rate": 5e-05, "loss": 0.8478, "mean_token_accuracy": 0.8779112100601196, "num_tokens": 399149966.0, "step": 10509 }, { "epoch": 1.3369800279862614, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.057024955749512, "learning_rate": 5e-05, "loss": 1.0645, "mean_token_accuracy": 0.8702600002288818, "num_tokens": 399188777.0, "step": 10510 }, { "epoch": 1.337107238264852, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 8.742650032043457, "learning_rate": 5e-05, "loss": 1.0812, "mean_token_accuracy": 0.8799893856048584, "num_tokens": 399229618.0, "step": 10511 }, { "epoch": 1.3372344485434424, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.761325359344482, "learning_rate": 5e-05, "loss": 0.7945, "mean_token_accuracy": 0.878526508808136, "num_tokens": 399270558.0, "step": 10512 }, { "epoch": 1.337361658822033, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 10.386677742004395, "learning_rate": 5e-05, "loss": 1.2728, "mean_token_accuracy": 0.8412797451019287, "num_tokens": 399310660.0, "step": 10513 }, { "epoch": 1.3374888691006233, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.536355018615723, "learning_rate": 5e-05, "loss": 0.7935, "mean_token_accuracy": 0.8695337772369385, "num_tokens": 399347838.0, "step": 10514 }, { "epoch": 1.3376160793792138, "ewc_loss": 1.0, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000949859619140625, "grad_norm": 12.398794174194336, "learning_rate": 5e-05, "loss": 1.3573, "mean_token_accuracy": 0.8780663013458252, "num_tokens": 399384790.0, "step": 10515 }, { "epoch": 1.3377432896578043, "ewc_loss": 0.9140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 9.461248397827148, "learning_rate": 5e-05, "loss": 1.2536, "mean_token_accuracy": 0.8804387450218201, "num_tokens": 399425479.0, "step": 10516 }, { "epoch": 1.3378704999363948, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 11.455486297607422, "learning_rate": 5e-05, "loss": 1.101, "mean_token_accuracy": 0.8745094537734985, "num_tokens": 399467161.0, "step": 10517 }, { "epoch": 1.3379977102149854, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 9.687566757202148, "learning_rate": 5e-05, "loss": 1.3093, "mean_token_accuracy": 0.8610764741897583, "num_tokens": 399500204.0, "step": 10518 }, { "epoch": 1.338124920493576, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.925015449523926, "learning_rate": 5e-05, "loss": 1.166, "mean_token_accuracy": 0.8573794364929199, "num_tokens": 399546142.0, "step": 10519 }, { "epoch": 1.3382521307721664, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.707558631896973, "learning_rate": 5e-05, "loss": 1.053, "mean_token_accuracy": 0.8566211462020874, "num_tokens": 399582446.0, "step": 10520 }, { "epoch": 1.338379341050757, "ewc_loss": 0.765625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 10.667153358459473, "learning_rate": 5e-05, "loss": 1.2276, "mean_token_accuracy": 0.8429427146911621, "num_tokens": 399616998.0, "step": 10521 }, { "epoch": 1.3385065513293475, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.203001976013184, "learning_rate": 5e-05, "loss": 0.9107, "mean_token_accuracy": 0.8771542310714722, "num_tokens": 399648649.0, "step": 10522 }, { "epoch": 1.3386337616079378, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000732421875, "grad_norm": 10.965890884399414, "learning_rate": 5e-05, "loss": 1.1124, "mean_token_accuracy": 0.8850721120834351, "num_tokens": 399690417.0, "step": 10523 }, { "epoch": 1.3387609718865283, "ewc_loss": 0.625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.569220542907715, "learning_rate": 5e-05, "loss": 1.0153, "mean_token_accuracy": 0.8649711012840271, "num_tokens": 399732346.0, "step": 10524 }, { "epoch": 1.3388881821651188, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 11.020611763000488, "learning_rate": 5e-05, "loss": 1.1727, "mean_token_accuracy": 0.8616129159927368, "num_tokens": 399770516.0, "step": 10525 }, { "epoch": 1.3390153924437094, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 9.186522483825684, "learning_rate": 5e-05, "loss": 1.0694, "mean_token_accuracy": 0.8696969747543335, "num_tokens": 399803266.0, "step": 10526 }, { "epoch": 1.3391426027223, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.376920700073242, "learning_rate": 5e-05, "loss": 1.0989, "mean_token_accuracy": 0.8673264980316162, "num_tokens": 399841940.0, "step": 10527 }, { "epoch": 1.3392698130008904, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.102339744567871, "learning_rate": 5e-05, "loss": 1.0484, "mean_token_accuracy": 0.8615356683731079, "num_tokens": 399880207.0, "step": 10528 }, { "epoch": 1.339397023279481, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.744503021240234, "learning_rate": 5e-05, "loss": 1.0494, "mean_token_accuracy": 0.8645604848861694, "num_tokens": 399915731.0, "step": 10529 }, { "epoch": 1.3395242335580715, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.639631271362305, "learning_rate": 5e-05, "loss": 0.8656, "mean_token_accuracy": 0.8947409987449646, "num_tokens": 399953398.0, "step": 10530 }, { "epoch": 1.339651443836662, "ewc_loss": 0.625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.453544616699219, "learning_rate": 5e-05, "loss": 1.0451, "mean_token_accuracy": 0.8539077639579773, "num_tokens": 399991132.0, "step": 10531 }, { "epoch": 1.3397786541152525, "ewc_loss": 0.5, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.089186668395996, "learning_rate": 5e-05, "loss": 0.9029, "mean_token_accuracy": 0.8637456297874451, "num_tokens": 400031615.0, "step": 10532 }, { "epoch": 1.339905864393843, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.662299156188965, "learning_rate": 5e-05, "loss": 1.0379, "mean_token_accuracy": 0.8544015884399414, "num_tokens": 400070804.0, "step": 10533 }, { "epoch": 1.3400330746724336, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.812615394592285, "learning_rate": 5e-05, "loss": 0.8806, "mean_token_accuracy": 0.8652125000953674, "num_tokens": 400114021.0, "step": 10534 }, { "epoch": 1.3401602849510241, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.074193000793457, "learning_rate": 5e-05, "loss": 1.0555, "mean_token_accuracy": 0.8758031725883484, "num_tokens": 400156526.0, "step": 10535 }, { "epoch": 1.3402874952296147, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.207818031311035, "learning_rate": 5e-05, "loss": 0.8433, "mean_token_accuracy": 0.8918377757072449, "num_tokens": 400188487.0, "step": 10536 }, { "epoch": 1.3404147055082052, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.902047157287598, "learning_rate": 5e-05, "loss": 1.1157, "mean_token_accuracy": 0.8491650819778442, "num_tokens": 400222223.0, "step": 10537 }, { "epoch": 1.3405419157867955, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.677977561950684, "learning_rate": 5e-05, "loss": 0.8961, "mean_token_accuracy": 0.8847407102584839, "num_tokens": 400259181.0, "step": 10538 }, { "epoch": 1.340669126065386, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.099302291870117, "learning_rate": 5e-05, "loss": 0.9165, "mean_token_accuracy": 0.8832961320877075, "num_tokens": 400295202.0, "step": 10539 }, { "epoch": 1.3407963363439765, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.501834869384766, "learning_rate": 5e-05, "loss": 0.8649, "mean_token_accuracy": 0.8844804763793945, "num_tokens": 400330936.0, "step": 10540 }, { "epoch": 1.340923546622567, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.59267807006836, "learning_rate": 5e-05, "loss": 0.9177, "mean_token_accuracy": 0.8730584383010864, "num_tokens": 400367971.0, "step": 10541 }, { "epoch": 1.3410507569011576, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.912382125854492, "learning_rate": 5e-05, "loss": 0.8402, "mean_token_accuracy": 0.8676630258560181, "num_tokens": 400406868.0, "step": 10542 }, { "epoch": 1.3411779671797481, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.73135757446289, "learning_rate": 5e-05, "loss": 0.9615, "mean_token_accuracy": 0.8581722974777222, "num_tokens": 400442907.0, "step": 10543 }, { "epoch": 1.3413051774583387, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.376908302307129, "learning_rate": 5e-05, "loss": 0.8101, "mean_token_accuracy": 0.8603839874267578, "num_tokens": 400479485.0, "step": 10544 }, { "epoch": 1.3414323877369292, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.134821891784668, "learning_rate": 5e-05, "loss": 0.9519, "mean_token_accuracy": 0.8675751090049744, "num_tokens": 400513790.0, "step": 10545 }, { "epoch": 1.3415595980155197, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.284413814544678, "learning_rate": 5e-05, "loss": 0.7759, "mean_token_accuracy": 0.8713147640228271, "num_tokens": 400557638.0, "step": 10546 }, { "epoch": 1.3416868082941102, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.3829927444458, "learning_rate": 5e-05, "loss": 0.9587, "mean_token_accuracy": 0.8814966082572937, "num_tokens": 400594121.0, "step": 10547 }, { "epoch": 1.3418140185727006, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.7218708992004395, "learning_rate": 5e-05, "loss": 0.8314, "mean_token_accuracy": 0.8712515830993652, "num_tokens": 400632427.0, "step": 10548 }, { "epoch": 1.341941228851291, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.018498420715332, "learning_rate": 5e-05, "loss": 0.9495, "mean_token_accuracy": 0.8744754791259766, "num_tokens": 400666938.0, "step": 10549 }, { "epoch": 1.3420684391298816, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.132810592651367, "learning_rate": 5e-05, "loss": 0.8295, "mean_token_accuracy": 0.8840898275375366, "num_tokens": 400703510.0, "step": 10550 }, { "epoch": 1.3421956494084721, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.260696411132812, "learning_rate": 5e-05, "loss": 0.8539, "mean_token_accuracy": 0.8762945532798767, "num_tokens": 400737082.0, "step": 10551 }, { "epoch": 1.3423228596870627, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.082437515258789, "learning_rate": 5e-05, "loss": 0.8421, "mean_token_accuracy": 0.8718183636665344, "num_tokens": 400787254.0, "step": 10552 }, { "epoch": 1.3424500699656532, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.832990646362305, "learning_rate": 5e-05, "loss": 0.8232, "mean_token_accuracy": 0.8725814819335938, "num_tokens": 400824384.0, "step": 10553 }, { "epoch": 1.3425772802442437, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.823919773101807, "learning_rate": 5e-05, "loss": 0.7876, "mean_token_accuracy": 0.8649573922157288, "num_tokens": 400859536.0, "step": 10554 }, { "epoch": 1.3427044905228342, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.826280117034912, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8824617266654968, "num_tokens": 400896022.0, "step": 10555 }, { "epoch": 1.3428317008014248, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.511439800262451, "learning_rate": 5e-05, "loss": 0.7188, "mean_token_accuracy": 0.878539502620697, "num_tokens": 400936291.0, "step": 10556 }, { "epoch": 1.3429589110800153, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00052642822265625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.955215930938721, "learning_rate": 5e-05, "loss": 0.7526, "mean_token_accuracy": 0.8811395168304443, "num_tokens": 400980749.0, "step": 10557 }, { "epoch": 1.3430861213586058, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.360772132873535, "learning_rate": 5e-05, "loss": 0.786, "mean_token_accuracy": 0.853645920753479, "num_tokens": 401016591.0, "step": 10558 }, { "epoch": 1.3432133316371964, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.944939136505127, "learning_rate": 5e-05, "loss": 0.7986, "mean_token_accuracy": 0.8698738813400269, "num_tokens": 401062520.0, "step": 10559 }, { "epoch": 1.343340541915787, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.272864818572998, "learning_rate": 5e-05, "loss": 0.7153, "mean_token_accuracy": 0.8752750158309937, "num_tokens": 401105968.0, "step": 10560 }, { "epoch": 1.3434677521943774, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.034855842590332, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.8833909034729004, "num_tokens": 401145290.0, "step": 10561 }, { "epoch": 1.343594962472968, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.095251560211182, "learning_rate": 5e-05, "loss": 0.6939, "mean_token_accuracy": 0.8802753686904907, "num_tokens": 401181253.0, "step": 10562 }, { "epoch": 1.3437221727515583, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.295047760009766, "learning_rate": 5e-05, "loss": 0.7918, "mean_token_accuracy": 0.8738349676132202, "num_tokens": 401218191.0, "step": 10563 }, { "epoch": 1.3438493830301488, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.889316082000732, "learning_rate": 5e-05, "loss": 0.6876, "mean_token_accuracy": 0.8721024990081787, "num_tokens": 401252049.0, "step": 10564 }, { "epoch": 1.3439765933087393, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.485432624816895, "learning_rate": 5e-05, "loss": 0.8927, "mean_token_accuracy": 0.8488872051239014, "num_tokens": 401294819.0, "step": 10565 }, { "epoch": 1.3441038035873298, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.898462295532227, "learning_rate": 5e-05, "loss": 0.6954, "mean_token_accuracy": 0.8750767707824707, "num_tokens": 401332007.0, "step": 10566 }, { "epoch": 1.3442310138659204, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.421555519104004, "learning_rate": 5e-05, "loss": 0.8027, "mean_token_accuracy": 0.8805220127105713, "num_tokens": 401371442.0, "step": 10567 }, { "epoch": 1.344358224144511, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.107664585113525, "learning_rate": 5e-05, "loss": 0.6974, "mean_token_accuracy": 0.8787650465965271, "num_tokens": 401409956.0, "step": 10568 }, { "epoch": 1.3444854344231014, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.141623497009277, "learning_rate": 5e-05, "loss": 0.7855, "mean_token_accuracy": 0.8801404237747192, "num_tokens": 401446445.0, "step": 10569 }, { "epoch": 1.344612644701692, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.421512126922607, "learning_rate": 5e-05, "loss": 0.7728, "mean_token_accuracy": 0.8628807067871094, "num_tokens": 401489366.0, "step": 10570 }, { "epoch": 1.3447398549802825, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.657445907592773, "learning_rate": 5e-05, "loss": 0.7749, "mean_token_accuracy": 0.8658084273338318, "num_tokens": 401533326.0, "step": 10571 }, { "epoch": 1.3448670652588728, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.866434097290039, "learning_rate": 5e-05, "loss": 0.7553, "mean_token_accuracy": 0.8763946294784546, "num_tokens": 401571607.0, "step": 10572 }, { "epoch": 1.3449942755374633, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.076935768127441, "learning_rate": 5e-05, "loss": 0.7692, "mean_token_accuracy": 0.8526567220687866, "num_tokens": 401607930.0, "step": 10573 }, { "epoch": 1.3451214858160538, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.3536376953125, "learning_rate": 5e-05, "loss": 0.7531, "mean_token_accuracy": 0.8860669136047363, "num_tokens": 401645841.0, "step": 10574 }, { "epoch": 1.3452486960946444, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.564945697784424, "learning_rate": 5e-05, "loss": 0.7043, "mean_token_accuracy": 0.8642011284828186, "num_tokens": 401680795.0, "step": 10575 }, { "epoch": 1.345375906373235, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.976898193359375, "learning_rate": 5e-05, "loss": 0.8605, "mean_token_accuracy": 0.8728495240211487, "num_tokens": 401719710.0, "step": 10576 }, { "epoch": 1.3455031166518254, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.164681434631348, "learning_rate": 5e-05, "loss": 0.6603, "mean_token_accuracy": 0.8762933611869812, "num_tokens": 401758956.0, "step": 10577 }, { "epoch": 1.345630326930416, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.572854995727539, "learning_rate": 5e-05, "loss": 0.9853, "mean_token_accuracy": 0.8616530895233154, "num_tokens": 401798108.0, "step": 10578 }, { "epoch": 1.3457575372090065, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.426699638366699, "learning_rate": 5e-05, "loss": 0.6853, "mean_token_accuracy": 0.8783156871795654, "num_tokens": 401833495.0, "step": 10579 }, { "epoch": 1.345884747487597, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.420869827270508, "learning_rate": 5e-05, "loss": 0.9985, "mean_token_accuracy": 0.8700476288795471, "num_tokens": 401875399.0, "step": 10580 }, { "epoch": 1.3460119577661875, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.643707752227783, "learning_rate": 5e-05, "loss": 0.8578, "mean_token_accuracy": 0.8731599450111389, "num_tokens": 401915540.0, "step": 10581 }, { "epoch": 1.346139168044778, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.04244327545166, "learning_rate": 5e-05, "loss": 0.8967, "mean_token_accuracy": 0.8563094139099121, "num_tokens": 401960013.0, "step": 10582 }, { "epoch": 1.3462663783233686, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.693533897399902, "learning_rate": 5e-05, "loss": 0.9527, "mean_token_accuracy": 0.8617882132530212, "num_tokens": 401990709.0, "step": 10583 }, { "epoch": 1.3463935886019591, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.7650275230407715, "learning_rate": 5e-05, "loss": 0.7033, "mean_token_accuracy": 0.8768273591995239, "num_tokens": 402031277.0, "step": 10584 }, { "epoch": 1.3465207988805497, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.219752311706543, "learning_rate": 5e-05, "loss": 0.933, "mean_token_accuracy": 0.8764728903770447, "num_tokens": 402073248.0, "step": 10585 }, { "epoch": 1.3466480091591402, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.822808265686035, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.866982102394104, "num_tokens": 402116072.0, "step": 10586 }, { "epoch": 1.3467752194377305, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.72314739227295, "learning_rate": 5e-05, "loss": 0.8985, "mean_token_accuracy": 0.8730044364929199, "num_tokens": 402153092.0, "step": 10587 }, { "epoch": 1.346902429716321, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.637455940246582, "learning_rate": 5e-05, "loss": 0.8131, "mean_token_accuracy": 0.8780828714370728, "num_tokens": 402192959.0, "step": 10588 }, { "epoch": 1.3470296399949115, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.6314263343811035, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8707997798919678, "num_tokens": 402232245.0, "step": 10589 }, { "epoch": 1.347156850273502, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.31553840637207, "learning_rate": 5e-05, "loss": 0.8128, "mean_token_accuracy": 0.8840163350105286, "num_tokens": 402265862.0, "step": 10590 }, { "epoch": 1.3472840605520926, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.005413055419922, "learning_rate": 5e-05, "loss": 0.7013, "mean_token_accuracy": 0.8784295916557312, "num_tokens": 402303747.0, "step": 10591 }, { "epoch": 1.3474112708306831, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.56810474395752, "learning_rate": 5e-05, "loss": 0.8528, "mean_token_accuracy": 0.8733307123184204, "num_tokens": 402342952.0, "step": 10592 }, { "epoch": 1.3475384811092737, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.98323917388916, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.849407434463501, "num_tokens": 402382388.0, "step": 10593 }, { "epoch": 1.3476656913878642, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.412002563476562, "learning_rate": 5e-05, "loss": 0.8303, "mean_token_accuracy": 0.8775327205657959, "num_tokens": 402418937.0, "step": 10594 }, { "epoch": 1.3477929016664547, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.228019714355469, "learning_rate": 5e-05, "loss": 0.7745, "mean_token_accuracy": 0.8621194362640381, "num_tokens": 402454347.0, "step": 10595 }, { "epoch": 1.3479201119450452, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.05436897277832, "learning_rate": 5e-05, "loss": 0.8757, "mean_token_accuracy": 0.8549748659133911, "num_tokens": 402490887.0, "step": 10596 }, { "epoch": 1.3480473222236355, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.408862590789795, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.8631502389907837, "num_tokens": 402531293.0, "step": 10597 }, { "epoch": 1.348174532502226, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.891600131988525, "learning_rate": 5e-05, "loss": 0.8412, "mean_token_accuracy": 0.858887791633606, "num_tokens": 402574396.0, "step": 10598 }, { "epoch": 1.3483017427808166, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.3211469650268555, "learning_rate": 5e-05, "loss": 0.7312, "mean_token_accuracy": 0.8714755773544312, "num_tokens": 402615609.0, "step": 10599 }, { "epoch": 1.3484289530594071, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.0384521484375, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.870485782623291, "num_tokens": 402650627.0, "step": 10600 }, { "epoch": 1.3485561633379977, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.984472274780273, "learning_rate": 5e-05, "loss": 0.7207, "mean_token_accuracy": 0.8678662180900574, "num_tokens": 402693036.0, "step": 10601 }, { "epoch": 1.3486833736165882, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.441210746765137, "learning_rate": 5e-05, "loss": 0.7962, "mean_token_accuracy": 0.880744457244873, "num_tokens": 402730748.0, "step": 10602 }, { "epoch": 1.3488105838951787, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.556722640991211, "learning_rate": 5e-05, "loss": 0.7198, "mean_token_accuracy": 0.8646194338798523, "num_tokens": 402762642.0, "step": 10603 }, { "epoch": 1.3489377941737692, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.98767375946045, "learning_rate": 5e-05, "loss": 0.8891, "mean_token_accuracy": 0.8653566241264343, "num_tokens": 402794265.0, "step": 10604 }, { "epoch": 1.3490650044523598, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.239459991455078, "learning_rate": 5e-05, "loss": 0.7601, "mean_token_accuracy": 0.8537664413452148, "num_tokens": 402830609.0, "step": 10605 }, { "epoch": 1.3491922147309503, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.491622924804688, "learning_rate": 5e-05, "loss": 0.9347, "mean_token_accuracy": 0.8777684569358826, "num_tokens": 402866608.0, "step": 10606 }, { "epoch": 1.3493194250095408, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.423626899719238, "learning_rate": 5e-05, "loss": 0.7491, "mean_token_accuracy": 0.8632621169090271, "num_tokens": 402904030.0, "step": 10607 }, { "epoch": 1.3494466352881314, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.520062446594238, "learning_rate": 5e-05, "loss": 1.0016, "mean_token_accuracy": 0.8724170923233032, "num_tokens": 402939552.0, "step": 10608 }, { "epoch": 1.3495738455667219, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.2940673828125, "learning_rate": 5e-05, "loss": 0.8412, "mean_token_accuracy": 0.8792442083358765, "num_tokens": 402981683.0, "step": 10609 }, { "epoch": 1.3497010558453124, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.72092056274414, "learning_rate": 5e-05, "loss": 1.0008, "mean_token_accuracy": 0.859438419342041, "num_tokens": 403019497.0, "step": 10610 }, { "epoch": 1.349828266123903, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.11351490020752, "learning_rate": 5e-05, "loss": 0.9259, "mean_token_accuracy": 0.8699918389320374, "num_tokens": 403053415.0, "step": 10611 }, { "epoch": 1.3499554764024932, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.606784820556641, "learning_rate": 5e-05, "loss": 0.8187, "mean_token_accuracy": 0.8716070055961609, "num_tokens": 403089676.0, "step": 10612 }, { "epoch": 1.3500826866810838, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.425426483154297, "learning_rate": 5e-05, "loss": 0.912, "mean_token_accuracy": 0.8671420812606812, "num_tokens": 403130429.0, "step": 10613 }, { "epoch": 1.3502098969596743, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.210238456726074, "learning_rate": 5e-05, "loss": 0.7338, "mean_token_accuracy": 0.8853778839111328, "num_tokens": 403171420.0, "step": 10614 }, { "epoch": 1.3503371072382648, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.256059646606445, "learning_rate": 5e-05, "loss": 0.8783, "mean_token_accuracy": 0.8683620691299438, "num_tokens": 403210700.0, "step": 10615 }, { "epoch": 1.3504643175168554, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.28619909286499, "learning_rate": 5e-05, "loss": 0.7728, "mean_token_accuracy": 0.8708105087280273, "num_tokens": 403249572.0, "step": 10616 }, { "epoch": 1.350591527795446, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.072635650634766, "learning_rate": 5e-05, "loss": 0.8373, "mean_token_accuracy": 0.866570234298706, "num_tokens": 403288176.0, "step": 10617 }, { "epoch": 1.3507187380740364, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.226887226104736, "learning_rate": 5e-05, "loss": 0.7166, "mean_token_accuracy": 0.8831497430801392, "num_tokens": 403329403.0, "step": 10618 }, { "epoch": 1.350845948352627, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.231581687927246, "learning_rate": 5e-05, "loss": 0.8506, "mean_token_accuracy": 0.8686102628707886, "num_tokens": 403362276.0, "step": 10619 }, { "epoch": 1.3509731586312175, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.942149639129639, "learning_rate": 5e-05, "loss": 0.7273, "mean_token_accuracy": 0.8726421594619751, "num_tokens": 403397797.0, "step": 10620 }, { "epoch": 1.3511003689098078, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.53746223449707, "learning_rate": 5e-05, "loss": 0.848, "mean_token_accuracy": 0.8715167045593262, "num_tokens": 403437943.0, "step": 10621 }, { "epoch": 1.3512275791883983, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.772444725036621, "learning_rate": 5e-05, "loss": 0.7161, "mean_token_accuracy": 0.878524661064148, "num_tokens": 403473575.0, "step": 10622 }, { "epoch": 1.3513547894669888, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.662407875061035, "learning_rate": 5e-05, "loss": 0.8876, "mean_token_accuracy": 0.8716992139816284, "num_tokens": 403504720.0, "step": 10623 }, { "epoch": 1.3514819997455794, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.849271297454834, "learning_rate": 5e-05, "loss": 0.7749, "mean_token_accuracy": 0.8613507747650146, "num_tokens": 403549244.0, "step": 10624 }, { "epoch": 1.35160921002417, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.585102081298828, "learning_rate": 5e-05, "loss": 0.9555, "mean_token_accuracy": 0.8552149534225464, "num_tokens": 403582881.0, "step": 10625 }, { "epoch": 1.3517364203027604, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.074469566345215, "learning_rate": 5e-05, "loss": 0.7477, "mean_token_accuracy": 0.8825520873069763, "num_tokens": 403623477.0, "step": 10626 }, { "epoch": 1.351863630581351, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.327346801757812, "learning_rate": 5e-05, "loss": 0.9162, "mean_token_accuracy": 0.8575661778450012, "num_tokens": 403662399.0, "step": 10627 }, { "epoch": 1.3519908408599415, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.301140308380127, "learning_rate": 5e-05, "loss": 0.7856, "mean_token_accuracy": 0.8724409341812134, "num_tokens": 403703908.0, "step": 10628 }, { "epoch": 1.352118051138532, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.047892570495605, "learning_rate": 5e-05, "loss": 0.8267, "mean_token_accuracy": 0.8750104904174805, "num_tokens": 403746385.0, "step": 10629 }, { "epoch": 1.3522452614171225, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.445906639099121, "learning_rate": 5e-05, "loss": 0.7798, "mean_token_accuracy": 0.8753682971000671, "num_tokens": 403785728.0, "step": 10630 }, { "epoch": 1.352372471695713, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.861219882965088, "learning_rate": 5e-05, "loss": 0.7945, "mean_token_accuracy": 0.8773794174194336, "num_tokens": 403827991.0, "step": 10631 }, { "epoch": 1.3524996819743036, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.435970306396484, "learning_rate": 5e-05, "loss": 0.7791, "mean_token_accuracy": 0.86656254529953, "num_tokens": 403861963.0, "step": 10632 }, { "epoch": 1.3526268922528941, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.860626697540283, "learning_rate": 5e-05, "loss": 0.742, "mean_token_accuracy": 0.885315477848053, "num_tokens": 403899744.0, "step": 10633 }, { "epoch": 1.3527541025314846, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.30076789855957, "learning_rate": 5e-05, "loss": 0.6966, "mean_token_accuracy": 0.888535737991333, "num_tokens": 403939482.0, "step": 10634 }, { "epoch": 1.3528813128100752, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.9683003425598145, "learning_rate": 5e-05, "loss": 0.7558, "mean_token_accuracy": 0.8843576908111572, "num_tokens": 403976233.0, "step": 10635 }, { "epoch": 1.3530085230886655, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.130568504333496, "learning_rate": 5e-05, "loss": 0.7053, "mean_token_accuracy": 0.882050633430481, "num_tokens": 404011181.0, "step": 10636 }, { "epoch": 1.353135733367256, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.091582298278809, "learning_rate": 5e-05, "loss": 0.7644, "mean_token_accuracy": 0.8828554749488831, "num_tokens": 404050519.0, "step": 10637 }, { "epoch": 1.3532629436458465, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.004306793212891, "learning_rate": 5e-05, "loss": 0.6926, "mean_token_accuracy": 0.879909873008728, "num_tokens": 404089407.0, "step": 10638 }, { "epoch": 1.353390153924437, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.2435884475708, "learning_rate": 5e-05, "loss": 0.7422, "mean_token_accuracy": 0.8930325508117676, "num_tokens": 404126564.0, "step": 10639 }, { "epoch": 1.3535173642030276, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.8713860511779785, "learning_rate": 5e-05, "loss": 0.7098, "mean_token_accuracy": 0.8708292245864868, "num_tokens": 404166113.0, "step": 10640 }, { "epoch": 1.3536445744816181, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.387839317321777, "learning_rate": 5e-05, "loss": 0.8323, "mean_token_accuracy": 0.8694514632225037, "num_tokens": 404204513.0, "step": 10641 }, { "epoch": 1.3537717847602087, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.828344345092773, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8539029359817505, "num_tokens": 404244475.0, "step": 10642 }, { "epoch": 1.3538989950387992, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.416889190673828, "learning_rate": 5e-05, "loss": 0.8571, "mean_token_accuracy": 0.8654729127883911, "num_tokens": 404278974.0, "step": 10643 }, { "epoch": 1.3540262053173897, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.949109077453613, "learning_rate": 5e-05, "loss": 0.7447, "mean_token_accuracy": 0.8644086122512817, "num_tokens": 404313123.0, "step": 10644 }, { "epoch": 1.3541534155959802, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.214856147766113, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.8762550354003906, "num_tokens": 404350195.0, "step": 10645 }, { "epoch": 1.3542806258745705, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.260220527648926, "learning_rate": 5e-05, "loss": 0.7546, "mean_token_accuracy": 0.8732903003692627, "num_tokens": 404387044.0, "step": 10646 }, { "epoch": 1.354407836153161, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.762465953826904, "learning_rate": 5e-05, "loss": 0.7817, "mean_token_accuracy": 0.8728549480438232, "num_tokens": 404423877.0, "step": 10647 }, { "epoch": 1.3545350464317516, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.660480976104736, "learning_rate": 5e-05, "loss": 0.7601, "mean_token_accuracy": 0.8770287036895752, "num_tokens": 404466428.0, "step": 10648 }, { "epoch": 1.3546622567103421, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.260697364807129, "learning_rate": 5e-05, "loss": 0.7336, "mean_token_accuracy": 0.8692873120307922, "num_tokens": 404503341.0, "step": 10649 }, { "epoch": 1.3547894669889327, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.060821533203125, "learning_rate": 5e-05, "loss": 0.799, "mean_token_accuracy": 0.8698164224624634, "num_tokens": 404543650.0, "step": 10650 }, { "epoch": 1.3549166772675232, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.82501745223999, "learning_rate": 5e-05, "loss": 0.7072, "mean_token_accuracy": 0.8673372864723206, "num_tokens": 404588834.0, "step": 10651 }, { "epoch": 1.3550438875461137, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000530242919921875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.498085975646973, "learning_rate": 5e-05, "loss": 0.8321, "mean_token_accuracy": 0.8671407699584961, "num_tokens": 404625903.0, "step": 10652 }, { "epoch": 1.3551710978247042, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.424314022064209, "learning_rate": 5e-05, "loss": 0.6721, "mean_token_accuracy": 0.875029444694519, "num_tokens": 404666299.0, "step": 10653 }, { "epoch": 1.3552983081032948, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 9.007197380065918, "learning_rate": 5e-05, "loss": 0.8227, "mean_token_accuracy": 0.8927076458930969, "num_tokens": 404702743.0, "step": 10654 }, { "epoch": 1.3554255183818853, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.207517623901367, "learning_rate": 5e-05, "loss": 0.7412, "mean_token_accuracy": 0.8535981178283691, "num_tokens": 404748010.0, "step": 10655 }, { "epoch": 1.3555527286604758, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.33836841583252, "learning_rate": 5e-05, "loss": 0.951, "mean_token_accuracy": 0.8722018599510193, "num_tokens": 404785886.0, "step": 10656 }, { "epoch": 1.3556799389390664, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.504892826080322, "learning_rate": 5e-05, "loss": 0.6728, "mean_token_accuracy": 0.8911601901054382, "num_tokens": 404820982.0, "step": 10657 }, { "epoch": 1.3558071492176569, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.104642868041992, "learning_rate": 5e-05, "loss": 0.9678, "mean_token_accuracy": 0.8692315816879272, "num_tokens": 404863943.0, "step": 10658 }, { "epoch": 1.3559343594962474, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.388798236846924, "learning_rate": 5e-05, "loss": 0.8854, "mean_token_accuracy": 0.8636146783828735, "num_tokens": 404905168.0, "step": 10659 }, { "epoch": 1.356061569774838, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.023677825927734, "learning_rate": 5e-05, "loss": 0.8763, "mean_token_accuracy": 0.8710838556289673, "num_tokens": 404941863.0, "step": 10660 }, { "epoch": 1.3561887800534282, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.291298866271973, "learning_rate": 5e-05, "loss": 0.9166, "mean_token_accuracy": 0.8685351610183716, "num_tokens": 404979780.0, "step": 10661 }, { "epoch": 1.3563159903320188, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.9156060218811035, "learning_rate": 5e-05, "loss": 0.7693, "mean_token_accuracy": 0.8632310628890991, "num_tokens": 405019194.0, "step": 10662 }, { "epoch": 1.3564432006106093, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.822425842285156, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.8894305229187012, "num_tokens": 405056021.0, "step": 10663 }, { "epoch": 1.3565704108891998, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.681815147399902, "learning_rate": 5e-05, "loss": 0.7721, "mean_token_accuracy": 0.8610327839851379, "num_tokens": 405099195.0, "step": 10664 }, { "epoch": 1.3566976211677904, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.698201179504395, "learning_rate": 5e-05, "loss": 0.9216, "mean_token_accuracy": 0.867962658405304, "num_tokens": 405142821.0, "step": 10665 }, { "epoch": 1.3568248314463809, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.116469383239746, "learning_rate": 5e-05, "loss": 0.8005, "mean_token_accuracy": 0.8678443431854248, "num_tokens": 405178434.0, "step": 10666 }, { "epoch": 1.3569520417249714, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.056191444396973, "learning_rate": 5e-05, "loss": 0.8874, "mean_token_accuracy": 0.8572331666946411, "num_tokens": 405211197.0, "step": 10667 }, { "epoch": 1.357079252003562, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.609145641326904, "learning_rate": 5e-05, "loss": 0.8009, "mean_token_accuracy": 0.8778495788574219, "num_tokens": 405250330.0, "step": 10668 }, { "epoch": 1.3572064622821525, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.553441047668457, "learning_rate": 5e-05, "loss": 0.7842, "mean_token_accuracy": 0.87351393699646, "num_tokens": 405291694.0, "step": 10669 }, { "epoch": 1.3573336725607428, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.787229061126709, "learning_rate": 5e-05, "loss": 0.7717, "mean_token_accuracy": 0.8809938430786133, "num_tokens": 405326162.0, "step": 10670 }, { "epoch": 1.3574608828393333, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.412397861480713, "learning_rate": 5e-05, "loss": 0.7335, "mean_token_accuracy": 0.8802757263183594, "num_tokens": 405361862.0, "step": 10671 }, { "epoch": 1.3575880931179238, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.695938587188721, "learning_rate": 5e-05, "loss": 0.7713, "mean_token_accuracy": 0.8729546070098877, "num_tokens": 405402934.0, "step": 10672 }, { "epoch": 1.3577153033965144, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.488989353179932, "learning_rate": 5e-05, "loss": 0.7758, "mean_token_accuracy": 0.8648232221603394, "num_tokens": 405442206.0, "step": 10673 }, { "epoch": 1.357842513675105, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.530643463134766, "learning_rate": 5e-05, "loss": 0.8371, "mean_token_accuracy": 0.8448101282119751, "num_tokens": 405483351.0, "step": 10674 }, { "epoch": 1.3579697239536954, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.607011318206787, "learning_rate": 5e-05, "loss": 0.7748, "mean_token_accuracy": 0.8662933111190796, "num_tokens": 405523763.0, "step": 10675 }, { "epoch": 1.358096934232286, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.3795905113220215, "learning_rate": 5e-05, "loss": 0.6905, "mean_token_accuracy": 0.8858978152275085, "num_tokens": 405559640.0, "step": 10676 }, { "epoch": 1.3582241445108765, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.685379981994629, "learning_rate": 5e-05, "loss": 0.7308, "mean_token_accuracy": 0.8763269186019897, "num_tokens": 405594907.0, "step": 10677 }, { "epoch": 1.358351354789467, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.275594711303711, "learning_rate": 5e-05, "loss": 0.6752, "mean_token_accuracy": 0.8842566013336182, "num_tokens": 405630346.0, "step": 10678 }, { "epoch": 1.3584785650680575, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.789958953857422, "learning_rate": 5e-05, "loss": 0.7584, "mean_token_accuracy": 0.8663457632064819, "num_tokens": 405661603.0, "step": 10679 }, { "epoch": 1.358605775346648, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.117411136627197, "learning_rate": 5e-05, "loss": 0.7202, "mean_token_accuracy": 0.8705501556396484, "num_tokens": 405698907.0, "step": 10680 }, { "epoch": 1.3587329856252386, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.9776835441589355, "learning_rate": 5e-05, "loss": 0.7988, "mean_token_accuracy": 0.8607654571533203, "num_tokens": 405744780.0, "step": 10681 }, { "epoch": 1.3588601959038291, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.85380744934082, "learning_rate": 5e-05, "loss": 0.6583, "mean_token_accuracy": 0.8806281089782715, "num_tokens": 405777883.0, "step": 10682 }, { "epoch": 1.3589874061824196, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.353425025939941, "learning_rate": 5e-05, "loss": 0.8128, "mean_token_accuracy": 0.8695569634437561, "num_tokens": 405814680.0, "step": 10683 }, { "epoch": 1.3591146164610102, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.425527095794678, "learning_rate": 5e-05, "loss": 0.7212, "mean_token_accuracy": 0.855862021446228, "num_tokens": 405847734.0, "step": 10684 }, { "epoch": 1.3592418267396005, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 9.029019355773926, "learning_rate": 5e-05, "loss": 0.8754, "mean_token_accuracy": 0.8693851828575134, "num_tokens": 405882365.0, "step": 10685 }, { "epoch": 1.359369037018191, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00023746490478515625, "grad_norm": 5.852089881896973, "learning_rate": 5e-05, "loss": 0.6534, "mean_token_accuracy": 0.8747750520706177, "num_tokens": 405920641.0, "step": 10686 }, { "epoch": 1.3594962472967815, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 10.054315567016602, "learning_rate": 5e-05, "loss": 1.0103, "mean_token_accuracy": 0.865443229675293, "num_tokens": 405958523.0, "step": 10687 }, { "epoch": 1.359623457575372, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 5.8303375244140625, "learning_rate": 5e-05, "loss": 0.7013, "mean_token_accuracy": 0.8763058185577393, "num_tokens": 405994338.0, "step": 10688 }, { "epoch": 1.3597506678539626, "ewc_loss": 0.8125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 10.883809089660645, "learning_rate": 5e-05, "loss": 1.2353, "mean_token_accuracy": 0.8567537069320679, "num_tokens": 406033285.0, "step": 10689 }, { "epoch": 1.3598778781325531, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 7.701498985290527, "learning_rate": 5e-05, "loss": 0.99, "mean_token_accuracy": 0.8785783052444458, "num_tokens": 406076933.0, "step": 10690 }, { "epoch": 1.3600050884111436, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.977173805236816, "learning_rate": 5e-05, "loss": 1.0551, "mean_token_accuracy": 0.8724140524864197, "num_tokens": 406118258.0, "step": 10691 }, { "epoch": 1.3601322986897342, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 9.208334922790527, "learning_rate": 5e-05, "loss": 1.204, "mean_token_accuracy": 0.8787009716033936, "num_tokens": 406155445.0, "step": 10692 }, { "epoch": 1.3602595089683247, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.705018520355225, "learning_rate": 5e-05, "loss": 0.8278, "mean_token_accuracy": 0.8674354553222656, "num_tokens": 406195128.0, "step": 10693 }, { "epoch": 1.3603867192469152, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 10.721435546875, "learning_rate": 5e-05, "loss": 1.2079, "mean_token_accuracy": 0.8784663677215576, "num_tokens": 406229311.0, "step": 10694 }, { "epoch": 1.3605139295255055, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.117315292358398, "learning_rate": 5e-05, "loss": 0.8972, "mean_token_accuracy": 0.8717108964920044, "num_tokens": 406266252.0, "step": 10695 }, { "epoch": 1.360641139804096, "ewc_loss": 0.96875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 12.13353157043457, "learning_rate": 5e-05, "loss": 1.3225, "mean_token_accuracy": 0.8777905702590942, "num_tokens": 406310555.0, "step": 10696 }, { "epoch": 1.3607683500826866, "ewc_loss": 0.95703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000904083251953125, "grad_norm": 9.477691650390625, "learning_rate": 5e-05, "loss": 1.2899, "mean_token_accuracy": 0.8818031549453735, "num_tokens": 406349620.0, "step": 10697 }, { "epoch": 1.3608955603612771, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 11.214825630187988, "learning_rate": 5e-05, "loss": 1.0933, "mean_token_accuracy": 0.8732348680496216, "num_tokens": 406380956.0, "step": 10698 }, { "epoch": 1.3610227706398677, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 9.367762565612793, "learning_rate": 5e-05, "loss": 1.2006, "mean_token_accuracy": 0.8846888542175293, "num_tokens": 406417274.0, "step": 10699 }, { "epoch": 1.3611499809184582, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.783248901367188, "learning_rate": 5e-05, "loss": 1.1716, "mean_token_accuracy": 0.8577808141708374, "num_tokens": 406453809.0, "step": 10700 }, { "epoch": 1.3612771911970487, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.52553653717041, "learning_rate": 5e-05, "loss": 0.9905, "mean_token_accuracy": 0.8768883347511292, "num_tokens": 406489227.0, "step": 10701 }, { "epoch": 1.3614044014756392, "ewc_loss": 0.75, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 10.577910423278809, "learning_rate": 5e-05, "loss": 1.1084, "mean_token_accuracy": 0.8766465187072754, "num_tokens": 406532736.0, "step": 10702 }, { "epoch": 1.3615316117542298, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.074141502380371, "learning_rate": 5e-05, "loss": 0.9496, "mean_token_accuracy": 0.8716893792152405, "num_tokens": 406574409.0, "step": 10703 }, { "epoch": 1.3616588220328203, "ewc_loss": 0.78125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.905621528625488, "learning_rate": 5e-05, "loss": 1.1847, "mean_token_accuracy": 0.8590341210365295, "num_tokens": 406612519.0, "step": 10704 }, { "epoch": 1.3617860323114108, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.450109481811523, "learning_rate": 5e-05, "loss": 1.0158, "mean_token_accuracy": 0.8654931783676147, "num_tokens": 406652578.0, "step": 10705 }, { "epoch": 1.3619132425900013, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 10.985058784484863, "learning_rate": 5e-05, "loss": 1.1725, "mean_token_accuracy": 0.8582131862640381, "num_tokens": 406688739.0, "step": 10706 }, { "epoch": 1.3620404528685919, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.073569297790527, "learning_rate": 5e-05, "loss": 1.0334, "mean_token_accuracy": 0.8774591088294983, "num_tokens": 406727721.0, "step": 10707 }, { "epoch": 1.3621676631471824, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.380894660949707, "learning_rate": 5e-05, "loss": 1.0847, "mean_token_accuracy": 0.8683698177337646, "num_tokens": 406762864.0, "step": 10708 }, { "epoch": 1.362294873425773, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.056814193725586, "learning_rate": 5e-05, "loss": 0.9793, "mean_token_accuracy": 0.8818162679672241, "num_tokens": 406803333.0, "step": 10709 }, { "epoch": 1.3624220837043632, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.727176666259766, "learning_rate": 5e-05, "loss": 1.0119, "mean_token_accuracy": 0.8773292899131775, "num_tokens": 406840942.0, "step": 10710 }, { "epoch": 1.3625492939829538, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.642446517944336, "learning_rate": 5e-05, "loss": 0.9468, "mean_token_accuracy": 0.875251054763794, "num_tokens": 406877293.0, "step": 10711 }, { "epoch": 1.3626765042615443, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.357117652893066, "learning_rate": 5e-05, "loss": 0.9854, "mean_token_accuracy": 0.8726651072502136, "num_tokens": 406915622.0, "step": 10712 }, { "epoch": 1.3628037145401348, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.086860656738281, "learning_rate": 5e-05, "loss": 0.8359, "mean_token_accuracy": 0.8817977905273438, "num_tokens": 406955378.0, "step": 10713 }, { "epoch": 1.3629309248187254, "ewc_loss": 0.625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.497538566589355, "learning_rate": 5e-05, "loss": 1.0215, "mean_token_accuracy": 0.8644992709159851, "num_tokens": 406997305.0, "step": 10714 }, { "epoch": 1.3630581350973159, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.693676948547363, "learning_rate": 5e-05, "loss": 0.7823, "mean_token_accuracy": 0.8860061168670654, "num_tokens": 407039121.0, "step": 10715 }, { "epoch": 1.3631853453759064, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.912182807922363, "learning_rate": 5e-05, "loss": 1.0599, "mean_token_accuracy": 0.8647128343582153, "num_tokens": 407081007.0, "step": 10716 }, { "epoch": 1.363312555654497, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.967191696166992, "learning_rate": 5e-05, "loss": 0.863, "mean_token_accuracy": 0.8734525442123413, "num_tokens": 407122498.0, "step": 10717 }, { "epoch": 1.3634397659330875, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.862584114074707, "learning_rate": 5e-05, "loss": 1.0147, "mean_token_accuracy": 0.8829282522201538, "num_tokens": 407169304.0, "step": 10718 }, { "epoch": 1.3635669762116778, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.499601364135742, "learning_rate": 5e-05, "loss": 0.9698, "mean_token_accuracy": 0.859252393245697, "num_tokens": 407204364.0, "step": 10719 }, { "epoch": 1.3636941864902683, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.134454727172852, "learning_rate": 5e-05, "loss": 1.0443, "mean_token_accuracy": 0.8491520881652832, "num_tokens": 407242467.0, "step": 10720 }, { "epoch": 1.3638213967688588, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.459680557250977, "learning_rate": 5e-05, "loss": 0.9107, "mean_token_accuracy": 0.874026894569397, "num_tokens": 407279965.0, "step": 10721 }, { "epoch": 1.3639486070474494, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.528904914855957, "learning_rate": 5e-05, "loss": 0.879, "mean_token_accuracy": 0.8818514943122864, "num_tokens": 407316086.0, "step": 10722 }, { "epoch": 1.3640758173260399, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.9809465408325195, "learning_rate": 5e-05, "loss": 0.8388, "mean_token_accuracy": 0.8620208501815796, "num_tokens": 407355433.0, "step": 10723 }, { "epoch": 1.3642030276046304, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.498766899108887, "learning_rate": 5e-05, "loss": 0.8709, "mean_token_accuracy": 0.8777112364768982, "num_tokens": 407392919.0, "step": 10724 }, { "epoch": 1.364330237883221, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.413244724273682, "learning_rate": 5e-05, "loss": 0.7814, "mean_token_accuracy": 0.8682472705841064, "num_tokens": 407428089.0, "step": 10725 }, { "epoch": 1.3644574481618115, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.84371566772461, "learning_rate": 5e-05, "loss": 0.8888, "mean_token_accuracy": 0.8813957571983337, "num_tokens": 407467232.0, "step": 10726 }, { "epoch": 1.364584658440402, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.1533331871032715, "learning_rate": 5e-05, "loss": 0.7788, "mean_token_accuracy": 0.8662399649620056, "num_tokens": 407504975.0, "step": 10727 }, { "epoch": 1.3647118687189925, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.144771575927734, "learning_rate": 5e-05, "loss": 0.8893, "mean_token_accuracy": 0.8924309015274048, "num_tokens": 407535308.0, "step": 10728 }, { "epoch": 1.364839078997583, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.330695629119873, "learning_rate": 5e-05, "loss": 0.7728, "mean_token_accuracy": 0.8758014440536499, "num_tokens": 407578956.0, "step": 10729 }, { "epoch": 1.3649662892761736, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.074583053588867, "learning_rate": 5e-05, "loss": 0.9405, "mean_token_accuracy": 0.8763217926025391, "num_tokens": 407617633.0, "step": 10730 }, { "epoch": 1.3650934995547641, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.803375720977783, "learning_rate": 5e-05, "loss": 0.8152, "mean_token_accuracy": 0.8750433921813965, "num_tokens": 407646322.0, "step": 10731 }, { "epoch": 1.3652207098333546, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.475335121154785, "learning_rate": 5e-05, "loss": 0.9232, "mean_token_accuracy": 0.8618783950805664, "num_tokens": 407684957.0, "step": 10732 }, { "epoch": 1.3653479201119452, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.017796516418457, "learning_rate": 5e-05, "loss": 0.8091, "mean_token_accuracy": 0.8764681816101074, "num_tokens": 407722917.0, "step": 10733 }, { "epoch": 1.3654751303905355, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.8433027267456055, "learning_rate": 5e-05, "loss": 0.8131, "mean_token_accuracy": 0.8728824853897095, "num_tokens": 407759599.0, "step": 10734 }, { "epoch": 1.365602340669126, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.934850215911865, "learning_rate": 5e-05, "loss": 0.8018, "mean_token_accuracy": 0.8708317875862122, "num_tokens": 407796416.0, "step": 10735 }, { "epoch": 1.3657295509477165, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.6714653968811035, "learning_rate": 5e-05, "loss": 0.8695, "mean_token_accuracy": 0.8491939306259155, "num_tokens": 407841211.0, "step": 10736 }, { "epoch": 1.365856761226307, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.6824631690979, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.8700640201568604, "num_tokens": 407879409.0, "step": 10737 }, { "epoch": 1.3659839715048976, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.749461650848389, "learning_rate": 5e-05, "loss": 0.7795, "mean_token_accuracy": 0.8671988248825073, "num_tokens": 407915275.0, "step": 10738 }, { "epoch": 1.3661111817834881, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.494448184967041, "learning_rate": 5e-05, "loss": 0.7459, "mean_token_accuracy": 0.8737980127334595, "num_tokens": 407953641.0, "step": 10739 }, { "epoch": 1.3662383920620786, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.791431427001953, "learning_rate": 5e-05, "loss": 0.7499, "mean_token_accuracy": 0.8781599998474121, "num_tokens": 407992662.0, "step": 10740 }, { "epoch": 1.3663656023406692, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.372485160827637, "learning_rate": 5e-05, "loss": 0.7043, "mean_token_accuracy": 0.8800942301750183, "num_tokens": 408026005.0, "step": 10741 }, { "epoch": 1.3664928126192597, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.862629413604736, "learning_rate": 5e-05, "loss": 0.7585, "mean_token_accuracy": 0.8741946220397949, "num_tokens": 408065250.0, "step": 10742 }, { "epoch": 1.3666200228978502, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.198585510253906, "learning_rate": 5e-05, "loss": 0.7156, "mean_token_accuracy": 0.8727544546127319, "num_tokens": 408109333.0, "step": 10743 }, { "epoch": 1.3667472331764405, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.081615447998047, "learning_rate": 5e-05, "loss": 0.7705, "mean_token_accuracy": 0.8784123659133911, "num_tokens": 408142416.0, "step": 10744 }, { "epoch": 1.366874443455031, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.939918041229248, "learning_rate": 5e-05, "loss": 0.6543, "mean_token_accuracy": 0.8875421285629272, "num_tokens": 408184335.0, "step": 10745 }, { "epoch": 1.3670016537336216, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.367307662963867, "learning_rate": 5e-05, "loss": 0.817, "mean_token_accuracy": 0.8719032406806946, "num_tokens": 408221585.0, "step": 10746 }, { "epoch": 1.3671288640122121, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.7540435791015625, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.8490548729896545, "num_tokens": 408262105.0, "step": 10747 }, { "epoch": 1.3672560742908026, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.547552108764648, "learning_rate": 5e-05, "loss": 0.8177, "mean_token_accuracy": 0.8740431666374207, "num_tokens": 408302365.0, "step": 10748 }, { "epoch": 1.3673832845693932, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.76035213470459, "learning_rate": 5e-05, "loss": 0.712, "mean_token_accuracy": 0.8661648035049438, "num_tokens": 408342033.0, "step": 10749 }, { "epoch": 1.3675104948479837, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.54775333404541, "learning_rate": 5e-05, "loss": 0.8302, "mean_token_accuracy": 0.8769256472587585, "num_tokens": 408381512.0, "step": 10750 }, { "epoch": 1.3676377051265742, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.988918304443359, "learning_rate": 5e-05, "loss": 0.7169, "mean_token_accuracy": 0.8716272711753845, "num_tokens": 408423867.0, "step": 10751 }, { "epoch": 1.3677649154051648, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.206302642822266, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.8749350309371948, "num_tokens": 408468417.0, "step": 10752 }, { "epoch": 1.3678921256837553, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.448076248168945, "learning_rate": 5e-05, "loss": 0.817, "mean_token_accuracy": 0.8515414595603943, "num_tokens": 408514628.0, "step": 10753 }, { "epoch": 1.3680193359623458, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.544480800628662, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.8586153388023376, "num_tokens": 408551936.0, "step": 10754 }, { "epoch": 1.3681465462409363, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.993806838989258, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.8670052289962769, "num_tokens": 408581713.0, "step": 10755 }, { "epoch": 1.3682737565195269, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.905445098876953, "learning_rate": 5e-05, "loss": 0.6756, "mean_token_accuracy": 0.880132794380188, "num_tokens": 408618265.0, "step": 10756 }, { "epoch": 1.3684009667981174, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.559224128723145, "learning_rate": 5e-05, "loss": 0.8418, "mean_token_accuracy": 0.8718273043632507, "num_tokens": 408655701.0, "step": 10757 }, { "epoch": 1.368528177076708, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.4113664627075195, "learning_rate": 5e-05, "loss": 0.7087, "mean_token_accuracy": 0.8626542091369629, "num_tokens": 408690281.0, "step": 10758 }, { "epoch": 1.3686553873552982, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 9.156444549560547, "learning_rate": 5e-05, "loss": 0.9549, "mean_token_accuracy": 0.8540502786636353, "num_tokens": 408733483.0, "step": 10759 }, { "epoch": 1.3687825976338888, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.200202465057373, "learning_rate": 5e-05, "loss": 0.6853, "mean_token_accuracy": 0.8732205629348755, "num_tokens": 408770596.0, "step": 10760 }, { "epoch": 1.3689098079124793, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.564885139465332, "learning_rate": 5e-05, "loss": 0.9286, "mean_token_accuracy": 0.8827719688415527, "num_tokens": 408808265.0, "step": 10761 }, { "epoch": 1.3690370181910698, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.788774490356445, "learning_rate": 5e-05, "loss": 0.7144, "mean_token_accuracy": 0.8824725151062012, "num_tokens": 408847205.0, "step": 10762 }, { "epoch": 1.3691642284696603, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.055232048034668, "learning_rate": 5e-05, "loss": 0.9743, "mean_token_accuracy": 0.8679630160331726, "num_tokens": 408885744.0, "step": 10763 }, { "epoch": 1.3692914387482509, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.920285224914551, "learning_rate": 5e-05, "loss": 0.8538, "mean_token_accuracy": 0.8789268732070923, "num_tokens": 408925393.0, "step": 10764 }, { "epoch": 1.3694186490268414, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.690984725952148, "learning_rate": 5e-05, "loss": 0.8242, "mean_token_accuracy": 0.8701877593994141, "num_tokens": 408964995.0, "step": 10765 }, { "epoch": 1.369545859305432, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.659734725952148, "learning_rate": 5e-05, "loss": 0.8805, "mean_token_accuracy": 0.8808059692382812, "num_tokens": 408996336.0, "step": 10766 }, { "epoch": 1.3696730695840225, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.919600963592529, "learning_rate": 5e-05, "loss": 0.7894, "mean_token_accuracy": 0.8545211553573608, "num_tokens": 409034972.0, "step": 10767 }, { "epoch": 1.3698002798626128, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.601659774780273, "learning_rate": 5e-05, "loss": 0.8423, "mean_token_accuracy": 0.8829998970031738, "num_tokens": 409069421.0, "step": 10768 }, { "epoch": 1.3699274901412033, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.344444751739502, "learning_rate": 5e-05, "loss": 0.7864, "mean_token_accuracy": 0.8662981986999512, "num_tokens": 409104763.0, "step": 10769 }, { "epoch": 1.3700547004197938, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.659612655639648, "learning_rate": 5e-05, "loss": 0.7681, "mean_token_accuracy": 0.8746492862701416, "num_tokens": 409142799.0, "step": 10770 }, { "epoch": 1.3701819106983844, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.177515983581543, "learning_rate": 5e-05, "loss": 0.836, "mean_token_accuracy": 0.8641440868377686, "num_tokens": 409178861.0, "step": 10771 }, { "epoch": 1.3703091209769749, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.745698928833008, "learning_rate": 5e-05, "loss": 0.7001, "mean_token_accuracy": 0.8732483386993408, "num_tokens": 409218628.0, "step": 10772 }, { "epoch": 1.3704363312555654, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.936666488647461, "learning_rate": 5e-05, "loss": 0.9359, "mean_token_accuracy": 0.8543062210083008, "num_tokens": 409255209.0, "step": 10773 }, { "epoch": 1.370563541534156, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.325821399688721, "learning_rate": 5e-05, "loss": 0.7027, "mean_token_accuracy": 0.870255708694458, "num_tokens": 409295807.0, "step": 10774 }, { "epoch": 1.3706907518127465, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.462374687194824, "learning_rate": 5e-05, "loss": 0.9456, "mean_token_accuracy": 0.8743261694908142, "num_tokens": 409329263.0, "step": 10775 }, { "epoch": 1.370817962091337, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.601012229919434, "learning_rate": 5e-05, "loss": 0.6846, "mean_token_accuracy": 0.888507068157196, "num_tokens": 409365571.0, "step": 10776 }, { "epoch": 1.3709451723699275, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.367355346679688, "learning_rate": 5e-05, "loss": 0.9851, "mean_token_accuracy": 0.8692339062690735, "num_tokens": 409406096.0, "step": 10777 }, { "epoch": 1.371072382648518, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.4050493240356445, "learning_rate": 5e-05, "loss": 0.888, "mean_token_accuracy": 0.8592090606689453, "num_tokens": 409450436.0, "step": 10778 }, { "epoch": 1.3711995929271086, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.663154602050781, "learning_rate": 5e-05, "loss": 0.9434, "mean_token_accuracy": 0.8717780113220215, "num_tokens": 409485134.0, "step": 10779 }, { "epoch": 1.371326803205699, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.851926326751709, "learning_rate": 5e-05, "loss": 0.8594, "mean_token_accuracy": 0.8739306330680847, "num_tokens": 409522890.0, "step": 10780 }, { "epoch": 1.3714540134842896, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.188322067260742, "learning_rate": 5e-05, "loss": 0.8946, "mean_token_accuracy": 0.8684202432632446, "num_tokens": 409561537.0, "step": 10781 }, { "epoch": 1.3715812237628802, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.526230812072754, "learning_rate": 5e-05, "loss": 0.7803, "mean_token_accuracy": 0.8809289932250977, "num_tokens": 409595259.0, "step": 10782 }, { "epoch": 1.3717084340414705, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.54006290435791, "learning_rate": 5e-05, "loss": 0.8802, "mean_token_accuracy": 0.8789792060852051, "num_tokens": 409633978.0, "step": 10783 }, { "epoch": 1.371835644320061, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.876147270202637, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8727034330368042, "num_tokens": 409671320.0, "step": 10784 }, { "epoch": 1.3719628545986515, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.15811538696289, "learning_rate": 5e-05, "loss": 0.9402, "mean_token_accuracy": 0.8756568431854248, "num_tokens": 409712816.0, "step": 10785 }, { "epoch": 1.372090064877242, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.889494895935059, "learning_rate": 5e-05, "loss": 0.7747, "mean_token_accuracy": 0.8697761297225952, "num_tokens": 409751918.0, "step": 10786 }, { "epoch": 1.3722172751558326, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.057141304016113, "learning_rate": 5e-05, "loss": 0.9288, "mean_token_accuracy": 0.8855314254760742, "num_tokens": 409786905.0, "step": 10787 }, { "epoch": 1.3723444854344231, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.5976786613464355, "learning_rate": 5e-05, "loss": 0.8538, "mean_token_accuracy": 0.8724087476730347, "num_tokens": 409827634.0, "step": 10788 }, { "epoch": 1.3724716957130136, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.140768051147461, "learning_rate": 5e-05, "loss": 0.8717, "mean_token_accuracy": 0.8752099275588989, "num_tokens": 409863060.0, "step": 10789 }, { "epoch": 1.3725989059916042, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.129944801330566, "learning_rate": 5e-05, "loss": 0.8598, "mean_token_accuracy": 0.8777044415473938, "num_tokens": 409896723.0, "step": 10790 }, { "epoch": 1.3727261162701947, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.411632061004639, "learning_rate": 5e-05, "loss": 0.7811, "mean_token_accuracy": 0.8714890480041504, "num_tokens": 409935912.0, "step": 10791 }, { "epoch": 1.3728533265487852, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.159223556518555, "learning_rate": 5e-05, "loss": 0.8439, "mean_token_accuracy": 0.8748955726623535, "num_tokens": 409973381.0, "step": 10792 }, { "epoch": 1.3729805368273755, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.336977958679199, "learning_rate": 5e-05, "loss": 0.7786, "mean_token_accuracy": 0.8654465675354004, "num_tokens": 410010548.0, "step": 10793 }, { "epoch": 1.373107747105966, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.865643501281738, "learning_rate": 5e-05, "loss": 0.7553, "mean_token_accuracy": 0.885771632194519, "num_tokens": 410046456.0, "step": 10794 }, { "epoch": 1.3732349573845566, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.514306545257568, "learning_rate": 5e-05, "loss": 0.7094, "mean_token_accuracy": 0.887209415435791, "num_tokens": 410080568.0, "step": 10795 }, { "epoch": 1.3733621676631471, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.663423538208008, "learning_rate": 5e-05, "loss": 0.7422, "mean_token_accuracy": 0.8774638175964355, "num_tokens": 410121008.0, "step": 10796 }, { "epoch": 1.3734893779417376, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.495196342468262, "learning_rate": 5e-05, "loss": 0.7354, "mean_token_accuracy": 0.875688910484314, "num_tokens": 410165872.0, "step": 10797 }, { "epoch": 1.3736165882203282, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.732634544372559, "learning_rate": 5e-05, "loss": 0.7611, "mean_token_accuracy": 0.876177191734314, "num_tokens": 410207379.0, "step": 10798 }, { "epoch": 1.3737437984989187, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.290868759155273, "learning_rate": 5e-05, "loss": 0.7634, "mean_token_accuracy": 0.8599903583526611, "num_tokens": 410246729.0, "step": 10799 }, { "epoch": 1.3738710087775092, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.9726457595825195, "learning_rate": 5e-05, "loss": 0.752, "mean_token_accuracy": 0.8769837021827698, "num_tokens": 410281237.0, "step": 10800 }, { "epoch": 1.3739982190560998, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.015275478363037, "learning_rate": 5e-05, "loss": 0.7143, "mean_token_accuracy": 0.8680528402328491, "num_tokens": 410318485.0, "step": 10801 }, { "epoch": 1.3741254293346903, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.237078666687012, "learning_rate": 5e-05, "loss": 0.8006, "mean_token_accuracy": 0.8732085227966309, "num_tokens": 410357995.0, "step": 10802 }, { "epoch": 1.3742526396132808, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.773454189300537, "learning_rate": 5e-05, "loss": 0.7084, "mean_token_accuracy": 0.8707469701766968, "num_tokens": 410395672.0, "step": 10803 }, { "epoch": 1.3743798498918713, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.575699806213379, "learning_rate": 5e-05, "loss": 0.902, "mean_token_accuracy": 0.851599931716919, "num_tokens": 410433566.0, "step": 10804 }, { "epoch": 1.3745070601704619, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.5093231201171875, "learning_rate": 5e-05, "loss": 0.6929, "mean_token_accuracy": 0.8709717988967896, "num_tokens": 410473622.0, "step": 10805 }, { "epoch": 1.3746342704490524, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.959024429321289, "learning_rate": 5e-05, "loss": 0.8523, "mean_token_accuracy": 0.8793972730636597, "num_tokens": 410507155.0, "step": 10806 }, { "epoch": 1.374761480727643, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.433507919311523, "learning_rate": 5e-05, "loss": 0.7224, "mean_token_accuracy": 0.8663505911827087, "num_tokens": 410543352.0, "step": 10807 }, { "epoch": 1.3748886910062332, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.15658187866211, "learning_rate": 5e-05, "loss": 0.918, "mean_token_accuracy": 0.8785403966903687, "num_tokens": 410584560.0, "step": 10808 }, { "epoch": 1.3750159012848238, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.831450462341309, "learning_rate": 5e-05, "loss": 0.7281, "mean_token_accuracy": 0.8802976608276367, "num_tokens": 410615948.0, "step": 10809 }, { "epoch": 1.3751431115634143, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.7247314453125, "learning_rate": 5e-05, "loss": 0.9114, "mean_token_accuracy": 0.8727734088897705, "num_tokens": 410649513.0, "step": 10810 }, { "epoch": 1.3752703218420048, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.573905944824219, "learning_rate": 5e-05, "loss": 0.8589, "mean_token_accuracy": 0.8701560497283936, "num_tokens": 410693087.0, "step": 10811 }, { "epoch": 1.3753975321205953, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.752124309539795, "learning_rate": 5e-05, "loss": 0.785, "mean_token_accuracy": 0.8836504817008972, "num_tokens": 410729200.0, "step": 10812 }, { "epoch": 1.3755247423991859, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.233250617980957, "learning_rate": 5e-05, "loss": 0.8938, "mean_token_accuracy": 0.8671207427978516, "num_tokens": 410769830.0, "step": 10813 }, { "epoch": 1.3756519526777764, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.939639091491699, "learning_rate": 5e-05, "loss": 0.7322, "mean_token_accuracy": 0.8692182302474976, "num_tokens": 410815570.0, "step": 10814 }, { "epoch": 1.375779162956367, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.62425422668457, "learning_rate": 5e-05, "loss": 0.7992, "mean_token_accuracy": 0.8959658145904541, "num_tokens": 410852777.0, "step": 10815 }, { "epoch": 1.3759063732349575, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.684542655944824, "learning_rate": 5e-05, "loss": 0.7134, "mean_token_accuracy": 0.8759344220161438, "num_tokens": 410886809.0, "step": 10816 }, { "epoch": 1.3760335835135478, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.680521965026855, "learning_rate": 5e-05, "loss": 0.875, "mean_token_accuracy": 0.8758213520050049, "num_tokens": 410924631.0, "step": 10817 }, { "epoch": 1.3761607937921383, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.835833549499512, "learning_rate": 5e-05, "loss": 0.7239, "mean_token_accuracy": 0.8736844062805176, "num_tokens": 410965212.0, "step": 10818 }, { "epoch": 1.3762880040707288, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.48555850982666, "learning_rate": 5e-05, "loss": 0.9209, "mean_token_accuracy": 0.8606902360916138, "num_tokens": 411003429.0, "step": 10819 }, { "epoch": 1.3764152143493193, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.122674942016602, "learning_rate": 5e-05, "loss": 0.8139, "mean_token_accuracy": 0.8580255508422852, "num_tokens": 411043121.0, "step": 10820 }, { "epoch": 1.3765424246279099, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.149750709533691, "learning_rate": 5e-05, "loss": 0.8806, "mean_token_accuracy": 0.8629189133644104, "num_tokens": 411082146.0, "step": 10821 }, { "epoch": 1.3766696349065004, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.381378650665283, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.8669143915176392, "num_tokens": 411116070.0, "step": 10822 }, { "epoch": 1.376796845185091, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.844446182250977, "learning_rate": 5e-05, "loss": 0.8256, "mean_token_accuracy": 0.8651162385940552, "num_tokens": 411162066.0, "step": 10823 }, { "epoch": 1.3769240554636815, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.4959259033203125, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8740875720977783, "num_tokens": 411200939.0, "step": 10824 }, { "epoch": 1.377051265742272, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.712318420410156, "learning_rate": 5e-05, "loss": 0.8023, "mean_token_accuracy": 0.8662895560264587, "num_tokens": 411241161.0, "step": 10825 }, { "epoch": 1.3771784760208625, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.431674480438232, "learning_rate": 5e-05, "loss": 0.7552, "mean_token_accuracy": 0.8695300817489624, "num_tokens": 411284316.0, "step": 10826 }, { "epoch": 1.377305686299453, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.798431396484375, "learning_rate": 5e-05, "loss": 0.8297, "mean_token_accuracy": 0.8534078598022461, "num_tokens": 411320795.0, "step": 10827 }, { "epoch": 1.3774328965780436, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.20682954788208, "learning_rate": 5e-05, "loss": 0.7156, "mean_token_accuracy": 0.8752481937408447, "num_tokens": 411357513.0, "step": 10828 }, { "epoch": 1.377560106856634, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.025385856628418, "learning_rate": 5e-05, "loss": 0.7337, "mean_token_accuracy": 0.8919438719749451, "num_tokens": 411395310.0, "step": 10829 }, { "epoch": 1.3776873171352246, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.935862064361572, "learning_rate": 5e-05, "loss": 0.7104, "mean_token_accuracy": 0.8694933652877808, "num_tokens": 411437774.0, "step": 10830 }, { "epoch": 1.3778145274138152, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.363175392150879, "learning_rate": 5e-05, "loss": 0.789, "mean_token_accuracy": 0.8806060552597046, "num_tokens": 411471399.0, "step": 10831 }, { "epoch": 1.3779417376924055, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.583128452301025, "learning_rate": 5e-05, "loss": 0.732, "mean_token_accuracy": 0.8607531189918518, "num_tokens": 411505289.0, "step": 10832 }, { "epoch": 1.378068947970996, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.827388763427734, "learning_rate": 5e-05, "loss": 0.8335, "mean_token_accuracy": 0.8830626010894775, "num_tokens": 411541463.0, "step": 10833 }, { "epoch": 1.3781961582495865, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.308950901031494, "learning_rate": 5e-05, "loss": 0.6978, "mean_token_accuracy": 0.870333194732666, "num_tokens": 411577276.0, "step": 10834 }, { "epoch": 1.378323368528177, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.276588439941406, "learning_rate": 5e-05, "loss": 1.0247, "mean_token_accuracy": 0.8440887928009033, "num_tokens": 411610417.0, "step": 10835 }, { "epoch": 1.3784505788067676, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.488018989562988, "learning_rate": 5e-05, "loss": 0.7765, "mean_token_accuracy": 0.8553113341331482, "num_tokens": 411646539.0, "step": 10836 }, { "epoch": 1.378577789085358, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.155064582824707, "learning_rate": 5e-05, "loss": 0.9846, "mean_token_accuracy": 0.864641547203064, "num_tokens": 411685890.0, "step": 10837 }, { "epoch": 1.3787049993639486, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.336226940155029, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.8812589645385742, "num_tokens": 411725566.0, "step": 10838 }, { "epoch": 1.3788322096425392, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.108074188232422, "learning_rate": 5e-05, "loss": 0.8632, "mean_token_accuracy": 0.8776137828826904, "num_tokens": 411767256.0, "step": 10839 }, { "epoch": 1.3789594199211297, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.284097671508789, "learning_rate": 5e-05, "loss": 0.91, "mean_token_accuracy": 0.8713763356208801, "num_tokens": 411802210.0, "step": 10840 }, { "epoch": 1.3790866301997202, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.848108768463135, "learning_rate": 5e-05, "loss": 0.6718, "mean_token_accuracy": 0.8908095955848694, "num_tokens": 411842011.0, "step": 10841 }, { "epoch": 1.3792138404783105, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.086297035217285, "learning_rate": 5e-05, "loss": 0.9373, "mean_token_accuracy": 0.8776422142982483, "num_tokens": 411877137.0, "step": 10842 }, { "epoch": 1.379341050756901, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.297316551208496, "learning_rate": 5e-05, "loss": 0.6858, "mean_token_accuracy": 0.8776744604110718, "num_tokens": 411909130.0, "step": 10843 }, { "epoch": 1.3794682610354916, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.531001091003418, "learning_rate": 5e-05, "loss": 0.9957, "mean_token_accuracy": 0.8667365908622742, "num_tokens": 411942542.0, "step": 10844 }, { "epoch": 1.3795954713140821, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.700226306915283, "learning_rate": 5e-05, "loss": 0.8096, "mean_token_accuracy": 0.8638607859611511, "num_tokens": 411974272.0, "step": 10845 }, { "epoch": 1.3797226815926726, "ewc_loss": 0.625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.278090476989746, "learning_rate": 5e-05, "loss": 0.9701, "mean_token_accuracy": 0.8806134462356567, "num_tokens": 412012748.0, "step": 10846 }, { "epoch": 1.3798498918712632, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.626918792724609, "learning_rate": 5e-05, "loss": 0.9368, "mean_token_accuracy": 0.8672960996627808, "num_tokens": 412051171.0, "step": 10847 }, { "epoch": 1.3799771021498537, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.313970565795898, "learning_rate": 5e-05, "loss": 0.9709, "mean_token_accuracy": 0.8593857884407043, "num_tokens": 412090267.0, "step": 10848 }, { "epoch": 1.3801043124284442, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.203174591064453, "learning_rate": 5e-05, "loss": 0.9631, "mean_token_accuracy": 0.8639568090438843, "num_tokens": 412118471.0, "step": 10849 }, { "epoch": 1.3802315227070348, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.529296398162842, "learning_rate": 5e-05, "loss": 0.8119, "mean_token_accuracy": 0.8773456811904907, "num_tokens": 412163639.0, "step": 10850 }, { "epoch": 1.3803587329856253, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.135129928588867, "learning_rate": 5e-05, "loss": 0.9293, "mean_token_accuracy": 0.8599441051483154, "num_tokens": 412203685.0, "step": 10851 }, { "epoch": 1.3804859432642158, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.6399431228637695, "learning_rate": 5e-05, "loss": 0.8117, "mean_token_accuracy": 0.8735994100570679, "num_tokens": 412238810.0, "step": 10852 }, { "epoch": 1.3806131535428063, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.467097282409668, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.8696917295455933, "num_tokens": 412273840.0, "step": 10853 }, { "epoch": 1.3807403638213969, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.146501541137695, "learning_rate": 5e-05, "loss": 0.8489, "mean_token_accuracy": 0.8703525066375732, "num_tokens": 412309931.0, "step": 10854 }, { "epoch": 1.3808675740999874, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.898006916046143, "learning_rate": 5e-05, "loss": 0.8198, "mean_token_accuracy": 0.8399724960327148, "num_tokens": 412346934.0, "step": 10855 }, { "epoch": 1.380994784378578, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.575814247131348, "learning_rate": 5e-05, "loss": 0.8778, "mean_token_accuracy": 0.8695915341377258, "num_tokens": 412387942.0, "step": 10856 }, { "epoch": 1.3811219946571682, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.598939895629883, "learning_rate": 5e-05, "loss": 0.7063, "mean_token_accuracy": 0.8751475811004639, "num_tokens": 412429726.0, "step": 10857 }, { "epoch": 1.3812492049357588, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.954716682434082, "learning_rate": 5e-05, "loss": 0.9064, "mean_token_accuracy": 0.8754512667655945, "num_tokens": 412472711.0, "step": 10858 }, { "epoch": 1.3813764152143493, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.5040740966796875, "learning_rate": 5e-05, "loss": 0.6907, "mean_token_accuracy": 0.8809517621994019, "num_tokens": 412511384.0, "step": 10859 }, { "epoch": 1.3815036254929398, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.28239917755127, "learning_rate": 5e-05, "loss": 1.0313, "mean_token_accuracy": 0.8567250370979309, "num_tokens": 412552991.0, "step": 10860 }, { "epoch": 1.3816308357715303, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 6.8231682777404785, "learning_rate": 5e-05, "loss": 0.9066, "mean_token_accuracy": 0.8366407752037048, "num_tokens": 412593794.0, "step": 10861 }, { "epoch": 1.3817580460501209, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.123875617980957, "learning_rate": 5e-05, "loss": 0.9499, "mean_token_accuracy": 0.8817723393440247, "num_tokens": 412625569.0, "step": 10862 }, { "epoch": 1.3818852563287114, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.513814449310303, "learning_rate": 5e-05, "loss": 0.8513, "mean_token_accuracy": 0.8859158754348755, "num_tokens": 412658086.0, "step": 10863 }, { "epoch": 1.382012466607302, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.364359855651855, "learning_rate": 5e-05, "loss": 0.9312, "mean_token_accuracy": 0.8716906905174255, "num_tokens": 412698642.0, "step": 10864 }, { "epoch": 1.3821396768858925, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.991985321044922, "learning_rate": 5e-05, "loss": 0.9034, "mean_token_accuracy": 0.8729763031005859, "num_tokens": 412738327.0, "step": 10865 }, { "epoch": 1.3822668871644828, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.659545421600342, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.8717484474182129, "num_tokens": 412773420.0, "step": 10866 }, { "epoch": 1.3823940974430733, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.021671295166016, "learning_rate": 5e-05, "loss": 0.8597, "mean_token_accuracy": 0.8777681589126587, "num_tokens": 412805383.0, "step": 10867 }, { "epoch": 1.3825213077216638, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.501674175262451, "learning_rate": 5e-05, "loss": 0.7782, "mean_token_accuracy": 0.8790603876113892, "num_tokens": 412846924.0, "step": 10868 }, { "epoch": 1.3826485180002543, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.793101787567139, "learning_rate": 5e-05, "loss": 0.8522, "mean_token_accuracy": 0.8613619804382324, "num_tokens": 412883232.0, "step": 10869 }, { "epoch": 1.3827757282788449, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.519895076751709, "learning_rate": 5e-05, "loss": 0.7559, "mean_token_accuracy": 0.8783044815063477, "num_tokens": 412922110.0, "step": 10870 }, { "epoch": 1.3829029385574354, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.760555267333984, "learning_rate": 5e-05, "loss": 0.8043, "mean_token_accuracy": 0.8665258884429932, "num_tokens": 412963661.0, "step": 10871 }, { "epoch": 1.383030148836026, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.300146579742432, "learning_rate": 5e-05, "loss": 0.6782, "mean_token_accuracy": 0.8964470624923706, "num_tokens": 413002468.0, "step": 10872 }, { "epoch": 1.3831573591146165, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.031282424926758, "learning_rate": 5e-05, "loss": 0.8617, "mean_token_accuracy": 0.855221152305603, "num_tokens": 413048135.0, "step": 10873 }, { "epoch": 1.383284569393207, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.905580043792725, "learning_rate": 5e-05, "loss": 0.7197, "mean_token_accuracy": 0.8709315061569214, "num_tokens": 413082552.0, "step": 10874 }, { "epoch": 1.3834117796717975, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.452421188354492, "learning_rate": 5e-05, "loss": 0.8764, "mean_token_accuracy": 0.864759087562561, "num_tokens": 413129230.0, "step": 10875 }, { "epoch": 1.383538989950388, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.60839319229126, "learning_rate": 5e-05, "loss": 0.674, "mean_token_accuracy": 0.8848994970321655, "num_tokens": 413168830.0, "step": 10876 }, { "epoch": 1.3836662002289786, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.780866622924805, "learning_rate": 5e-05, "loss": 0.874, "mean_token_accuracy": 0.8715147972106934, "num_tokens": 413208423.0, "step": 10877 }, { "epoch": 1.383793410507569, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.511308193206787, "learning_rate": 5e-05, "loss": 0.7176, "mean_token_accuracy": 0.8712441921234131, "num_tokens": 413238133.0, "step": 10878 }, { "epoch": 1.3839206207861596, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.996660232543945, "learning_rate": 5e-05, "loss": 0.9081, "mean_token_accuracy": 0.8771617412567139, "num_tokens": 413273653.0, "step": 10879 }, { "epoch": 1.3840478310647502, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.695359230041504, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.8720093369483948, "num_tokens": 413316492.0, "step": 10880 }, { "epoch": 1.3841750413433405, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.900703430175781, "learning_rate": 5e-05, "loss": 0.9286, "mean_token_accuracy": 0.8807897567749023, "num_tokens": 413359101.0, "step": 10881 }, { "epoch": 1.384302251621931, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.221532821655273, "learning_rate": 5e-05, "loss": 0.7967, "mean_token_accuracy": 0.882293701171875, "num_tokens": 413396618.0, "step": 10882 }, { "epoch": 1.3844294619005215, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.292754173278809, "learning_rate": 5e-05, "loss": 0.8629, "mean_token_accuracy": 0.8837348222732544, "num_tokens": 413431863.0, "step": 10883 }, { "epoch": 1.384556672179112, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.773257732391357, "learning_rate": 5e-05, "loss": 0.8754, "mean_token_accuracy": 0.8679687976837158, "num_tokens": 413470137.0, "step": 10884 }, { "epoch": 1.3846838824577026, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.558271408081055, "learning_rate": 5e-05, "loss": 0.775, "mean_token_accuracy": 0.8838316202163696, "num_tokens": 413508475.0, "step": 10885 }, { "epoch": 1.384811092736293, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.035513877868652, "learning_rate": 5e-05, "loss": 0.8711, "mean_token_accuracy": 0.870017945766449, "num_tokens": 413548846.0, "step": 10886 }, { "epoch": 1.3849383030148836, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.287290096282959, "learning_rate": 5e-05, "loss": 0.773, "mean_token_accuracy": 0.8681174516677856, "num_tokens": 413587375.0, "step": 10887 }, { "epoch": 1.3850655132934742, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.874736309051514, "learning_rate": 5e-05, "loss": 0.8096, "mean_token_accuracy": 0.8751646876335144, "num_tokens": 413626701.0, "step": 10888 }, { "epoch": 1.3851927235720647, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.466713905334473, "learning_rate": 5e-05, "loss": 0.8208, "mean_token_accuracy": 0.8554129600524902, "num_tokens": 413665207.0, "step": 10889 }, { "epoch": 1.385319933850655, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.512231826782227, "learning_rate": 5e-05, "loss": 0.7157, "mean_token_accuracy": 0.8865039348602295, "num_tokens": 413700549.0, "step": 10890 }, { "epoch": 1.3854471441292455, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.7316575050354, "learning_rate": 5e-05, "loss": 0.871, "mean_token_accuracy": 0.8427737951278687, "num_tokens": 413738308.0, "step": 10891 }, { "epoch": 1.385574354407836, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.235263347625732, "learning_rate": 5e-05, "loss": 0.7247, "mean_token_accuracy": 0.8737685680389404, "num_tokens": 413779540.0, "step": 10892 }, { "epoch": 1.3857015646864266, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.917872428894043, "learning_rate": 5e-05, "loss": 0.7258, "mean_token_accuracy": 0.8876115679740906, "num_tokens": 413818677.0, "step": 10893 }, { "epoch": 1.385828774965017, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.031851291656494, "learning_rate": 5e-05, "loss": 0.7342, "mean_token_accuracy": 0.8681038618087769, "num_tokens": 413861397.0, "step": 10894 }, { "epoch": 1.3859559852436076, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.139700889587402, "learning_rate": 5e-05, "loss": 0.7692, "mean_token_accuracy": 0.8778964281082153, "num_tokens": 413897629.0, "step": 10895 }, { "epoch": 1.3860831955221982, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.778280735015869, "learning_rate": 5e-05, "loss": 0.756, "mean_token_accuracy": 0.855696439743042, "num_tokens": 413935711.0, "step": 10896 }, { "epoch": 1.3862104058007887, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.478428840637207, "learning_rate": 5e-05, "loss": 0.8059, "mean_token_accuracy": 0.8793160915374756, "num_tokens": 413977340.0, "step": 10897 }, { "epoch": 1.3863376160793792, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.471715450286865, "learning_rate": 5e-05, "loss": 0.6752, "mean_token_accuracy": 0.8818596601486206, "num_tokens": 414014429.0, "step": 10898 }, { "epoch": 1.3864648263579697, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.937516212463379, "learning_rate": 5e-05, "loss": 0.9208, "mean_token_accuracy": 0.8610053062438965, "num_tokens": 414055426.0, "step": 10899 }, { "epoch": 1.3865920366365603, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.279630184173584, "learning_rate": 5e-05, "loss": 0.7205, "mean_token_accuracy": 0.8654736876487732, "num_tokens": 414086549.0, "step": 10900 }, { "epoch": 1.3867192469151508, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.312946319580078, "learning_rate": 5e-05, "loss": 0.9379, "mean_token_accuracy": 0.8789185285568237, "num_tokens": 414121767.0, "step": 10901 }, { "epoch": 1.3868464571937413, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.6171698570251465, "learning_rate": 5e-05, "loss": 0.7745, "mean_token_accuracy": 0.8650553226470947, "num_tokens": 414154201.0, "step": 10902 }, { "epoch": 1.3869736674723319, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.027180671691895, "learning_rate": 5e-05, "loss": 0.9992, "mean_token_accuracy": 0.8576698303222656, "num_tokens": 414189898.0, "step": 10903 }, { "epoch": 1.3871008777509224, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.534432411193848, "learning_rate": 5e-05, "loss": 0.882, "mean_token_accuracy": 0.8735268115997314, "num_tokens": 414226507.0, "step": 10904 }, { "epoch": 1.387228088029513, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.835328578948975, "learning_rate": 5e-05, "loss": 0.8576, "mean_token_accuracy": 0.8690268397331238, "num_tokens": 414261909.0, "step": 10905 }, { "epoch": 1.3873552983081032, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.453749656677246, "learning_rate": 5e-05, "loss": 0.9401, "mean_token_accuracy": 0.8647019863128662, "num_tokens": 414303296.0, "step": 10906 }, { "epoch": 1.3874825085866938, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.696647644042969, "learning_rate": 5e-05, "loss": 0.7478, "mean_token_accuracy": 0.8700094819068909, "num_tokens": 414340638.0, "step": 10907 }, { "epoch": 1.3876097188652843, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.119718551635742, "learning_rate": 5e-05, "loss": 0.9537, "mean_token_accuracy": 0.8676084876060486, "num_tokens": 414376900.0, "step": 10908 }, { "epoch": 1.3877369291438748, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.407329082489014, "learning_rate": 5e-05, "loss": 0.718, "mean_token_accuracy": 0.8740414977073669, "num_tokens": 414410657.0, "step": 10909 }, { "epoch": 1.3878641394224653, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.310515403747559, "learning_rate": 5e-05, "loss": 1.0141, "mean_token_accuracy": 0.8551326990127563, "num_tokens": 414443165.0, "step": 10910 }, { "epoch": 1.3879913497010559, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 6.92054557800293, "learning_rate": 5e-05, "loss": 0.858, "mean_token_accuracy": 0.8572084903717041, "num_tokens": 414482185.0, "step": 10911 }, { "epoch": 1.3881185599796464, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.814199447631836, "learning_rate": 5e-05, "loss": 1.0127, "mean_token_accuracy": 0.857380211353302, "num_tokens": 414517509.0, "step": 10912 }, { "epoch": 1.388245770258237, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 7.675153732299805, "learning_rate": 5e-05, "loss": 0.9053, "mean_token_accuracy": 0.8711063265800476, "num_tokens": 414558650.0, "step": 10913 }, { "epoch": 1.3883729805368275, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.9335479736328125, "learning_rate": 5e-05, "loss": 0.8582, "mean_token_accuracy": 0.8761107921600342, "num_tokens": 414594301.0, "step": 10914 }, { "epoch": 1.3885001908154178, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.038362503051758, "learning_rate": 5e-05, "loss": 0.8767, "mean_token_accuracy": 0.880181074142456, "num_tokens": 414626895.0, "step": 10915 }, { "epoch": 1.3886274010940083, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.491150856018066, "learning_rate": 5e-05, "loss": 0.7968, "mean_token_accuracy": 0.8712047338485718, "num_tokens": 414664708.0, "step": 10916 }, { "epoch": 1.3887546113725988, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.816412448883057, "learning_rate": 5e-05, "loss": 0.8932, "mean_token_accuracy": 0.8543129563331604, "num_tokens": 414706243.0, "step": 10917 }, { "epoch": 1.3888818216511893, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.74349308013916, "learning_rate": 5e-05, "loss": 0.815, "mean_token_accuracy": 0.8703614473342896, "num_tokens": 414740681.0, "step": 10918 }, { "epoch": 1.3890090319297799, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.218292236328125, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.8730838298797607, "num_tokens": 414774942.0, "step": 10919 }, { "epoch": 1.3891362422083704, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.231861114501953, "learning_rate": 5e-05, "loss": 0.8106, "mean_token_accuracy": 0.8807631731033325, "num_tokens": 414817369.0, "step": 10920 }, { "epoch": 1.389263452486961, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.727543830871582, "learning_rate": 5e-05, "loss": 0.7249, "mean_token_accuracy": 0.8676609992980957, "num_tokens": 414855503.0, "step": 10921 }, { "epoch": 1.3893906627655515, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.685141563415527, "learning_rate": 5e-05, "loss": 0.8944, "mean_token_accuracy": 0.8660703897476196, "num_tokens": 414891974.0, "step": 10922 }, { "epoch": 1.389517873044142, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.450064659118652, "learning_rate": 5e-05, "loss": 0.6726, "mean_token_accuracy": 0.8817996382713318, "num_tokens": 414926773.0, "step": 10923 }, { "epoch": 1.3896450833227325, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.090202331542969, "learning_rate": 5e-05, "loss": 0.9084, "mean_token_accuracy": 0.8789852857589722, "num_tokens": 414962848.0, "step": 10924 }, { "epoch": 1.389772293601323, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.48391580581665, "learning_rate": 5e-05, "loss": 0.7149, "mean_token_accuracy": 0.8768490552902222, "num_tokens": 415003341.0, "step": 10925 }, { "epoch": 1.3898995038799136, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.253981590270996, "learning_rate": 5e-05, "loss": 0.9417, "mean_token_accuracy": 0.8824459314346313, "num_tokens": 415036836.0, "step": 10926 }, { "epoch": 1.390026714158504, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.036471366882324, "learning_rate": 5e-05, "loss": 0.8395, "mean_token_accuracy": 0.8709797859191895, "num_tokens": 415078589.0, "step": 10927 }, { "epoch": 1.3901539244370946, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.786197662353516, "learning_rate": 5e-05, "loss": 0.9144, "mean_token_accuracy": 0.8842325210571289, "num_tokens": 415114916.0, "step": 10928 }, { "epoch": 1.3902811347156852, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.739264011383057, "learning_rate": 5e-05, "loss": 0.9203, "mean_token_accuracy": 0.8684813380241394, "num_tokens": 415153774.0, "step": 10929 }, { "epoch": 1.3904083449942755, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.886007785797119, "learning_rate": 5e-05, "loss": 0.8937, "mean_token_accuracy": 0.8602363467216492, "num_tokens": 415188359.0, "step": 10930 }, { "epoch": 1.390535555272866, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.107770919799805, "learning_rate": 5e-05, "loss": 0.8825, "mean_token_accuracy": 0.8780021667480469, "num_tokens": 415224423.0, "step": 10931 }, { "epoch": 1.3906627655514565, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.33988618850708, "learning_rate": 5e-05, "loss": 0.8065, "mean_token_accuracy": 0.8643994331359863, "num_tokens": 415263219.0, "step": 10932 }, { "epoch": 1.390789975830047, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.059958457946777, "learning_rate": 5e-05, "loss": 0.8723, "mean_token_accuracy": 0.8712396621704102, "num_tokens": 415301043.0, "step": 10933 }, { "epoch": 1.3909171861086376, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.30109167098999, "learning_rate": 5e-05, "loss": 0.7803, "mean_token_accuracy": 0.8699520230293274, "num_tokens": 415339128.0, "step": 10934 }, { "epoch": 1.391044396387228, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.915201663970947, "learning_rate": 5e-05, "loss": 0.822, "mean_token_accuracy": 0.871414303779602, "num_tokens": 415374574.0, "step": 10935 }, { "epoch": 1.3911716066658186, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.235587120056152, "learning_rate": 5e-05, "loss": 0.834, "mean_token_accuracy": 0.8449135422706604, "num_tokens": 415410133.0, "step": 10936 }, { "epoch": 1.3912988169444092, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.0374116897583, "learning_rate": 5e-05, "loss": 0.8553, "mean_token_accuracy": 0.8599724769592285, "num_tokens": 415451329.0, "step": 10937 }, { "epoch": 1.3914260272229997, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.937459468841553, "learning_rate": 5e-05, "loss": 0.732, "mean_token_accuracy": 0.8719288110733032, "num_tokens": 415491448.0, "step": 10938 }, { "epoch": 1.39155323750159, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.440587997436523, "learning_rate": 5e-05, "loss": 0.8509, "mean_token_accuracy": 0.8706279993057251, "num_tokens": 415533583.0, "step": 10939 }, { "epoch": 1.3916804477801805, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.5860795974731445, "learning_rate": 5e-05, "loss": 0.7461, "mean_token_accuracy": 0.860122799873352, "num_tokens": 415567985.0, "step": 10940 }, { "epoch": 1.391807658058771, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.862578392028809, "learning_rate": 5e-05, "loss": 0.893, "mean_token_accuracy": 0.8746622800827026, "num_tokens": 415603130.0, "step": 10941 }, { "epoch": 1.3919348683373616, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.490428447723389, "learning_rate": 5e-05, "loss": 0.7489, "mean_token_accuracy": 0.8628635406494141, "num_tokens": 415641284.0, "step": 10942 }, { "epoch": 1.392062078615952, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.030679702758789, "learning_rate": 5e-05, "loss": 0.9989, "mean_token_accuracy": 0.8584269881248474, "num_tokens": 415682483.0, "step": 10943 }, { "epoch": 1.3921892888945426, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.8806047439575195, "learning_rate": 5e-05, "loss": 0.8439, "mean_token_accuracy": 0.8545486927032471, "num_tokens": 415723364.0, "step": 10944 }, { "epoch": 1.3923164991731332, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.647021293640137, "learning_rate": 5e-05, "loss": 0.9798, "mean_token_accuracy": 0.8608589172363281, "num_tokens": 415762750.0, "step": 10945 }, { "epoch": 1.3924437094517237, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.5065999031066895, "learning_rate": 5e-05, "loss": 0.8944, "mean_token_accuracy": 0.8573635220527649, "num_tokens": 415801363.0, "step": 10946 }, { "epoch": 1.3925709197303142, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.873941421508789, "learning_rate": 5e-05, "loss": 0.8359, "mean_token_accuracy": 0.8752143383026123, "num_tokens": 415836121.0, "step": 10947 }, { "epoch": 1.3926981300089047, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.9491729736328125, "learning_rate": 5e-05, "loss": 0.9092, "mean_token_accuracy": 0.8591623902320862, "num_tokens": 415880145.0, "step": 10948 }, { "epoch": 1.3928253402874953, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.3267598152160645, "learning_rate": 5e-05, "loss": 0.7783, "mean_token_accuracy": 0.8703532218933105, "num_tokens": 415921244.0, "step": 10949 }, { "epoch": 1.3929525505660858, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.991172790527344, "learning_rate": 5e-05, "loss": 0.8201, "mean_token_accuracy": 0.8723804950714111, "num_tokens": 415957166.0, "step": 10950 }, { "epoch": 1.3930797608446763, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.330982685089111, "learning_rate": 5e-05, "loss": 0.7795, "mean_token_accuracy": 0.8625830411911011, "num_tokens": 415999462.0, "step": 10951 }, { "epoch": 1.3932069711232669, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.691090106964111, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.8708402514457703, "num_tokens": 416038619.0, "step": 10952 }, { "epoch": 1.3933341814018574, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.604099273681641, "learning_rate": 5e-05, "loss": 0.7819, "mean_token_accuracy": 0.8673492670059204, "num_tokens": 416084885.0, "step": 10953 }, { "epoch": 1.393461391680448, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.3272552490234375, "learning_rate": 5e-05, "loss": 0.7827, "mean_token_accuracy": 0.862925112247467, "num_tokens": 416126863.0, "step": 10954 }, { "epoch": 1.3935886019590382, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.853339195251465, "learning_rate": 5e-05, "loss": 0.7174, "mean_token_accuracy": 0.8929756283760071, "num_tokens": 416160888.0, "step": 10955 }, { "epoch": 1.3937158122376287, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.055714130401611, "learning_rate": 5e-05, "loss": 0.6983, "mean_token_accuracy": 0.8766517043113708, "num_tokens": 416199012.0, "step": 10956 }, { "epoch": 1.3938430225162193, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.08405590057373, "learning_rate": 5e-05, "loss": 0.7626, "mean_token_accuracy": 0.8793257474899292, "num_tokens": 416234078.0, "step": 10957 }, { "epoch": 1.3939702327948098, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.8155341148376465, "learning_rate": 5e-05, "loss": 0.6803, "mean_token_accuracy": 0.8788455128669739, "num_tokens": 416272912.0, "step": 10958 }, { "epoch": 1.3940974430734003, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.388720512390137, "learning_rate": 5e-05, "loss": 0.7974, "mean_token_accuracy": 0.8786293864250183, "num_tokens": 416312829.0, "step": 10959 }, { "epoch": 1.3942246533519909, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.513710975646973, "learning_rate": 5e-05, "loss": 0.6884, "mean_token_accuracy": 0.8704137206077576, "num_tokens": 416353337.0, "step": 10960 }, { "epoch": 1.3943518636305814, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.850706100463867, "learning_rate": 5e-05, "loss": 0.8854, "mean_token_accuracy": 0.8661855459213257, "num_tokens": 416392425.0, "step": 10961 }, { "epoch": 1.394479073909172, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.259064674377441, "learning_rate": 5e-05, "loss": 0.7282, "mean_token_accuracy": 0.8595800995826721, "num_tokens": 416434641.0, "step": 10962 }, { "epoch": 1.3946062841877624, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.263666152954102, "learning_rate": 5e-05, "loss": 0.948, "mean_token_accuracy": 0.8690227270126343, "num_tokens": 416479107.0, "step": 10963 }, { "epoch": 1.3947334944663528, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.494544982910156, "learning_rate": 5e-05, "loss": 0.7819, "mean_token_accuracy": 0.8629934787750244, "num_tokens": 416512523.0, "step": 10964 }, { "epoch": 1.3948607047449433, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.147845268249512, "learning_rate": 5e-05, "loss": 0.9802, "mean_token_accuracy": 0.8689691424369812, "num_tokens": 416550176.0, "step": 10965 }, { "epoch": 1.3949879150235338, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.327880382537842, "learning_rate": 5e-05, "loss": 0.8658, "mean_token_accuracy": 0.8715417385101318, "num_tokens": 416583878.0, "step": 10966 }, { "epoch": 1.3951151253021243, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.167082786560059, "learning_rate": 5e-05, "loss": 0.8868, "mean_token_accuracy": 0.8691842555999756, "num_tokens": 416612863.0, "step": 10967 }, { "epoch": 1.3952423355807149, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.202652931213379, "learning_rate": 5e-05, "loss": 0.9748, "mean_token_accuracy": 0.8585704565048218, "num_tokens": 416650691.0, "step": 10968 }, { "epoch": 1.3953695458593054, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.972928524017334, "learning_rate": 5e-05, "loss": 0.731, "mean_token_accuracy": 0.876099705696106, "num_tokens": 416687227.0, "step": 10969 }, { "epoch": 1.395496756137896, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.811700820922852, "learning_rate": 5e-05, "loss": 0.9502, "mean_token_accuracy": 0.8639205694198608, "num_tokens": 416725712.0, "step": 10970 }, { "epoch": 1.3956239664164865, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.551816463470459, "learning_rate": 5e-05, "loss": 0.7197, "mean_token_accuracy": 0.878656804561615, "num_tokens": 416756169.0, "step": 10971 }, { "epoch": 1.395751176695077, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.955711364746094, "learning_rate": 5e-05, "loss": 1.0068, "mean_token_accuracy": 0.8509339690208435, "num_tokens": 416794351.0, "step": 10972 }, { "epoch": 1.3958783869736675, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.826013565063477, "learning_rate": 5e-05, "loss": 0.7302, "mean_token_accuracy": 0.8841533660888672, "num_tokens": 416829201.0, "step": 10973 }, { "epoch": 1.396005597252258, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.615117073059082, "learning_rate": 5e-05, "loss": 0.9035, "mean_token_accuracy": 0.8773753643035889, "num_tokens": 416862171.0, "step": 10974 }, { "epoch": 1.3961328075308486, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.319604396820068, "learning_rate": 5e-05, "loss": 0.8659, "mean_token_accuracy": 0.8610981702804565, "num_tokens": 416895923.0, "step": 10975 }, { "epoch": 1.396260017809439, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.083108901977539, "learning_rate": 5e-05, "loss": 0.8977, "mean_token_accuracy": 0.864587128162384, "num_tokens": 416927441.0, "step": 10976 }, { "epoch": 1.3963872280880296, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.582113742828369, "learning_rate": 5e-05, "loss": 0.8378, "mean_token_accuracy": 0.873711884021759, "num_tokens": 416962911.0, "step": 10977 }, { "epoch": 1.3965144383666201, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.786842346191406, "learning_rate": 5e-05, "loss": 0.848, "mean_token_accuracy": 0.8661117553710938, "num_tokens": 417001403.0, "step": 10978 }, { "epoch": 1.3966416486452105, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.4972310066223145, "learning_rate": 5e-05, "loss": 0.8317, "mean_token_accuracy": 0.8622092008590698, "num_tokens": 417038043.0, "step": 10979 }, { "epoch": 1.396768858923801, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.883806228637695, "learning_rate": 5e-05, "loss": 0.8262, "mean_token_accuracy": 0.8649888634681702, "num_tokens": 417073516.0, "step": 10980 }, { "epoch": 1.3968960692023915, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.116946220397949, "learning_rate": 5e-05, "loss": 0.7497, "mean_token_accuracy": 0.8687067031860352, "num_tokens": 417112334.0, "step": 10981 }, { "epoch": 1.397023279480982, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.249444961547852, "learning_rate": 5e-05, "loss": 0.837, "mean_token_accuracy": 0.8723833560943604, "num_tokens": 417155273.0, "step": 10982 }, { "epoch": 1.3971504897595726, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.737576007843018, "learning_rate": 5e-05, "loss": 0.6911, "mean_token_accuracy": 0.879631519317627, "num_tokens": 417190145.0, "step": 10983 }, { "epoch": 1.397277700038163, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.648946762084961, "learning_rate": 5e-05, "loss": 0.8224, "mean_token_accuracy": 0.8846055865287781, "num_tokens": 417228184.0, "step": 10984 }, { "epoch": 1.3974049103167536, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.520401477813721, "learning_rate": 5e-05, "loss": 0.6859, "mean_token_accuracy": 0.8798563480377197, "num_tokens": 417258590.0, "step": 10985 }, { "epoch": 1.3975321205953442, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.919610977172852, "learning_rate": 5e-05, "loss": 0.8883, "mean_token_accuracy": 0.8808482885360718, "num_tokens": 417295969.0, "step": 10986 }, { "epoch": 1.3976593308739347, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.656335830688477, "learning_rate": 5e-05, "loss": 0.7951, "mean_token_accuracy": 0.8543870449066162, "num_tokens": 417334428.0, "step": 10987 }, { "epoch": 1.397786541152525, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.842430114746094, "learning_rate": 5e-05, "loss": 0.9343, "mean_token_accuracy": 0.8737984895706177, "num_tokens": 417370074.0, "step": 10988 }, { "epoch": 1.3979137514311155, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.151443004608154, "learning_rate": 5e-05, "loss": 0.7976, "mean_token_accuracy": 0.8741805553436279, "num_tokens": 417413544.0, "step": 10989 }, { "epoch": 1.398040961709706, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.225499153137207, "learning_rate": 5e-05, "loss": 0.8998, "mean_token_accuracy": 0.8660982847213745, "num_tokens": 417445893.0, "step": 10990 }, { "epoch": 1.3981681719882966, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.760880470275879, "learning_rate": 5e-05, "loss": 0.8346, "mean_token_accuracy": 0.8823069334030151, "num_tokens": 417481469.0, "step": 10991 }, { "epoch": 1.398295382266887, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.383318901062012, "learning_rate": 5e-05, "loss": 0.8048, "mean_token_accuracy": 0.8628067970275879, "num_tokens": 417517070.0, "step": 10992 }, { "epoch": 1.3984225925454776, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.180405616760254, "learning_rate": 5e-05, "loss": 0.8663, "mean_token_accuracy": 0.8666306734085083, "num_tokens": 417558265.0, "step": 10993 }, { "epoch": 1.3985498028240682, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.931528091430664, "learning_rate": 5e-05, "loss": 0.7149, "mean_token_accuracy": 0.8752449154853821, "num_tokens": 417599083.0, "step": 10994 }, { "epoch": 1.3986770131026587, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.313725471496582, "learning_rate": 5e-05, "loss": 0.8417, "mean_token_accuracy": 0.8788788318634033, "num_tokens": 417632425.0, "step": 10995 }, { "epoch": 1.3988042233812492, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.860443592071533, "learning_rate": 5e-05, "loss": 0.6999, "mean_token_accuracy": 0.8784580230712891, "num_tokens": 417667209.0, "step": 10996 }, { "epoch": 1.3989314336598397, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.318204879760742, "learning_rate": 5e-05, "loss": 0.833, "mean_token_accuracy": 0.8752785921096802, "num_tokens": 417706168.0, "step": 10997 }, { "epoch": 1.3990586439384303, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.863891124725342, "learning_rate": 5e-05, "loss": 0.7045, "mean_token_accuracy": 0.8796099424362183, "num_tokens": 417745225.0, "step": 10998 }, { "epoch": 1.3991858542170208, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.376867294311523, "learning_rate": 5e-05, "loss": 0.8392, "mean_token_accuracy": 0.8764817714691162, "num_tokens": 417782688.0, "step": 10999 }, { "epoch": 1.3993130644956113, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.861959934234619, "learning_rate": 5e-05, "loss": 0.7356, "mean_token_accuracy": 0.8688421845436096, "num_tokens": 417814157.0, "step": 11000 }, { "epoch": 1.3994402747742019, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.386920928955078, "learning_rate": 5e-05, "loss": 0.8834, "mean_token_accuracy": 0.8648796081542969, "num_tokens": 417845460.0, "step": 11001 }, { "epoch": 1.3995674850527924, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.932936668395996, "learning_rate": 5e-05, "loss": 0.7437, "mean_token_accuracy": 0.8748371005058289, "num_tokens": 417881460.0, "step": 11002 }, { "epoch": 1.399694695331383, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.271377563476562, "learning_rate": 5e-05, "loss": 0.7991, "mean_token_accuracy": 0.8877288103103638, "num_tokens": 417919690.0, "step": 11003 }, { "epoch": 1.3998219056099732, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.100862503051758, "learning_rate": 5e-05, "loss": 0.7148, "mean_token_accuracy": 0.8873043656349182, "num_tokens": 417959474.0, "step": 11004 }, { "epoch": 1.3999491158885637, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.06303596496582, "learning_rate": 5e-05, "loss": 0.8345, "mean_token_accuracy": 0.8691941499710083, "num_tokens": 417992342.0, "step": 11005 }, { "epoch": 1.4000763261671543, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.247547626495361, "learning_rate": 5e-05, "loss": 0.832, "mean_token_accuracy": 0.8530367612838745, "num_tokens": 418032006.0, "step": 11006 }, { "epoch": 1.4002035364457448, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.861428737640381, "learning_rate": 5e-05, "loss": 0.859, "mean_token_accuracy": 0.8573598861694336, "num_tokens": 418069318.0, "step": 11007 }, { "epoch": 1.4003307467243353, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.31596565246582, "learning_rate": 5e-05, "loss": 0.7249, "mean_token_accuracy": 0.8858612775802612, "num_tokens": 418103605.0, "step": 11008 }, { "epoch": 1.4004579570029259, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.791635990142822, "learning_rate": 5e-05, "loss": 0.7986, "mean_token_accuracy": 0.8674808144569397, "num_tokens": 418136363.0, "step": 11009 }, { "epoch": 1.4005851672815164, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.240466594696045, "learning_rate": 5e-05, "loss": 0.7492, "mean_token_accuracy": 0.8698235750198364, "num_tokens": 418178096.0, "step": 11010 }, { "epoch": 1.400712377560107, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.87526798248291, "learning_rate": 5e-05, "loss": 0.7508, "mean_token_accuracy": 0.8826636075973511, "num_tokens": 418211025.0, "step": 11011 }, { "epoch": 1.4008395878386974, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.059300422668457, "learning_rate": 5e-05, "loss": 0.7222, "mean_token_accuracy": 0.8738958239555359, "num_tokens": 418251210.0, "step": 11012 }, { "epoch": 1.4009667981172877, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.093006134033203, "learning_rate": 5e-05, "loss": 0.7856, "mean_token_accuracy": 0.878322958946228, "num_tokens": 418286593.0, "step": 11013 }, { "epoch": 1.4010940083958783, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.81995153427124, "learning_rate": 5e-05, "loss": 0.6901, "mean_token_accuracy": 0.8818830251693726, "num_tokens": 418324794.0, "step": 11014 }, { "epoch": 1.4012212186744688, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.357427597045898, "learning_rate": 5e-05, "loss": 0.8116, "mean_token_accuracy": 0.8770513534545898, "num_tokens": 418366067.0, "step": 11015 }, { "epoch": 1.4013484289530593, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.626828670501709, "learning_rate": 5e-05, "loss": 0.6551, "mean_token_accuracy": 0.8881222009658813, "num_tokens": 418406600.0, "step": 11016 }, { "epoch": 1.4014756392316499, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.615705490112305, "learning_rate": 5e-05, "loss": 0.8455, "mean_token_accuracy": 0.8757719397544861, "num_tokens": 418440168.0, "step": 11017 }, { "epoch": 1.4016028495102404, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.5959649085998535, "learning_rate": 5e-05, "loss": 0.7646, "mean_token_accuracy": 0.8523463010787964, "num_tokens": 418482247.0, "step": 11018 }, { "epoch": 1.401730059788831, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.650134086608887, "learning_rate": 5e-05, "loss": 0.8536, "mean_token_accuracy": 0.8821359872817993, "num_tokens": 418517376.0, "step": 11019 }, { "epoch": 1.4018572700674214, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.898777961730957, "learning_rate": 5e-05, "loss": 0.7843, "mean_token_accuracy": 0.8661018013954163, "num_tokens": 418555841.0, "step": 11020 }, { "epoch": 1.401984480346012, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.188494682312012, "learning_rate": 5e-05, "loss": 0.8552, "mean_token_accuracy": 0.8730664253234863, "num_tokens": 418594232.0, "step": 11021 }, { "epoch": 1.4021116906246025, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.490896701812744, "learning_rate": 5e-05, "loss": 0.7824, "mean_token_accuracy": 0.8796728849411011, "num_tokens": 418631726.0, "step": 11022 }, { "epoch": 1.402238900903193, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.387455940246582, "learning_rate": 5e-05, "loss": 0.7077, "mean_token_accuracy": 0.8937853574752808, "num_tokens": 418671172.0, "step": 11023 }, { "epoch": 1.4023661111817836, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.139791488647461, "learning_rate": 5e-05, "loss": 0.8145, "mean_token_accuracy": 0.8779139518737793, "num_tokens": 418706541.0, "step": 11024 }, { "epoch": 1.402493321460374, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.6398444175720215, "learning_rate": 5e-05, "loss": 0.7, "mean_token_accuracy": 0.8780745267868042, "num_tokens": 418743557.0, "step": 11025 }, { "epoch": 1.4026205317389646, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.759913444519043, "learning_rate": 5e-05, "loss": 0.8782, "mean_token_accuracy": 0.8694584965705872, "num_tokens": 418783741.0, "step": 11026 }, { "epoch": 1.4027477420175551, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.171870708465576, "learning_rate": 5e-05, "loss": 0.6578, "mean_token_accuracy": 0.8791719675064087, "num_tokens": 418822208.0, "step": 11027 }, { "epoch": 1.4028749522961454, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.344802856445312, "learning_rate": 5e-05, "loss": 0.9391, "mean_token_accuracy": 0.8778032064437866, "num_tokens": 418862717.0, "step": 11028 }, { "epoch": 1.403002162574736, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.232000350952148, "learning_rate": 5e-05, "loss": 0.6994, "mean_token_accuracy": 0.877082347869873, "num_tokens": 418892961.0, "step": 11029 }, { "epoch": 1.4031293728533265, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.445164680480957, "learning_rate": 5e-05, "loss": 1.0332, "mean_token_accuracy": 0.8556423187255859, "num_tokens": 418933384.0, "step": 11030 }, { "epoch": 1.403256583131917, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.096649169921875, "learning_rate": 5e-05, "loss": 0.8461, "mean_token_accuracy": 0.8779549598693848, "num_tokens": 418971042.0, "step": 11031 }, { "epoch": 1.4033837934105076, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.60612964630127, "learning_rate": 5e-05, "loss": 0.9242, "mean_token_accuracy": 0.8769546747207642, "num_tokens": 419009949.0, "step": 11032 }, { "epoch": 1.403511003689098, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.099715232849121, "learning_rate": 5e-05, "loss": 0.9325, "mean_token_accuracy": 0.8772100210189819, "num_tokens": 419044607.0, "step": 11033 }, { "epoch": 1.4036382139676886, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.248793601989746, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8742051124572754, "num_tokens": 419081407.0, "step": 11034 }, { "epoch": 1.4037654242462791, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.787137031555176, "learning_rate": 5e-05, "loss": 0.9158, "mean_token_accuracy": 0.8820910453796387, "num_tokens": 419125007.0, "step": 11035 }, { "epoch": 1.4038926345248697, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.689797878265381, "learning_rate": 5e-05, "loss": 0.7759, "mean_token_accuracy": 0.8751180171966553, "num_tokens": 419166009.0, "step": 11036 }, { "epoch": 1.40401984480346, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.789541244506836, "learning_rate": 5e-05, "loss": 0.9315, "mean_token_accuracy": 0.8709889054298401, "num_tokens": 419194185.0, "step": 11037 }, { "epoch": 1.4041470550820505, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.088753700256348, "learning_rate": 5e-05, "loss": 0.7999, "mean_token_accuracy": 0.8779606819152832, "num_tokens": 419236742.0, "step": 11038 }, { "epoch": 1.404274265360641, "ewc_loss": 0.5, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.065347671508789, "learning_rate": 5e-05, "loss": 0.8523, "mean_token_accuracy": 0.8739007711410522, "num_tokens": 419272477.0, "step": 11039 }, { "epoch": 1.4044014756392316, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.705170631408691, "learning_rate": 5e-05, "loss": 0.8788, "mean_token_accuracy": 0.8650354146957397, "num_tokens": 419309198.0, "step": 11040 }, { "epoch": 1.404528685917822, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.347399711608887, "learning_rate": 5e-05, "loss": 0.7328, "mean_token_accuracy": 0.8812925815582275, "num_tokens": 419343065.0, "step": 11041 }, { "epoch": 1.4046558961964126, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.022014617919922, "learning_rate": 5e-05, "loss": 0.8781, "mean_token_accuracy": 0.862850546836853, "num_tokens": 419382826.0, "step": 11042 }, { "epoch": 1.4047831064750032, "ewc_loss": 0.375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.112127304077148, "learning_rate": 5e-05, "loss": 0.763, "mean_token_accuracy": 0.8641181588172913, "num_tokens": 419416839.0, "step": 11043 }, { "epoch": 1.4049103167535937, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.039942741394043, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.8762644529342651, "num_tokens": 419448041.0, "step": 11044 }, { "epoch": 1.4050375270321842, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.0916852951049805, "learning_rate": 5e-05, "loss": 0.8209, "mean_token_accuracy": 0.8489338159561157, "num_tokens": 419484917.0, "step": 11045 }, { "epoch": 1.4051647373107747, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.013041496276855, "learning_rate": 5e-05, "loss": 0.8056, "mean_token_accuracy": 0.8718260526657104, "num_tokens": 419521329.0, "step": 11046 }, { "epoch": 1.4052919475893653, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.057203769683838, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8708457946777344, "num_tokens": 419556141.0, "step": 11047 }, { "epoch": 1.4054191578679558, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.050248146057129, "learning_rate": 5e-05, "loss": 0.8031, "mean_token_accuracy": 0.8757705688476562, "num_tokens": 419593027.0, "step": 11048 }, { "epoch": 1.4055463681465463, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.950286388397217, "learning_rate": 5e-05, "loss": 0.7062, "mean_token_accuracy": 0.8823624849319458, "num_tokens": 419632419.0, "step": 11049 }, { "epoch": 1.4056735784251368, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.204943656921387, "learning_rate": 5e-05, "loss": 0.8377, "mean_token_accuracy": 0.868878960609436, "num_tokens": 419677625.0, "step": 11050 }, { "epoch": 1.4058007887037274, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.7458577156066895, "learning_rate": 5e-05, "loss": 0.713, "mean_token_accuracy": 0.8705804347991943, "num_tokens": 419721546.0, "step": 11051 }, { "epoch": 1.405927998982318, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.482346534729004, "learning_rate": 5e-05, "loss": 0.895, "mean_token_accuracy": 0.8548435568809509, "num_tokens": 419759345.0, "step": 11052 }, { "epoch": 1.4060552092609082, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.509916305541992, "learning_rate": 5e-05, "loss": 0.7295, "mean_token_accuracy": 0.8680676221847534, "num_tokens": 419789819.0, "step": 11053 }, { "epoch": 1.4061824195394987, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.815790176391602, "learning_rate": 5e-05, "loss": 0.9019, "mean_token_accuracy": 0.8689321279525757, "num_tokens": 419830525.0, "step": 11054 }, { "epoch": 1.4063096298180893, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.393897533416748, "learning_rate": 5e-05, "loss": 0.7174, "mean_token_accuracy": 0.871284008026123, "num_tokens": 419875592.0, "step": 11055 }, { "epoch": 1.4064368400966798, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.111778259277344, "learning_rate": 5e-05, "loss": 0.9026, "mean_token_accuracy": 0.8888371586799622, "num_tokens": 419913530.0, "step": 11056 }, { "epoch": 1.4065640503752703, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.603259563446045, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.8695526719093323, "num_tokens": 419955947.0, "step": 11057 }, { "epoch": 1.4066912606538609, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.001446723937988, "learning_rate": 5e-05, "loss": 0.9271, "mean_token_accuracy": 0.8832789063453674, "num_tokens": 419986931.0, "step": 11058 }, { "epoch": 1.4068184709324514, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.232189655303955, "learning_rate": 5e-05, "loss": 0.8627, "mean_token_accuracy": 0.8711799383163452, "num_tokens": 420024037.0, "step": 11059 }, { "epoch": 1.406945681211042, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.290635108947754, "learning_rate": 5e-05, "loss": 0.918, "mean_token_accuracy": 0.8671998381614685, "num_tokens": 420062843.0, "step": 11060 }, { "epoch": 1.4070728914896324, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.8791823387146, "learning_rate": 5e-05, "loss": 0.9476, "mean_token_accuracy": 0.8565700054168701, "num_tokens": 420102311.0, "step": 11061 }, { "epoch": 1.4072001017682227, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.351396560668945, "learning_rate": 5e-05, "loss": 0.8088, "mean_token_accuracy": 0.8652894496917725, "num_tokens": 420143364.0, "step": 11062 }, { "epoch": 1.4073273120468133, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.274849891662598, "learning_rate": 5e-05, "loss": 0.8684, "mean_token_accuracy": 0.8761544823646545, "num_tokens": 420180930.0, "step": 11063 }, { "epoch": 1.4074545223254038, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.953689098358154, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.872452437877655, "num_tokens": 420216048.0, "step": 11064 }, { "epoch": 1.4075817326039943, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.21595287322998, "learning_rate": 5e-05, "loss": 0.8761, "mean_token_accuracy": 0.871211051940918, "num_tokens": 420250792.0, "step": 11065 }, { "epoch": 1.4077089428825849, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.108859539031982, "learning_rate": 5e-05, "loss": 0.743, "mean_token_accuracy": 0.8776603937149048, "num_tokens": 420280518.0, "step": 11066 }, { "epoch": 1.4078361531611754, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.8573899269104, "learning_rate": 5e-05, "loss": 0.8575, "mean_token_accuracy": 0.8562382459640503, "num_tokens": 420318274.0, "step": 11067 }, { "epoch": 1.407963363439766, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.372710704803467, "learning_rate": 5e-05, "loss": 0.7537, "mean_token_accuracy": 0.8738813400268555, "num_tokens": 420353277.0, "step": 11068 }, { "epoch": 1.4080905737183564, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.576066493988037, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8768417835235596, "num_tokens": 420384457.0, "step": 11069 }, { "epoch": 1.408217783996947, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.491932392120361, "learning_rate": 5e-05, "loss": 0.7932, "mean_token_accuracy": 0.8613187074661255, "num_tokens": 420419229.0, "step": 11070 }, { "epoch": 1.4083449942755375, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.468947887420654, "learning_rate": 5e-05, "loss": 0.7738, "mean_token_accuracy": 0.8613172769546509, "num_tokens": 420457002.0, "step": 11071 }, { "epoch": 1.408472204554128, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.497306823730469, "learning_rate": 5e-05, "loss": 0.7221, "mean_token_accuracy": 0.8809609413146973, "num_tokens": 420490802.0, "step": 11072 }, { "epoch": 1.4085994148327186, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.393438339233398, "learning_rate": 5e-05, "loss": 0.7164, "mean_token_accuracy": 0.8801125288009644, "num_tokens": 420531782.0, "step": 11073 }, { "epoch": 1.408726625111309, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.544075012207031, "learning_rate": 5e-05, "loss": 0.7826, "mean_token_accuracy": 0.8623243570327759, "num_tokens": 420575577.0, "step": 11074 }, { "epoch": 1.4088538353898996, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.304286003112793, "learning_rate": 5e-05, "loss": 0.7235, "mean_token_accuracy": 0.8745815753936768, "num_tokens": 420612963.0, "step": 11075 }, { "epoch": 1.4089810456684901, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.619319438934326, "learning_rate": 5e-05, "loss": 0.7112, "mean_token_accuracy": 0.8802078366279602, "num_tokens": 420649518.0, "step": 11076 }, { "epoch": 1.4091082559470804, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.215324401855469, "learning_rate": 5e-05, "loss": 0.7466, "mean_token_accuracy": 0.8617864847183228, "num_tokens": 420684377.0, "step": 11077 }, { "epoch": 1.409235466225671, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.691095352172852, "learning_rate": 5e-05, "loss": 0.777, "mean_token_accuracy": 0.8636767864227295, "num_tokens": 420721566.0, "step": 11078 }, { "epoch": 1.4093626765042615, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.114079475402832, "learning_rate": 5e-05, "loss": 0.6983, "mean_token_accuracy": 0.8733088374137878, "num_tokens": 420758591.0, "step": 11079 }, { "epoch": 1.409489886782852, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.819577217102051, "learning_rate": 5e-05, "loss": 0.7764, "mean_token_accuracy": 0.8646138310432434, "num_tokens": 420796864.0, "step": 11080 }, { "epoch": 1.4096170970614426, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.917067050933838, "learning_rate": 5e-05, "loss": 0.6799, "mean_token_accuracy": 0.8805335760116577, "num_tokens": 420834689.0, "step": 11081 }, { "epoch": 1.409744307340033, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.112286567687988, "learning_rate": 5e-05, "loss": 0.7775, "mean_token_accuracy": 0.8724733591079712, "num_tokens": 420877795.0, "step": 11082 }, { "epoch": 1.4098715176186236, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.55199670791626, "learning_rate": 5e-05, "loss": 0.6698, "mean_token_accuracy": 0.8755970597267151, "num_tokens": 420917217.0, "step": 11083 }, { "epoch": 1.4099987278972141, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.679537773132324, "learning_rate": 5e-05, "loss": 0.8404, "mean_token_accuracy": 0.871375560760498, "num_tokens": 420952607.0, "step": 11084 }, { "epoch": 1.4101259381758047, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 5.966281414031982, "learning_rate": 5e-05, "loss": 0.678, "mean_token_accuracy": 0.8630390167236328, "num_tokens": 420986570.0, "step": 11085 }, { "epoch": 1.410253148454395, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.66218090057373, "learning_rate": 5e-05, "loss": 0.951, "mean_token_accuracy": 0.871696949005127, "num_tokens": 421024482.0, "step": 11086 }, { "epoch": 1.4103803587329855, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.565890312194824, "learning_rate": 5e-05, "loss": 0.6095, "mean_token_accuracy": 0.8927873969078064, "num_tokens": 421058185.0, "step": 11087 }, { "epoch": 1.410507569011576, "ewc_loss": 0.765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 10.829207420349121, "learning_rate": 5e-05, "loss": 1.144, "mean_token_accuracy": 0.8698877692222595, "num_tokens": 421089796.0, "step": 11088 }, { "epoch": 1.4106347792901666, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 6.9856133460998535, "learning_rate": 5e-05, "loss": 0.9488, "mean_token_accuracy": 0.8673750758171082, "num_tokens": 421132644.0, "step": 11089 }, { "epoch": 1.410761989568757, "ewc_loss": 0.765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 10.737689971923828, "learning_rate": 5e-05, "loss": 1.143, "mean_token_accuracy": 0.8696597814559937, "num_tokens": 421167790.0, "step": 11090 }, { "epoch": 1.4108891998473476, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000843048095703125, "grad_norm": 9.028401374816895, "learning_rate": 5e-05, "loss": 1.2908, "mean_token_accuracy": 0.8643785715103149, "num_tokens": 421206765.0, "step": 11091 }, { "epoch": 1.4110164101259381, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.651545524597168, "learning_rate": 5e-05, "loss": 0.8714, "mean_token_accuracy": 0.8791409134864807, "num_tokens": 421249997.0, "step": 11092 }, { "epoch": 1.4111436204045287, "ewc_loss": 0.875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 10.164223670959473, "learning_rate": 5e-05, "loss": 1.2339, "mean_token_accuracy": 0.8739039897918701, "num_tokens": 421287674.0, "step": 11093 }, { "epoch": 1.4112708306831192, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.155232906341553, "learning_rate": 5e-05, "loss": 0.813, "mean_token_accuracy": 0.8761633634567261, "num_tokens": 421325494.0, "step": 11094 }, { "epoch": 1.4113980409617097, "ewc_loss": 0.98828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000934600830078125, "grad_norm": 12.069045066833496, "learning_rate": 5e-05, "loss": 1.3254, "mean_token_accuracy": 0.8795992136001587, "num_tokens": 421359934.0, "step": 11095 }, { "epoch": 1.4115252512403003, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 9.298229217529297, "learning_rate": 5e-05, "loss": 1.3246, "mean_token_accuracy": 0.8566243648529053, "num_tokens": 421399195.0, "step": 11096 }, { "epoch": 1.4116524615188908, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 11.491616249084473, "learning_rate": 5e-05, "loss": 1.0743, "mean_token_accuracy": 0.8886391520500183, "num_tokens": 421428822.0, "step": 11097 }, { "epoch": 1.4117796717974813, "ewc_loss": 0.90234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00084686279296875, "grad_norm": 9.544717788696289, "learning_rate": 5e-05, "loss": 1.3016, "mean_token_accuracy": 0.8672968149185181, "num_tokens": 421468409.0, "step": 11098 }, { "epoch": 1.4119068820760718, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 11.181377410888672, "learning_rate": 5e-05, "loss": 1.2087, "mean_token_accuracy": 0.8513649702072144, "num_tokens": 421506905.0, "step": 11099 }, { "epoch": 1.4120340923546624, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.612503051757812, "learning_rate": 5e-05, "loss": 1.0702, "mean_token_accuracy": 0.8582048416137695, "num_tokens": 421545347.0, "step": 11100 }, { "epoch": 1.412161302633253, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00075531005859375, "grad_norm": 11.316557884216309, "learning_rate": 5e-05, "loss": 1.1857, "mean_token_accuracy": 0.8730566501617432, "num_tokens": 421580819.0, "step": 11101 }, { "epoch": 1.4122885129118432, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.309656143188477, "learning_rate": 5e-05, "loss": 0.951, "mean_token_accuracy": 0.8728681802749634, "num_tokens": 421628474.0, "step": 11102 }, { "epoch": 1.4124157231904337, "ewc_loss": 0.859375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 11.92917251586914, "learning_rate": 5e-05, "loss": 1.2442, "mean_token_accuracy": 0.8637098073959351, "num_tokens": 421670437.0, "step": 11103 }, { "epoch": 1.4125429334690243, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 9.350521087646484, "learning_rate": 5e-05, "loss": 1.1489, "mean_token_accuracy": 0.8542026281356812, "num_tokens": 421709157.0, "step": 11104 }, { "epoch": 1.4126701437476148, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 11.725894927978516, "learning_rate": 5e-05, "loss": 1.1646, "mean_token_accuracy": 0.8767661452293396, "num_tokens": 421743783.0, "step": 11105 }, { "epoch": 1.4127973540262053, "ewc_loss": 0.765625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 9.979381561279297, "learning_rate": 5e-05, "loss": 1.09, "mean_token_accuracy": 0.8851252794265747, "num_tokens": 421781376.0, "step": 11106 }, { "epoch": 1.4129245643047958, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 10.96308708190918, "learning_rate": 5e-05, "loss": 1.0821, "mean_token_accuracy": 0.8812124729156494, "num_tokens": 421819367.0, "step": 11107 }, { "epoch": 1.4130517745833864, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.633249282836914, "learning_rate": 5e-05, "loss": 1.0052, "mean_token_accuracy": 0.8826386332511902, "num_tokens": 421862830.0, "step": 11108 }, { "epoch": 1.413178984861977, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.37463665008545, "learning_rate": 5e-05, "loss": 1.0273, "mean_token_accuracy": 0.8805297017097473, "num_tokens": 421898353.0, "step": 11109 }, { "epoch": 1.4133061951405674, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.104297637939453, "learning_rate": 5e-05, "loss": 0.9555, "mean_token_accuracy": 0.870063841342926, "num_tokens": 421932911.0, "step": 11110 }, { "epoch": 1.4134334054191577, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 10.016639709472656, "learning_rate": 5e-05, "loss": 1.0163, "mean_token_accuracy": 0.8742181658744812, "num_tokens": 421974409.0, "step": 11111 }, { "epoch": 1.4135606156977483, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.523947715759277, "learning_rate": 5e-05, "loss": 0.9532, "mean_token_accuracy": 0.8535845279693604, "num_tokens": 422011913.0, "step": 11112 }, { "epoch": 1.4136878259763388, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 10.093600273132324, "learning_rate": 5e-05, "loss": 1.0248, "mean_token_accuracy": 0.8731776475906372, "num_tokens": 422054487.0, "step": 11113 }, { "epoch": 1.4138150362549293, "ewc_loss": 0.5, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.19726276397705, "learning_rate": 5e-05, "loss": 0.874, "mean_token_accuracy": 0.8734222650527954, "num_tokens": 422092980.0, "step": 11114 }, { "epoch": 1.4139422465335199, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.422517776489258, "learning_rate": 5e-05, "loss": 1.088, "mean_token_accuracy": 0.8626964688301086, "num_tokens": 422139368.0, "step": 11115 }, { "epoch": 1.4140694568121104, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.439818382263184, "learning_rate": 5e-05, "loss": 0.8568, "mean_token_accuracy": 0.8854889869689941, "num_tokens": 422177342.0, "step": 11116 }, { "epoch": 1.414196667090701, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.44040298461914, "learning_rate": 5e-05, "loss": 1.0581, "mean_token_accuracy": 0.8768674731254578, "num_tokens": 422221378.0, "step": 11117 }, { "epoch": 1.4143238773692914, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.936697006225586, "learning_rate": 5e-05, "loss": 0.9528, "mean_token_accuracy": 0.8714419007301331, "num_tokens": 422260503.0, "step": 11118 }, { "epoch": 1.414451087647882, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.78706169128418, "learning_rate": 5e-05, "loss": 1.0171, "mean_token_accuracy": 0.8665684461593628, "num_tokens": 422301218.0, "step": 11119 }, { "epoch": 1.4145782979264725, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.853495597839355, "learning_rate": 5e-05, "loss": 0.8999, "mean_token_accuracy": 0.8787571787834167, "num_tokens": 422337366.0, "step": 11120 }, { "epoch": 1.414705508205063, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.122647285461426, "learning_rate": 5e-05, "loss": 0.8949, "mean_token_accuracy": 0.8848189115524292, "num_tokens": 422376087.0, "step": 11121 }, { "epoch": 1.4148327184836536, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.3429594039917, "learning_rate": 5e-05, "loss": 0.845, "mean_token_accuracy": 0.8719214200973511, "num_tokens": 422413871.0, "step": 11122 }, { "epoch": 1.414959928762244, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.884613037109375, "learning_rate": 5e-05, "loss": 0.9012, "mean_token_accuracy": 0.8751896619796753, "num_tokens": 422450834.0, "step": 11123 }, { "epoch": 1.4150871390408346, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.733854293823242, "learning_rate": 5e-05, "loss": 0.7437, "mean_token_accuracy": 0.8821789622306824, "num_tokens": 422494566.0, "step": 11124 }, { "epoch": 1.4152143493194251, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.111676216125488, "learning_rate": 5e-05, "loss": 0.951, "mean_token_accuracy": 0.8652886152267456, "num_tokens": 422530107.0, "step": 11125 }, { "epoch": 1.4153415595980154, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.4054999351501465, "learning_rate": 5e-05, "loss": 0.8128, "mean_token_accuracy": 0.8533788919448853, "num_tokens": 422567598.0, "step": 11126 }, { "epoch": 1.415468769876606, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.443532943725586, "learning_rate": 5e-05, "loss": 0.9473, "mean_token_accuracy": 0.8728393316268921, "num_tokens": 422611705.0, "step": 11127 }, { "epoch": 1.4155959801551965, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.60015344619751, "learning_rate": 5e-05, "loss": 0.8015, "mean_token_accuracy": 0.8649806976318359, "num_tokens": 422648082.0, "step": 11128 }, { "epoch": 1.415723190433787, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.458922386169434, "learning_rate": 5e-05, "loss": 0.9673, "mean_token_accuracy": 0.8731290698051453, "num_tokens": 422691974.0, "step": 11129 }, { "epoch": 1.4158504007123776, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.085845947265625, "learning_rate": 5e-05, "loss": 0.8489, "mean_token_accuracy": 0.8649390935897827, "num_tokens": 422728245.0, "step": 11130 }, { "epoch": 1.415977610990968, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.847381591796875, "learning_rate": 5e-05, "loss": 0.8963, "mean_token_accuracy": 0.8759967684745789, "num_tokens": 422760920.0, "step": 11131 }, { "epoch": 1.4161048212695586, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.24765396118164, "learning_rate": 5e-05, "loss": 0.831, "mean_token_accuracy": 0.8710825443267822, "num_tokens": 422796188.0, "step": 11132 }, { "epoch": 1.4162320315481491, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.069089889526367, "learning_rate": 5e-05, "loss": 0.7916, "mean_token_accuracy": 0.883844792842865, "num_tokens": 422836232.0, "step": 11133 }, { "epoch": 1.4163592418267397, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.017217636108398, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.8782258629798889, "num_tokens": 422872855.0, "step": 11134 }, { "epoch": 1.41648645210533, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.766438961029053, "learning_rate": 5e-05, "loss": 0.7933, "mean_token_accuracy": 0.8692238330841064, "num_tokens": 422912826.0, "step": 11135 }, { "epoch": 1.4166136623839205, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.733191967010498, "learning_rate": 5e-05, "loss": 0.7952, "mean_token_accuracy": 0.8621565699577332, "num_tokens": 422950034.0, "step": 11136 }, { "epoch": 1.416740872662511, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.6981916427612305, "learning_rate": 5e-05, "loss": 0.7371, "mean_token_accuracy": 0.8773250579833984, "num_tokens": 422987764.0, "step": 11137 }, { "epoch": 1.4168680829411016, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.612124919891357, "learning_rate": 5e-05, "loss": 0.6893, "mean_token_accuracy": 0.8888116478919983, "num_tokens": 423024274.0, "step": 11138 }, { "epoch": 1.416995293219692, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.565598011016846, "learning_rate": 5e-05, "loss": 0.7713, "mean_token_accuracy": 0.8612164258956909, "num_tokens": 423058630.0, "step": 11139 }, { "epoch": 1.4171225034982826, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.641343116760254, "learning_rate": 5e-05, "loss": 0.7163, "mean_token_accuracy": 0.8811125755310059, "num_tokens": 423099829.0, "step": 11140 }, { "epoch": 1.4172497137768731, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.441434860229492, "learning_rate": 5e-05, "loss": 0.715, "mean_token_accuracy": 0.873468816280365, "num_tokens": 423130369.0, "step": 11141 }, { "epoch": 1.4173769240554637, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.667881488800049, "learning_rate": 5e-05, "loss": 0.723, "mean_token_accuracy": 0.8737637400627136, "num_tokens": 423166437.0, "step": 11142 }, { "epoch": 1.4175041343340542, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.36010217666626, "learning_rate": 5e-05, "loss": 0.6761, "mean_token_accuracy": 0.8851833343505859, "num_tokens": 423202220.0, "step": 11143 }, { "epoch": 1.4176313446126447, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.689019680023193, "learning_rate": 5e-05, "loss": 0.7438, "mean_token_accuracy": 0.8705142736434937, "num_tokens": 423243417.0, "step": 11144 }, { "epoch": 1.4177585548912353, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.2854905128479, "learning_rate": 5e-05, "loss": 0.7179, "mean_token_accuracy": 0.8688738346099854, "num_tokens": 423282337.0, "step": 11145 }, { "epoch": 1.4178857651698258, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.7456440925598145, "learning_rate": 5e-05, "loss": 0.7603, "mean_token_accuracy": 0.8674609661102295, "num_tokens": 423318666.0, "step": 11146 }, { "epoch": 1.4180129754484163, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.1994524002075195, "learning_rate": 5e-05, "loss": 0.7264, "mean_token_accuracy": 0.8653481006622314, "num_tokens": 423357895.0, "step": 11147 }, { "epoch": 1.4181401857270068, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.827844619750977, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.8729084730148315, "num_tokens": 423395330.0, "step": 11148 }, { "epoch": 1.4182673960055974, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.093493938446045, "learning_rate": 5e-05, "loss": 0.7018, "mean_token_accuracy": 0.8733595609664917, "num_tokens": 423436824.0, "step": 11149 }, { "epoch": 1.418394606284188, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.956681728363037, "learning_rate": 5e-05, "loss": 0.77, "mean_token_accuracy": 0.8700755834579468, "num_tokens": 423471906.0, "step": 11150 }, { "epoch": 1.4185218165627782, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.935663223266602, "learning_rate": 5e-05, "loss": 0.6852, "mean_token_accuracy": 0.8729188442230225, "num_tokens": 423510774.0, "step": 11151 }, { "epoch": 1.4186490268413687, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.170638084411621, "learning_rate": 5e-05, "loss": 0.777, "mean_token_accuracy": 0.8724711537361145, "num_tokens": 423549825.0, "step": 11152 }, { "epoch": 1.4187762371199593, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.731600284576416, "learning_rate": 5e-05, "loss": 0.6557, "mean_token_accuracy": 0.8781298995018005, "num_tokens": 423577546.0, "step": 11153 }, { "epoch": 1.4189034473985498, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.471602439880371, "learning_rate": 5e-05, "loss": 0.7647, "mean_token_accuracy": 0.883757472038269, "num_tokens": 423613167.0, "step": 11154 }, { "epoch": 1.4190306576771403, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.498429298400879, "learning_rate": 5e-05, "loss": 0.6891, "mean_token_accuracy": 0.861324667930603, "num_tokens": 423649601.0, "step": 11155 }, { "epoch": 1.4191578679557308, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.815834999084473, "learning_rate": 5e-05, "loss": 0.8695, "mean_token_accuracy": 0.8618428111076355, "num_tokens": 423685429.0, "step": 11156 }, { "epoch": 1.4192850782343214, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.409667491912842, "learning_rate": 5e-05, "loss": 0.6638, "mean_token_accuracy": 0.8804858326911926, "num_tokens": 423724270.0, "step": 11157 }, { "epoch": 1.419412288512912, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.941567420959473, "learning_rate": 5e-05, "loss": 0.8459, "mean_token_accuracy": 0.8833969831466675, "num_tokens": 423760935.0, "step": 11158 }, { "epoch": 1.4195394987915024, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.790966987609863, "learning_rate": 5e-05, "loss": 0.7156, "mean_token_accuracy": 0.8708964586257935, "num_tokens": 423801949.0, "step": 11159 }, { "epoch": 1.4196667090700927, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.471339225769043, "learning_rate": 5e-05, "loss": 0.8092, "mean_token_accuracy": 0.8821653723716736, "num_tokens": 423840532.0, "step": 11160 }, { "epoch": 1.4197939193486833, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.528984069824219, "learning_rate": 5e-05, "loss": 0.7906, "mean_token_accuracy": 0.8621393442153931, "num_tokens": 423877912.0, "step": 11161 }, { "epoch": 1.4199211296272738, "ewc_loss": 0.375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.4384284019470215, "learning_rate": 5e-05, "loss": 0.722, "mean_token_accuracy": 0.8785715699195862, "num_tokens": 423917836.0, "step": 11162 }, { "epoch": 1.4200483399058643, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.318100929260254, "learning_rate": 5e-05, "loss": 0.8259, "mean_token_accuracy": 0.865216076374054, "num_tokens": 423957307.0, "step": 11163 }, { "epoch": 1.4201755501844548, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.5752153396606445, "learning_rate": 5e-05, "loss": 0.691, "mean_token_accuracy": 0.8706601858139038, "num_tokens": 423994280.0, "step": 11164 }, { "epoch": 1.4203027604630454, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.959373474121094, "learning_rate": 5e-05, "loss": 0.8802, "mean_token_accuracy": 0.8695966601371765, "num_tokens": 424033682.0, "step": 11165 }, { "epoch": 1.420429970741636, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.240325927734375, "learning_rate": 5e-05, "loss": 0.6926, "mean_token_accuracy": 0.8650140762329102, "num_tokens": 424072432.0, "step": 11166 }, { "epoch": 1.4205571810202264, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.318953514099121, "learning_rate": 5e-05, "loss": 0.9603, "mean_token_accuracy": 0.8616333603858948, "num_tokens": 424110074.0, "step": 11167 }, { "epoch": 1.420684391298817, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.620430946350098, "learning_rate": 5e-05, "loss": 0.7353, "mean_token_accuracy": 0.8654310703277588, "num_tokens": 424148263.0, "step": 11168 }, { "epoch": 1.4208116015774075, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.955832481384277, "learning_rate": 5e-05, "loss": 0.9671, "mean_token_accuracy": 0.8611412048339844, "num_tokens": 424190342.0, "step": 11169 }, { "epoch": 1.420938811855998, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.53975248336792, "learning_rate": 5e-05, "loss": 0.8541, "mean_token_accuracy": 0.8626399040222168, "num_tokens": 424236136.0, "step": 11170 }, { "epoch": 1.4210660221345885, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.847794532775879, "learning_rate": 5e-05, "loss": 0.8511, "mean_token_accuracy": 0.8589269518852234, "num_tokens": 424274436.0, "step": 11171 }, { "epoch": 1.421193232413179, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.259801864624023, "learning_rate": 5e-05, "loss": 0.8101, "mean_token_accuracy": 0.8882565498352051, "num_tokens": 424309661.0, "step": 11172 }, { "epoch": 1.4213204426917696, "ewc_loss": 0.375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.093507289886475, "learning_rate": 5e-05, "loss": 0.7751, "mean_token_accuracy": 0.8622617721557617, "num_tokens": 424347127.0, "step": 11173 }, { "epoch": 1.4214476529703601, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.349672317504883, "learning_rate": 5e-05, "loss": 0.8138, "mean_token_accuracy": 0.8792766332626343, "num_tokens": 424391287.0, "step": 11174 }, { "epoch": 1.4215748632489504, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.260462284088135, "learning_rate": 5e-05, "loss": 0.711, "mean_token_accuracy": 0.8814425468444824, "num_tokens": 424428091.0, "step": 11175 }, { "epoch": 1.421702073527541, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.69169282913208, "learning_rate": 5e-05, "loss": 0.7769, "mean_token_accuracy": 0.8705629110336304, "num_tokens": 424469270.0, "step": 11176 }, { "epoch": 1.4218292838061315, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.8998799324035645, "learning_rate": 5e-05, "loss": 0.8288, "mean_token_accuracy": 0.8581505417823792, "num_tokens": 424509116.0, "step": 11177 }, { "epoch": 1.421956494084722, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.940704345703125, "learning_rate": 5e-05, "loss": 0.7064, "mean_token_accuracy": 0.8741999864578247, "num_tokens": 424542777.0, "step": 11178 }, { "epoch": 1.4220837043633126, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.476563453674316, "learning_rate": 5e-05, "loss": 0.8286, "mean_token_accuracy": 0.8755345344543457, "num_tokens": 424582299.0, "step": 11179 }, { "epoch": 1.422210914641903, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.490671157836914, "learning_rate": 5e-05, "loss": 0.646, "mean_token_accuracy": 0.8873571157455444, "num_tokens": 424617543.0, "step": 11180 }, { "epoch": 1.4223381249204936, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.990076065063477, "learning_rate": 5e-05, "loss": 0.8471, "mean_token_accuracy": 0.8824930787086487, "num_tokens": 424657512.0, "step": 11181 }, { "epoch": 1.4224653351990841, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.363037109375, "learning_rate": 5e-05, "loss": 0.7321, "mean_token_accuracy": 0.8622132539749146, "num_tokens": 424695954.0, "step": 11182 }, { "epoch": 1.4225925454776747, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.264311790466309, "learning_rate": 5e-05, "loss": 0.9889, "mean_token_accuracy": 0.8581951856613159, "num_tokens": 424736174.0, "step": 11183 }, { "epoch": 1.422719755756265, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.767881870269775, "learning_rate": 5e-05, "loss": 0.7574, "mean_token_accuracy": 0.8708040714263916, "num_tokens": 424776371.0, "step": 11184 }, { "epoch": 1.4228469660348555, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.019771575927734, "learning_rate": 5e-05, "loss": 0.981, "mean_token_accuracy": 0.8638272881507874, "num_tokens": 424816874.0, "step": 11185 }, { "epoch": 1.422974176313446, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.37202787399292, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8875254988670349, "num_tokens": 424847924.0, "step": 11186 }, { "epoch": 1.4231013865920366, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.459692001342773, "learning_rate": 5e-05, "loss": 0.8692, "mean_token_accuracy": 0.8873423337936401, "num_tokens": 424887950.0, "step": 11187 }, { "epoch": 1.423228596870627, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.5401740074157715, "learning_rate": 5e-05, "loss": 0.842, "mean_token_accuracy": 0.8634894490242004, "num_tokens": 424922608.0, "step": 11188 }, { "epoch": 1.4233558071492176, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.36683464050293, "learning_rate": 5e-05, "loss": 0.8725, "mean_token_accuracy": 0.8797339200973511, "num_tokens": 424958564.0, "step": 11189 }, { "epoch": 1.4234830174278081, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.1129608154296875, "learning_rate": 5e-05, "loss": 0.7681, "mean_token_accuracy": 0.8706350326538086, "num_tokens": 424993351.0, "step": 11190 }, { "epoch": 1.4236102277063987, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.879365921020508, "learning_rate": 5e-05, "loss": 0.9454, "mean_token_accuracy": 0.8727479577064514, "num_tokens": 425028018.0, "step": 11191 }, { "epoch": 1.4237374379849892, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.818069934844971, "learning_rate": 5e-05, "loss": 0.7547, "mean_token_accuracy": 0.875723123550415, "num_tokens": 425065023.0, "step": 11192 }, { "epoch": 1.4238646482635797, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.156051635742188, "learning_rate": 5e-05, "loss": 0.9511, "mean_token_accuracy": 0.8772298097610474, "num_tokens": 425101152.0, "step": 11193 }, { "epoch": 1.4239918585421703, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.330941200256348, "learning_rate": 5e-05, "loss": 0.8607, "mean_token_accuracy": 0.8662523031234741, "num_tokens": 425137756.0, "step": 11194 }, { "epoch": 1.4241190688207608, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.467854499816895, "learning_rate": 5e-05, "loss": 0.9066, "mean_token_accuracy": 0.876843273639679, "num_tokens": 425171947.0, "step": 11195 }, { "epoch": 1.4242462790993513, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.033012390136719, "learning_rate": 5e-05, "loss": 0.8886, "mean_token_accuracy": 0.8709368705749512, "num_tokens": 425207219.0, "step": 11196 }, { "epoch": 1.4243734893779418, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.42819356918335, "learning_rate": 5e-05, "loss": 0.8131, "mean_token_accuracy": 0.8647755980491638, "num_tokens": 425242515.0, "step": 11197 }, { "epoch": 1.4245006996565324, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.350484848022461, "learning_rate": 5e-05, "loss": 0.9197, "mean_token_accuracy": 0.8571298718452454, "num_tokens": 425283409.0, "step": 11198 }, { "epoch": 1.424627909935123, "ewc_loss": 0.375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.047986030578613, "learning_rate": 5e-05, "loss": 0.7634, "mean_token_accuracy": 0.8636990785598755, "num_tokens": 425321862.0, "step": 11199 }, { "epoch": 1.4247551202137132, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.175345420837402, "learning_rate": 5e-05, "loss": 0.8171, "mean_token_accuracy": 0.8813226222991943, "num_tokens": 425362034.0, "step": 11200 }, { "epoch": 1.4248823304923037, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.2575364112854, "learning_rate": 5e-05, "loss": 0.8033, "mean_token_accuracy": 0.85665363073349, "num_tokens": 425397900.0, "step": 11201 }, { "epoch": 1.4250095407708943, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.805137634277344, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8724291920661926, "num_tokens": 425435458.0, "step": 11202 }, { "epoch": 1.4251367510494848, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.426812171936035, "learning_rate": 5e-05, "loss": 0.7175, "mean_token_accuracy": 0.8821629285812378, "num_tokens": 425465687.0, "step": 11203 }, { "epoch": 1.4252639613280753, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.734019756317139, "learning_rate": 5e-05, "loss": 0.7715, "mean_token_accuracy": 0.8696932196617126, "num_tokens": 425500702.0, "step": 11204 }, { "epoch": 1.4253911716066658, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.2417097091674805, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.873900294303894, "num_tokens": 425542827.0, "step": 11205 }, { "epoch": 1.4255183818852564, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.007853507995605, "learning_rate": 5e-05, "loss": 0.8186, "mean_token_accuracy": 0.8641317486763, "num_tokens": 425585146.0, "step": 11206 }, { "epoch": 1.425645592163847, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.913078308105469, "learning_rate": 5e-05, "loss": 0.6918, "mean_token_accuracy": 0.8797816038131714, "num_tokens": 425623790.0, "step": 11207 }, { "epoch": 1.4257728024424374, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.352069854736328, "learning_rate": 5e-05, "loss": 0.8221, "mean_token_accuracy": 0.8722566366195679, "num_tokens": 425663290.0, "step": 11208 }, { "epoch": 1.4259000127210277, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.672171115875244, "learning_rate": 5e-05, "loss": 0.6895, "mean_token_accuracy": 0.875385046005249, "num_tokens": 425701226.0, "step": 11209 }, { "epoch": 1.4260272229996183, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.637491226196289, "learning_rate": 5e-05, "loss": 0.8466, "mean_token_accuracy": 0.8735440969467163, "num_tokens": 425737524.0, "step": 11210 }, { "epoch": 1.4261544332782088, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.583059787750244, "learning_rate": 5e-05, "loss": 0.6922, "mean_token_accuracy": 0.8800750970840454, "num_tokens": 425771010.0, "step": 11211 }, { "epoch": 1.4262816435567993, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.755212783813477, "learning_rate": 5e-05, "loss": 0.9142, "mean_token_accuracy": 0.8641586303710938, "num_tokens": 425813077.0, "step": 11212 }, { "epoch": 1.4264088538353898, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.782081604003906, "learning_rate": 5e-05, "loss": 0.6891, "mean_token_accuracy": 0.8854634761810303, "num_tokens": 425852718.0, "step": 11213 }, { "epoch": 1.4265360641139804, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.542742729187012, "learning_rate": 5e-05, "loss": 0.887, "mean_token_accuracy": 0.8725194931030273, "num_tokens": 425888196.0, "step": 11214 }, { "epoch": 1.426663274392571, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.211905002593994, "learning_rate": 5e-05, "loss": 0.7615, "mean_token_accuracy": 0.8784574866294861, "num_tokens": 425926774.0, "step": 11215 }, { "epoch": 1.4267904846711614, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.982125759124756, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8874586820602417, "num_tokens": 425966207.0, "step": 11216 }, { "epoch": 1.426917694949752, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.637393951416016, "learning_rate": 5e-05, "loss": 0.8344, "mean_token_accuracy": 0.8623760938644409, "num_tokens": 426008792.0, "step": 11217 }, { "epoch": 1.4270449052283425, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.460618019104004, "learning_rate": 5e-05, "loss": 0.7787, "mean_token_accuracy": 0.8712248802185059, "num_tokens": 426049650.0, "step": 11218 }, { "epoch": 1.427172115506933, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.865499973297119, "learning_rate": 5e-05, "loss": 0.7696, "mean_token_accuracy": 0.8816856145858765, "num_tokens": 426092143.0, "step": 11219 }, { "epoch": 1.4272993257855235, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.232156276702881, "learning_rate": 5e-05, "loss": 0.7831, "mean_token_accuracy": 0.8598089814186096, "num_tokens": 426135393.0, "step": 11220 }, { "epoch": 1.427426536064114, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.857309341430664, "learning_rate": 5e-05, "loss": 0.7772, "mean_token_accuracy": 0.871212363243103, "num_tokens": 426171858.0, "step": 11221 }, { "epoch": 1.4275537463427046, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.261667728424072, "learning_rate": 5e-05, "loss": 0.6984, "mean_token_accuracy": 0.8819067478179932, "num_tokens": 426206635.0, "step": 11222 }, { "epoch": 1.4276809566212951, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.724808692932129, "learning_rate": 5e-05, "loss": 0.7495, "mean_token_accuracy": 0.874813437461853, "num_tokens": 426240942.0, "step": 11223 }, { "epoch": 1.4278081668998854, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.341076374053955, "learning_rate": 5e-05, "loss": 0.6667, "mean_token_accuracy": 0.8930946588516235, "num_tokens": 426279592.0, "step": 11224 }, { "epoch": 1.427935377178476, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.586843967437744, "learning_rate": 5e-05, "loss": 0.6961, "mean_token_accuracy": 0.8854327201843262, "num_tokens": 426314218.0, "step": 11225 }, { "epoch": 1.4280625874570665, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.429689407348633, "learning_rate": 5e-05, "loss": 0.7244, "mean_token_accuracy": 0.8721829652786255, "num_tokens": 426351514.0, "step": 11226 }, { "epoch": 1.428189797735657, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.470174312591553, "learning_rate": 5e-05, "loss": 0.7868, "mean_token_accuracy": 0.8605908155441284, "num_tokens": 426389553.0, "step": 11227 }, { "epoch": 1.4283170080142475, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.537398338317871, "learning_rate": 5e-05, "loss": 0.7095, "mean_token_accuracy": 0.8800774812698364, "num_tokens": 426424026.0, "step": 11228 }, { "epoch": 1.428444218292838, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.331089019775391, "learning_rate": 5e-05, "loss": 0.7087, "mean_token_accuracy": 0.8719242811203003, "num_tokens": 426466711.0, "step": 11229 }, { "epoch": 1.4285714285714286, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.663431644439697, "learning_rate": 5e-05, "loss": 0.7576, "mean_token_accuracy": 0.8687930107116699, "num_tokens": 426507636.0, "step": 11230 }, { "epoch": 1.4286986388500191, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.144882678985596, "learning_rate": 5e-05, "loss": 0.7109, "mean_token_accuracy": 0.8692953586578369, "num_tokens": 426551811.0, "step": 11231 }, { "epoch": 1.4288258491286097, "ewc_loss": 0.375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.890312671661377, "learning_rate": 5e-05, "loss": 0.7429, "mean_token_accuracy": 0.8719378709793091, "num_tokens": 426585814.0, "step": 11232 }, { "epoch": 1.4289530594072, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.870964527130127, "learning_rate": 5e-05, "loss": 0.6628, "mean_token_accuracy": 0.8788334727287292, "num_tokens": 426619327.0, "step": 11233 }, { "epoch": 1.4290802696857905, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.279853820800781, "learning_rate": 5e-05, "loss": 0.7939, "mean_token_accuracy": 0.8643027544021606, "num_tokens": 426656976.0, "step": 11234 }, { "epoch": 1.429207479964381, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.418477535247803, "learning_rate": 5e-05, "loss": 0.6449, "mean_token_accuracy": 0.8793058395385742, "num_tokens": 426693930.0, "step": 11235 }, { "epoch": 1.4293346902429716, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.954195976257324, "learning_rate": 5e-05, "loss": 0.8, "mean_token_accuracy": 0.8893535137176514, "num_tokens": 426731655.0, "step": 11236 }, { "epoch": 1.429461900521562, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 5.816764831542969, "learning_rate": 5e-05, "loss": 0.6645, "mean_token_accuracy": 0.8682646751403809, "num_tokens": 426768384.0, "step": 11237 }, { "epoch": 1.4295891108001526, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 10.013375282287598, "learning_rate": 5e-05, "loss": 0.998, "mean_token_accuracy": 0.8692982792854309, "num_tokens": 426806319.0, "step": 11238 }, { "epoch": 1.4297163210787431, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 5.75058126449585, "learning_rate": 5e-05, "loss": 0.7675, "mean_token_accuracy": 0.848618745803833, "num_tokens": 426841076.0, "step": 11239 }, { "epoch": 1.4298435313573337, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 10.8823881149292, "learning_rate": 5e-05, "loss": 1.1246, "mean_token_accuracy": 0.8778812885284424, "num_tokens": 426877647.0, "step": 11240 }, { "epoch": 1.4299707416359242, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 7.662290573120117, "learning_rate": 5e-05, "loss": 0.9852, "mean_token_accuracy": 0.8744335174560547, "num_tokens": 426917975.0, "step": 11241 }, { "epoch": 1.4300979519145147, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.920445442199707, "learning_rate": 5e-05, "loss": 1.0266, "mean_token_accuracy": 0.8717032074928284, "num_tokens": 426958975.0, "step": 11242 }, { "epoch": 1.4302251621931052, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00075531005859375, "grad_norm": 9.242949485778809, "learning_rate": 5e-05, "loss": 1.1816, "mean_token_accuracy": 0.8717747926712036, "num_tokens": 426997205.0, "step": 11243 }, { "epoch": 1.4303523724716958, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.527487754821777, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.8718570470809937, "num_tokens": 427037455.0, "step": 11244 }, { "epoch": 1.4304795827502863, "ewc_loss": 0.890625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000835418701171875, "grad_norm": 11.104080200195312, "learning_rate": 5e-05, "loss": 1.2784, "mean_token_accuracy": 0.8679627180099487, "num_tokens": 427071443.0, "step": 11245 }, { "epoch": 1.4306067930288768, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 7.351062297821045, "learning_rate": 5e-05, "loss": 0.8966, "mean_token_accuracy": 0.8758773803710938, "num_tokens": 427113258.0, "step": 11246 }, { "epoch": 1.4307340033074674, "ewc_loss": 0.984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0009307861328125, "grad_norm": 12.695873260498047, "learning_rate": 5e-05, "loss": 1.329, "mean_token_accuracy": 0.8759196996688843, "num_tokens": 427153523.0, "step": 11247 }, { "epoch": 1.430861213586058, "ewc_loss": 1.015625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0009613037109375, "grad_norm": 10.229888916015625, "learning_rate": 5e-05, "loss": 1.379, "mean_token_accuracy": 0.8723182082176208, "num_tokens": 427186818.0, "step": 11248 }, { "epoch": 1.4309884238646482, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 11.590928077697754, "learning_rate": 5e-05, "loss": 1.1776, "mean_token_accuracy": 0.8574976325035095, "num_tokens": 427226817.0, "step": 11249 }, { "epoch": 1.4311156341432387, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 9.485033988952637, "learning_rate": 5e-05, "loss": 1.1657, "mean_token_accuracy": 0.8703310489654541, "num_tokens": 427268471.0, "step": 11250 }, { "epoch": 1.4312428444218293, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00079345703125, "grad_norm": 11.805352210998535, "learning_rate": 5e-05, "loss": 1.1759, "mean_token_accuracy": 0.883857786655426, "num_tokens": 427301242.0, "step": 11251 }, { "epoch": 1.4313700547004198, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.712677001953125, "learning_rate": 5e-05, "loss": 1.0017, "mean_token_accuracy": 0.8644985556602478, "num_tokens": 427335230.0, "step": 11252 }, { "epoch": 1.4314972649790103, "ewc_loss": 0.9140625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 12.297364234924316, "learning_rate": 5e-05, "loss": 1.3091, "mean_token_accuracy": 0.8626758456230164, "num_tokens": 427371993.0, "step": 11253 }, { "epoch": 1.4316244752576008, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 9.668696403503418, "learning_rate": 5e-05, "loss": 1.113, "mean_token_accuracy": 0.8659013509750366, "num_tokens": 427408810.0, "step": 11254 }, { "epoch": 1.4317516855361914, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00079345703125, "grad_norm": 12.034675598144531, "learning_rate": 5e-05, "loss": 1.2588, "mean_token_accuracy": 0.8577299118041992, "num_tokens": 427448740.0, "step": 11255 }, { "epoch": 1.431878895814782, "ewc_loss": 0.796875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000743865966796875, "grad_norm": 10.35348892211914, "learning_rate": 5e-05, "loss": 1.1638, "mean_token_accuracy": 0.8735150098800659, "num_tokens": 427488234.0, "step": 11256 }, { "epoch": 1.4320061060933724, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 11.213038444519043, "learning_rate": 5e-05, "loss": 1.1116, "mean_token_accuracy": 0.8685608506202698, "num_tokens": 427528053.0, "step": 11257 }, { "epoch": 1.4321333163719627, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.819690704345703, "learning_rate": 5e-05, "loss": 1.0186, "mean_token_accuracy": 0.8789488673210144, "num_tokens": 427566882.0, "step": 11258 }, { "epoch": 1.4322605266505533, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.813079833984375, "learning_rate": 5e-05, "loss": 1.0474, "mean_token_accuracy": 0.8753021955490112, "num_tokens": 427600464.0, "step": 11259 }, { "epoch": 1.4323877369291438, "ewc_loss": 0.625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.309298515319824, "learning_rate": 5e-05, "loss": 0.9713, "mean_token_accuracy": 0.8783607482910156, "num_tokens": 427633394.0, "step": 11260 }, { "epoch": 1.4325149472077343, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.553701400756836, "learning_rate": 5e-05, "loss": 1.0424, "mean_token_accuracy": 0.8724303245544434, "num_tokens": 427670591.0, "step": 11261 }, { "epoch": 1.4326421574863248, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.878639221191406, "learning_rate": 5e-05, "loss": 0.8998, "mean_token_accuracy": 0.8847047090530396, "num_tokens": 427711752.0, "step": 11262 }, { "epoch": 1.4327693677649154, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.58654499053955, "learning_rate": 5e-05, "loss": 1.0342, "mean_token_accuracy": 0.8721929788589478, "num_tokens": 427744655.0, "step": 11263 }, { "epoch": 1.432896578043506, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.69320011138916, "learning_rate": 5e-05, "loss": 0.903, "mean_token_accuracy": 0.8774745464324951, "num_tokens": 427784388.0, "step": 11264 }, { "epoch": 1.4330237883220964, "ewc_loss": 0.703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.751789093017578, "learning_rate": 5e-05, "loss": 1.0956, "mean_token_accuracy": 0.8628511428833008, "num_tokens": 427823498.0, "step": 11265 }, { "epoch": 1.433150998600687, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.853981018066406, "learning_rate": 5e-05, "loss": 0.9187, "mean_token_accuracy": 0.873847484588623, "num_tokens": 427868185.0, "step": 11266 }, { "epoch": 1.4332782088792775, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.716340065002441, "learning_rate": 5e-05, "loss": 1.0937, "mean_token_accuracy": 0.8672155141830444, "num_tokens": 427903315.0, "step": 11267 }, { "epoch": 1.433405419157868, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.15813159942627, "learning_rate": 5e-05, "loss": 0.9915, "mean_token_accuracy": 0.8613468408584595, "num_tokens": 427950206.0, "step": 11268 }, { "epoch": 1.4335326294364585, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.220132827758789, "learning_rate": 5e-05, "loss": 1.1139, "mean_token_accuracy": 0.8517807722091675, "num_tokens": 427983038.0, "step": 11269 }, { "epoch": 1.433659839715049, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.07259750366211, "learning_rate": 5e-05, "loss": 0.9093, "mean_token_accuracy": 0.8779584765434265, "num_tokens": 428015148.0, "step": 11270 }, { "epoch": 1.4337870499936396, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.65238094329834, "learning_rate": 5e-05, "loss": 0.9772, "mean_token_accuracy": 0.8725563287734985, "num_tokens": 428054300.0, "step": 11271 }, { "epoch": 1.4339142602722301, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.60270881652832, "learning_rate": 5e-05, "loss": 0.9156, "mean_token_accuracy": 0.8603376150131226, "num_tokens": 428093802.0, "step": 11272 }, { "epoch": 1.4340414705508204, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.455975532531738, "learning_rate": 5e-05, "loss": 0.9312, "mean_token_accuracy": 0.8818179965019226, "num_tokens": 428133283.0, "step": 11273 }, { "epoch": 1.434168680829411, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.08635139465332, "learning_rate": 5e-05, "loss": 0.8463, "mean_token_accuracy": 0.8651671409606934, "num_tokens": 428171125.0, "step": 11274 }, { "epoch": 1.4342958911080015, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.640006065368652, "learning_rate": 5e-05, "loss": 1.024, "mean_token_accuracy": 0.8563745617866516, "num_tokens": 428211731.0, "step": 11275 }, { "epoch": 1.434423101386592, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.922799110412598, "learning_rate": 5e-05, "loss": 0.8204, "mean_token_accuracy": 0.8681804537773132, "num_tokens": 428243737.0, "step": 11276 }, { "epoch": 1.4345503116651825, "ewc_loss": 0.625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.860060691833496, "learning_rate": 5e-05, "loss": 0.9674, "mean_token_accuracy": 0.8805608749389648, "num_tokens": 428278761.0, "step": 11277 }, { "epoch": 1.434677521943773, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.210262298583984, "learning_rate": 5e-05, "loss": 0.7876, "mean_token_accuracy": 0.8877810835838318, "num_tokens": 428316158.0, "step": 11278 }, { "epoch": 1.4348047322223636, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.65261173248291, "learning_rate": 5e-05, "loss": 0.9756, "mean_token_accuracy": 0.8718184232711792, "num_tokens": 428347180.0, "step": 11279 }, { "epoch": 1.4349319425009541, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.504986763000488, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8816171288490295, "num_tokens": 428389962.0, "step": 11280 }, { "epoch": 1.4350591527795447, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.971745491027832, "learning_rate": 5e-05, "loss": 0.9633, "mean_token_accuracy": 0.852213978767395, "num_tokens": 428432950.0, "step": 11281 }, { "epoch": 1.435186363058135, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.338715553283691, "learning_rate": 5e-05, "loss": 0.8772, "mean_token_accuracy": 0.8580899238586426, "num_tokens": 428470745.0, "step": 11282 }, { "epoch": 1.4353135733367255, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.461518287658691, "learning_rate": 5e-05, "loss": 0.884, "mean_token_accuracy": 0.8671169281005859, "num_tokens": 428508836.0, "step": 11283 }, { "epoch": 1.435440783615316, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.887820243835449, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8741322755813599, "num_tokens": 428546371.0, "step": 11284 }, { "epoch": 1.4355679938939065, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.372650146484375, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8855366110801697, "num_tokens": 428578878.0, "step": 11285 }, { "epoch": 1.435695204172497, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.492010593414307, "learning_rate": 5e-05, "loss": 0.759, "mean_token_accuracy": 0.8659574389457703, "num_tokens": 428611218.0, "step": 11286 }, { "epoch": 1.4358224144510876, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.417724609375, "learning_rate": 5e-05, "loss": 0.9102, "mean_token_accuracy": 0.8443409204483032, "num_tokens": 428649381.0, "step": 11287 }, { "epoch": 1.4359496247296781, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.3077192306518555, "learning_rate": 5e-05, "loss": 0.7373, "mean_token_accuracy": 0.8701854348182678, "num_tokens": 428685221.0, "step": 11288 }, { "epoch": 1.4360768350082687, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.4483642578125, "learning_rate": 5e-05, "loss": 0.8278, "mean_token_accuracy": 0.8723638653755188, "num_tokens": 428724194.0, "step": 11289 }, { "epoch": 1.4362040452868592, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.221463680267334, "learning_rate": 5e-05, "loss": 0.7293, "mean_token_accuracy": 0.872585654258728, "num_tokens": 428760501.0, "step": 11290 }, { "epoch": 1.4363312555654497, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.510852813720703, "learning_rate": 5e-05, "loss": 0.8666, "mean_token_accuracy": 0.8609610795974731, "num_tokens": 428793917.0, "step": 11291 }, { "epoch": 1.4364584658440402, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.222530841827393, "learning_rate": 5e-05, "loss": 0.7127, "mean_token_accuracy": 0.8802518844604492, "num_tokens": 428828998.0, "step": 11292 }, { "epoch": 1.4365856761226308, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.46121597290039, "learning_rate": 5e-05, "loss": 0.8042, "mean_token_accuracy": 0.88575279712677, "num_tokens": 428864739.0, "step": 11293 }, { "epoch": 1.4367128864012213, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.3991522789001465, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8612709045410156, "num_tokens": 428901026.0, "step": 11294 }, { "epoch": 1.4368400966798118, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.107250213623047, "learning_rate": 5e-05, "loss": 0.7603, "mean_token_accuracy": 0.8808383941650391, "num_tokens": 428937698.0, "step": 11295 }, { "epoch": 1.4369673069584024, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.635363578796387, "learning_rate": 5e-05, "loss": 0.8064, "mean_token_accuracy": 0.8551695942878723, "num_tokens": 428978288.0, "step": 11296 }, { "epoch": 1.4370945172369929, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.63768196105957, "learning_rate": 5e-05, "loss": 0.6913, "mean_token_accuracy": 0.8875864148139954, "num_tokens": 429014299.0, "step": 11297 }, { "epoch": 1.4372217275155832, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.860434532165527, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8639606237411499, "num_tokens": 429052455.0, "step": 11298 }, { "epoch": 1.4373489377941737, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.263847351074219, "learning_rate": 5e-05, "loss": 0.7214, "mean_token_accuracy": 0.870987594127655, "num_tokens": 429094168.0, "step": 11299 }, { "epoch": 1.4374761480727642, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.126946449279785, "learning_rate": 5e-05, "loss": 0.7756, "mean_token_accuracy": 0.8707077503204346, "num_tokens": 429137147.0, "step": 11300 }, { "epoch": 1.4376033583513548, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.921780109405518, "learning_rate": 5e-05, "loss": 0.6765, "mean_token_accuracy": 0.8776395320892334, "num_tokens": 429181679.0, "step": 11301 }, { "epoch": 1.4377305686299453, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.460591316223145, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.872398316860199, "num_tokens": 429222378.0, "step": 11302 }, { "epoch": 1.4378577789085358, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.710652828216553, "learning_rate": 5e-05, "loss": 0.6982, "mean_token_accuracy": 0.8695961236953735, "num_tokens": 429262044.0, "step": 11303 }, { "epoch": 1.4379849891871264, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.644396781921387, "learning_rate": 5e-05, "loss": 0.834, "mean_token_accuracy": 0.8724162578582764, "num_tokens": 429301643.0, "step": 11304 }, { "epoch": 1.438112199465717, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.854072570800781, "learning_rate": 5e-05, "loss": 0.7734, "mean_token_accuracy": 0.850013017654419, "num_tokens": 429339190.0, "step": 11305 }, { "epoch": 1.4382394097443074, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.369913101196289, "learning_rate": 5e-05, "loss": 0.8227, "mean_token_accuracy": 0.869683027267456, "num_tokens": 429378569.0, "step": 11306 }, { "epoch": 1.4383666200228977, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.332988262176514, "learning_rate": 5e-05, "loss": 0.7233, "mean_token_accuracy": 0.8715977668762207, "num_tokens": 429416012.0, "step": 11307 }, { "epoch": 1.4384938303014883, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.725907325744629, "learning_rate": 5e-05, "loss": 0.6891, "mean_token_accuracy": 0.8939146399497986, "num_tokens": 429453242.0, "step": 11308 }, { "epoch": 1.4386210405800788, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.870097637176514, "learning_rate": 5e-05, "loss": 0.7518, "mean_token_accuracy": 0.8780970573425293, "num_tokens": 429493908.0, "step": 11309 }, { "epoch": 1.4387482508586693, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.119041919708252, "learning_rate": 5e-05, "loss": 0.6805, "mean_token_accuracy": 0.8774142265319824, "num_tokens": 429530335.0, "step": 11310 }, { "epoch": 1.4388754611372598, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.290410041809082, "learning_rate": 5e-05, "loss": 0.7915, "mean_token_accuracy": 0.8699955344200134, "num_tokens": 429567523.0, "step": 11311 }, { "epoch": 1.4390026714158504, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.794802665710449, "learning_rate": 5e-05, "loss": 0.6833, "mean_token_accuracy": 0.8786894083023071, "num_tokens": 429605887.0, "step": 11312 }, { "epoch": 1.439129881694441, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.499794960021973, "learning_rate": 5e-05, "loss": 0.7767, "mean_token_accuracy": 0.8825302124023438, "num_tokens": 429644834.0, "step": 11313 }, { "epoch": 1.4392570919730314, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.785567283630371, "learning_rate": 5e-05, "loss": 0.6656, "mean_token_accuracy": 0.8822819590568542, "num_tokens": 429681151.0, "step": 11314 }, { "epoch": 1.439384302251622, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.38395881652832, "learning_rate": 5e-05, "loss": 0.8365, "mean_token_accuracy": 0.8622670769691467, "num_tokens": 429719135.0, "step": 11315 }, { "epoch": 1.4395115125302125, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.094668388366699, "learning_rate": 5e-05, "loss": 0.6732, "mean_token_accuracy": 0.8798084259033203, "num_tokens": 429753443.0, "step": 11316 }, { "epoch": 1.439638722808803, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.9316301345825195, "learning_rate": 5e-05, "loss": 0.7713, "mean_token_accuracy": 0.8768068552017212, "num_tokens": 429788762.0, "step": 11317 }, { "epoch": 1.4397659330873935, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.534612655639648, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.8625236749649048, "num_tokens": 429824054.0, "step": 11318 }, { "epoch": 1.439893143365984, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.455170154571533, "learning_rate": 5e-05, "loss": 0.7494, "mean_token_accuracy": 0.8698785305023193, "num_tokens": 429857772.0, "step": 11319 }, { "epoch": 1.4400203536445746, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.827923774719238, "learning_rate": 5e-05, "loss": 0.7346, "mean_token_accuracy": 0.8768072724342346, "num_tokens": 429891184.0, "step": 11320 }, { "epoch": 1.4401475639231651, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.253891944885254, "learning_rate": 5e-05, "loss": 0.6922, "mean_token_accuracy": 0.878520667552948, "num_tokens": 429930322.0, "step": 11321 }, { "epoch": 1.4402747742017554, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.8268208503723145, "learning_rate": 5e-05, "loss": 0.7118, "mean_token_accuracy": 0.8759767413139343, "num_tokens": 429965708.0, "step": 11322 }, { "epoch": 1.440401984480346, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.368410587310791, "learning_rate": 5e-05, "loss": 0.7234, "mean_token_accuracy": 0.8691991567611694, "num_tokens": 430007266.0, "step": 11323 }, { "epoch": 1.4405291947589365, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.5312418937683105, "learning_rate": 5e-05, "loss": 0.741, "mean_token_accuracy": 0.8650493025779724, "num_tokens": 430044470.0, "step": 11324 }, { "epoch": 1.440656405037527, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.720742225646973, "learning_rate": 5e-05, "loss": 0.739, "mean_token_accuracy": 0.8737344741821289, "num_tokens": 430079776.0, "step": 11325 }, { "epoch": 1.4407836153161175, "ewc_loss": 0.328125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.069061279296875, "learning_rate": 5e-05, "loss": 0.6681, "mean_token_accuracy": 0.8794807195663452, "num_tokens": 430110697.0, "step": 11326 }, { "epoch": 1.440910825594708, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.255982398986816, "learning_rate": 5e-05, "loss": 0.7785, "mean_token_accuracy": 0.8720104694366455, "num_tokens": 430150350.0, "step": 11327 }, { "epoch": 1.4410380358732986, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0002384185791015625, "grad_norm": 6.516147613525391, "learning_rate": 5e-05, "loss": 0.695, "mean_token_accuracy": 0.8552178740501404, "num_tokens": 430185415.0, "step": 11328 }, { "epoch": 1.4411652461518891, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.961562156677246, "learning_rate": 5e-05, "loss": 0.8643, "mean_token_accuracy": 0.8665579557418823, "num_tokens": 430227865.0, "step": 11329 }, { "epoch": 1.4412924564304797, "ewc_loss": 0.275390625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.00022220611572265625, "grad_norm": 5.963027000427246, "learning_rate": 5e-05, "loss": 0.6192, "mean_token_accuracy": 0.8789463043212891, "num_tokens": 430267821.0, "step": 11330 }, { "epoch": 1.44141966670907, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.845808029174805, "learning_rate": 5e-05, "loss": 0.9348, "mean_token_accuracy": 0.8769059777259827, "num_tokens": 430307251.0, "step": 11331 }, { "epoch": 1.4415468769876605, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.867915630340576, "learning_rate": 5e-05, "loss": 0.7126, "mean_token_accuracy": 0.8593628406524658, "num_tokens": 430347842.0, "step": 11332 }, { "epoch": 1.441674087266251, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.0005340576171875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.407866477966309, "learning_rate": 5e-05, "loss": 1.0953, "mean_token_accuracy": 0.8635889291763306, "num_tokens": 430383175.0, "step": 11333 }, { "epoch": 1.4418012975448415, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.2107977867126465, "learning_rate": 5e-05, "loss": 0.8789, "mean_token_accuracy": 0.8661985993385315, "num_tokens": 430414319.0, "step": 11334 }, { "epoch": 1.441928507823432, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.651576042175293, "learning_rate": 5e-05, "loss": 0.9914, "mean_token_accuracy": 0.8827234506607056, "num_tokens": 430454142.0, "step": 11335 }, { "epoch": 1.4420557181020226, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.649080276489258, "learning_rate": 5e-05, "loss": 1.0334, "mean_token_accuracy": 0.861618161201477, "num_tokens": 430496980.0, "step": 11336 }, { "epoch": 1.4421829283806131, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.714693546295166, "learning_rate": 5e-05, "loss": 0.8629, "mean_token_accuracy": 0.8612079620361328, "num_tokens": 430533645.0, "step": 11337 }, { "epoch": 1.4423101386592037, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.583067893981934, "learning_rate": 5e-05, "loss": 1.0144, "mean_token_accuracy": 0.8740707635879517, "num_tokens": 430567744.0, "step": 11338 }, { "epoch": 1.4424373489377942, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.747824192047119, "learning_rate": 5e-05, "loss": 0.8431, "mean_token_accuracy": 0.8474521040916443, "num_tokens": 430603887.0, "step": 11339 }, { "epoch": 1.4425645592163847, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.903769493103027, "learning_rate": 5e-05, "loss": 1.0339, "mean_token_accuracy": 0.8698461055755615, "num_tokens": 430638748.0, "step": 11340 }, { "epoch": 1.4426917694949752, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.369186878204346, "learning_rate": 5e-05, "loss": 0.8842, "mean_token_accuracy": 0.8705728650093079, "num_tokens": 430677799.0, "step": 11341 }, { "epoch": 1.4428189797735658, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.141167640686035, "learning_rate": 5e-05, "loss": 0.9884, "mean_token_accuracy": 0.8676252961158752, "num_tokens": 430720816.0, "step": 11342 }, { "epoch": 1.4429461900521563, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.137125968933105, "learning_rate": 5e-05, "loss": 0.9383, "mean_token_accuracy": 0.8726196885108948, "num_tokens": 430755220.0, "step": 11343 }, { "epoch": 1.4430734003307468, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.06639575958252, "learning_rate": 5e-05, "loss": 0.8381, "mean_token_accuracy": 0.8734062910079956, "num_tokens": 430787737.0, "step": 11344 }, { "epoch": 1.4432006106093374, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.298978805541992, "learning_rate": 5e-05, "loss": 0.9168, "mean_token_accuracy": 0.8720149397850037, "num_tokens": 430823817.0, "step": 11345 }, { "epoch": 1.4433278208879279, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.616061687469482, "learning_rate": 5e-05, "loss": 0.8303, "mean_token_accuracy": 0.8661051988601685, "num_tokens": 430863416.0, "step": 11346 }, { "epoch": 1.4434550311665182, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.131180763244629, "learning_rate": 5e-05, "loss": 0.8779, "mean_token_accuracy": 0.8616123199462891, "num_tokens": 430894017.0, "step": 11347 }, { "epoch": 1.4435822414451087, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.503066539764404, "learning_rate": 5e-05, "loss": 0.8082, "mean_token_accuracy": 0.8665249943733215, "num_tokens": 430935023.0, "step": 11348 }, { "epoch": 1.4437094517236992, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.025114059448242, "learning_rate": 5e-05, "loss": 0.8255, "mean_token_accuracy": 0.8716893196105957, "num_tokens": 430971663.0, "step": 11349 }, { "epoch": 1.4438366620022898, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.344379425048828, "learning_rate": 5e-05, "loss": 0.787, "mean_token_accuracy": 0.8667330741882324, "num_tokens": 431018919.0, "step": 11350 }, { "epoch": 1.4439638722808803, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.1370849609375, "learning_rate": 5e-05, "loss": 0.7985, "mean_token_accuracy": 0.8792159557342529, "num_tokens": 431050076.0, "step": 11351 }, { "epoch": 1.4440910825594708, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.087721824645996, "learning_rate": 5e-05, "loss": 0.7774, "mean_token_accuracy": 0.8598872423171997, "num_tokens": 431085535.0, "step": 11352 }, { "epoch": 1.4442182928380614, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.351914405822754, "learning_rate": 5e-05, "loss": 0.8149, "mean_token_accuracy": 0.875153660774231, "num_tokens": 431123187.0, "step": 11353 }, { "epoch": 1.4443455031166519, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.891881465911865, "learning_rate": 5e-05, "loss": 0.7645, "mean_token_accuracy": 0.8611020445823669, "num_tokens": 431157867.0, "step": 11354 }, { "epoch": 1.4444727133952424, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.519149780273438, "learning_rate": 5e-05, "loss": 0.8963, "mean_token_accuracy": 0.8650722503662109, "num_tokens": 431191344.0, "step": 11355 }, { "epoch": 1.4445999236738327, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.817322254180908, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.8734360933303833, "num_tokens": 431231019.0, "step": 11356 }, { "epoch": 1.4447271339524232, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.585100173950195, "learning_rate": 5e-05, "loss": 0.8592, "mean_token_accuracy": 0.8705936670303345, "num_tokens": 431265660.0, "step": 11357 }, { "epoch": 1.4448543442310138, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.8656840324401855, "learning_rate": 5e-05, "loss": 0.7075, "mean_token_accuracy": 0.8786723613739014, "num_tokens": 431298666.0, "step": 11358 }, { "epoch": 1.4449815545096043, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.510629653930664, "learning_rate": 5e-05, "loss": 0.8233, "mean_token_accuracy": 0.885748028755188, "num_tokens": 431327516.0, "step": 11359 }, { "epoch": 1.4451087647881948, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.039580821990967, "learning_rate": 5e-05, "loss": 0.7588, "mean_token_accuracy": 0.8687102794647217, "num_tokens": 431366244.0, "step": 11360 }, { "epoch": 1.4452359750667854, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.277809143066406, "learning_rate": 5e-05, "loss": 0.8526, "mean_token_accuracy": 0.8649780750274658, "num_tokens": 431402440.0, "step": 11361 }, { "epoch": 1.445363185345376, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.242242813110352, "learning_rate": 5e-05, "loss": 0.7336, "mean_token_accuracy": 0.8860256671905518, "num_tokens": 431436628.0, "step": 11362 }, { "epoch": 1.4454903956239664, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.985753536224365, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8849502801895142, "num_tokens": 431476343.0, "step": 11363 }, { "epoch": 1.445617605902557, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.441098690032959, "learning_rate": 5e-05, "loss": 0.7477, "mean_token_accuracy": 0.87864089012146, "num_tokens": 431513741.0, "step": 11364 }, { "epoch": 1.4457448161811475, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.715484142303467, "learning_rate": 5e-05, "loss": 0.8108, "mean_token_accuracy": 0.8610922694206238, "num_tokens": 431556227.0, "step": 11365 }, { "epoch": 1.445872026459738, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.596870422363281, "learning_rate": 5e-05, "loss": 0.7552, "mean_token_accuracy": 0.8727201223373413, "num_tokens": 431592004.0, "step": 11366 }, { "epoch": 1.4459992367383285, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.489668369293213, "learning_rate": 5e-05, "loss": 0.7662, "mean_token_accuracy": 0.867087721824646, "num_tokens": 431631286.0, "step": 11367 }, { "epoch": 1.446126447016919, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.719832897186279, "learning_rate": 5e-05, "loss": 0.7802, "mean_token_accuracy": 0.8699187636375427, "num_tokens": 431669269.0, "step": 11368 }, { "epoch": 1.4462536572955096, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.328850746154785, "learning_rate": 5e-05, "loss": 0.7596, "mean_token_accuracy": 0.862037181854248, "num_tokens": 431708061.0, "step": 11369 }, { "epoch": 1.4463808675741001, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.832710266113281, "learning_rate": 5e-05, "loss": 0.7608, "mean_token_accuracy": 0.8747677803039551, "num_tokens": 431745855.0, "step": 11370 }, { "epoch": 1.4465080778526904, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.1705708503723145, "learning_rate": 5e-05, "loss": 0.7254, "mean_token_accuracy": 0.8722286820411682, "num_tokens": 431784481.0, "step": 11371 }, { "epoch": 1.446635288131281, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.984655380249023, "learning_rate": 5e-05, "loss": 0.7832, "mean_token_accuracy": 0.8660848140716553, "num_tokens": 431822724.0, "step": 11372 }, { "epoch": 1.4467624984098715, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.0075860023498535, "learning_rate": 5e-05, "loss": 0.7094, "mean_token_accuracy": 0.877354621887207, "num_tokens": 431859415.0, "step": 11373 }, { "epoch": 1.446889708688462, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.175718307495117, "learning_rate": 5e-05, "loss": 0.7392, "mean_token_accuracy": 0.8911798000335693, "num_tokens": 431899432.0, "step": 11374 }, { "epoch": 1.4470169189670525, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.786427974700928, "learning_rate": 5e-05, "loss": 0.6823, "mean_token_accuracy": 0.8778101801872253, "num_tokens": 431935546.0, "step": 11375 }, { "epoch": 1.447144129245643, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.476943969726562, "learning_rate": 5e-05, "loss": 0.8251, "mean_token_accuracy": 0.8724076747894287, "num_tokens": 431972366.0, "step": 11376 }, { "epoch": 1.4472713395242336, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.529434680938721, "learning_rate": 5e-05, "loss": 0.719, "mean_token_accuracy": 0.8602442741394043, "num_tokens": 432007795.0, "step": 11377 }, { "epoch": 1.4473985498028241, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.862724304199219, "learning_rate": 5e-05, "loss": 0.9282, "mean_token_accuracy": 0.8504983186721802, "num_tokens": 432047589.0, "step": 11378 }, { "epoch": 1.4475257600814146, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.320878505706787, "learning_rate": 5e-05, "loss": 0.6565, "mean_token_accuracy": 0.8806412220001221, "num_tokens": 432085615.0, "step": 11379 }, { "epoch": 1.447652970360005, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.223650932312012, "learning_rate": 5e-05, "loss": 0.9365, "mean_token_accuracy": 0.873629093170166, "num_tokens": 432121112.0, "step": 11380 }, { "epoch": 1.4477801806385955, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.512469291687012, "learning_rate": 5e-05, "loss": 0.7132, "mean_token_accuracy": 0.8705059885978699, "num_tokens": 432156429.0, "step": 11381 }, { "epoch": 1.447907390917186, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.066152572631836, "learning_rate": 5e-05, "loss": 1.0024, "mean_token_accuracy": 0.8590261936187744, "num_tokens": 432195653.0, "step": 11382 }, { "epoch": 1.4480346011957765, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.263221740722656, "learning_rate": 5e-05, "loss": 0.8435, "mean_token_accuracy": 0.8689675331115723, "num_tokens": 432230997.0, "step": 11383 }, { "epoch": 1.448161811474367, "ewc_loss": 0.5, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.104307174682617, "learning_rate": 5e-05, "loss": 0.88, "mean_token_accuracy": 0.8638325929641724, "num_tokens": 432270868.0, "step": 11384 }, { "epoch": 1.4482890217529576, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.136275291442871, "learning_rate": 5e-05, "loss": 0.8944, "mean_token_accuracy": 0.8741650581359863, "num_tokens": 432309999.0, "step": 11385 }, { "epoch": 1.4484162320315481, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.998419284820557, "learning_rate": 5e-05, "loss": 0.7142, "mean_token_accuracy": 0.8820293545722961, "num_tokens": 432345240.0, "step": 11386 }, { "epoch": 1.4485434423101387, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.787602424621582, "learning_rate": 5e-05, "loss": 0.9402, "mean_token_accuracy": 0.8630901575088501, "num_tokens": 432391478.0, "step": 11387 }, { "epoch": 1.4486706525887292, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.523809909820557, "learning_rate": 5e-05, "loss": 0.6649, "mean_token_accuracy": 0.8865638375282288, "num_tokens": 432430601.0, "step": 11388 }, { "epoch": 1.4487978628673197, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.008927345275879, "learning_rate": 5e-05, "loss": 0.9602, "mean_token_accuracy": 0.868075430393219, "num_tokens": 432467260.0, "step": 11389 }, { "epoch": 1.4489250731459102, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.738315105438232, "learning_rate": 5e-05, "loss": 0.7937, "mean_token_accuracy": 0.8557482957839966, "num_tokens": 432506270.0, "step": 11390 }, { "epoch": 1.4490522834245008, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.694560050964355, "learning_rate": 5e-05, "loss": 0.9264, "mean_token_accuracy": 0.868740439414978, "num_tokens": 432543688.0, "step": 11391 }, { "epoch": 1.4491794937030913, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.281123161315918, "learning_rate": 5e-05, "loss": 0.8097, "mean_token_accuracy": 0.8792918920516968, "num_tokens": 432583522.0, "step": 11392 }, { "epoch": 1.4493067039816818, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.07053279876709, "learning_rate": 5e-05, "loss": 0.8268, "mean_token_accuracy": 0.8847271203994751, "num_tokens": 432622471.0, "step": 11393 }, { "epoch": 1.4494339142602723, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.7213850021362305, "learning_rate": 5e-05, "loss": 0.8181, "mean_token_accuracy": 0.8819847702980042, "num_tokens": 432666689.0, "step": 11394 }, { "epoch": 1.4495611245388629, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.577869415283203, "learning_rate": 5e-05, "loss": 0.78, "mean_token_accuracy": 0.8755861520767212, "num_tokens": 432707816.0, "step": 11395 }, { "epoch": 1.4496883348174532, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.820019245147705, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.8850683569908142, "num_tokens": 432746873.0, "step": 11396 }, { "epoch": 1.4498155450960437, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.481696605682373, "learning_rate": 5e-05, "loss": 0.73, "mean_token_accuracy": 0.8843486309051514, "num_tokens": 432777825.0, "step": 11397 }, { "epoch": 1.4499427553746342, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.660684585571289, "learning_rate": 5e-05, "loss": 0.7908, "mean_token_accuracy": 0.8697453737258911, "num_tokens": 432816806.0, "step": 11398 }, { "epoch": 1.4500699656532248, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.603687763214111, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.880682110786438, "num_tokens": 432855726.0, "step": 11399 }, { "epoch": 1.4501971759318153, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.429551124572754, "learning_rate": 5e-05, "loss": 0.7739, "mean_token_accuracy": 0.8617535829544067, "num_tokens": 432891526.0, "step": 11400 }, { "epoch": 1.4503243862104058, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.778706073760986, "learning_rate": 5e-05, "loss": 0.7043, "mean_token_accuracy": 0.8913438320159912, "num_tokens": 432921995.0, "step": 11401 }, { "epoch": 1.4504515964889964, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.194861888885498, "learning_rate": 5e-05, "loss": 0.741, "mean_token_accuracy": 0.866851806640625, "num_tokens": 432962120.0, "step": 11402 }, { "epoch": 1.4505788067675869, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.988330364227295, "learning_rate": 5e-05, "loss": 0.7934, "mean_token_accuracy": 0.8705190420150757, "num_tokens": 433001944.0, "step": 11403 }, { "epoch": 1.4507060170461774, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.9263811111450195, "learning_rate": 5e-05, "loss": 0.7648, "mean_token_accuracy": 0.8514806032180786, "num_tokens": 433040984.0, "step": 11404 }, { "epoch": 1.4508332273247677, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.343076705932617, "learning_rate": 5e-05, "loss": 0.822, "mean_token_accuracy": 0.8674449324607849, "num_tokens": 433077914.0, "step": 11405 }, { "epoch": 1.4509604376033582, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.5365095138549805, "learning_rate": 5e-05, "loss": 0.6563, "mean_token_accuracy": 0.8823279142379761, "num_tokens": 433111219.0, "step": 11406 }, { "epoch": 1.4510876478819488, "ewc_loss": 0.5, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.928056716918945, "learning_rate": 5e-05, "loss": 0.822, "mean_token_accuracy": 0.8879157304763794, "num_tokens": 433143204.0, "step": 11407 }, { "epoch": 1.4512148581605393, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.100668907165527, "learning_rate": 5e-05, "loss": 0.6413, "mean_token_accuracy": 0.8867831230163574, "num_tokens": 433175361.0, "step": 11408 }, { "epoch": 1.4513420684391298, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.651626586914062, "learning_rate": 5e-05, "loss": 0.9469, "mean_token_accuracy": 0.881186842918396, "num_tokens": 433215303.0, "step": 11409 }, { "epoch": 1.4514692787177204, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.18621826171875, "learning_rate": 5e-05, "loss": 0.7007, "mean_token_accuracy": 0.8790290355682373, "num_tokens": 433255301.0, "step": 11410 }, { "epoch": 1.4515964889963109, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.9697265625, "learning_rate": 5e-05, "loss": 1.0748, "mean_token_accuracy": 0.8639635443687439, "num_tokens": 433291110.0, "step": 11411 }, { "epoch": 1.4517236992749014, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.348156929016113, "learning_rate": 5e-05, "loss": 0.9325, "mean_token_accuracy": 0.8725267052650452, "num_tokens": 433330910.0, "step": 11412 }, { "epoch": 1.451850909553492, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.057831764221191, "learning_rate": 5e-05, "loss": 0.9896, "mean_token_accuracy": 0.8730015754699707, "num_tokens": 433364297.0, "step": 11413 }, { "epoch": 1.4519781198320825, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.559419631958008, "learning_rate": 5e-05, "loss": 1.022, "mean_token_accuracy": 0.8677797317504883, "num_tokens": 433405900.0, "step": 11414 }, { "epoch": 1.452105330110673, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.277616024017334, "learning_rate": 5e-05, "loss": 0.827, "mean_token_accuracy": 0.8640309572219849, "num_tokens": 433442639.0, "step": 11415 }, { "epoch": 1.4522325403892635, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.479325294494629, "learning_rate": 5e-05, "loss": 1.0147, "mean_token_accuracy": 0.866484522819519, "num_tokens": 433471657.0, "step": 11416 }, { "epoch": 1.452359750667854, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.589996337890625, "learning_rate": 5e-05, "loss": 0.8239, "mean_token_accuracy": 0.8537483215332031, "num_tokens": 433510936.0, "step": 11417 }, { "epoch": 1.4524869609464446, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.76468563079834, "learning_rate": 5e-05, "loss": 1.023, "mean_token_accuracy": 0.8711605668067932, "num_tokens": 433546940.0, "step": 11418 }, { "epoch": 1.4526141712250351, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.382025241851807, "learning_rate": 5e-05, "loss": 0.9142, "mean_token_accuracy": 0.8735352158546448, "num_tokens": 433588701.0, "step": 11419 }, { "epoch": 1.4527413815036254, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.909923553466797, "learning_rate": 5e-05, "loss": 0.9507, "mean_token_accuracy": 0.8710761070251465, "num_tokens": 433625678.0, "step": 11420 }, { "epoch": 1.452868591782216, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.241706848144531, "learning_rate": 5e-05, "loss": 1.0479, "mean_token_accuracy": 0.8551604151725769, "num_tokens": 433660624.0, "step": 11421 }, { "epoch": 1.4529958020608065, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.718160152435303, "learning_rate": 5e-05, "loss": 0.8093, "mean_token_accuracy": 0.8782581686973572, "num_tokens": 433702737.0, "step": 11422 }, { "epoch": 1.453123012339397, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.544110298156738, "learning_rate": 5e-05, "loss": 0.954, "mean_token_accuracy": 0.8695999979972839, "num_tokens": 433741866.0, "step": 11423 }, { "epoch": 1.4532502226179875, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.293394565582275, "learning_rate": 5e-05, "loss": 0.8359, "mean_token_accuracy": 0.8637178540229797, "num_tokens": 433779869.0, "step": 11424 }, { "epoch": 1.453377432896578, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.388842582702637, "learning_rate": 5e-05, "loss": 0.8813, "mean_token_accuracy": 0.8827346563339233, "num_tokens": 433819629.0, "step": 11425 }, { "epoch": 1.4535046431751686, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.393356800079346, "learning_rate": 5e-05, "loss": 0.8062, "mean_token_accuracy": 0.8775152564048767, "num_tokens": 433855330.0, "step": 11426 }, { "epoch": 1.4536318534537591, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.063077926635742, "learning_rate": 5e-05, "loss": 0.8632, "mean_token_accuracy": 0.8650758266448975, "num_tokens": 433896130.0, "step": 11427 }, { "epoch": 1.4537590637323496, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.494959354400635, "learning_rate": 5e-05, "loss": 0.7922, "mean_token_accuracy": 0.8817780017852783, "num_tokens": 433928514.0, "step": 11428 }, { "epoch": 1.45388627401094, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.849567413330078, "learning_rate": 5e-05, "loss": 0.8034, "mean_token_accuracy": 0.8686161637306213, "num_tokens": 433969213.0, "step": 11429 }, { "epoch": 1.4540134842895305, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.4885735511779785, "learning_rate": 5e-05, "loss": 0.8487, "mean_token_accuracy": 0.8577043414115906, "num_tokens": 434007902.0, "step": 11430 }, { "epoch": 1.454140694568121, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.727145195007324, "learning_rate": 5e-05, "loss": 0.8374, "mean_token_accuracy": 0.8593471050262451, "num_tokens": 434043243.0, "step": 11431 }, { "epoch": 1.4542679048467115, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.491581916809082, "learning_rate": 5e-05, "loss": 0.731, "mean_token_accuracy": 0.8852310180664062, "num_tokens": 434081075.0, "step": 11432 }, { "epoch": 1.454395115125302, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.604107856750488, "learning_rate": 5e-05, "loss": 0.7655, "mean_token_accuracy": 0.8724093437194824, "num_tokens": 434121230.0, "step": 11433 }, { "epoch": 1.4545223254038926, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.56345272064209, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8674241304397583, "num_tokens": 434158324.0, "step": 11434 }, { "epoch": 1.4546495356824831, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.4676923751831055, "learning_rate": 5e-05, "loss": 0.7422, "mean_token_accuracy": 0.8731226921081543, "num_tokens": 434203910.0, "step": 11435 }, { "epoch": 1.4547767459610736, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.638002395629883, "learning_rate": 5e-05, "loss": 0.7302, "mean_token_accuracy": 0.8822417855262756, "num_tokens": 434238915.0, "step": 11436 }, { "epoch": 1.4549039562396642, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.352522373199463, "learning_rate": 5e-05, "loss": 0.7271, "mean_token_accuracy": 0.8735041618347168, "num_tokens": 434280049.0, "step": 11437 }, { "epoch": 1.4550311665182547, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.723590850830078, "learning_rate": 5e-05, "loss": 0.7315, "mean_token_accuracy": 0.8801524639129639, "num_tokens": 434323980.0, "step": 11438 }, { "epoch": 1.4551583767968452, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.1953959465026855, "learning_rate": 5e-05, "loss": 0.7354, "mean_token_accuracy": 0.8670632243156433, "num_tokens": 434363386.0, "step": 11439 }, { "epoch": 1.4552855870754358, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.930963516235352, "learning_rate": 5e-05, "loss": 0.7298, "mean_token_accuracy": 0.8832355737686157, "num_tokens": 434399480.0, "step": 11440 }, { "epoch": 1.4554127973540263, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.919216156005859, "learning_rate": 5e-05, "loss": 0.7156, "mean_token_accuracy": 0.867516279220581, "num_tokens": 434445662.0, "step": 11441 }, { "epoch": 1.4555400076326168, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.285454750061035, "learning_rate": 5e-05, "loss": 0.7846, "mean_token_accuracy": 0.8739190101623535, "num_tokens": 434480926.0, "step": 11442 }, { "epoch": 1.4556672179112073, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.521317481994629, "learning_rate": 5e-05, "loss": 0.7121, "mean_token_accuracy": 0.8623864054679871, "num_tokens": 434520018.0, "step": 11443 }, { "epoch": 1.4557944281897979, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.863959312438965, "learning_rate": 5e-05, "loss": 0.8896, "mean_token_accuracy": 0.8593374490737915, "num_tokens": 434555163.0, "step": 11444 }, { "epoch": 1.4559216384683882, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.007061958312988, "learning_rate": 5e-05, "loss": 0.6259, "mean_token_accuracy": 0.883550226688385, "num_tokens": 434586353.0, "step": 11445 }, { "epoch": 1.4560488487469787, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.769868850708008, "learning_rate": 5e-05, "loss": 0.969, "mean_token_accuracy": 0.8771090507507324, "num_tokens": 434630558.0, "step": 11446 }, { "epoch": 1.4561760590255692, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 5.845494747161865, "learning_rate": 5e-05, "loss": 0.7039, "mean_token_accuracy": 0.8709807991981506, "num_tokens": 434672949.0, "step": 11447 }, { "epoch": 1.4563032693041598, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.559908866882324, "learning_rate": 5e-05, "loss": 1.1193, "mean_token_accuracy": 0.8661895394325256, "num_tokens": 434709207.0, "step": 11448 }, { "epoch": 1.4564304795827503, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.210525035858154, "learning_rate": 5e-05, "loss": 0.9055, "mean_token_accuracy": 0.8814777731895447, "num_tokens": 434745787.0, "step": 11449 }, { "epoch": 1.4565576898613408, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.027729988098145, "learning_rate": 5e-05, "loss": 1.0508, "mean_token_accuracy": 0.8725159168243408, "num_tokens": 434785419.0, "step": 11450 }, { "epoch": 1.4566849001399313, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 8.810872077941895, "learning_rate": 5e-05, "loss": 1.1421, "mean_token_accuracy": 0.8636982440948486, "num_tokens": 434818217.0, "step": 11451 }, { "epoch": 1.4568121104185219, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.986282825469971, "learning_rate": 5e-05, "loss": 0.8937, "mean_token_accuracy": 0.8625766038894653, "num_tokens": 434855654.0, "step": 11452 }, { "epoch": 1.4569393206971124, "ewc_loss": 0.75, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 9.903512001037598, "learning_rate": 5e-05, "loss": 1.1669, "mean_token_accuracy": 0.8600208759307861, "num_tokens": 434890657.0, "step": 11453 }, { "epoch": 1.4570665309757027, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 6.865710735321045, "learning_rate": 5e-05, "loss": 0.7847, "mean_token_accuracy": 0.8823248147964478, "num_tokens": 434929164.0, "step": 11454 }, { "epoch": 1.4571937412542932, "ewc_loss": 0.796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000743865966796875, "grad_norm": 10.864311218261719, "learning_rate": 5e-05, "loss": 1.1959, "mean_token_accuracy": 0.8645505309104919, "num_tokens": 434966017.0, "step": 11455 }, { "epoch": 1.4573209515328838, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 8.11464786529541, "learning_rate": 5e-05, "loss": 1.0118, "mean_token_accuracy": 0.8802736401557922, "num_tokens": 435002076.0, "step": 11456 }, { "epoch": 1.4574481618114743, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.126001358032227, "learning_rate": 5e-05, "loss": 1.0042, "mean_token_accuracy": 0.8775396347045898, "num_tokens": 435030058.0, "step": 11457 }, { "epoch": 1.4575753720900648, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 8.888498306274414, "learning_rate": 5e-05, "loss": 1.1653, "mean_token_accuracy": 0.8660633563995361, "num_tokens": 435067816.0, "step": 11458 }, { "epoch": 1.4577025823686554, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.801448822021484, "learning_rate": 5e-05, "loss": 0.9292, "mean_token_accuracy": 0.8746801614761353, "num_tokens": 435106078.0, "step": 11459 }, { "epoch": 1.4578297926472459, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.886140823364258, "learning_rate": 5e-05, "loss": 0.978, "mean_token_accuracy": 0.882729172706604, "num_tokens": 435144558.0, "step": 11460 }, { "epoch": 1.4579570029258364, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.11822509765625, "learning_rate": 5e-05, "loss": 0.9344, "mean_token_accuracy": 0.8673298358917236, "num_tokens": 435183083.0, "step": 11461 }, { "epoch": 1.458084213204427, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.47701358795166, "learning_rate": 5e-05, "loss": 0.8689, "mean_token_accuracy": 0.8876502513885498, "num_tokens": 435223459.0, "step": 11462 }, { "epoch": 1.4582114234830175, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.067734718322754, "learning_rate": 5e-05, "loss": 0.9126, "mean_token_accuracy": 0.8789288997650146, "num_tokens": 435261491.0, "step": 11463 }, { "epoch": 1.458338633761608, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.735461711883545, "learning_rate": 5e-05, "loss": 0.8446, "mean_token_accuracy": 0.8622792959213257, "num_tokens": 435299634.0, "step": 11464 }, { "epoch": 1.4584658440401985, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.35970401763916, "learning_rate": 5e-05, "loss": 0.9323, "mean_token_accuracy": 0.8572531342506409, "num_tokens": 435339645.0, "step": 11465 }, { "epoch": 1.458593054318789, "ewc_loss": 0.375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.081548690795898, "learning_rate": 5e-05, "loss": 0.7541, "mean_token_accuracy": 0.8716061115264893, "num_tokens": 435382635.0, "step": 11466 }, { "epoch": 1.4587202645973796, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.80589485168457, "learning_rate": 5e-05, "loss": 0.928, "mean_token_accuracy": 0.8635537624359131, "num_tokens": 435421653.0, "step": 11467 }, { "epoch": 1.45884747487597, "ewc_loss": 0.375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.737356185913086, "learning_rate": 5e-05, "loss": 0.7693, "mean_token_accuracy": 0.8637771606445312, "num_tokens": 435460135.0, "step": 11468 }, { "epoch": 1.4589746851545604, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.197859764099121, "learning_rate": 5e-05, "loss": 0.906, "mean_token_accuracy": 0.885373592376709, "num_tokens": 435497717.0, "step": 11469 }, { "epoch": 1.459101895433151, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.766870975494385, "learning_rate": 5e-05, "loss": 0.8345, "mean_token_accuracy": 0.8608294725418091, "num_tokens": 435535247.0, "step": 11470 }, { "epoch": 1.4592291057117415, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.33127498626709, "learning_rate": 5e-05, "loss": 0.9343, "mean_token_accuracy": 0.8789693117141724, "num_tokens": 435576841.0, "step": 11471 }, { "epoch": 1.459356315990332, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.1683244705200195, "learning_rate": 5e-05, "loss": 0.8694, "mean_token_accuracy": 0.8765102624893188, "num_tokens": 435618197.0, "step": 11472 }, { "epoch": 1.4594835262689225, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.94256591796875, "learning_rate": 5e-05, "loss": 1.0176, "mean_token_accuracy": 0.8483709096908569, "num_tokens": 435663327.0, "step": 11473 }, { "epoch": 1.459610736547513, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 7.604393005371094, "learning_rate": 5e-05, "loss": 0.9107, "mean_token_accuracy": 0.8730076551437378, "num_tokens": 435697795.0, "step": 11474 }, { "epoch": 1.4597379468261036, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.315645217895508, "learning_rate": 5e-05, "loss": 0.9132, "mean_token_accuracy": 0.8632062077522278, "num_tokens": 435734308.0, "step": 11475 }, { "epoch": 1.4598651571046941, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.714941024780273, "learning_rate": 5e-05, "loss": 0.893, "mean_token_accuracy": 0.8651325702667236, "num_tokens": 435772575.0, "step": 11476 }, { "epoch": 1.4599923673832846, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.963543891906738, "learning_rate": 5e-05, "loss": 0.8002, "mean_token_accuracy": 0.883552610874176, "num_tokens": 435807578.0, "step": 11477 }, { "epoch": 1.460119577661875, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.586737632751465, "learning_rate": 5e-05, "loss": 0.8278, "mean_token_accuracy": 0.8716099262237549, "num_tokens": 435849035.0, "step": 11478 }, { "epoch": 1.4602467879404655, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.900999069213867, "learning_rate": 5e-05, "loss": 0.8085, "mean_token_accuracy": 0.8748243451118469, "num_tokens": 435884912.0, "step": 11479 }, { "epoch": 1.460373998219056, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.402180194854736, "learning_rate": 5e-05, "loss": 0.7564, "mean_token_accuracy": 0.8806463479995728, "num_tokens": 435928822.0, "step": 11480 }, { "epoch": 1.4605012084976465, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.965991973876953, "learning_rate": 5e-05, "loss": 0.7982, "mean_token_accuracy": 0.8793396949768066, "num_tokens": 435963906.0, "step": 11481 }, { "epoch": 1.460628418776237, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.196320056915283, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8568578958511353, "num_tokens": 436002826.0, "step": 11482 }, { "epoch": 1.4607556290548276, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.141888618469238, "learning_rate": 5e-05, "loss": 0.7749, "mean_token_accuracy": 0.8872626423835754, "num_tokens": 436039043.0, "step": 11483 }, { "epoch": 1.4608828393334181, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.973443508148193, "learning_rate": 5e-05, "loss": 0.6882, "mean_token_accuracy": 0.8819695115089417, "num_tokens": 436078081.0, "step": 11484 }, { "epoch": 1.4610100496120086, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.390466690063477, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8573343753814697, "num_tokens": 436110415.0, "step": 11485 }, { "epoch": 1.4611372598905992, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.774203777313232, "learning_rate": 5e-05, "loss": 0.7066, "mean_token_accuracy": 0.8812296390533447, "num_tokens": 436152473.0, "step": 11486 }, { "epoch": 1.4612644701691897, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.673999786376953, "learning_rate": 5e-05, "loss": 0.9056, "mean_token_accuracy": 0.8681319952011108, "num_tokens": 436192410.0, "step": 11487 }, { "epoch": 1.4613916804477802, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.682616233825684, "learning_rate": 5e-05, "loss": 0.7565, "mean_token_accuracy": 0.8607113361358643, "num_tokens": 436230605.0, "step": 11488 }, { "epoch": 1.4615188907263708, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.824384689331055, "learning_rate": 5e-05, "loss": 0.8979, "mean_token_accuracy": 0.8795822858810425, "num_tokens": 436267951.0, "step": 11489 }, { "epoch": 1.4616461010049613, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.882205963134766, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8561134338378906, "num_tokens": 436306685.0, "step": 11490 }, { "epoch": 1.4617733112835518, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.598434448242188, "learning_rate": 5e-05, "loss": 0.8898, "mean_token_accuracy": 0.875080943107605, "num_tokens": 436345568.0, "step": 11491 }, { "epoch": 1.4619005215621423, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.348879814147949, "learning_rate": 5e-05, "loss": 0.8223, "mean_token_accuracy": 0.8719806671142578, "num_tokens": 436387580.0, "step": 11492 }, { "epoch": 1.4620277318407329, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.985357284545898, "learning_rate": 5e-05, "loss": 0.8579, "mean_token_accuracy": 0.8677679896354675, "num_tokens": 436423000.0, "step": 11493 }, { "epoch": 1.4621549421193232, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.792174339294434, "learning_rate": 5e-05, "loss": 0.8391, "mean_token_accuracy": 0.874086856842041, "num_tokens": 436464366.0, "step": 11494 }, { "epoch": 1.4622821523979137, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.355370998382568, "learning_rate": 5e-05, "loss": 0.8095, "mean_token_accuracy": 0.8599222898483276, "num_tokens": 436505385.0, "step": 11495 }, { "epoch": 1.4624093626765042, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.125330924987793, "learning_rate": 5e-05, "loss": 0.8612, "mean_token_accuracy": 0.8607050180435181, "num_tokens": 436546251.0, "step": 11496 }, { "epoch": 1.4625365729550948, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.975005626678467, "learning_rate": 5e-05, "loss": 0.7582, "mean_token_accuracy": 0.8607461452484131, "num_tokens": 436582136.0, "step": 11497 }, { "epoch": 1.4626637832336853, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.358016014099121, "learning_rate": 5e-05, "loss": 0.827, "mean_token_accuracy": 0.8750025629997253, "num_tokens": 436620027.0, "step": 11498 }, { "epoch": 1.4627909935122758, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.77365255355835, "learning_rate": 5e-05, "loss": 0.7265, "mean_token_accuracy": 0.8691641092300415, "num_tokens": 436656720.0, "step": 11499 }, { "epoch": 1.4629182037908663, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.561760902404785, "learning_rate": 5e-05, "loss": 0.8466, "mean_token_accuracy": 0.8793268203735352, "num_tokens": 436693813.0, "step": 11500 }, { "epoch": 1.4630454140694569, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.647070407867432, "learning_rate": 5e-05, "loss": 0.6977, "mean_token_accuracy": 0.8764297962188721, "num_tokens": 436732023.0, "step": 11501 }, { "epoch": 1.4631726243480474, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.794602394104004, "learning_rate": 5e-05, "loss": 0.9076, "mean_token_accuracy": 0.8670940399169922, "num_tokens": 436772330.0, "step": 11502 }, { "epoch": 1.4632998346266377, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.650823593139648, "learning_rate": 5e-05, "loss": 0.7338, "mean_token_accuracy": 0.8695993423461914, "num_tokens": 436814593.0, "step": 11503 }, { "epoch": 1.4634270449052282, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.859113693237305, "learning_rate": 5e-05, "loss": 0.9275, "mean_token_accuracy": 0.8725013732910156, "num_tokens": 436849951.0, "step": 11504 }, { "epoch": 1.4635542551838188, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.957244873046875, "learning_rate": 5e-05, "loss": 0.8019, "mean_token_accuracy": 0.8573703765869141, "num_tokens": 436889113.0, "step": 11505 }, { "epoch": 1.4636814654624093, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.542281150817871, "learning_rate": 5e-05, "loss": 0.8667, "mean_token_accuracy": 0.8801644444465637, "num_tokens": 436921009.0, "step": 11506 }, { "epoch": 1.4638086757409998, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.3860554695129395, "learning_rate": 5e-05, "loss": 0.8339, "mean_token_accuracy": 0.8744585514068604, "num_tokens": 436958425.0, "step": 11507 }, { "epoch": 1.4639358860195903, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.041509628295898, "learning_rate": 5e-05, "loss": 0.8473, "mean_token_accuracy": 0.8744568228721619, "num_tokens": 436992995.0, "step": 11508 }, { "epoch": 1.4640630962981809, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.603542804718018, "learning_rate": 5e-05, "loss": 0.7995, "mean_token_accuracy": 0.8801862001419067, "num_tokens": 437028884.0, "step": 11509 }, { "epoch": 1.4641903065767714, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.778626441955566, "learning_rate": 5e-05, "loss": 0.7931, "mean_token_accuracy": 0.8755573034286499, "num_tokens": 437071753.0, "step": 11510 }, { "epoch": 1.464317516855362, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.486291408538818, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8809236288070679, "num_tokens": 437108783.0, "step": 11511 }, { "epoch": 1.4644447271339525, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.9221086502075195, "learning_rate": 5e-05, "loss": 0.799, "mean_token_accuracy": 0.877056360244751, "num_tokens": 437143234.0, "step": 11512 }, { "epoch": 1.464571937412543, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.142175197601318, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.866949200630188, "num_tokens": 437182363.0, "step": 11513 }, { "epoch": 1.4646991476911335, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000537872314453125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.255602836608887, "learning_rate": 5e-05, "loss": 0.8263, "mean_token_accuracy": 0.8772474527359009, "num_tokens": 437224680.0, "step": 11514 }, { "epoch": 1.464826357969724, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.834443092346191, "learning_rate": 5e-05, "loss": 0.7112, "mean_token_accuracy": 0.8735672235488892, "num_tokens": 437261154.0, "step": 11515 }, { "epoch": 1.4649535682483146, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.519787788391113, "learning_rate": 5e-05, "loss": 0.8517, "mean_token_accuracy": 0.8710443377494812, "num_tokens": 437301654.0, "step": 11516 }, { "epoch": 1.465080778526905, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.753890514373779, "learning_rate": 5e-05, "loss": 0.728, "mean_token_accuracy": 0.8709312677383423, "num_tokens": 437338437.0, "step": 11517 }, { "epoch": 1.4652079888054954, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.559104919433594, "learning_rate": 5e-05, "loss": 0.8967, "mean_token_accuracy": 0.8665817975997925, "num_tokens": 437384258.0, "step": 11518 }, { "epoch": 1.465335199084086, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.965997219085693, "learning_rate": 5e-05, "loss": 0.786, "mean_token_accuracy": 0.8600028157234192, "num_tokens": 437421762.0, "step": 11519 }, { "epoch": 1.4654624093626765, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.268001556396484, "learning_rate": 5e-05, "loss": 0.819, "mean_token_accuracy": 0.8797990083694458, "num_tokens": 437456797.0, "step": 11520 }, { "epoch": 1.465589619641267, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.3538641929626465, "learning_rate": 5e-05, "loss": 0.7561, "mean_token_accuracy": 0.8821125626564026, "num_tokens": 437484435.0, "step": 11521 }, { "epoch": 1.4657168299198575, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.730640888214111, "learning_rate": 5e-05, "loss": 0.8118, "mean_token_accuracy": 0.8720600605010986, "num_tokens": 437522573.0, "step": 11522 }, { "epoch": 1.465844040198448, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.7554497718811035, "learning_rate": 5e-05, "loss": 0.8291, "mean_token_accuracy": 0.8607470393180847, "num_tokens": 437560204.0, "step": 11523 }, { "epoch": 1.4659712504770386, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.249871730804443, "learning_rate": 5e-05, "loss": 0.7596, "mean_token_accuracy": 0.8680235147476196, "num_tokens": 437595845.0, "step": 11524 }, { "epoch": 1.466098460755629, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.027515411376953, "learning_rate": 5e-05, "loss": 0.7757, "mean_token_accuracy": 0.8787120580673218, "num_tokens": 437631985.0, "step": 11525 }, { "epoch": 1.4662256710342196, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.966121673583984, "learning_rate": 5e-05, "loss": 0.7574, "mean_token_accuracy": 0.8646630048751831, "num_tokens": 437677024.0, "step": 11526 }, { "epoch": 1.46635288131281, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.231074333190918, "learning_rate": 5e-05, "loss": 0.7586, "mean_token_accuracy": 0.8873528242111206, "num_tokens": 437714980.0, "step": 11527 }, { "epoch": 1.4664800915914005, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.771342754364014, "learning_rate": 5e-05, "loss": 0.7187, "mean_token_accuracy": 0.8696555495262146, "num_tokens": 437752444.0, "step": 11528 }, { "epoch": 1.466607301869991, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.462380409240723, "learning_rate": 5e-05, "loss": 0.8364, "mean_token_accuracy": 0.8658190369606018, "num_tokens": 437787032.0, "step": 11529 }, { "epoch": 1.4667345121485815, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.585529804229736, "learning_rate": 5e-05, "loss": 0.7062, "mean_token_accuracy": 0.8718694448471069, "num_tokens": 437827560.0, "step": 11530 }, { "epoch": 1.466861722427172, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.708367347717285, "learning_rate": 5e-05, "loss": 0.8796, "mean_token_accuracy": 0.867465615272522, "num_tokens": 437865339.0, "step": 11531 }, { "epoch": 1.4669889327057626, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.539758682250977, "learning_rate": 5e-05, "loss": 0.7088, "mean_token_accuracy": 0.8707833290100098, "num_tokens": 437899769.0, "step": 11532 }, { "epoch": 1.467116142984353, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.821581840515137, "learning_rate": 5e-05, "loss": 0.8825, "mean_token_accuracy": 0.8748066425323486, "num_tokens": 437934738.0, "step": 11533 }, { "epoch": 1.4672433532629436, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.764705181121826, "learning_rate": 5e-05, "loss": 0.7608, "mean_token_accuracy": 0.8657580614089966, "num_tokens": 437975948.0, "step": 11534 }, { "epoch": 1.4673705635415342, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.520596504211426, "learning_rate": 5e-05, "loss": 0.87, "mean_token_accuracy": 0.8786418437957764, "num_tokens": 438010081.0, "step": 11535 }, { "epoch": 1.4674977738201247, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.286978244781494, "learning_rate": 5e-05, "loss": 0.8024, "mean_token_accuracy": 0.8755007386207581, "num_tokens": 438044069.0, "step": 11536 }, { "epoch": 1.4676249840987152, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.874939918518066, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.874319314956665, "num_tokens": 438080866.0, "step": 11537 }, { "epoch": 1.4677521943773058, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.769315719604492, "learning_rate": 5e-05, "loss": 0.8195, "mean_token_accuracy": 0.8704776167869568, "num_tokens": 438120550.0, "step": 11538 }, { "epoch": 1.4678794046558963, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.310909271240234, "learning_rate": 5e-05, "loss": 0.7734, "mean_token_accuracy": 0.868640661239624, "num_tokens": 438156968.0, "step": 11539 }, { "epoch": 1.4680066149344868, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.971519470214844, "learning_rate": 5e-05, "loss": 0.812, "mean_token_accuracy": 0.8727227449417114, "num_tokens": 438191470.0, "step": 11540 }, { "epoch": 1.4681338252130773, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.201811790466309, "learning_rate": 5e-05, "loss": 0.7932, "mean_token_accuracy": 0.8590484857559204, "num_tokens": 438225944.0, "step": 11541 }, { "epoch": 1.4682610354916679, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.81553840637207, "learning_rate": 5e-05, "loss": 0.7851, "mean_token_accuracy": 0.8716887831687927, "num_tokens": 438263759.0, "step": 11542 }, { "epoch": 1.4683882457702582, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.3850507736206055, "learning_rate": 5e-05, "loss": 0.7314, "mean_token_accuracy": 0.873038649559021, "num_tokens": 438302371.0, "step": 11543 }, { "epoch": 1.4685154560488487, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.497605323791504, "learning_rate": 5e-05, "loss": 0.8027, "mean_token_accuracy": 0.8551371097564697, "num_tokens": 438342006.0, "step": 11544 }, { "epoch": 1.4686426663274392, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.666054725646973, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8640456199645996, "num_tokens": 438375649.0, "step": 11545 }, { "epoch": 1.4687698766060298, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.193424701690674, "learning_rate": 5e-05, "loss": 0.7055, "mean_token_accuracy": 0.879160463809967, "num_tokens": 438417566.0, "step": 11546 }, { "epoch": 1.4688970868846203, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.945474147796631, "learning_rate": 5e-05, "loss": 0.7456, "mean_token_accuracy": 0.8791924715042114, "num_tokens": 438452219.0, "step": 11547 }, { "epoch": 1.4690242971632108, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.85561466217041, "learning_rate": 5e-05, "loss": 0.6667, "mean_token_accuracy": 0.8824437856674194, "num_tokens": 438490503.0, "step": 11548 }, { "epoch": 1.4691515074418013, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.323935508728027, "learning_rate": 5e-05, "loss": 0.7612, "mean_token_accuracy": 0.8847664594650269, "num_tokens": 438521792.0, "step": 11549 }, { "epoch": 1.4692787177203919, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.488513946533203, "learning_rate": 5e-05, "loss": 0.6625, "mean_token_accuracy": 0.8805617094039917, "num_tokens": 438557408.0, "step": 11550 }, { "epoch": 1.4694059279989824, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.860580444335938, "learning_rate": 5e-05, "loss": 0.8206, "mean_token_accuracy": 0.8859025239944458, "num_tokens": 438601424.0, "step": 11551 }, { "epoch": 1.4695331382775727, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.099843502044678, "learning_rate": 5e-05, "loss": 0.667, "mean_token_accuracy": 0.8753865361213684, "num_tokens": 438638772.0, "step": 11552 }, { "epoch": 1.4696603485561632, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.51612377166748, "learning_rate": 5e-05, "loss": 0.9608, "mean_token_accuracy": 0.8667161464691162, "num_tokens": 438676906.0, "step": 11553 }, { "epoch": 1.4697875588347538, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.12965202331543, "learning_rate": 5e-05, "loss": 0.7538, "mean_token_accuracy": 0.8567020893096924, "num_tokens": 438715790.0, "step": 11554 }, { "epoch": 1.4699147691133443, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.816853523254395, "learning_rate": 5e-05, "loss": 1.0118, "mean_token_accuracy": 0.8762136697769165, "num_tokens": 438757739.0, "step": 11555 }, { "epoch": 1.4700419793919348, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.169428825378418, "learning_rate": 5e-05, "loss": 0.8759, "mean_token_accuracy": 0.8737545013427734, "num_tokens": 438797336.0, "step": 11556 }, { "epoch": 1.4701691896705253, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.934590339660645, "learning_rate": 5e-05, "loss": 0.9834, "mean_token_accuracy": 0.8673931956291199, "num_tokens": 438833452.0, "step": 11557 }, { "epoch": 1.4702963999491159, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.398797988891602, "learning_rate": 5e-05, "loss": 0.9819, "mean_token_accuracy": 0.8730435371398926, "num_tokens": 438874346.0, "step": 11558 }, { "epoch": 1.4704236102277064, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.169672012329102, "learning_rate": 5e-05, "loss": 0.8227, "mean_token_accuracy": 0.8590517640113831, "num_tokens": 438908888.0, "step": 11559 }, { "epoch": 1.470550820506297, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.446901321411133, "learning_rate": 5e-05, "loss": 0.9904, "mean_token_accuracy": 0.8763522505760193, "num_tokens": 438953127.0, "step": 11560 }, { "epoch": 1.4706780307848875, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.354625225067139, "learning_rate": 5e-05, "loss": 0.7399, "mean_token_accuracy": 0.8695331811904907, "num_tokens": 438986485.0, "step": 11561 }, { "epoch": 1.470805241063478, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.044733047485352, "learning_rate": 5e-05, "loss": 1.0606, "mean_token_accuracy": 0.8702837824821472, "num_tokens": 439030958.0, "step": 11562 }, { "epoch": 1.4709324513420685, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.222397804260254, "learning_rate": 5e-05, "loss": 0.9306, "mean_token_accuracy": 0.8652446269989014, "num_tokens": 439069139.0, "step": 11563 }, { "epoch": 1.471059661620659, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.40982723236084, "learning_rate": 5e-05, "loss": 0.9915, "mean_token_accuracy": 0.878838837146759, "num_tokens": 439106526.0, "step": 11564 }, { "epoch": 1.4711868718992496, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.323612213134766, "learning_rate": 5e-05, "loss": 1.0139, "mean_token_accuracy": 0.8735440969467163, "num_tokens": 439146037.0, "step": 11565 }, { "epoch": 1.47131408217784, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.008914947509766, "learning_rate": 5e-05, "loss": 0.8425, "mean_token_accuracy": 0.8816382884979248, "num_tokens": 439177239.0, "step": 11566 }, { "epoch": 1.4714412924564304, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.74700927734375, "learning_rate": 5e-05, "loss": 0.9897, "mean_token_accuracy": 0.8726404905319214, "num_tokens": 439211009.0, "step": 11567 }, { "epoch": 1.471568502735021, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.340209484100342, "learning_rate": 5e-05, "loss": 0.8088, "mean_token_accuracy": 0.8746154308319092, "num_tokens": 439250656.0, "step": 11568 }, { "epoch": 1.4716957130136115, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.603584289550781, "learning_rate": 5e-05, "loss": 0.9876, "mean_token_accuracy": 0.8537151217460632, "num_tokens": 439285103.0, "step": 11569 }, { "epoch": 1.471822923292202, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.533341884613037, "learning_rate": 5e-05, "loss": 0.9029, "mean_token_accuracy": 0.8637580871582031, "num_tokens": 439325104.0, "step": 11570 }, { "epoch": 1.4719501335707925, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.940208435058594, "learning_rate": 5e-05, "loss": 0.8762, "mean_token_accuracy": 0.8601006269454956, "num_tokens": 439357760.0, "step": 11571 }, { "epoch": 1.472077343849383, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.833358287811279, "learning_rate": 5e-05, "loss": 0.8818, "mean_token_accuracy": 0.8676360845565796, "num_tokens": 439397823.0, "step": 11572 }, { "epoch": 1.4722045541279736, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.471194267272949, "learning_rate": 5e-05, "loss": 0.7954, "mean_token_accuracy": 0.8632670640945435, "num_tokens": 439435989.0, "step": 11573 }, { "epoch": 1.472331764406564, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.938430309295654, "learning_rate": 5e-05, "loss": 0.8288, "mean_token_accuracy": 0.8798401355743408, "num_tokens": 439474272.0, "step": 11574 }, { "epoch": 1.4724589746851546, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.225474834442139, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.8828741312026978, "num_tokens": 439512729.0, "step": 11575 }, { "epoch": 1.472586184963745, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.162128448486328, "learning_rate": 5e-05, "loss": 0.8633, "mean_token_accuracy": 0.863560676574707, "num_tokens": 439550368.0, "step": 11576 }, { "epoch": 1.4727133952423355, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.887257099151611, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.8700546622276306, "num_tokens": 439591244.0, "step": 11577 }, { "epoch": 1.472840605520926, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.569007873535156, "learning_rate": 5e-05, "loss": 0.8763, "mean_token_accuracy": 0.8679351806640625, "num_tokens": 439632389.0, "step": 11578 }, { "epoch": 1.4729678157995165, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.599920272827148, "learning_rate": 5e-05, "loss": 0.7555, "mean_token_accuracy": 0.8603627681732178, "num_tokens": 439674543.0, "step": 11579 }, { "epoch": 1.473095026078107, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.892680168151855, "learning_rate": 5e-05, "loss": 0.9152, "mean_token_accuracy": 0.8707737922668457, "num_tokens": 439709009.0, "step": 11580 }, { "epoch": 1.4732222363566976, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.575165748596191, "learning_rate": 5e-05, "loss": 0.7279, "mean_token_accuracy": 0.8774157166481018, "num_tokens": 439754154.0, "step": 11581 }, { "epoch": 1.473349446635288, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.026771545410156, "learning_rate": 5e-05, "loss": 0.8889, "mean_token_accuracy": 0.8912283778190613, "num_tokens": 439794986.0, "step": 11582 }, { "epoch": 1.4734766569138786, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.749215126037598, "learning_rate": 5e-05, "loss": 0.7916, "mean_token_accuracy": 0.8662493228912354, "num_tokens": 439833402.0, "step": 11583 }, { "epoch": 1.4736038671924692, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.017982482910156, "learning_rate": 5e-05, "loss": 0.911, "mean_token_accuracy": 0.8835712671279907, "num_tokens": 439878670.0, "step": 11584 }, { "epoch": 1.4737310774710597, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.08172082901001, "learning_rate": 5e-05, "loss": 0.8256, "mean_token_accuracy": 0.8789135217666626, "num_tokens": 439916505.0, "step": 11585 }, { "epoch": 1.4738582877496502, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.736828804016113, "learning_rate": 5e-05, "loss": 0.9334, "mean_token_accuracy": 0.8746075630187988, "num_tokens": 439955833.0, "step": 11586 }, { "epoch": 1.4739854980282407, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.373744964599609, "learning_rate": 5e-05, "loss": 0.8192, "mean_token_accuracy": 0.8838078379631042, "num_tokens": 439994105.0, "step": 11587 }, { "epoch": 1.4741127083068313, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.362649917602539, "learning_rate": 5e-05, "loss": 0.8873, "mean_token_accuracy": 0.8775817155838013, "num_tokens": 440030298.0, "step": 11588 }, { "epoch": 1.4742399185854218, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.45435905456543, "learning_rate": 5e-05, "loss": 0.8273, "mean_token_accuracy": 0.8786978125572205, "num_tokens": 440064792.0, "step": 11589 }, { "epoch": 1.4743671288640123, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.188207626342773, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.8716039657592773, "num_tokens": 440097886.0, "step": 11590 }, { "epoch": 1.4744943391426026, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.3014655113220215, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8780126571655273, "num_tokens": 440133488.0, "step": 11591 }, { "epoch": 1.4746215494211932, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.284064292907715, "learning_rate": 5e-05, "loss": 0.8653, "mean_token_accuracy": 0.878615140914917, "num_tokens": 440169448.0, "step": 11592 }, { "epoch": 1.4747487596997837, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.106223106384277, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.881315290927887, "num_tokens": 440204755.0, "step": 11593 }, { "epoch": 1.4748759699783742, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.370805740356445, "learning_rate": 5e-05, "loss": 0.8981, "mean_token_accuracy": 0.86968594789505, "num_tokens": 440239129.0, "step": 11594 }, { "epoch": 1.4750031802569648, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.067394256591797, "learning_rate": 5e-05, "loss": 0.772, "mean_token_accuracy": 0.8729556798934937, "num_tokens": 440280962.0, "step": 11595 }, { "epoch": 1.4751303905355553, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.323139190673828, "learning_rate": 5e-05, "loss": 0.8639, "mean_token_accuracy": 0.878937304019928, "num_tokens": 440315900.0, "step": 11596 }, { "epoch": 1.4752576008141458, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.173397541046143, "learning_rate": 5e-05, "loss": 0.7498, "mean_token_accuracy": 0.8801614046096802, "num_tokens": 440356699.0, "step": 11597 }, { "epoch": 1.4753848110927363, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.10140323638916, "learning_rate": 5e-05, "loss": 0.8288, "mean_token_accuracy": 0.8793108463287354, "num_tokens": 440398655.0, "step": 11598 }, { "epoch": 1.4755120213713269, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.371309280395508, "learning_rate": 5e-05, "loss": 0.8262, "mean_token_accuracy": 0.8626098036766052, "num_tokens": 440436204.0, "step": 11599 }, { "epoch": 1.4756392316499174, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.798550128936768, "learning_rate": 5e-05, "loss": 0.7996, "mean_token_accuracy": 0.8738707304000854, "num_tokens": 440478363.0, "step": 11600 }, { "epoch": 1.4757664419285077, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.501652240753174, "learning_rate": 5e-05, "loss": 0.7433, "mean_token_accuracy": 0.8827073574066162, "num_tokens": 440515454.0, "step": 11601 }, { "epoch": 1.4758936522070982, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.6284403800964355, "learning_rate": 5e-05, "loss": 0.7915, "mean_token_accuracy": 0.8735138177871704, "num_tokens": 440553526.0, "step": 11602 }, { "epoch": 1.4760208624856888, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.500696659088135, "learning_rate": 5e-05, "loss": 0.7903, "mean_token_accuracy": 0.8616769313812256, "num_tokens": 440591446.0, "step": 11603 }, { "epoch": 1.4761480727642793, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.583827018737793, "learning_rate": 5e-05, "loss": 0.7372, "mean_token_accuracy": 0.8806540369987488, "num_tokens": 440630505.0, "step": 11604 }, { "epoch": 1.4762752830428698, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.450806617736816, "learning_rate": 5e-05, "loss": 0.6985, "mean_token_accuracy": 0.8849477171897888, "num_tokens": 440669318.0, "step": 11605 }, { "epoch": 1.4764024933214603, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.598116874694824, "learning_rate": 5e-05, "loss": 0.7261, "mean_token_accuracy": 0.8805068135261536, "num_tokens": 440711218.0, "step": 11606 }, { "epoch": 1.4765297036000509, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.378611087799072, "learning_rate": 5e-05, "loss": 0.726, "mean_token_accuracy": 0.8720591068267822, "num_tokens": 440754581.0, "step": 11607 }, { "epoch": 1.4766569138786414, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.65448522567749, "learning_rate": 5e-05, "loss": 0.7513, "mean_token_accuracy": 0.8725708723068237, "num_tokens": 440795714.0, "step": 11608 }, { "epoch": 1.476784124157232, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.273101806640625, "learning_rate": 5e-05, "loss": 0.6705, "mean_token_accuracy": 0.8861178159713745, "num_tokens": 440827423.0, "step": 11609 }, { "epoch": 1.4769113344358225, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.726113319396973, "learning_rate": 5e-05, "loss": 0.7842, "mean_token_accuracy": 0.8625892400741577, "num_tokens": 440868857.0, "step": 11610 }, { "epoch": 1.477038544714413, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.1907196044921875, "learning_rate": 5e-05, "loss": 0.7544, "mean_token_accuracy": 0.8589589595794678, "num_tokens": 440904788.0, "step": 11611 }, { "epoch": 1.4771657549930035, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.7939324378967285, "learning_rate": 5e-05, "loss": 0.7531, "mean_token_accuracy": 0.870221734046936, "num_tokens": 440949013.0, "step": 11612 }, { "epoch": 1.477292965271594, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.09732723236084, "learning_rate": 5e-05, "loss": 0.7453, "mean_token_accuracy": 0.8615896105766296, "num_tokens": 440987392.0, "step": 11613 }, { "epoch": 1.4774201755501846, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.929421424865723, "learning_rate": 5e-05, "loss": 0.6745, "mean_token_accuracy": 0.8986145257949829, "num_tokens": 441024917.0, "step": 11614 }, { "epoch": 1.477547385828775, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.919158458709717, "learning_rate": 5e-05, "loss": 0.6824, "mean_token_accuracy": 0.8788384199142456, "num_tokens": 441066355.0, "step": 11615 }, { "epoch": 1.4776745961073654, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.187629699707031, "learning_rate": 5e-05, "loss": 0.795, "mean_token_accuracy": 0.8653109073638916, "num_tokens": 441100192.0, "step": 11616 }, { "epoch": 1.477801806385956, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.629277229309082, "learning_rate": 5e-05, "loss": 0.7127, "mean_token_accuracy": 0.8630634546279907, "num_tokens": 441144627.0, "step": 11617 }, { "epoch": 1.4779290166645465, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.584639549255371, "learning_rate": 5e-05, "loss": 0.8053, "mean_token_accuracy": 0.8777891993522644, "num_tokens": 441181656.0, "step": 11618 }, { "epoch": 1.478056226943137, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.332967758178711, "learning_rate": 5e-05, "loss": 0.6967, "mean_token_accuracy": 0.8687154054641724, "num_tokens": 441215476.0, "step": 11619 }, { "epoch": 1.4781834372217275, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 9.03191089630127, "learning_rate": 5e-05, "loss": 0.9047, "mean_token_accuracy": 0.8661655187606812, "num_tokens": 441262919.0, "step": 11620 }, { "epoch": 1.478310647500318, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.252559661865234, "learning_rate": 5e-05, "loss": 0.7103, "mean_token_accuracy": 0.8693547248840332, "num_tokens": 441302602.0, "step": 11621 }, { "epoch": 1.4784378577789086, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.207475662231445, "learning_rate": 5e-05, "loss": 0.9384, "mean_token_accuracy": 0.8755357265472412, "num_tokens": 441335432.0, "step": 11622 }, { "epoch": 1.478565068057499, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.795844554901123, "learning_rate": 5e-05, "loss": 0.8068, "mean_token_accuracy": 0.8639225959777832, "num_tokens": 441377849.0, "step": 11623 }, { "epoch": 1.4786922783360896, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.527313232421875, "learning_rate": 5e-05, "loss": 0.9045, "mean_token_accuracy": 0.8719502687454224, "num_tokens": 441413225.0, "step": 11624 }, { "epoch": 1.47881948861468, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.848501205444336, "learning_rate": 5e-05, "loss": 0.8651, "mean_token_accuracy": 0.8780779242515564, "num_tokens": 441448041.0, "step": 11625 }, { "epoch": 1.4789466988932705, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.028653144836426, "learning_rate": 5e-05, "loss": 0.7147, "mean_token_accuracy": 0.8831868171691895, "num_tokens": 441483919.0, "step": 11626 }, { "epoch": 1.479073909171861, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.136772155761719, "learning_rate": 5e-05, "loss": 0.9481, "mean_token_accuracy": 0.8687191009521484, "num_tokens": 441518313.0, "step": 11627 }, { "epoch": 1.4792011194504515, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 5.8251471519470215, "learning_rate": 5e-05, "loss": 0.6895, "mean_token_accuracy": 0.869076132774353, "num_tokens": 441557248.0, "step": 11628 }, { "epoch": 1.479328329729042, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.594449996948242, "learning_rate": 5e-05, "loss": 1.1621, "mean_token_accuracy": 0.8432055711746216, "num_tokens": 441599759.0, "step": 11629 }, { "epoch": 1.4794555400076326, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.401708602905273, "learning_rate": 5e-05, "loss": 0.8047, "mean_token_accuracy": 0.8682658672332764, "num_tokens": 441631643.0, "step": 11630 }, { "epoch": 1.479582750286223, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 11.269415855407715, "learning_rate": 5e-05, "loss": 1.264, "mean_token_accuracy": 0.8611127734184265, "num_tokens": 441669804.0, "step": 11631 }, { "epoch": 1.4797099605648136, "ewc_loss": 0.84375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 8.944551467895508, "learning_rate": 5e-05, "loss": 1.2476, "mean_token_accuracy": 0.8603261113166809, "num_tokens": 441708661.0, "step": 11632 }, { "epoch": 1.4798371708434042, "ewc_loss": 0.625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.626917839050293, "learning_rate": 5e-05, "loss": 0.9863, "mean_token_accuracy": 0.8721573948860168, "num_tokens": 441745510.0, "step": 11633 }, { "epoch": 1.4799643811219947, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 9.52197551727295, "learning_rate": 5e-05, "loss": 1.196, "mean_token_accuracy": 0.8716068267822266, "num_tokens": 441781482.0, "step": 11634 }, { "epoch": 1.4800915914005852, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.022000312805176, "learning_rate": 5e-05, "loss": 0.8624, "mean_token_accuracy": 0.8822616934776306, "num_tokens": 441814299.0, "step": 11635 }, { "epoch": 1.4802188016791757, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.019536972045898, "learning_rate": 5e-05, "loss": 1.1111, "mean_token_accuracy": 0.8577220439910889, "num_tokens": 441859742.0, "step": 11636 }, { "epoch": 1.4803460119577663, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.7608232498168945, "learning_rate": 5e-05, "loss": 0.9246, "mean_token_accuracy": 0.8821045160293579, "num_tokens": 441896563.0, "step": 11637 }, { "epoch": 1.4804732222363568, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.745100975036621, "learning_rate": 5e-05, "loss": 0.9955, "mean_token_accuracy": 0.8797914981842041, "num_tokens": 441935380.0, "step": 11638 }, { "epoch": 1.4806004325149473, "ewc_loss": 0.625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.125635147094727, "learning_rate": 5e-05, "loss": 1.0262, "mean_token_accuracy": 0.8626905083656311, "num_tokens": 441977077.0, "step": 11639 }, { "epoch": 1.4807276427935376, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.709546089172363, "learning_rate": 5e-05, "loss": 0.9195, "mean_token_accuracy": 0.8724340200424194, "num_tokens": 442020538.0, "step": 11640 }, { "epoch": 1.4808548530721282, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.35232162475586, "learning_rate": 5e-05, "loss": 0.9517, "mean_token_accuracy": 0.8736999034881592, "num_tokens": 442057093.0, "step": 11641 }, { "epoch": 1.4809820633507187, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.835127353668213, "learning_rate": 5e-05, "loss": 0.853, "mean_token_accuracy": 0.8673579096794128, "num_tokens": 442092364.0, "step": 11642 }, { "epoch": 1.4811092736293092, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.369833946228027, "learning_rate": 5e-05, "loss": 0.9167, "mean_token_accuracy": 0.8665083050727844, "num_tokens": 442134181.0, "step": 11643 }, { "epoch": 1.4812364839078997, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.491775035858154, "learning_rate": 5e-05, "loss": 0.8629, "mean_token_accuracy": 0.856110692024231, "num_tokens": 442174987.0, "step": 11644 }, { "epoch": 1.4813636941864903, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.21371841430664, "learning_rate": 5e-05, "loss": 0.8217, "mean_token_accuracy": 0.8763846755027771, "num_tokens": 442204214.0, "step": 11645 }, { "epoch": 1.4814909044650808, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.423637390136719, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.8813067674636841, "num_tokens": 442243227.0, "step": 11646 }, { "epoch": 1.4816181147436713, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.942698001861572, "learning_rate": 5e-05, "loss": 0.8114, "mean_token_accuracy": 0.8742307424545288, "num_tokens": 442282107.0, "step": 11647 }, { "epoch": 1.4817453250222619, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.524518013000488, "learning_rate": 5e-05, "loss": 0.783, "mean_token_accuracy": 0.8809806108474731, "num_tokens": 442320445.0, "step": 11648 }, { "epoch": 1.4818725353008524, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.597455024719238, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8675386905670166, "num_tokens": 442358626.0, "step": 11649 }, { "epoch": 1.4819997455794427, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.744818687438965, "learning_rate": 5e-05, "loss": 0.7932, "mean_token_accuracy": 0.8712800741195679, "num_tokens": 442405095.0, "step": 11650 }, { "epoch": 1.4821269558580332, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.224771499633789, "learning_rate": 5e-05, "loss": 0.7295, "mean_token_accuracy": 0.8703736066818237, "num_tokens": 442442131.0, "step": 11651 }, { "epoch": 1.4822541661366238, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.133462905883789, "learning_rate": 5e-05, "loss": 0.8522, "mean_token_accuracy": 0.8599574565887451, "num_tokens": 442479660.0, "step": 11652 }, { "epoch": 1.4823813764152143, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.759232997894287, "learning_rate": 5e-05, "loss": 0.7185, "mean_token_accuracy": 0.8651291131973267, "num_tokens": 442523659.0, "step": 11653 }, { "epoch": 1.4825085866938048, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.68156909942627, "learning_rate": 5e-05, "loss": 0.8659, "mean_token_accuracy": 0.8666672706604004, "num_tokens": 442564445.0, "step": 11654 }, { "epoch": 1.4826357969723953, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.357537269592285, "learning_rate": 5e-05, "loss": 0.6907, "mean_token_accuracy": 0.8733212351799011, "num_tokens": 442597662.0, "step": 11655 }, { "epoch": 1.4827630072509859, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.271224975585938, "learning_rate": 5e-05, "loss": 0.9346, "mean_token_accuracy": 0.8677300214767456, "num_tokens": 442633365.0, "step": 11656 }, { "epoch": 1.4828902175295764, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.253366947174072, "learning_rate": 5e-05, "loss": 0.7108, "mean_token_accuracy": 0.8711476922035217, "num_tokens": 442668389.0, "step": 11657 }, { "epoch": 1.483017427808167, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.65512752532959, "learning_rate": 5e-05, "loss": 0.9822, "mean_token_accuracy": 0.8726259469985962, "num_tokens": 442695089.0, "step": 11658 }, { "epoch": 1.4831446380867574, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 6.832946300506592, "learning_rate": 5e-05, "loss": 0.823, "mean_token_accuracy": 0.8730235695838928, "num_tokens": 442738602.0, "step": 11659 }, { "epoch": 1.483271848365348, "ewc_loss": 0.625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.360751152038574, "learning_rate": 5e-05, "loss": 1.007, "mean_token_accuracy": 0.8694522976875305, "num_tokens": 442777400.0, "step": 11660 }, { "epoch": 1.4833990586439385, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.72791051864624, "learning_rate": 5e-05, "loss": 0.9357, "mean_token_accuracy": 0.8752887845039368, "num_tokens": 442815502.0, "step": 11661 }, { "epoch": 1.483526268922529, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.328753471374512, "learning_rate": 5e-05, "loss": 0.9183, "mean_token_accuracy": 0.8628923892974854, "num_tokens": 442858418.0, "step": 11662 }, { "epoch": 1.4836534792011196, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.21888542175293, "learning_rate": 5e-05, "loss": 0.9006, "mean_token_accuracy": 0.8867592811584473, "num_tokens": 442896073.0, "step": 11663 }, { "epoch": 1.48378068947971, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.487452983856201, "learning_rate": 5e-05, "loss": 0.8024, "mean_token_accuracy": 0.870769739151001, "num_tokens": 442940719.0, "step": 11664 }, { "epoch": 1.4839078997583004, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.244275093078613, "learning_rate": 5e-05, "loss": 0.8539, "mean_token_accuracy": 0.8809146881103516, "num_tokens": 442977724.0, "step": 11665 }, { "epoch": 1.484035110036891, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.4065728187561035, "learning_rate": 5e-05, "loss": 0.7718, "mean_token_accuracy": 0.8812367916107178, "num_tokens": 443017754.0, "step": 11666 }, { "epoch": 1.4841623203154815, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.821644306182861, "learning_rate": 5e-05, "loss": 0.8096, "mean_token_accuracy": 0.874057412147522, "num_tokens": 443048661.0, "step": 11667 }, { "epoch": 1.484289530594072, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.698702812194824, "learning_rate": 5e-05, "loss": 0.8211, "mean_token_accuracy": 0.8715865015983582, "num_tokens": 443094445.0, "step": 11668 }, { "epoch": 1.4844167408726625, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.341042995452881, "learning_rate": 5e-05, "loss": 0.7178, "mean_token_accuracy": 0.8828511238098145, "num_tokens": 443134468.0, "step": 11669 }, { "epoch": 1.484543951151253, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.012446403503418, "learning_rate": 5e-05, "loss": 0.8013, "mean_token_accuracy": 0.8749530911445618, "num_tokens": 443174458.0, "step": 11670 }, { "epoch": 1.4846711614298436, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.000129222869873, "learning_rate": 5e-05, "loss": 0.6948, "mean_token_accuracy": 0.8802481889724731, "num_tokens": 443212421.0, "step": 11671 }, { "epoch": 1.484798371708434, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.27145004272461, "learning_rate": 5e-05, "loss": 0.9081, "mean_token_accuracy": 0.8471672534942627, "num_tokens": 443252083.0, "step": 11672 }, { "epoch": 1.4849255819870246, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.775139331817627, "learning_rate": 5e-05, "loss": 0.728, "mean_token_accuracy": 0.8649731874465942, "num_tokens": 443286776.0, "step": 11673 }, { "epoch": 1.485052792265615, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.54212474822998, "learning_rate": 5e-05, "loss": 0.8465, "mean_token_accuracy": 0.8769762516021729, "num_tokens": 443325734.0, "step": 11674 }, { "epoch": 1.4851800025442055, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.630285263061523, "learning_rate": 5e-05, "loss": 0.7122, "mean_token_accuracy": 0.8718907237052917, "num_tokens": 443362637.0, "step": 11675 }, { "epoch": 1.485307212822796, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.759390830993652, "learning_rate": 5e-05, "loss": 0.8623, "mean_token_accuracy": 0.8797646164894104, "num_tokens": 443401629.0, "step": 11676 }, { "epoch": 1.4854344231013865, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.660501480102539, "learning_rate": 5e-05, "loss": 0.6799, "mean_token_accuracy": 0.8868753910064697, "num_tokens": 443437619.0, "step": 11677 }, { "epoch": 1.485561633379977, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.747856140136719, "learning_rate": 5e-05, "loss": 0.9376, "mean_token_accuracy": 0.8574424982070923, "num_tokens": 443472175.0, "step": 11678 }, { "epoch": 1.4856888436585676, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.930171966552734, "learning_rate": 5e-05, "loss": 0.7897, "mean_token_accuracy": 0.8660669922828674, "num_tokens": 443506904.0, "step": 11679 }, { "epoch": 1.485816053937158, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.473706245422363, "learning_rate": 5e-05, "loss": 0.9188, "mean_token_accuracy": 0.8609116077423096, "num_tokens": 443548141.0, "step": 11680 }, { "epoch": 1.4859432642157486, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.244416236877441, "learning_rate": 5e-05, "loss": 0.8118, "mean_token_accuracy": 0.8722152709960938, "num_tokens": 443586609.0, "step": 11681 }, { "epoch": 1.4860704744943392, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.086572647094727, "learning_rate": 5e-05, "loss": 0.8513, "mean_token_accuracy": 0.8755767941474915, "num_tokens": 443628792.0, "step": 11682 }, { "epoch": 1.4861976847729297, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.446217060089111, "learning_rate": 5e-05, "loss": 0.8685, "mean_token_accuracy": 0.8550726175308228, "num_tokens": 443671336.0, "step": 11683 }, { "epoch": 1.4863248950515202, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.803643703460693, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8717041611671448, "num_tokens": 443708987.0, "step": 11684 }, { "epoch": 1.4864521053301107, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.462719440460205, "learning_rate": 5e-05, "loss": 0.7846, "mean_token_accuracy": 0.8727403879165649, "num_tokens": 443749968.0, "step": 11685 }, { "epoch": 1.4865793156087013, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.736886978149414, "learning_rate": 5e-05, "loss": 0.7878, "mean_token_accuracy": 0.8746099472045898, "num_tokens": 443786287.0, "step": 11686 }, { "epoch": 1.4867065258872918, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.360505104064941, "learning_rate": 5e-05, "loss": 0.7231, "mean_token_accuracy": 0.885010838508606, "num_tokens": 443820729.0, "step": 11687 }, { "epoch": 1.4868337361658823, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.767825603485107, "learning_rate": 5e-05, "loss": 0.7405, "mean_token_accuracy": 0.8826938271522522, "num_tokens": 443859401.0, "step": 11688 }, { "epoch": 1.4869609464444726, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.276778221130371, "learning_rate": 5e-05, "loss": 0.7041, "mean_token_accuracy": 0.890068769454956, "num_tokens": 443895697.0, "step": 11689 }, { "epoch": 1.4870881567230632, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.754297733306885, "learning_rate": 5e-05, "loss": 0.7615, "mean_token_accuracy": 0.8767211437225342, "num_tokens": 443935229.0, "step": 11690 }, { "epoch": 1.4872153670016537, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.264350414276123, "learning_rate": 5e-05, "loss": 0.7559, "mean_token_accuracy": 0.8709384202957153, "num_tokens": 443967544.0, "step": 11691 }, { "epoch": 1.4873425772802442, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.723159313201904, "learning_rate": 5e-05, "loss": 0.7763, "mean_token_accuracy": 0.8692417144775391, "num_tokens": 444006973.0, "step": 11692 }, { "epoch": 1.4874697875588347, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.2632222175598145, "learning_rate": 5e-05, "loss": 0.7009, "mean_token_accuracy": 0.8807424306869507, "num_tokens": 444041143.0, "step": 11693 }, { "epoch": 1.4875969978374253, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.703484058380127, "learning_rate": 5e-05, "loss": 0.7874, "mean_token_accuracy": 0.8640848398208618, "num_tokens": 444086279.0, "step": 11694 }, { "epoch": 1.4877242081160158, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.240061283111572, "learning_rate": 5e-05, "loss": 0.7254, "mean_token_accuracy": 0.8732023239135742, "num_tokens": 444123559.0, "step": 11695 }, { "epoch": 1.4878514183946063, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.721409320831299, "learning_rate": 5e-05, "loss": 0.7842, "mean_token_accuracy": 0.8621985912322998, "num_tokens": 444159197.0, "step": 11696 }, { "epoch": 1.4879786286731969, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.1875319480896, "learning_rate": 5e-05, "loss": 0.7245, "mean_token_accuracy": 0.8704552054405212, "num_tokens": 444190217.0, "step": 11697 }, { "epoch": 1.4881058389517874, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.773654937744141, "learning_rate": 5e-05, "loss": 0.7305, "mean_token_accuracy": 0.8881211280822754, "num_tokens": 444225241.0, "step": 11698 }, { "epoch": 1.4882330492303777, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.102578639984131, "learning_rate": 5e-05, "loss": 0.7492, "mean_token_accuracy": 0.8592813014984131, "num_tokens": 444263702.0, "step": 11699 }, { "epoch": 1.4883602595089682, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.898914813995361, "learning_rate": 5e-05, "loss": 0.7466, "mean_token_accuracy": 0.8735532760620117, "num_tokens": 444303911.0, "step": 11700 }, { "epoch": 1.4884874697875587, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.912722587585449, "learning_rate": 5e-05, "loss": 0.6954, "mean_token_accuracy": 0.8715616464614868, "num_tokens": 444344705.0, "step": 11701 }, { "epoch": 1.4886146800661493, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.13573169708252, "learning_rate": 5e-05, "loss": 0.7436, "mean_token_accuracy": 0.8857944011688232, "num_tokens": 444381598.0, "step": 11702 }, { "epoch": 1.4887418903447398, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.659235000610352, "learning_rate": 5e-05, "loss": 0.6721, "mean_token_accuracy": 0.8757492899894714, "num_tokens": 444413077.0, "step": 11703 }, { "epoch": 1.4888691006233303, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.50893783569336, "learning_rate": 5e-05, "loss": 0.8076, "mean_token_accuracy": 0.8751212358474731, "num_tokens": 444453028.0, "step": 11704 }, { "epoch": 1.4889963109019209, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.342066287994385, "learning_rate": 5e-05, "loss": 0.6883, "mean_token_accuracy": 0.8700845837593079, "num_tokens": 444486382.0, "step": 11705 }, { "epoch": 1.4891235211805114, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.9945650100708, "learning_rate": 5e-05, "loss": 0.8464, "mean_token_accuracy": 0.8804377913475037, "num_tokens": 444523299.0, "step": 11706 }, { "epoch": 1.489250731459102, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.1566996574401855, "learning_rate": 5e-05, "loss": 0.6783, "mean_token_accuracy": 0.8752342462539673, "num_tokens": 444561772.0, "step": 11707 }, { "epoch": 1.4893779417376924, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.346410751342773, "learning_rate": 5e-05, "loss": 0.9962, "mean_token_accuracy": 0.8592888712882996, "num_tokens": 444603056.0, "step": 11708 }, { "epoch": 1.489505152016283, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.581920623779297, "learning_rate": 5e-05, "loss": 0.7259, "mean_token_accuracy": 0.883073091506958, "num_tokens": 444640469.0, "step": 11709 }, { "epoch": 1.4896323622948735, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.885326385498047, "learning_rate": 5e-05, "loss": 0.9411, "mean_token_accuracy": 0.8725313544273376, "num_tokens": 444680737.0, "step": 11710 }, { "epoch": 1.489759572573464, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.675840377807617, "learning_rate": 5e-05, "loss": 0.8496, "mean_token_accuracy": 0.8845243453979492, "num_tokens": 444713828.0, "step": 11711 }, { "epoch": 1.4898867828520546, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.346078872680664, "learning_rate": 5e-05, "loss": 0.7998, "mean_token_accuracy": 0.8633835315704346, "num_tokens": 444755823.0, "step": 11712 }, { "epoch": 1.490013993130645, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.022945404052734, "learning_rate": 5e-05, "loss": 0.9459, "mean_token_accuracy": 0.8767493367195129, "num_tokens": 444791558.0, "step": 11713 }, { "epoch": 1.4901412034092354, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 5.8850250244140625, "learning_rate": 5e-05, "loss": 0.6682, "mean_token_accuracy": 0.8779993653297424, "num_tokens": 444832037.0, "step": 11714 }, { "epoch": 1.490268413687826, "ewc_loss": 0.71875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.710346221923828, "learning_rate": 5e-05, "loss": 1.1303, "mean_token_accuracy": 0.8597661256790161, "num_tokens": 444870092.0, "step": 11715 }, { "epoch": 1.4903956239664164, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 6.354718208312988, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8721129894256592, "num_tokens": 444909267.0, "step": 11716 }, { "epoch": 1.490522834245007, "ewc_loss": 0.93359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00087738037109375, "grad_norm": 11.744118690490723, "learning_rate": 5e-05, "loss": 1.313, "mean_token_accuracy": 0.8695775270462036, "num_tokens": 444946851.0, "step": 11717 }, { "epoch": 1.4906500445235975, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 9.292447090148926, "learning_rate": 5e-05, "loss": 1.2303, "mean_token_accuracy": 0.8875499367713928, "num_tokens": 444986324.0, "step": 11718 }, { "epoch": 1.490777254802188, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.155521392822266, "learning_rate": 5e-05, "loss": 0.9989, "mean_token_accuracy": 0.8801707029342651, "num_tokens": 445020443.0, "step": 11719 }, { "epoch": 1.4909044650807786, "ewc_loss": 0.86328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0008087158203125, "grad_norm": 9.579787254333496, "learning_rate": 5e-05, "loss": 1.2735, "mean_token_accuracy": 0.8564558029174805, "num_tokens": 445057592.0, "step": 11720 }, { "epoch": 1.491031675359369, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.764083862304688, "learning_rate": 5e-05, "loss": 0.9292, "mean_token_accuracy": 0.886625349521637, "num_tokens": 445091513.0, "step": 11721 }, { "epoch": 1.4911588856379596, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.62618637084961, "learning_rate": 5e-05, "loss": 1.0878, "mean_token_accuracy": 0.8574342727661133, "num_tokens": 445132579.0, "step": 11722 }, { "epoch": 1.49128609591655, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.015961647033691, "learning_rate": 5e-05, "loss": 1.0157, "mean_token_accuracy": 0.8521288633346558, "num_tokens": 445167336.0, "step": 11723 }, { "epoch": 1.4914133061951405, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.381409645080566, "learning_rate": 5e-05, "loss": 1.017, "mean_token_accuracy": 0.8642743825912476, "num_tokens": 445208873.0, "step": 11724 }, { "epoch": 1.491540516473731, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 7.92694091796875, "learning_rate": 5e-05, "loss": 0.9471, "mean_token_accuracy": 0.8716408014297485, "num_tokens": 445245649.0, "step": 11725 }, { "epoch": 1.4916677267523215, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.605236053466797, "learning_rate": 5e-05, "loss": 0.8653, "mean_token_accuracy": 0.8871186971664429, "num_tokens": 445285890.0, "step": 11726 }, { "epoch": 1.491794937030912, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.9838972091674805, "learning_rate": 5e-05, "loss": 0.9132, "mean_token_accuracy": 0.8651121854782104, "num_tokens": 445323408.0, "step": 11727 }, { "epoch": 1.4919221473095026, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.982182025909424, "learning_rate": 5e-05, "loss": 0.8252, "mean_token_accuracy": 0.8804479837417603, "num_tokens": 445361086.0, "step": 11728 }, { "epoch": 1.492049357588093, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.881953716278076, "learning_rate": 5e-05, "loss": 0.8339, "mean_token_accuracy": 0.8748785257339478, "num_tokens": 445404726.0, "step": 11729 }, { "epoch": 1.4921765678666836, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.751874923706055, "learning_rate": 5e-05, "loss": 0.8099, "mean_token_accuracy": 0.8734362125396729, "num_tokens": 445444190.0, "step": 11730 }, { "epoch": 1.4923037781452742, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.646454811096191, "learning_rate": 5e-05, "loss": 0.8183, "mean_token_accuracy": 0.8649296760559082, "num_tokens": 445484395.0, "step": 11731 }, { "epoch": 1.4924309884238647, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.770577907562256, "learning_rate": 5e-05, "loss": 0.8031, "mean_token_accuracy": 0.8686492443084717, "num_tokens": 445518104.0, "step": 11732 }, { "epoch": 1.4925581987024552, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.3721022605896, "learning_rate": 5e-05, "loss": 0.7733, "mean_token_accuracy": 0.8651036620140076, "num_tokens": 445554761.0, "step": 11733 }, { "epoch": 1.4926854089810457, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.967644214630127, "learning_rate": 5e-05, "loss": 0.8226, "mean_token_accuracy": 0.8663969039916992, "num_tokens": 445595331.0, "step": 11734 }, { "epoch": 1.4928126192596363, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.008816242218018, "learning_rate": 5e-05, "loss": 0.6967, "mean_token_accuracy": 0.8779882192611694, "num_tokens": 445639280.0, "step": 11735 }, { "epoch": 1.4929398295382268, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.3731107711792, "learning_rate": 5e-05, "loss": 0.8276, "mean_token_accuracy": 0.873268723487854, "num_tokens": 445675437.0, "step": 11736 }, { "epoch": 1.4930670398168173, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.576916694641113, "learning_rate": 5e-05, "loss": 0.6923, "mean_token_accuracy": 0.8751794099807739, "num_tokens": 445712419.0, "step": 11737 }, { "epoch": 1.4931942500954076, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.962104797363281, "learning_rate": 5e-05, "loss": 0.9399, "mean_token_accuracy": 0.8541306853294373, "num_tokens": 445751465.0, "step": 11738 }, { "epoch": 1.4933214603739982, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.247325897216797, "learning_rate": 5e-05, "loss": 0.6688, "mean_token_accuracy": 0.8839284777641296, "num_tokens": 445788145.0, "step": 11739 }, { "epoch": 1.4934486706525887, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.550247192382812, "learning_rate": 5e-05, "loss": 0.9561, "mean_token_accuracy": 0.8774482011795044, "num_tokens": 445825954.0, "step": 11740 }, { "epoch": 1.4935758809311792, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.491157531738281, "learning_rate": 5e-05, "loss": 0.8035, "mean_token_accuracy": 0.8594392538070679, "num_tokens": 445866207.0, "step": 11741 }, { "epoch": 1.4937030912097697, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.663698196411133, "learning_rate": 5e-05, "loss": 1.0281, "mean_token_accuracy": 0.8627981543540955, "num_tokens": 445904679.0, "step": 11742 }, { "epoch": 1.4938303014883603, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.495726108551025, "learning_rate": 5e-05, "loss": 0.9627, "mean_token_accuracy": 0.8579673767089844, "num_tokens": 445936001.0, "step": 11743 }, { "epoch": 1.4939575117669508, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.77768325805664, "learning_rate": 5e-05, "loss": 0.9652, "mean_token_accuracy": 0.8652242422103882, "num_tokens": 445969212.0, "step": 11744 }, { "epoch": 1.4940847220455413, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.314727783203125, "learning_rate": 5e-05, "loss": 0.9873, "mean_token_accuracy": 0.8716282844543457, "num_tokens": 446013455.0, "step": 11745 }, { "epoch": 1.4942119323241319, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.414116859436035, "learning_rate": 5e-05, "loss": 0.8137, "mean_token_accuracy": 0.8659799098968506, "num_tokens": 446044152.0, "step": 11746 }, { "epoch": 1.4943391426027224, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.944181442260742, "learning_rate": 5e-05, "loss": 1.0183, "mean_token_accuracy": 0.8600260019302368, "num_tokens": 446087047.0, "step": 11747 }, { "epoch": 1.4944663528813127, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.752946376800537, "learning_rate": 5e-05, "loss": 0.7177, "mean_token_accuracy": 0.8794307708740234, "num_tokens": 446120095.0, "step": 11748 }, { "epoch": 1.4945935631599032, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.307165145874023, "learning_rate": 5e-05, "loss": 0.9586, "mean_token_accuracy": 0.877981960773468, "num_tokens": 446156827.0, "step": 11749 }, { "epoch": 1.4947207734384937, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 6.969746112823486, "learning_rate": 5e-05, "loss": 0.825, "mean_token_accuracy": 0.8702669739723206, "num_tokens": 446197101.0, "step": 11750 }, { "epoch": 1.4948479837170843, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.175313949584961, "learning_rate": 5e-05, "loss": 0.9834, "mean_token_accuracy": 0.8698456287384033, "num_tokens": 446230442.0, "step": 11751 }, { "epoch": 1.4949751939956748, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.5297088623046875, "learning_rate": 5e-05, "loss": 0.8721, "mean_token_accuracy": 0.8825665712356567, "num_tokens": 446268395.0, "step": 11752 }, { "epoch": 1.4951024042742653, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.602141380310059, "learning_rate": 5e-05, "loss": 0.9063, "mean_token_accuracy": 0.8761682510375977, "num_tokens": 446307855.0, "step": 11753 }, { "epoch": 1.4952296145528559, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.7726240158081055, "learning_rate": 5e-05, "loss": 0.9355, "mean_token_accuracy": 0.8710155487060547, "num_tokens": 446348808.0, "step": 11754 }, { "epoch": 1.4953568248314464, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.122979164123535, "learning_rate": 5e-05, "loss": 0.8638, "mean_token_accuracy": 0.8775382041931152, "num_tokens": 446380580.0, "step": 11755 }, { "epoch": 1.495484035110037, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.597171783447266, "learning_rate": 5e-05, "loss": 0.8402, "mean_token_accuracy": 0.8789774179458618, "num_tokens": 446414420.0, "step": 11756 }, { "epoch": 1.4956112453886274, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.12891960144043, "learning_rate": 5e-05, "loss": 0.8027, "mean_token_accuracy": 0.8909640312194824, "num_tokens": 446450220.0, "step": 11757 }, { "epoch": 1.495738455667218, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.17439079284668, "learning_rate": 5e-05, "loss": 0.7535, "mean_token_accuracy": 0.8860868215560913, "num_tokens": 446483823.0, "step": 11758 }, { "epoch": 1.4958656659458085, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.40184497833252, "learning_rate": 5e-05, "loss": 0.8642, "mean_token_accuracy": 0.8753949403762817, "num_tokens": 446522317.0, "step": 11759 }, { "epoch": 1.495992876224399, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.865108966827393, "learning_rate": 5e-05, "loss": 0.77, "mean_token_accuracy": 0.8656874895095825, "num_tokens": 446564491.0, "step": 11760 }, { "epoch": 1.4961200865029896, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.605449676513672, "learning_rate": 5e-05, "loss": 0.8754, "mean_token_accuracy": 0.8727182149887085, "num_tokens": 446605490.0, "step": 11761 }, { "epoch": 1.49624729678158, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.791579246520996, "learning_rate": 5e-05, "loss": 0.7753, "mean_token_accuracy": 0.8614267706871033, "num_tokens": 446647385.0, "step": 11762 }, { "epoch": 1.4963745070601704, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.73232650756836, "learning_rate": 5e-05, "loss": 0.9206, "mean_token_accuracy": 0.8676560521125793, "num_tokens": 446682895.0, "step": 11763 }, { "epoch": 1.496501717338761, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.891844749450684, "learning_rate": 5e-05, "loss": 0.7995, "mean_token_accuracy": 0.8629311323165894, "num_tokens": 446724724.0, "step": 11764 }, { "epoch": 1.4966289276173514, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.696231842041016, "learning_rate": 5e-05, "loss": 0.9188, "mean_token_accuracy": 0.873267650604248, "num_tokens": 446763433.0, "step": 11765 }, { "epoch": 1.496756137895942, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.176896572113037, "learning_rate": 5e-05, "loss": 0.8916, "mean_token_accuracy": 0.8477309942245483, "num_tokens": 446809275.0, "step": 11766 }, { "epoch": 1.4968833481745325, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.318228721618652, "learning_rate": 5e-05, "loss": 0.8375, "mean_token_accuracy": 0.8869531750679016, "num_tokens": 446840619.0, "step": 11767 }, { "epoch": 1.497010558453123, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.544748306274414, "learning_rate": 5e-05, "loss": 0.8793, "mean_token_accuracy": 0.8596729040145874, "num_tokens": 446879979.0, "step": 11768 }, { "epoch": 1.4971377687317136, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.7407121658325195, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8827398419380188, "num_tokens": 446913729.0, "step": 11769 }, { "epoch": 1.497264979010304, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.803842067718506, "learning_rate": 5e-05, "loss": 0.9009, "mean_token_accuracy": 0.860260546207428, "num_tokens": 446951843.0, "step": 11770 }, { "epoch": 1.4973921892888946, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.360064506530762, "learning_rate": 5e-05, "loss": 0.7756, "mean_token_accuracy": 0.8689159154891968, "num_tokens": 446988091.0, "step": 11771 }, { "epoch": 1.497519399567485, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.910702228546143, "learning_rate": 5e-05, "loss": 0.8369, "mean_token_accuracy": 0.8629578351974487, "num_tokens": 447023833.0, "step": 11772 }, { "epoch": 1.4976466098460754, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.183166027069092, "learning_rate": 5e-05, "loss": 0.7045, "mean_token_accuracy": 0.8786565065383911, "num_tokens": 447056232.0, "step": 11773 }, { "epoch": 1.497773820124666, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.962186336517334, "learning_rate": 5e-05, "loss": 0.7762, "mean_token_accuracy": 0.8829966187477112, "num_tokens": 447094524.0, "step": 11774 }, { "epoch": 1.4979010304032565, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.0802202224731445, "learning_rate": 5e-05, "loss": 0.7411, "mean_token_accuracy": 0.8676257133483887, "num_tokens": 447129076.0, "step": 11775 }, { "epoch": 1.498028240681847, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.028812408447266, "learning_rate": 5e-05, "loss": 0.8064, "mean_token_accuracy": 0.8723204731941223, "num_tokens": 447168651.0, "step": 11776 }, { "epoch": 1.4981554509604376, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.999330997467041, "learning_rate": 5e-05, "loss": 0.7289, "mean_token_accuracy": 0.8721640706062317, "num_tokens": 447207491.0, "step": 11777 }, { "epoch": 1.498282661239028, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.100461959838867, "learning_rate": 5e-05, "loss": 0.7925, "mean_token_accuracy": 0.8796840906143188, "num_tokens": 447242955.0, "step": 11778 }, { "epoch": 1.4984098715176186, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.933194160461426, "learning_rate": 5e-05, "loss": 0.6576, "mean_token_accuracy": 0.8905922174453735, "num_tokens": 447277634.0, "step": 11779 }, { "epoch": 1.4985370817962091, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.174606323242188, "learning_rate": 5e-05, "loss": 0.8301, "mean_token_accuracy": 0.8663432598114014, "num_tokens": 447315255.0, "step": 11780 }, { "epoch": 1.4986642920747997, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.868159770965576, "learning_rate": 5e-05, "loss": 0.6798, "mean_token_accuracy": 0.8842839002609253, "num_tokens": 447353801.0, "step": 11781 }, { "epoch": 1.4987915023533902, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.270899772644043, "learning_rate": 5e-05, "loss": 0.8186, "mean_token_accuracy": 0.8677505850791931, "num_tokens": 447388098.0, "step": 11782 }, { "epoch": 1.4989187126319807, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.795691013336182, "learning_rate": 5e-05, "loss": 0.7286, "mean_token_accuracy": 0.8709891438484192, "num_tokens": 447424860.0, "step": 11783 }, { "epoch": 1.4990459229105713, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.362932205200195, "learning_rate": 5e-05, "loss": 0.9014, "mean_token_accuracy": 0.853448212146759, "num_tokens": 447465179.0, "step": 11784 }, { "epoch": 1.4991731331891618, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.814550876617432, "learning_rate": 5e-05, "loss": 0.7537, "mean_token_accuracy": 0.8581151962280273, "num_tokens": 447505960.0, "step": 11785 }, { "epoch": 1.4993003434677523, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.36327075958252, "learning_rate": 5e-05, "loss": 0.8672, "mean_token_accuracy": 0.8637481331825256, "num_tokens": 447545410.0, "step": 11786 }, { "epoch": 1.4994275537463426, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.885374069213867, "learning_rate": 5e-05, "loss": 0.7411, "mean_token_accuracy": 0.8678423166275024, "num_tokens": 447586294.0, "step": 11787 }, { "epoch": 1.4995547640249332, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.28596305847168, "learning_rate": 5e-05, "loss": 0.8338, "mean_token_accuracy": 0.8723574876785278, "num_tokens": 447628841.0, "step": 11788 }, { "epoch": 1.4996819743035237, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.0473480224609375, "learning_rate": 5e-05, "loss": 0.7631, "mean_token_accuracy": 0.8708674907684326, "num_tokens": 447669456.0, "step": 11789 }, { "epoch": 1.4998091845821142, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.047249794006348, "learning_rate": 5e-05, "loss": 0.8165, "mean_token_accuracy": 0.878470778465271, "num_tokens": 447703297.0, "step": 11790 }, { "epoch": 1.4999363948607047, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.275573253631592, "learning_rate": 5e-05, "loss": 0.822, "mean_token_accuracy": 0.8563624620437622, "num_tokens": 447742497.0, "step": 11791 }, { "epoch": 1.5000636051392953, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.724733352661133, "learning_rate": 5e-05, "loss": 0.7798, "mean_token_accuracy": 0.8724748492240906, "num_tokens": 447783956.0, "step": 11792 }, { "epoch": 1.5001908154178858, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.486050605773926, "learning_rate": 5e-05, "loss": 0.775, "mean_token_accuracy": 0.8715969324111938, "num_tokens": 447824527.0, "step": 11793 }, { "epoch": 1.5003180256964763, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.481195449829102, "learning_rate": 5e-05, "loss": 0.7617, "mean_token_accuracy": 0.8698470592498779, "num_tokens": 447863676.0, "step": 11794 }, { "epoch": 1.5004452359750666, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.58038330078125, "learning_rate": 5e-05, "loss": 0.7999, "mean_token_accuracy": 0.8573049902915955, "num_tokens": 447906157.0, "step": 11795 }, { "epoch": 1.5005724462536572, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.369114398956299, "learning_rate": 5e-05, "loss": 0.6924, "mean_token_accuracy": 0.8843567371368408, "num_tokens": 447944016.0, "step": 11796 }, { "epoch": 1.5006996565322477, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.618183135986328, "learning_rate": 5e-05, "loss": 0.7273, "mean_token_accuracy": 0.8794879913330078, "num_tokens": 447979500.0, "step": 11797 }, { "epoch": 1.5008268668108382, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.307896614074707, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.8607320785522461, "num_tokens": 448012770.0, "step": 11798 }, { "epoch": 1.5009540770894287, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.601293087005615, "learning_rate": 5e-05, "loss": 0.7442, "mean_token_accuracy": 0.8713392615318298, "num_tokens": 448048227.0, "step": 11799 }, { "epoch": 1.5010812873680193, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.321279048919678, "learning_rate": 5e-05, "loss": 0.7648, "mean_token_accuracy": 0.8613796234130859, "num_tokens": 448081672.0, "step": 11800 }, { "epoch": 1.5012084976466098, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.547743320465088, "learning_rate": 5e-05, "loss": 0.7373, "mean_token_accuracy": 0.8702548742294312, "num_tokens": 448122693.0, "step": 11801 }, { "epoch": 1.5013357079252003, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.35999870300293, "learning_rate": 5e-05, "loss": 0.7676, "mean_token_accuracy": 0.8616617918014526, "num_tokens": 448161246.0, "step": 11802 }, { "epoch": 1.5014629182037909, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.47210168838501, "learning_rate": 5e-05, "loss": 0.7631, "mean_token_accuracy": 0.8682568073272705, "num_tokens": 448205233.0, "step": 11803 }, { "epoch": 1.5015901284823814, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.387262344360352, "learning_rate": 5e-05, "loss": 0.7846, "mean_token_accuracy": 0.8508484363555908, "num_tokens": 448247404.0, "step": 11804 }, { "epoch": 1.501717338760972, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.4345383644104, "learning_rate": 5e-05, "loss": 0.7168, "mean_token_accuracy": 0.8731751441955566, "num_tokens": 448286874.0, "step": 11805 }, { "epoch": 1.5018445490395624, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.422788143157959, "learning_rate": 5e-05, "loss": 0.7857, "mean_token_accuracy": 0.8522326946258545, "num_tokens": 448325461.0, "step": 11806 }, { "epoch": 1.501971759318153, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.370902061462402, "learning_rate": 5e-05, "loss": 0.6913, "mean_token_accuracy": 0.8802381753921509, "num_tokens": 448366871.0, "step": 11807 }, { "epoch": 1.5020989695967435, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.485198020935059, "learning_rate": 5e-05, "loss": 0.6983, "mean_token_accuracy": 0.877608060836792, "num_tokens": 448396798.0, "step": 11808 }, { "epoch": 1.502226179875334, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.286646366119385, "learning_rate": 5e-05, "loss": 0.7028, "mean_token_accuracy": 0.8722575902938843, "num_tokens": 448437782.0, "step": 11809 }, { "epoch": 1.5023533901539246, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.591732025146484, "learning_rate": 5e-05, "loss": 0.7054, "mean_token_accuracy": 0.8748211860656738, "num_tokens": 448469892.0, "step": 11810 }, { "epoch": 1.502480600432515, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 7.118349075317383, "learning_rate": 5e-05, "loss": 0.7137, "mean_token_accuracy": 0.8670450448989868, "num_tokens": 448508233.0, "step": 11811 }, { "epoch": 1.5026078107111056, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.81360387802124, "learning_rate": 5e-05, "loss": 0.7656, "mean_token_accuracy": 0.8579185009002686, "num_tokens": 448541071.0, "step": 11812 }, { "epoch": 1.502735020989696, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.835576057434082, "learning_rate": 5e-05, "loss": 0.7151, "mean_token_accuracy": 0.8634934425354004, "num_tokens": 448571348.0, "step": 11813 }, { "epoch": 1.5028622312682864, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.216607093811035, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.8739467263221741, "num_tokens": 448611889.0, "step": 11814 }, { "epoch": 1.502989441546877, "ewc_loss": 0.27734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002231597900390625, "grad_norm": 6.311242580413818, "learning_rate": 5e-05, "loss": 0.6332, "mean_token_accuracy": 0.8776137232780457, "num_tokens": 448646683.0, "step": 11815 }, { "epoch": 1.5031166518254675, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 9.076326370239258, "learning_rate": 5e-05, "loss": 0.8425, "mean_token_accuracy": 0.8726032376289368, "num_tokens": 448688154.0, "step": 11816 }, { "epoch": 1.503243862104058, "ewc_loss": 0.248046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000194549560546875, "grad_norm": 5.450536251068115, "learning_rate": 5e-05, "loss": 0.6199, "mean_token_accuracy": 0.8689187169075012, "num_tokens": 448726343.0, "step": 11817 }, { "epoch": 1.5033710723826486, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.721073150634766, "learning_rate": 5e-05, "loss": 0.9897, "mean_token_accuracy": 0.8877865076065063, "num_tokens": 448770376.0, "step": 11818 }, { "epoch": 1.503498282661239, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 5.352509498596191, "learning_rate": 5e-05, "loss": 0.6642, "mean_token_accuracy": 0.881625771522522, "num_tokens": 448799247.0, "step": 11819 }, { "epoch": 1.5036254929398294, "ewc_loss": 1.109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0010528564453125, "grad_norm": 12.96035099029541, "learning_rate": 5e-05, "loss": 1.5398, "mean_token_accuracy": 0.8603668808937073, "num_tokens": 448839379.0, "step": 11820 }, { "epoch": 1.50375270321842, "ewc_loss": 0.98046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000926971435546875, "grad_norm": 9.428802490234375, "learning_rate": 5e-05, "loss": 1.415, "mean_token_accuracy": 0.8530046343803406, "num_tokens": 448873924.0, "step": 11821 }, { "epoch": 1.5038799134970104, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 12.051667213439941, "learning_rate": 5e-05, "loss": 1.2012, "mean_token_accuracy": 0.8694863319396973, "num_tokens": 448915765.0, "step": 11822 }, { "epoch": 1.504007123775601, "ewc_loss": 1.046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0009918212890625, "grad_norm": 10.1521635055542, "learning_rate": 5e-05, "loss": 1.4047, "mean_token_accuracy": 0.8791821002960205, "num_tokens": 448953116.0, "step": 11823 }, { "epoch": 1.5041343340541915, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 11.457601547241211, "learning_rate": 5e-05, "loss": 1.1839, "mean_token_accuracy": 0.8651862144470215, "num_tokens": 448989065.0, "step": 11824 }, { "epoch": 1.504261544332782, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 9.079242706298828, "learning_rate": 5e-05, "loss": 1.1302, "mean_token_accuracy": 0.8650033473968506, "num_tokens": 449029608.0, "step": 11825 }, { "epoch": 1.5043887546113726, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 10.993607521057129, "learning_rate": 5e-05, "loss": 1.2079, "mean_token_accuracy": 0.8609088063240051, "num_tokens": 449068288.0, "step": 11826 }, { "epoch": 1.504515964889963, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.621601104736328, "learning_rate": 5e-05, "loss": 1.0189, "mean_token_accuracy": 0.8675882816314697, "num_tokens": 449107754.0, "step": 11827 }, { "epoch": 1.5046431751685536, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.590448379516602, "learning_rate": 5e-05, "loss": 1.0156, "mean_token_accuracy": 0.8936781883239746, "num_tokens": 449148680.0, "step": 11828 }, { "epoch": 1.5047703854471441, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.506756782531738, "learning_rate": 5e-05, "loss": 1.0103, "mean_token_accuracy": 0.8696226477622986, "num_tokens": 449191616.0, "step": 11829 }, { "epoch": 1.5048975957257347, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.364877700805664, "learning_rate": 5e-05, "loss": 1.0345, "mean_token_accuracy": 0.8797745108604431, "num_tokens": 449227149.0, "step": 11830 }, { "epoch": 1.5050248060043252, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.504654884338379, "learning_rate": 5e-05, "loss": 1.0155, "mean_token_accuracy": 0.8663334250450134, "num_tokens": 449259186.0, "step": 11831 }, { "epoch": 1.5051520162829157, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.050044059753418, "learning_rate": 5e-05, "loss": 1.0125, "mean_token_accuracy": 0.8804726004600525, "num_tokens": 449294403.0, "step": 11832 }, { "epoch": 1.5052792265615063, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.375258445739746, "learning_rate": 5e-05, "loss": 0.9221, "mean_token_accuracy": 0.8907698392868042, "num_tokens": 449334205.0, "step": 11833 }, { "epoch": 1.5054064368400968, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.870523452758789, "learning_rate": 5e-05, "loss": 1.1032, "mean_token_accuracy": 0.8545248508453369, "num_tokens": 449372501.0, "step": 11834 }, { "epoch": 1.5055336471186873, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.150282859802246, "learning_rate": 5e-05, "loss": 0.9636, "mean_token_accuracy": 0.8613324165344238, "num_tokens": 449410394.0, "step": 11835 }, { "epoch": 1.5056608573972778, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.908706665039062, "learning_rate": 5e-05, "loss": 1.0445, "mean_token_accuracy": 0.8724657893180847, "num_tokens": 449449494.0, "step": 11836 }, { "epoch": 1.5057880676758684, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.121678352355957, "learning_rate": 5e-05, "loss": 0.936, "mean_token_accuracy": 0.8675310015678406, "num_tokens": 449485086.0, "step": 11837 }, { "epoch": 1.5059152779544587, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.921006202697754, "learning_rate": 5e-05, "loss": 1.0674, "mean_token_accuracy": 0.8715571761131287, "num_tokens": 449518681.0, "step": 11838 }, { "epoch": 1.5060424882330492, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.280197143554688, "learning_rate": 5e-05, "loss": 0.9301, "mean_token_accuracy": 0.8778947591781616, "num_tokens": 449557439.0, "step": 11839 }, { "epoch": 1.5061696985116397, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.73647403717041, "learning_rate": 5e-05, "loss": 1.0926, "mean_token_accuracy": 0.8564494848251343, "num_tokens": 449600571.0, "step": 11840 }, { "epoch": 1.5062969087902303, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.4093599319458, "learning_rate": 5e-05, "loss": 0.9961, "mean_token_accuracy": 0.8609099984169006, "num_tokens": 449639323.0, "step": 11841 }, { "epoch": 1.5064241190688208, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.326827049255371, "learning_rate": 5e-05, "loss": 1.0074, "mean_token_accuracy": 0.8690693974494934, "num_tokens": 449680082.0, "step": 11842 }, { "epoch": 1.5065513293474113, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.311055183410645, "learning_rate": 5e-05, "loss": 0.976, "mean_token_accuracy": 0.8578943610191345, "num_tokens": 449717655.0, "step": 11843 }, { "epoch": 1.5066785396260016, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.957918167114258, "learning_rate": 5e-05, "loss": 0.9356, "mean_token_accuracy": 0.8823776245117188, "num_tokens": 449756833.0, "step": 11844 }, { "epoch": 1.5068057499045922, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.96193265914917, "learning_rate": 5e-05, "loss": 0.8744, "mean_token_accuracy": 0.8708697557449341, "num_tokens": 449796525.0, "step": 11845 }, { "epoch": 1.5069329601831827, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.995224952697754, "learning_rate": 5e-05, "loss": 0.9861, "mean_token_accuracy": 0.8692247867584229, "num_tokens": 449837731.0, "step": 11846 }, { "epoch": 1.5070601704617732, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.577350616455078, "learning_rate": 5e-05, "loss": 0.8177, "mean_token_accuracy": 0.8782826662063599, "num_tokens": 449876235.0, "step": 11847 }, { "epoch": 1.5071873807403637, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.350950241088867, "learning_rate": 5e-05, "loss": 1.0129, "mean_token_accuracy": 0.8727174997329712, "num_tokens": 449916153.0, "step": 11848 }, { "epoch": 1.5073145910189543, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.66144323348999, "learning_rate": 5e-05, "loss": 0.9059, "mean_token_accuracy": 0.8569014072418213, "num_tokens": 449957969.0, "step": 11849 }, { "epoch": 1.5074418012975448, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.37297534942627, "learning_rate": 5e-05, "loss": 1.0121, "mean_token_accuracy": 0.8737620115280151, "num_tokens": 449997464.0, "step": 11850 }, { "epoch": 1.5075690115761353, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.094109535217285, "learning_rate": 5e-05, "loss": 0.9413, "mean_token_accuracy": 0.8619657754898071, "num_tokens": 450038102.0, "step": 11851 }, { "epoch": 1.5076962218547258, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.776094436645508, "learning_rate": 5e-05, "loss": 0.951, "mean_token_accuracy": 0.8777185678482056, "num_tokens": 450073832.0, "step": 11852 }, { "epoch": 1.5078234321333164, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.174779891967773, "learning_rate": 5e-05, "loss": 0.8788, "mean_token_accuracy": 0.8820260763168335, "num_tokens": 450108662.0, "step": 11853 }, { "epoch": 1.507950642411907, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.227550506591797, "learning_rate": 5e-05, "loss": 0.8618, "mean_token_accuracy": 0.886175811290741, "num_tokens": 450144601.0, "step": 11854 }, { "epoch": 1.5080778526904974, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.821231365203857, "learning_rate": 5e-05, "loss": 0.8618, "mean_token_accuracy": 0.8662872910499573, "num_tokens": 450190112.0, "step": 11855 }, { "epoch": 1.508205062969088, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.212204933166504, "learning_rate": 5e-05, "loss": 0.8458, "mean_token_accuracy": 0.8836809396743774, "num_tokens": 450227660.0, "step": 11856 }, { "epoch": 1.5083322732476785, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.361049652099609, "learning_rate": 5e-05, "loss": 0.7689, "mean_token_accuracy": 0.8783209323883057, "num_tokens": 450265243.0, "step": 11857 }, { "epoch": 1.508459483526269, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.5217924118042, "learning_rate": 5e-05, "loss": 0.9174, "mean_token_accuracy": 0.8642256259918213, "num_tokens": 450303382.0, "step": 11858 }, { "epoch": 1.5085866938048595, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.0907368659973145, "learning_rate": 5e-05, "loss": 0.7585, "mean_token_accuracy": 0.8754523992538452, "num_tokens": 450335458.0, "step": 11859 }, { "epoch": 1.50871390408345, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.771157264709473, "learning_rate": 5e-05, "loss": 0.944, "mean_token_accuracy": 0.8675556182861328, "num_tokens": 450373375.0, "step": 11860 }, { "epoch": 1.5088411143620406, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.1622514724731445, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.8795441389083862, "num_tokens": 450406258.0, "step": 11861 }, { "epoch": 1.508968324640631, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.76392936706543, "learning_rate": 5e-05, "loss": 0.9269, "mean_token_accuracy": 0.8768953084945679, "num_tokens": 450441558.0, "step": 11862 }, { "epoch": 1.5090955349192214, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.484411716461182, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.8772369027137756, "num_tokens": 450477685.0, "step": 11863 }, { "epoch": 1.509222745197812, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.35546588897705, "learning_rate": 5e-05, "loss": 0.943, "mean_token_accuracy": 0.8572962284088135, "num_tokens": 450516661.0, "step": 11864 }, { "epoch": 1.5093499554764025, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.73997163772583, "learning_rate": 5e-05, "loss": 0.8179, "mean_token_accuracy": 0.8741048574447632, "num_tokens": 450555091.0, "step": 11865 }, { "epoch": 1.509477165754993, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.826934814453125, "learning_rate": 5e-05, "loss": 0.8399, "mean_token_accuracy": 0.8691500425338745, "num_tokens": 450596172.0, "step": 11866 }, { "epoch": 1.5096043760335836, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.795708656311035, "learning_rate": 5e-05, "loss": 0.8254, "mean_token_accuracy": 0.8654435873031616, "num_tokens": 450636902.0, "step": 11867 }, { "epoch": 1.509731586312174, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.556320667266846, "learning_rate": 5e-05, "loss": 0.86, "mean_token_accuracy": 0.8567990660667419, "num_tokens": 450682826.0, "step": 11868 }, { "epoch": 1.5098587965907644, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.687037467956543, "learning_rate": 5e-05, "loss": 0.8095, "mean_token_accuracy": 0.8610132932662964, "num_tokens": 450722995.0, "step": 11869 }, { "epoch": 1.509986006869355, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.539194583892822, "learning_rate": 5e-05, "loss": 0.7516, "mean_token_accuracy": 0.8774325847625732, "num_tokens": 450762109.0, "step": 11870 }, { "epoch": 1.5101132171479454, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.578925609588623, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.8746408224105835, "num_tokens": 450799543.0, "step": 11871 }, { "epoch": 1.510240427426536, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.51763391494751, "learning_rate": 5e-05, "loss": 0.7784, "mean_token_accuracy": 0.8677524328231812, "num_tokens": 450840424.0, "step": 11872 }, { "epoch": 1.5103676377051265, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.543551445007324, "learning_rate": 5e-05, "loss": 0.7326, "mean_token_accuracy": 0.8766651153564453, "num_tokens": 450877652.0, "step": 11873 }, { "epoch": 1.510494847983717, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.461674690246582, "learning_rate": 5e-05, "loss": 0.8055, "mean_token_accuracy": 0.8524717688560486, "num_tokens": 450916171.0, "step": 11874 }, { "epoch": 1.5106220582623076, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.553245544433594, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.8766253590583801, "num_tokens": 450950455.0, "step": 11875 }, { "epoch": 1.510749268540898, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.439975738525391, "learning_rate": 5e-05, "loss": 0.7461, "mean_token_accuracy": 0.870046079158783, "num_tokens": 450990747.0, "step": 11876 }, { "epoch": 1.5108764788194886, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.492028713226318, "learning_rate": 5e-05, "loss": 0.7256, "mean_token_accuracy": 0.876549243927002, "num_tokens": 451029396.0, "step": 11877 }, { "epoch": 1.5110036890980791, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.49161434173584, "learning_rate": 5e-05, "loss": 0.7315, "mean_token_accuracy": 0.8711803555488586, "num_tokens": 451065071.0, "step": 11878 }, { "epoch": 1.5111308993766697, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.431310176849365, "learning_rate": 5e-05, "loss": 0.7274, "mean_token_accuracy": 0.8759782910346985, "num_tokens": 451105487.0, "step": 11879 }, { "epoch": 1.5112581096552602, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.5230278968811035, "learning_rate": 5e-05, "loss": 0.7643, "mean_token_accuracy": 0.8626940250396729, "num_tokens": 451141538.0, "step": 11880 }, { "epoch": 1.5113853199338507, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.362424373626709, "learning_rate": 5e-05, "loss": 0.6823, "mean_token_accuracy": 0.8857138752937317, "num_tokens": 451175080.0, "step": 11881 }, { "epoch": 1.5115125302124413, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.600896835327148, "learning_rate": 5e-05, "loss": 0.7406, "mean_token_accuracy": 0.8697513341903687, "num_tokens": 451216117.0, "step": 11882 }, { "epoch": 1.5116397404910318, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.218286037445068, "learning_rate": 5e-05, "loss": 0.6975, "mean_token_accuracy": 0.8747504949569702, "num_tokens": 451251585.0, "step": 11883 }, { "epoch": 1.5117669507696223, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.794766902923584, "learning_rate": 5e-05, "loss": 0.7904, "mean_token_accuracy": 0.8577842712402344, "num_tokens": 451290401.0, "step": 11884 }, { "epoch": 1.5118941610482128, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.974554061889648, "learning_rate": 5e-05, "loss": 0.7246, "mean_token_accuracy": 0.8616095781326294, "num_tokens": 451331896.0, "step": 11885 }, { "epoch": 1.5120213713268034, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.076712608337402, "learning_rate": 5e-05, "loss": 0.7518, "mean_token_accuracy": 0.8745236992835999, "num_tokens": 451368802.0, "step": 11886 }, { "epoch": 1.5121485816053937, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.657751560211182, "learning_rate": 5e-05, "loss": 0.6357, "mean_token_accuracy": 0.8842674493789673, "num_tokens": 451400101.0, "step": 11887 }, { "epoch": 1.5122757918839842, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.565716743469238, "learning_rate": 5e-05, "loss": 0.8101, "mean_token_accuracy": 0.8713441491127014, "num_tokens": 451444732.0, "step": 11888 }, { "epoch": 1.5124030021625747, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00022602081298828125, "grad_norm": 6.14109992980957, "learning_rate": 5e-05, "loss": 0.5858, "mean_token_accuracy": 0.8929761052131653, "num_tokens": 451481467.0, "step": 11889 }, { "epoch": 1.5125302124411653, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 9.434548377990723, "learning_rate": 5e-05, "loss": 0.9031, "mean_token_accuracy": 0.8709355592727661, "num_tokens": 451520048.0, "step": 11890 }, { "epoch": 1.5126574227197558, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 5.703891754150391, "learning_rate": 5e-05, "loss": 0.6512, "mean_token_accuracy": 0.8737767934799194, "num_tokens": 451555121.0, "step": 11891 }, { "epoch": 1.5127846329983463, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.380772590637207, "learning_rate": 5e-05, "loss": 1.05, "mean_token_accuracy": 0.8697113990783691, "num_tokens": 451594894.0, "step": 11892 }, { "epoch": 1.5129118432769366, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 6.6697678565979, "learning_rate": 5e-05, "loss": 0.7843, "mean_token_accuracy": 0.8770250082015991, "num_tokens": 451627175.0, "step": 11893 }, { "epoch": 1.5130390535555271, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.086575508117676, "learning_rate": 5e-05, "loss": 1.0297, "mean_token_accuracy": 0.8784858584403992, "num_tokens": 451664177.0, "step": 11894 }, { "epoch": 1.5131662638341177, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 8.625192642211914, "learning_rate": 5e-05, "loss": 1.035, "mean_token_accuracy": 0.8798437714576721, "num_tokens": 451696964.0, "step": 11895 }, { "epoch": 1.5132934741127082, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.768406867980957, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8871035575866699, "num_tokens": 451735349.0, "step": 11896 }, { "epoch": 1.5134206843912987, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 10.371199607849121, "learning_rate": 5e-05, "loss": 1.0738, "mean_token_accuracy": 0.8814198970794678, "num_tokens": 451770410.0, "step": 11897 }, { "epoch": 1.5135478946698893, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.372687339782715, "learning_rate": 5e-05, "loss": 0.7203, "mean_token_accuracy": 0.8767667412757874, "num_tokens": 451812170.0, "step": 11898 }, { "epoch": 1.5136751049484798, "ewc_loss": 1.015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0009613037109375, "grad_norm": 12.81552791595459, "learning_rate": 5e-05, "loss": 1.3787, "mean_token_accuracy": 0.8730568885803223, "num_tokens": 451858079.0, "step": 11899 }, { "epoch": 1.5138023152270703, "ewc_loss": 0.90625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 9.917498588562012, "learning_rate": 5e-05, "loss": 1.3072, "mean_token_accuracy": 0.863359808921814, "num_tokens": 451901120.0, "step": 11900 }, { "epoch": 1.5139295255056608, "ewc_loss": 0.734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 11.644448280334473, "learning_rate": 5e-05, "loss": 1.1574, "mean_token_accuracy": 0.8584206104278564, "num_tokens": 451944990.0, "step": 11901 }, { "epoch": 1.5140567357842514, "ewc_loss": 0.82421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00077056884765625, "grad_norm": 9.823209762573242, "learning_rate": 5e-05, "loss": 1.1759, "mean_token_accuracy": 0.8802541494369507, "num_tokens": 451983261.0, "step": 11902 }, { "epoch": 1.514183946062842, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 11.49009895324707, "learning_rate": 5e-05, "loss": 1.1883, "mean_token_accuracy": 0.8663370013237, "num_tokens": 452019438.0, "step": 11903 }, { "epoch": 1.5143111563414324, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.71219539642334, "learning_rate": 5e-05, "loss": 1.0162, "mean_token_accuracy": 0.8568602800369263, "num_tokens": 452055941.0, "step": 11904 }, { "epoch": 1.514438366620023, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 11.657333374023438, "learning_rate": 5e-05, "loss": 1.1933, "mean_token_accuracy": 0.8811620473861694, "num_tokens": 452092625.0, "step": 11905 }, { "epoch": 1.5145655768986135, "ewc_loss": 0.625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.974906921386719, "learning_rate": 5e-05, "loss": 1.0099, "mean_token_accuracy": 0.8667754530906677, "num_tokens": 452126279.0, "step": 11906 }, { "epoch": 1.514692787177204, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 11.719807624816895, "learning_rate": 5e-05, "loss": 1.1679, "mean_token_accuracy": 0.8772469758987427, "num_tokens": 452163506.0, "step": 11907 }, { "epoch": 1.5148199974557945, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 9.891478538513184, "learning_rate": 5e-05, "loss": 1.1344, "mean_token_accuracy": 0.861694872379303, "num_tokens": 452200898.0, "step": 11908 }, { "epoch": 1.514947207734385, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.902905464172363, "learning_rate": 5e-05, "loss": 1.1139, "mean_token_accuracy": 0.8614275455474854, "num_tokens": 452237772.0, "step": 11909 }, { "epoch": 1.5150744180129756, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.598453521728516, "learning_rate": 5e-05, "loss": 1.0474, "mean_token_accuracy": 0.8695131540298462, "num_tokens": 452272741.0, "step": 11910 }, { "epoch": 1.515201628291566, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.348212242126465, "learning_rate": 5e-05, "loss": 1.0505, "mean_token_accuracy": 0.868598222732544, "num_tokens": 452307674.0, "step": 11911 }, { "epoch": 1.5153288385701564, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.983266830444336, "learning_rate": 5e-05, "loss": 0.983, "mean_token_accuracy": 0.8625911474227905, "num_tokens": 452352186.0, "step": 11912 }, { "epoch": 1.515456048848747, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.107977867126465, "learning_rate": 5e-05, "loss": 0.9975, "mean_token_accuracy": 0.879000186920166, "num_tokens": 452390371.0, "step": 11913 }, { "epoch": 1.5155832591273375, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.453722953796387, "learning_rate": 5e-05, "loss": 0.8441, "mean_token_accuracy": 0.8857688307762146, "num_tokens": 452424545.0, "step": 11914 }, { "epoch": 1.515710469405928, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.252091407775879, "learning_rate": 5e-05, "loss": 1.0359, "mean_token_accuracy": 0.868118405342102, "num_tokens": 452456543.0, "step": 11915 }, { "epoch": 1.5158376796845185, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.333126068115234, "learning_rate": 5e-05, "loss": 0.9133, "mean_token_accuracy": 0.8603053092956543, "num_tokens": 452497884.0, "step": 11916 }, { "epoch": 1.515964889963109, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.44059944152832, "learning_rate": 5e-05, "loss": 1.0881, "mean_token_accuracy": 0.865293025970459, "num_tokens": 452542596.0, "step": 11917 }, { "epoch": 1.5160921002416994, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.605308532714844, "learning_rate": 5e-05, "loss": 0.8971, "mean_token_accuracy": 0.8786317706108093, "num_tokens": 452581385.0, "step": 11918 }, { "epoch": 1.51621931052029, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.364638328552246, "learning_rate": 5e-05, "loss": 1.0683, "mean_token_accuracy": 0.8703664541244507, "num_tokens": 452621399.0, "step": 11919 }, { "epoch": 1.5163465207988804, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.890329360961914, "learning_rate": 5e-05, "loss": 0.9637, "mean_token_accuracy": 0.8670114874839783, "num_tokens": 452657782.0, "step": 11920 }, { "epoch": 1.516473731077471, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.850760459899902, "learning_rate": 5e-05, "loss": 0.9704, "mean_token_accuracy": 0.8843288421630859, "num_tokens": 452696239.0, "step": 11921 }, { "epoch": 1.5166009413560615, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.804732322692871, "learning_rate": 5e-05, "loss": 0.965, "mean_token_accuracy": 0.8622182607650757, "num_tokens": 452734235.0, "step": 11922 }, { "epoch": 1.516728151634652, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.295499801635742, "learning_rate": 5e-05, "loss": 0.9324, "mean_token_accuracy": 0.8784253597259521, "num_tokens": 452769151.0, "step": 11923 }, { "epoch": 1.5168553619132426, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.367377281188965, "learning_rate": 5e-05, "loss": 0.8303, "mean_token_accuracy": 0.877511739730835, "num_tokens": 452808944.0, "step": 11924 }, { "epoch": 1.516982572191833, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.139497756958008, "learning_rate": 5e-05, "loss": 0.9455, "mean_token_accuracy": 0.8693361878395081, "num_tokens": 452845815.0, "step": 11925 }, { "epoch": 1.5171097824704236, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.844231128692627, "learning_rate": 5e-05, "loss": 0.8044, "mean_token_accuracy": 0.8731416463851929, "num_tokens": 452888010.0, "step": 11926 }, { "epoch": 1.5172369927490141, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.402403831481934, "learning_rate": 5e-05, "loss": 0.9277, "mean_token_accuracy": 0.8832634687423706, "num_tokens": 452924149.0, "step": 11927 }, { "epoch": 1.5173642030276047, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.686203956604004, "learning_rate": 5e-05, "loss": 0.7951, "mean_token_accuracy": 0.8772856593132019, "num_tokens": 452964436.0, "step": 11928 }, { "epoch": 1.5174914133061952, "ewc_loss": 0.625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.65494155883789, "learning_rate": 5e-05, "loss": 1.0148, "mean_token_accuracy": 0.8659787774085999, "num_tokens": 453000455.0, "step": 11929 }, { "epoch": 1.5176186235847857, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.01147747039795, "learning_rate": 5e-05, "loss": 0.7834, "mean_token_accuracy": 0.8873633146286011, "num_tokens": 453040419.0, "step": 11930 }, { "epoch": 1.5177458338633762, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.401251792907715, "learning_rate": 5e-05, "loss": 0.979, "mean_token_accuracy": 0.8706568479537964, "num_tokens": 453077668.0, "step": 11931 }, { "epoch": 1.5178730441419668, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.335403442382812, "learning_rate": 5e-05, "loss": 0.8543, "mean_token_accuracy": 0.8782857656478882, "num_tokens": 453115423.0, "step": 11932 }, { "epoch": 1.5180002544205573, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.721607208251953, "learning_rate": 5e-05, "loss": 0.9054, "mean_token_accuracy": 0.8707835674285889, "num_tokens": 453150529.0, "step": 11933 }, { "epoch": 1.5181274646991478, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.184374809265137, "learning_rate": 5e-05, "loss": 0.8353, "mean_token_accuracy": 0.8735979199409485, "num_tokens": 453189123.0, "step": 11934 }, { "epoch": 1.5182546749777384, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.285767555236816, "learning_rate": 5e-05, "loss": 0.8222, "mean_token_accuracy": 0.8849090337753296, "num_tokens": 453231381.0, "step": 11935 }, { "epoch": 1.5183818852563287, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.782231330871582, "learning_rate": 5e-05, "loss": 0.8385, "mean_token_accuracy": 0.860878050327301, "num_tokens": 453272134.0, "step": 11936 }, { "epoch": 1.5185090955349192, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.285699844360352, "learning_rate": 5e-05, "loss": 0.8675, "mean_token_accuracy": 0.8669956922531128, "num_tokens": 453306950.0, "step": 11937 }, { "epoch": 1.5186363058135097, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.386982440948486, "learning_rate": 5e-05, "loss": 0.7748, "mean_token_accuracy": 0.8615013957023621, "num_tokens": 453354206.0, "step": 11938 }, { "epoch": 1.5187635160921003, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.432236671447754, "learning_rate": 5e-05, "loss": 0.8605, "mean_token_accuracy": 0.8680362701416016, "num_tokens": 453388033.0, "step": 11939 }, { "epoch": 1.5188907263706908, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.230842113494873, "learning_rate": 5e-05, "loss": 0.7497, "mean_token_accuracy": 0.8670201897621155, "num_tokens": 453421636.0, "step": 11940 }, { "epoch": 1.5190179366492813, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.465812683105469, "learning_rate": 5e-05, "loss": 0.8415, "mean_token_accuracy": 0.8747646808624268, "num_tokens": 453467994.0, "step": 11941 }, { "epoch": 1.5191451469278716, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.186944007873535, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8589191436767578, "num_tokens": 453508627.0, "step": 11942 }, { "epoch": 1.5192723572064621, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.499115943908691, "learning_rate": 5e-05, "loss": 0.8591, "mean_token_accuracy": 0.8725305795669556, "num_tokens": 453547286.0, "step": 11943 }, { "epoch": 1.5193995674850527, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.212612628936768, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8497366905212402, "num_tokens": 453588485.0, "step": 11944 }, { "epoch": 1.5195267777636432, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.44060230255127, "learning_rate": 5e-05, "loss": 0.856, "mean_token_accuracy": 0.8683686852455139, "num_tokens": 453633103.0, "step": 11945 }, { "epoch": 1.5196539880422337, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.4125261306762695, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8774172067642212, "num_tokens": 453671817.0, "step": 11946 }, { "epoch": 1.5197811983208243, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.035372734069824, "learning_rate": 5e-05, "loss": 0.8097, "mean_token_accuracy": 0.8703712224960327, "num_tokens": 453704704.0, "step": 11947 }, { "epoch": 1.5199084085994148, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.667840957641602, "learning_rate": 5e-05, "loss": 0.7764, "mean_token_accuracy": 0.8651783466339111, "num_tokens": 453743899.0, "step": 11948 }, { "epoch": 1.5200356188780053, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.604031562805176, "learning_rate": 5e-05, "loss": 0.7308, "mean_token_accuracy": 0.8765673637390137, "num_tokens": 453783052.0, "step": 11949 }, { "epoch": 1.5201628291565958, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.826230049133301, "learning_rate": 5e-05, "loss": 0.7719, "mean_token_accuracy": 0.8658302426338196, "num_tokens": 453825522.0, "step": 11950 }, { "epoch": 1.5202900394351864, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.342618942260742, "learning_rate": 5e-05, "loss": 0.7106, "mean_token_accuracy": 0.8770754337310791, "num_tokens": 453859353.0, "step": 11951 }, { "epoch": 1.520417249713777, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.959604263305664, "learning_rate": 5e-05, "loss": 0.7061, "mean_token_accuracy": 0.8907131552696228, "num_tokens": 453896792.0, "step": 11952 }, { "epoch": 1.5205444599923674, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.097438335418701, "learning_rate": 5e-05, "loss": 0.6911, "mean_token_accuracy": 0.8749575018882751, "num_tokens": 453939679.0, "step": 11953 }, { "epoch": 1.520671670270958, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.201387405395508, "learning_rate": 5e-05, "loss": 0.8204, "mean_token_accuracy": 0.8561357259750366, "num_tokens": 453970171.0, "step": 11954 }, { "epoch": 1.5207988805495485, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.875263214111328, "learning_rate": 5e-05, "loss": 0.7322, "mean_token_accuracy": 0.8623443245887756, "num_tokens": 454008470.0, "step": 11955 }, { "epoch": 1.520926090828139, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.376038551330566, "learning_rate": 5e-05, "loss": 0.8292, "mean_token_accuracy": 0.8650149703025818, "num_tokens": 454047711.0, "step": 11956 }, { "epoch": 1.5210533011067295, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.876263618469238, "learning_rate": 5e-05, "loss": 0.7178, "mean_token_accuracy": 0.8702850341796875, "num_tokens": 454083494.0, "step": 11957 }, { "epoch": 1.52118051138532, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.280991554260254, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8779265880584717, "num_tokens": 454115220.0, "step": 11958 }, { "epoch": 1.5213077216639106, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.11410665512085, "learning_rate": 5e-05, "loss": 0.694, "mean_token_accuracy": 0.8761900663375854, "num_tokens": 454152754.0, "step": 11959 }, { "epoch": 1.521434931942501, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.953563213348389, "learning_rate": 5e-05, "loss": 0.7666, "mean_token_accuracy": 0.8724774122238159, "num_tokens": 454185357.0, "step": 11960 }, { "epoch": 1.5215621422210914, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.459193229675293, "learning_rate": 5e-05, "loss": 0.699, "mean_token_accuracy": 0.8792701959609985, "num_tokens": 454217277.0, "step": 11961 }, { "epoch": 1.521689352499682, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.515575885772705, "learning_rate": 5e-05, "loss": 0.7863, "mean_token_accuracy": 0.8539978265762329, "num_tokens": 454253252.0, "step": 11962 }, { "epoch": 1.5218165627782725, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.83040714263916, "learning_rate": 5e-05, "loss": 0.7222, "mean_token_accuracy": 0.8810688257217407, "num_tokens": 454292779.0, "step": 11963 }, { "epoch": 1.521943773056863, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.111095905303955, "learning_rate": 5e-05, "loss": 0.7163, "mean_token_accuracy": 0.8717831373214722, "num_tokens": 454336077.0, "step": 11964 }, { "epoch": 1.5220709833354535, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 8.145021438598633, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.8823027014732361, "num_tokens": 454369088.0, "step": 11965 }, { "epoch": 1.5221981936140438, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.845950126647949, "learning_rate": 5e-05, "loss": 0.7202, "mean_token_accuracy": 0.8626885414123535, "num_tokens": 454413057.0, "step": 11966 }, { "epoch": 1.5223254038926344, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.363171577453613, "learning_rate": 5e-05, "loss": 0.7544, "mean_token_accuracy": 0.8853734731674194, "num_tokens": 454447458.0, "step": 11967 }, { "epoch": 1.522452614171225, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.7418293952941895, "learning_rate": 5e-05, "loss": 0.6568, "mean_token_accuracy": 0.8824309706687927, "num_tokens": 454479057.0, "step": 11968 }, { "epoch": 1.5225798244498154, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.422857284545898, "learning_rate": 5e-05, "loss": 0.8377, "mean_token_accuracy": 0.8604794144630432, "num_tokens": 454520676.0, "step": 11969 }, { "epoch": 1.522707034728406, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.847774505615234, "learning_rate": 5e-05, "loss": 0.7222, "mean_token_accuracy": 0.8651202321052551, "num_tokens": 454558021.0, "step": 11970 }, { "epoch": 1.5228342450069965, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.260506629943848, "learning_rate": 5e-05, "loss": 0.7742, "mean_token_accuracy": 0.8770439028739929, "num_tokens": 454594701.0, "step": 11971 }, { "epoch": 1.522961455285587, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.062098503112793, "learning_rate": 5e-05, "loss": 0.7774, "mean_token_accuracy": 0.8491407036781311, "num_tokens": 454638746.0, "step": 11972 }, { "epoch": 1.5230886655641775, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.988433361053467, "learning_rate": 5e-05, "loss": 0.7605, "mean_token_accuracy": 0.8754204511642456, "num_tokens": 454676040.0, "step": 11973 }, { "epoch": 1.523215875842768, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.383995532989502, "learning_rate": 5e-05, "loss": 0.7302, "mean_token_accuracy": 0.8693200945854187, "num_tokens": 454713262.0, "step": 11974 }, { "epoch": 1.5233430861213586, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.587549686431885, "learning_rate": 5e-05, "loss": 0.7465, "mean_token_accuracy": 0.8684566020965576, "num_tokens": 454748613.0, "step": 11975 }, { "epoch": 1.5234702963999491, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.640050888061523, "learning_rate": 5e-05, "loss": 0.6788, "mean_token_accuracy": 0.889778733253479, "num_tokens": 454782040.0, "step": 11976 }, { "epoch": 1.5235975066785397, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.441237926483154, "learning_rate": 5e-05, "loss": 0.7518, "mean_token_accuracy": 0.8609867095947266, "num_tokens": 454828394.0, "step": 11977 }, { "epoch": 1.5237247169571302, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.534367561340332, "learning_rate": 5e-05, "loss": 0.7497, "mean_token_accuracy": 0.8615838885307312, "num_tokens": 454866324.0, "step": 11978 }, { "epoch": 1.5238519272357207, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.6774773597717285, "learning_rate": 5e-05, "loss": 0.735, "mean_token_accuracy": 0.8746941089630127, "num_tokens": 454907458.0, "step": 11979 }, { "epoch": 1.5239791375143112, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.09896993637085, "learning_rate": 5e-05, "loss": 0.6565, "mean_token_accuracy": 0.8851497173309326, "num_tokens": 454948159.0, "step": 11980 }, { "epoch": 1.5241063477929018, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.25231647491455, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8795567750930786, "num_tokens": 454984760.0, "step": 11981 }, { "epoch": 1.5242335580714923, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00023746490478515625, "grad_norm": 6.426265716552734, "learning_rate": 5e-05, "loss": 0.6652, "mean_token_accuracy": 0.8701348304748535, "num_tokens": 455026670.0, "step": 11982 }, { "epoch": 1.5243607683500828, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 9.189704895019531, "learning_rate": 5e-05, "loss": 0.8855, "mean_token_accuracy": 0.8695800304412842, "num_tokens": 455063470.0, "step": 11983 }, { "epoch": 1.5244879786286734, "ewc_loss": 0.265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00021076202392578125, "grad_norm": 5.679957389831543, "learning_rate": 5e-05, "loss": 0.594, "mean_token_accuracy": 0.8845700025558472, "num_tokens": 455097787.0, "step": 11984 }, { "epoch": 1.5246151889072637, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 10.457552909851074, "learning_rate": 5e-05, "loss": 0.9739, "mean_token_accuracy": 0.8873275518417358, "num_tokens": 455137987.0, "step": 11985 }, { "epoch": 1.5247423991858542, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 5.954651355743408, "learning_rate": 5e-05, "loss": 0.7557, "mean_token_accuracy": 0.8556444048881531, "num_tokens": 455184048.0, "step": 11986 }, { "epoch": 1.5248696094644447, "ewc_loss": 0.82421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00077056884765625, "grad_norm": 11.222582817077637, "learning_rate": 5e-05, "loss": 1.174, "mean_token_accuracy": 0.8768091201782227, "num_tokens": 455219876.0, "step": 11987 }, { "epoch": 1.5249968197430352, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 8.423344612121582, "learning_rate": 5e-05, "loss": 1.0536, "mean_token_accuracy": 0.8722174763679504, "num_tokens": 455253491.0, "step": 11988 }, { "epoch": 1.5251240300216258, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.629364013671875, "learning_rate": 5e-05, "loss": 1.0018, "mean_token_accuracy": 0.867958664894104, "num_tokens": 455293183.0, "step": 11989 }, { "epoch": 1.5252512403002163, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000732421875, "grad_norm": 9.707422256469727, "learning_rate": 5e-05, "loss": 1.1186, "mean_token_accuracy": 0.8860602378845215, "num_tokens": 455333970.0, "step": 11990 }, { "epoch": 1.5253784505788066, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.370863914489746, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8693838119506836, "num_tokens": 455371547.0, "step": 11991 }, { "epoch": 1.5255056608573971, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000843048095703125, "grad_norm": 11.706299781799316, "learning_rate": 5e-05, "loss": 1.2807, "mean_token_accuracy": 0.871696412563324, "num_tokens": 455403427.0, "step": 11992 }, { "epoch": 1.5256328711359877, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.37031078338623, "learning_rate": 5e-05, "loss": 0.9518, "mean_token_accuracy": 0.8933030962944031, "num_tokens": 455434707.0, "step": 11993 }, { "epoch": 1.5257600814145782, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00079345703125, "grad_norm": 12.171120643615723, "learning_rate": 5e-05, "loss": 1.2016, "mean_token_accuracy": 0.876708984375, "num_tokens": 455473255.0, "step": 11994 }, { "epoch": 1.5258872916931687, "ewc_loss": 0.8828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000827789306640625, "grad_norm": 9.904624938964844, "learning_rate": 5e-05, "loss": 1.2519, "mean_token_accuracy": 0.8706775903701782, "num_tokens": 455504631.0, "step": 11995 }, { "epoch": 1.5260145019717593, "ewc_loss": 0.75, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 11.465415000915527, "learning_rate": 5e-05, "loss": 1.0899, "mean_token_accuracy": 0.8816964626312256, "num_tokens": 455542652.0, "step": 11996 }, { "epoch": 1.5261417122503498, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.051030158996582, "learning_rate": 5e-05, "loss": 1.0447, "mean_token_accuracy": 0.8757603764533997, "num_tokens": 455577912.0, "step": 11997 }, { "epoch": 1.5262689225289403, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 11.550069808959961, "learning_rate": 5e-05, "loss": 1.1564, "mean_token_accuracy": 0.8780720829963684, "num_tokens": 455617900.0, "step": 11998 }, { "epoch": 1.5263961328075308, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.425482749938965, "learning_rate": 5e-05, "loss": 0.9398, "mean_token_accuracy": 0.8783150911331177, "num_tokens": 455656779.0, "step": 11999 }, { "epoch": 1.5265233430861214, "ewc_loss": 0.875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 12.11949634552002, "learning_rate": 5e-05, "loss": 1.2308, "mean_token_accuracy": 0.876563549041748, "num_tokens": 455692620.0, "step": 12000 }, { "epoch": 1.526650553364712, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.30524730682373, "learning_rate": 5e-05, "loss": 1.0328, "mean_token_accuracy": 0.8763476610183716, "num_tokens": 455729805.0, "step": 12001 }, { "epoch": 1.5267777636433024, "ewc_loss": 0.859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 12.153278350830078, "learning_rate": 5e-05, "loss": 1.217, "mean_token_accuracy": 0.8784338235855103, "num_tokens": 455767529.0, "step": 12002 }, { "epoch": 1.526904973921893, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 10.340923309326172, "learning_rate": 5e-05, "loss": 1.1454, "mean_token_accuracy": 0.8715690970420837, "num_tokens": 455807898.0, "step": 12003 }, { "epoch": 1.5270321842004835, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 11.145580291748047, "learning_rate": 5e-05, "loss": 1.1025, "mean_token_accuracy": 0.8696989417076111, "num_tokens": 455848726.0, "step": 12004 }, { "epoch": 1.527159394479074, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 9.967077255249023, "learning_rate": 5e-05, "loss": 1.0737, "mean_token_accuracy": 0.8704884052276611, "num_tokens": 455892363.0, "step": 12005 }, { "epoch": 1.5272866047576645, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.399709701538086, "learning_rate": 5e-05, "loss": 1.0698, "mean_token_accuracy": 0.8615601658821106, "num_tokens": 455930104.0, "step": 12006 }, { "epoch": 1.527413815036255, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.310724258422852, "learning_rate": 5e-05, "loss": 0.9845, "mean_token_accuracy": 0.8682210445404053, "num_tokens": 455964272.0, "step": 12007 }, { "epoch": 1.5275410253148456, "ewc_loss": 0.625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.93769645690918, "learning_rate": 5e-05, "loss": 1.0, "mean_token_accuracy": 0.8702166676521301, "num_tokens": 456003020.0, "step": 12008 }, { "epoch": 1.527668235593436, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.558662414550781, "learning_rate": 5e-05, "loss": 0.8689, "mean_token_accuracy": 0.8730301260948181, "num_tokens": 456040305.0, "step": 12009 }, { "epoch": 1.5277954458720264, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.992020606994629, "learning_rate": 5e-05, "loss": 1.0251, "mean_token_accuracy": 0.8649249076843262, "num_tokens": 456080477.0, "step": 12010 }, { "epoch": 1.527922656150617, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.043061256408691, "learning_rate": 5e-05, "loss": 0.8102, "mean_token_accuracy": 0.880318284034729, "num_tokens": 456116941.0, "step": 12011 }, { "epoch": 1.5280498664292075, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.489167213439941, "learning_rate": 5e-05, "loss": 1.0622, "mean_token_accuracy": 0.8665973544120789, "num_tokens": 456156184.0, "step": 12012 }, { "epoch": 1.528177076707798, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.235807418823242, "learning_rate": 5e-05, "loss": 0.9087, "mean_token_accuracy": 0.8573899865150452, "num_tokens": 456200920.0, "step": 12013 }, { "epoch": 1.5283042869863885, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.822991371154785, "learning_rate": 5e-05, "loss": 1.0723, "mean_token_accuracy": 0.8801397681236267, "num_tokens": 456235632.0, "step": 12014 }, { "epoch": 1.5284314972649788, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.092080116271973, "learning_rate": 5e-05, "loss": 0.9286, "mean_token_accuracy": 0.8748692274093628, "num_tokens": 456264898.0, "step": 12015 }, { "epoch": 1.5285587075435694, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.111260414123535, "learning_rate": 5e-05, "loss": 1.0359, "mean_token_accuracy": 0.869620144367218, "num_tokens": 456304972.0, "step": 12016 }, { "epoch": 1.52868591782216, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.21054744720459, "learning_rate": 5e-05, "loss": 0.9591, "mean_token_accuracy": 0.8708347082138062, "num_tokens": 456344496.0, "step": 12017 }, { "epoch": 1.5288131281007504, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.137845993041992, "learning_rate": 5e-05, "loss": 0.9118, "mean_token_accuracy": 0.8780993223190308, "num_tokens": 456382408.0, "step": 12018 }, { "epoch": 1.528940338379341, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.713801383972168, "learning_rate": 5e-05, "loss": 0.892, "mean_token_accuracy": 0.8658676147460938, "num_tokens": 456424946.0, "step": 12019 }, { "epoch": 1.5290675486579315, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.646081924438477, "learning_rate": 5e-05, "loss": 0.846, "mean_token_accuracy": 0.8813756704330444, "num_tokens": 456467432.0, "step": 12020 }, { "epoch": 1.529194758936522, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.070538520812988, "learning_rate": 5e-05, "loss": 0.7775, "mean_token_accuracy": 0.878296434879303, "num_tokens": 456501557.0, "step": 12021 }, { "epoch": 1.5293219692151125, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.629899978637695, "learning_rate": 5e-05, "loss": 0.9514, "mean_token_accuracy": 0.8442015051841736, "num_tokens": 456540500.0, "step": 12022 }, { "epoch": 1.529449179493703, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.4039106369018555, "learning_rate": 5e-05, "loss": 0.7209, "mean_token_accuracy": 0.877839207649231, "num_tokens": 456574403.0, "step": 12023 }, { "epoch": 1.5295763897722936, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.977853775024414, "learning_rate": 5e-05, "loss": 0.8825, "mean_token_accuracy": 0.87722247838974, "num_tokens": 456610044.0, "step": 12024 }, { "epoch": 1.5297036000508841, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.050119400024414, "learning_rate": 5e-05, "loss": 0.7032, "mean_token_accuracy": 0.8763349652290344, "num_tokens": 456650289.0, "step": 12025 }, { "epoch": 1.5298308103294747, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.508172035217285, "learning_rate": 5e-05, "loss": 0.9615, "mean_token_accuracy": 0.8707753419876099, "num_tokens": 456687193.0, "step": 12026 }, { "epoch": 1.5299580206080652, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.241128444671631, "learning_rate": 5e-05, "loss": 0.7604, "mean_token_accuracy": 0.8727544546127319, "num_tokens": 456720553.0, "step": 12027 }, { "epoch": 1.5300852308866557, "ewc_loss": 0.625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.72402286529541, "learning_rate": 5e-05, "loss": 0.9969, "mean_token_accuracy": 0.8671495914459229, "num_tokens": 456754736.0, "step": 12028 }, { "epoch": 1.5302124411652462, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.0911283493042, "learning_rate": 5e-05, "loss": 0.7862, "mean_token_accuracy": 0.8861649632453918, "num_tokens": 456791268.0, "step": 12029 }, { "epoch": 1.5303396514438368, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.897238731384277, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.8824423551559448, "num_tokens": 456831728.0, "step": 12030 }, { "epoch": 1.5304668617224273, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.5457124710083, "learning_rate": 5e-05, "loss": 0.8298, "mean_token_accuracy": 0.8776290416717529, "num_tokens": 456869375.0, "step": 12031 }, { "epoch": 1.5305940720010178, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.720855236053467, "learning_rate": 5e-05, "loss": 0.8102, "mean_token_accuracy": 0.8640496134757996, "num_tokens": 456909376.0, "step": 12032 }, { "epoch": 1.5307212822796084, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.569035530090332, "learning_rate": 5e-05, "loss": 0.8015, "mean_token_accuracy": 0.8847317695617676, "num_tokens": 456945851.0, "step": 12033 }, { "epoch": 1.5308484925581987, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.208175182342529, "learning_rate": 5e-05, "loss": 0.7423, "mean_token_accuracy": 0.8716509938240051, "num_tokens": 456979833.0, "step": 12034 }, { "epoch": 1.5309757028367892, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.672844886779785, "learning_rate": 5e-05, "loss": 0.7778, "mean_token_accuracy": 0.890268087387085, "num_tokens": 457016032.0, "step": 12035 }, { "epoch": 1.5311029131153797, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.012693881988525, "learning_rate": 5e-05, "loss": 0.6932, "mean_token_accuracy": 0.8876955509185791, "num_tokens": 457050015.0, "step": 12036 }, { "epoch": 1.5312301233939702, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.940641403198242, "learning_rate": 5e-05, "loss": 0.8716, "mean_token_accuracy": 0.8746041059494019, "num_tokens": 457087287.0, "step": 12037 }, { "epoch": 1.5313573336725608, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.1280364990234375, "learning_rate": 5e-05, "loss": 0.7332, "mean_token_accuracy": 0.8707960247993469, "num_tokens": 457123237.0, "step": 12038 }, { "epoch": 1.5314845439511513, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.768658638000488, "learning_rate": 5e-05, "loss": 0.8741, "mean_token_accuracy": 0.8702664375305176, "num_tokens": 457161834.0, "step": 12039 }, { "epoch": 1.5316117542297416, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.617931842803955, "learning_rate": 5e-05, "loss": 0.808, "mean_token_accuracy": 0.8618010878562927, "num_tokens": 457203672.0, "step": 12040 }, { "epoch": 1.5317389645083321, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.028331756591797, "learning_rate": 5e-05, "loss": 0.7739, "mean_token_accuracy": 0.883820652961731, "num_tokens": 457238486.0, "step": 12041 }, { "epoch": 1.5318661747869227, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.019876480102539, "learning_rate": 5e-05, "loss": 0.814, "mean_token_accuracy": 0.8578968048095703, "num_tokens": 457270817.0, "step": 12042 }, { "epoch": 1.5319933850655132, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.4831719398498535, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8723419308662415, "num_tokens": 457306812.0, "step": 12043 }, { "epoch": 1.5321205953441037, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.064901351928711, "learning_rate": 5e-05, "loss": 0.7386, "mean_token_accuracy": 0.8821138739585876, "num_tokens": 457340893.0, "step": 12044 }, { "epoch": 1.5322478056226942, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.458817481994629, "learning_rate": 5e-05, "loss": 0.7285, "mean_token_accuracy": 0.8756083250045776, "num_tokens": 457381668.0, "step": 12045 }, { "epoch": 1.5323750159012848, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.883277416229248, "learning_rate": 5e-05, "loss": 0.7791, "mean_token_accuracy": 0.8614523410797119, "num_tokens": 457417498.0, "step": 12046 }, { "epoch": 1.5325022261798753, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.58795166015625, "learning_rate": 5e-05, "loss": 0.7349, "mean_token_accuracy": 0.8724308013916016, "num_tokens": 457460300.0, "step": 12047 }, { "epoch": 1.5326294364584658, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.620832443237305, "learning_rate": 5e-05, "loss": 0.7721, "mean_token_accuracy": 0.86300128698349, "num_tokens": 457502030.0, "step": 12048 }, { "epoch": 1.5327566467370564, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.792654514312744, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.881088137626648, "num_tokens": 457540170.0, "step": 12049 }, { "epoch": 1.532883857015647, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.279183864593506, "learning_rate": 5e-05, "loss": 0.696, "mean_token_accuracy": 0.877777099609375, "num_tokens": 457585834.0, "step": 12050 }, { "epoch": 1.5330110672942374, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 8.141678810119629, "learning_rate": 5e-05, "loss": 0.7315, "mean_token_accuracy": 0.8828295469284058, "num_tokens": 457622027.0, "step": 12051 }, { "epoch": 1.533138277572828, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.937036514282227, "learning_rate": 5e-05, "loss": 0.7142, "mean_token_accuracy": 0.8651362657546997, "num_tokens": 457659129.0, "step": 12052 }, { "epoch": 1.5332654878514185, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.482276916503906, "learning_rate": 5e-05, "loss": 0.7473, "mean_token_accuracy": 0.8894264698028564, "num_tokens": 457695046.0, "step": 12053 }, { "epoch": 1.533392698130009, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.714641571044922, "learning_rate": 5e-05, "loss": 0.6549, "mean_token_accuracy": 0.8797735571861267, "num_tokens": 457735650.0, "step": 12054 }, { "epoch": 1.5335199084085995, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.730042457580566, "learning_rate": 5e-05, "loss": 0.8099, "mean_token_accuracy": 0.8803635239601135, "num_tokens": 457767395.0, "step": 12055 }, { "epoch": 1.53364711868719, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.59833288192749, "learning_rate": 5e-05, "loss": 0.6262, "mean_token_accuracy": 0.8879793882369995, "num_tokens": 457800100.0, "step": 12056 }, { "epoch": 1.5337743289657806, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.949712753295898, "learning_rate": 5e-05, "loss": 0.8596, "mean_token_accuracy": 0.8766599893569946, "num_tokens": 457837497.0, "step": 12057 }, { "epoch": 1.533901539244371, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.5848822593688965, "learning_rate": 5e-05, "loss": 0.6772, "mean_token_accuracy": 0.8740599155426025, "num_tokens": 457875711.0, "step": 12058 }, { "epoch": 1.5340287495229614, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 9.127620697021484, "learning_rate": 5e-05, "loss": 0.9316, "mean_token_accuracy": 0.8616458177566528, "num_tokens": 457914888.0, "step": 12059 }, { "epoch": 1.534155959801552, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.733884811401367, "learning_rate": 5e-05, "loss": 0.7533, "mean_token_accuracy": 0.8580611944198608, "num_tokens": 457953315.0, "step": 12060 }, { "epoch": 1.5342831700801425, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.129075050354004, "learning_rate": 5e-05, "loss": 0.971, "mean_token_accuracy": 0.8602429628372192, "num_tokens": 457990857.0, "step": 12061 }, { "epoch": 1.534410380358733, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.096681118011475, "learning_rate": 5e-05, "loss": 0.7711, "mean_token_accuracy": 0.8676140308380127, "num_tokens": 458034060.0, "step": 12062 }, { "epoch": 1.5345375906373235, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.866751670837402, "learning_rate": 5e-05, "loss": 0.9397, "mean_token_accuracy": 0.8660605549812317, "num_tokens": 458074093.0, "step": 12063 }, { "epoch": 1.5346648009159138, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.501865863800049, "learning_rate": 5e-05, "loss": 0.7746, "mean_token_accuracy": 0.878768801689148, "num_tokens": 458107374.0, "step": 12064 }, { "epoch": 1.5347920111945044, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.392202377319336, "learning_rate": 5e-05, "loss": 0.8791, "mean_token_accuracy": 0.8743468523025513, "num_tokens": 458141873.0, "step": 12065 }, { "epoch": 1.534919221473095, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.709906578063965, "learning_rate": 5e-05, "loss": 0.8174, "mean_token_accuracy": 0.8723629713058472, "num_tokens": 458181483.0, "step": 12066 }, { "epoch": 1.5350464317516854, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.05705451965332, "learning_rate": 5e-05, "loss": 0.8617, "mean_token_accuracy": 0.8653625249862671, "num_tokens": 458223644.0, "step": 12067 }, { "epoch": 1.535173642030276, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.630388259887695, "learning_rate": 5e-05, "loss": 0.7816, "mean_token_accuracy": 0.8738059997558594, "num_tokens": 458263518.0, "step": 12068 }, { "epoch": 1.5353008523088665, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.023834228515625, "learning_rate": 5e-05, "loss": 0.7904, "mean_token_accuracy": 0.8785157203674316, "num_tokens": 458296621.0, "step": 12069 }, { "epoch": 1.535428062587457, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.4389238357543945, "learning_rate": 5e-05, "loss": 0.7788, "mean_token_accuracy": 0.8676895499229431, "num_tokens": 458335161.0, "step": 12070 }, { "epoch": 1.5355552728660475, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.087533950805664, "learning_rate": 5e-05, "loss": 0.8472, "mean_token_accuracy": 0.8615232706069946, "num_tokens": 458373758.0, "step": 12071 }, { "epoch": 1.535682483144638, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.331231117248535, "learning_rate": 5e-05, "loss": 0.7273, "mean_token_accuracy": 0.8757598400115967, "num_tokens": 458412662.0, "step": 12072 }, { "epoch": 1.5358096934232286, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.051013946533203, "learning_rate": 5e-05, "loss": 0.8309, "mean_token_accuracy": 0.8642584681510925, "num_tokens": 458455314.0, "step": 12073 }, { "epoch": 1.5359369037018191, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.350413799285889, "learning_rate": 5e-05, "loss": 0.8037, "mean_token_accuracy": 0.8519445657730103, "num_tokens": 458498633.0, "step": 12074 }, { "epoch": 1.5360641139804097, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.930459499359131, "learning_rate": 5e-05, "loss": 0.7939, "mean_token_accuracy": 0.8697491884231567, "num_tokens": 458538920.0, "step": 12075 }, { "epoch": 1.5361913242590002, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.389922618865967, "learning_rate": 5e-05, "loss": 0.7392, "mean_token_accuracy": 0.8708903789520264, "num_tokens": 458574333.0, "step": 12076 }, { "epoch": 1.5363185345375907, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.879997730255127, "learning_rate": 5e-05, "loss": 0.7978, "mean_token_accuracy": 0.8644058108329773, "num_tokens": 458612131.0, "step": 12077 }, { "epoch": 1.5364457448161812, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.35262393951416, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8686549067497253, "num_tokens": 458652169.0, "step": 12078 }, { "epoch": 1.5365729550947718, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.897505760192871, "learning_rate": 5e-05, "loss": 0.8095, "mean_token_accuracy": 0.8667629361152649, "num_tokens": 458687816.0, "step": 12079 }, { "epoch": 1.5367001653733623, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.289669513702393, "learning_rate": 5e-05, "loss": 0.7006, "mean_token_accuracy": 0.8760777711868286, "num_tokens": 458723600.0, "step": 12080 }, { "epoch": 1.5368273756519528, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.941096782684326, "learning_rate": 5e-05, "loss": 0.8114, "mean_token_accuracy": 0.8573943376541138, "num_tokens": 458761863.0, "step": 12081 }, { "epoch": 1.5369545859305433, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.224434852600098, "learning_rate": 5e-05, "loss": 0.7531, "mean_token_accuracy": 0.8615891337394714, "num_tokens": 458794986.0, "step": 12082 }, { "epoch": 1.5370817962091337, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.983006954193115, "learning_rate": 5e-05, "loss": 0.8003, "mean_token_accuracy": 0.8625255823135376, "num_tokens": 458833939.0, "step": 12083 }, { "epoch": 1.5372090064877242, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.2119245529174805, "learning_rate": 5e-05, "loss": 0.7555, "mean_token_accuracy": 0.8605890870094299, "num_tokens": 458874522.0, "step": 12084 }, { "epoch": 1.5373362167663147, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.969437122344971, "learning_rate": 5e-05, "loss": 0.8246, "mean_token_accuracy": 0.8528895378112793, "num_tokens": 458911794.0, "step": 12085 }, { "epoch": 1.5374634270449052, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.2230682373046875, "learning_rate": 5e-05, "loss": 0.7221, "mean_token_accuracy": 0.870826244354248, "num_tokens": 458951979.0, "step": 12086 }, { "epoch": 1.5375906373234958, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.944095611572266, "learning_rate": 5e-05, "loss": 0.7865, "mean_token_accuracy": 0.8629130125045776, "num_tokens": 458993657.0, "step": 12087 }, { "epoch": 1.5377178476020863, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.241909503936768, "learning_rate": 5e-05, "loss": 0.6704, "mean_token_accuracy": 0.8826867341995239, "num_tokens": 459030216.0, "step": 12088 }, { "epoch": 1.5378450578806766, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.897512912750244, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.8602323532104492, "num_tokens": 459066451.0, "step": 12089 }, { "epoch": 1.5379722681592671, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.274441242218018, "learning_rate": 5e-05, "loss": 0.6895, "mean_token_accuracy": 0.8813658952713013, "num_tokens": 459106928.0, "step": 12090 }, { "epoch": 1.5380994784378577, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.831669330596924, "learning_rate": 5e-05, "loss": 0.7861, "mean_token_accuracy": 0.8602291941642761, "num_tokens": 459142769.0, "step": 12091 }, { "epoch": 1.5382266887164482, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.325263500213623, "learning_rate": 5e-05, "loss": 0.7064, "mean_token_accuracy": 0.8744705319404602, "num_tokens": 459178895.0, "step": 12092 }, { "epoch": 1.5383538989950387, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.734899997711182, "learning_rate": 5e-05, "loss": 0.6987, "mean_token_accuracy": 0.8848273754119873, "num_tokens": 459214973.0, "step": 12093 }, { "epoch": 1.5384811092736292, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.429472923278809, "learning_rate": 5e-05, "loss": 0.7219, "mean_token_accuracy": 0.8692708015441895, "num_tokens": 459251279.0, "step": 12094 }, { "epoch": 1.5386083195522198, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.573507785797119, "learning_rate": 5e-05, "loss": 0.7734, "mean_token_accuracy": 0.8513529300689697, "num_tokens": 459285587.0, "step": 12095 }, { "epoch": 1.5387355298308103, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.587661266326904, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8588680028915405, "num_tokens": 459326103.0, "step": 12096 }, { "epoch": 1.5388627401094008, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.37688684463501, "learning_rate": 5e-05, "loss": 0.7211, "mean_token_accuracy": 0.869938850402832, "num_tokens": 459367075.0, "step": 12097 }, { "epoch": 1.5389899503879914, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.801444053649902, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8639733791351318, "num_tokens": 459408164.0, "step": 12098 }, { "epoch": 1.5391171606665819, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 7.098602294921875, "learning_rate": 5e-05, "loss": 0.7135, "mean_token_accuracy": 0.8626641035079956, "num_tokens": 459443081.0, "step": 12099 }, { "epoch": 1.5392443709451724, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 8.135765075683594, "learning_rate": 5e-05, "loss": 0.7851, "mean_token_accuracy": 0.8662678003311157, "num_tokens": 459481469.0, "step": 12100 }, { "epoch": 1.539371581223763, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.6918792724609375, "learning_rate": 5e-05, "loss": 0.6747, "mean_token_accuracy": 0.8691970109939575, "num_tokens": 459519319.0, "step": 12101 }, { "epoch": 1.5394987915023535, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.726000785827637, "learning_rate": 5e-05, "loss": 0.8111, "mean_token_accuracy": 0.8722858428955078, "num_tokens": 459559998.0, "step": 12102 }, { "epoch": 1.539626001780944, "ewc_loss": 0.267578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000213623046875, "grad_norm": 6.035563945770264, "learning_rate": 5e-05, "loss": 0.5947, "mean_token_accuracy": 0.8848974704742432, "num_tokens": 459596765.0, "step": 12103 }, { "epoch": 1.5397532120595345, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.798129081726074, "learning_rate": 5e-05, "loss": 0.9902, "mean_token_accuracy": 0.8539725542068481, "num_tokens": 459634023.0, "step": 12104 }, { "epoch": 1.539880422338125, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 5.3790059089660645, "learning_rate": 5e-05, "loss": 0.5935, "mean_token_accuracy": 0.8864644765853882, "num_tokens": 459675355.0, "step": 12105 }, { "epoch": 1.5400076326167156, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000732421875, "grad_norm": 11.30343246459961, "learning_rate": 5e-05, "loss": 1.1978, "mean_token_accuracy": 0.861066460609436, "num_tokens": 459716788.0, "step": 12106 }, { "epoch": 1.5401348428953059, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 6.7578511238098145, "learning_rate": 5e-05, "loss": 0.866, "mean_token_accuracy": 0.8753084540367126, "num_tokens": 459752580.0, "step": 12107 }, { "epoch": 1.5402620531738964, "ewc_loss": 0.90625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 11.776366233825684, "learning_rate": 5e-05, "loss": 1.2622, "mean_token_accuracy": 0.8760513067245483, "num_tokens": 459788594.0, "step": 12108 }, { "epoch": 1.540389263452487, "ewc_loss": 0.9609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00090789794921875, "grad_norm": 9.530876159667969, "learning_rate": 5e-05, "loss": 1.3024, "mean_token_accuracy": 0.8825057744979858, "num_tokens": 459826806.0, "step": 12109 }, { "epoch": 1.5405164737310775, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.799824714660645, "learning_rate": 5e-05, "loss": 0.9585, "mean_token_accuracy": 0.8764645457267761, "num_tokens": 459870157.0, "step": 12110 }, { "epoch": 1.540643684009668, "ewc_loss": 0.9140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 10.143451690673828, "learning_rate": 5e-05, "loss": 1.2735, "mean_token_accuracy": 0.8768705725669861, "num_tokens": 459909081.0, "step": 12111 }, { "epoch": 1.5407708942882585, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.15864086151123, "learning_rate": 5e-05, "loss": 0.8817, "mean_token_accuracy": 0.8766458630561829, "num_tokens": 459942886.0, "step": 12112 }, { "epoch": 1.5408981045668488, "ewc_loss": 0.88671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 11.410734176635742, "learning_rate": 5e-05, "loss": 1.2627, "mean_token_accuracy": 0.8713960647583008, "num_tokens": 459979351.0, "step": 12113 }, { "epoch": 1.5410253148454394, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 8.906454086303711, "learning_rate": 5e-05, "loss": 1.1055, "mean_token_accuracy": 0.8671364784240723, "num_tokens": 460026058.0, "step": 12114 }, { "epoch": 1.54115252512403, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 11.16609001159668, "learning_rate": 5e-05, "loss": 1.1018, "mean_token_accuracy": 0.8687514066696167, "num_tokens": 460066414.0, "step": 12115 }, { "epoch": 1.5412797354026204, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 9.240314483642578, "learning_rate": 5e-05, "loss": 1.1926, "mean_token_accuracy": 0.8669735193252563, "num_tokens": 460107747.0, "step": 12116 }, { "epoch": 1.541406945681211, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.416014671325684, "learning_rate": 5e-05, "loss": 1.0238, "mean_token_accuracy": 0.8751497268676758, "num_tokens": 460144418.0, "step": 12117 }, { "epoch": 1.5415341559598015, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 8.692405700683594, "learning_rate": 5e-05, "loss": 1.0677, "mean_token_accuracy": 0.8617516756057739, "num_tokens": 460186061.0, "step": 12118 }, { "epoch": 1.541661366238392, "ewc_loss": 0.625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.640853881835938, "learning_rate": 5e-05, "loss": 1.011, "mean_token_accuracy": 0.8644028902053833, "num_tokens": 460222672.0, "step": 12119 }, { "epoch": 1.5417885765169825, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.236151695251465, "learning_rate": 5e-05, "loss": 0.9195, "mean_token_accuracy": 0.8798213005065918, "num_tokens": 460265992.0, "step": 12120 }, { "epoch": 1.541915786795573, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.897272109985352, "learning_rate": 5e-05, "loss": 0.9189, "mean_token_accuracy": 0.8716005086898804, "num_tokens": 460305498.0, "step": 12121 }, { "epoch": 1.5420429970741636, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.082127571105957, "learning_rate": 5e-05, "loss": 0.9079, "mean_token_accuracy": 0.8685941696166992, "num_tokens": 460336013.0, "step": 12122 }, { "epoch": 1.5421702073527541, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.217264175415039, "learning_rate": 5e-05, "loss": 0.8423, "mean_token_accuracy": 0.8717614412307739, "num_tokens": 460373452.0, "step": 12123 }, { "epoch": 1.5422974176313446, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.160072326660156, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.8734204769134521, "num_tokens": 460414059.0, "step": 12124 }, { "epoch": 1.5424246279099352, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.51356315612793, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.8860020041465759, "num_tokens": 460446098.0, "step": 12125 }, { "epoch": 1.5425518381885257, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.620471954345703, "learning_rate": 5e-05, "loss": 0.8625, "mean_token_accuracy": 0.8720861077308655, "num_tokens": 460485384.0, "step": 12126 }, { "epoch": 1.5426790484671162, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.8029961585998535, "learning_rate": 5e-05, "loss": 0.7397, "mean_token_accuracy": 0.8644043207168579, "num_tokens": 460517200.0, "step": 12127 }, { "epoch": 1.5428062587457068, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.428110122680664, "learning_rate": 5e-05, "loss": 0.9252, "mean_token_accuracy": 0.8799534440040588, "num_tokens": 460547557.0, "step": 12128 }, { "epoch": 1.5429334690242973, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.423144817352295, "learning_rate": 5e-05, "loss": 0.7129, "mean_token_accuracy": 0.8796460628509521, "num_tokens": 460583189.0, "step": 12129 }, { "epoch": 1.5430606793028878, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.292913436889648, "learning_rate": 5e-05, "loss": 1.0303, "mean_token_accuracy": 0.8767236471176147, "num_tokens": 460623619.0, "step": 12130 }, { "epoch": 1.5431878895814783, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.202335834503174, "learning_rate": 5e-05, "loss": 0.877, "mean_token_accuracy": 0.8688914179801941, "num_tokens": 460666667.0, "step": 12131 }, { "epoch": 1.5433150998600687, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.31982135772705, "learning_rate": 5e-05, "loss": 1.0563, "mean_token_accuracy": 0.8727518320083618, "num_tokens": 460704292.0, "step": 12132 }, { "epoch": 1.5434423101386592, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 8.565692901611328, "learning_rate": 5e-05, "loss": 1.0384, "mean_token_accuracy": 0.8747676014900208, "num_tokens": 460748153.0, "step": 12133 }, { "epoch": 1.5435695204172497, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.927536010742188, "learning_rate": 5e-05, "loss": 0.9022, "mean_token_accuracy": 0.8765969276428223, "num_tokens": 460787663.0, "step": 12134 }, { "epoch": 1.5436967306958402, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.13266658782959, "learning_rate": 5e-05, "loss": 1.0177, "mean_token_accuracy": 0.8782418966293335, "num_tokens": 460821643.0, "step": 12135 }, { "epoch": 1.5438239409744308, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.468991279602051, "learning_rate": 5e-05, "loss": 0.8187, "mean_token_accuracy": 0.8647561073303223, "num_tokens": 460857180.0, "step": 12136 }, { "epoch": 1.5439511512530213, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.817793846130371, "learning_rate": 5e-05, "loss": 1.0263, "mean_token_accuracy": 0.8696964383125305, "num_tokens": 460891573.0, "step": 12137 }, { "epoch": 1.5440783615316116, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.27438497543335, "learning_rate": 5e-05, "loss": 0.7942, "mean_token_accuracy": 0.883424699306488, "num_tokens": 460927727.0, "step": 12138 }, { "epoch": 1.5442055718102021, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.993898391723633, "learning_rate": 5e-05, "loss": 1.0796, "mean_token_accuracy": 0.8620070815086365, "num_tokens": 460965531.0, "step": 12139 }, { "epoch": 1.5443327820887927, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.130949020385742, "learning_rate": 5e-05, "loss": 0.9883, "mean_token_accuracy": 0.8720401525497437, "num_tokens": 461002900.0, "step": 12140 }, { "epoch": 1.5444599923673832, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.091985702514648, "learning_rate": 5e-05, "loss": 0.958, "mean_token_accuracy": 0.87098228931427, "num_tokens": 461041569.0, "step": 12141 }, { "epoch": 1.5445872026459737, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.571518898010254, "learning_rate": 5e-05, "loss": 1.0258, "mean_token_accuracy": 0.8627262115478516, "num_tokens": 461080453.0, "step": 12142 }, { "epoch": 1.5447144129245642, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.979705333709717, "learning_rate": 5e-05, "loss": 0.8321, "mean_token_accuracy": 0.8691168427467346, "num_tokens": 461120963.0, "step": 12143 }, { "epoch": 1.5448416232031548, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.76508903503418, "learning_rate": 5e-05, "loss": 0.9427, "mean_token_accuracy": 0.8769199848175049, "num_tokens": 461156441.0, "step": 12144 }, { "epoch": 1.5449688334817453, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.406423091888428, "learning_rate": 5e-05, "loss": 0.7504, "mean_token_accuracy": 0.8804978132247925, "num_tokens": 461189058.0, "step": 12145 }, { "epoch": 1.5450960437603358, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.94875431060791, "learning_rate": 5e-05, "loss": 0.9129, "mean_token_accuracy": 0.8791933059692383, "num_tokens": 461223962.0, "step": 12146 }, { "epoch": 1.5452232540389264, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.262345314025879, "learning_rate": 5e-05, "loss": 0.7816, "mean_token_accuracy": 0.8799886107444763, "num_tokens": 461262114.0, "step": 12147 }, { "epoch": 1.5453504643175169, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.05678653717041, "learning_rate": 5e-05, "loss": 0.9368, "mean_token_accuracy": 0.8725889921188354, "num_tokens": 461294633.0, "step": 12148 }, { "epoch": 1.5454776745961074, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.371981143951416, "learning_rate": 5e-05, "loss": 0.8578, "mean_token_accuracy": 0.8667224049568176, "num_tokens": 461338549.0, "step": 12149 }, { "epoch": 1.545604884874698, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.89748764038086, "learning_rate": 5e-05, "loss": 0.9319, "mean_token_accuracy": 0.8691734075546265, "num_tokens": 461372612.0, "step": 12150 }, { "epoch": 1.5457320951532885, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.550098896026611, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8619662523269653, "num_tokens": 461404060.0, "step": 12151 }, { "epoch": 1.545859305431879, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.482559204101562, "learning_rate": 5e-05, "loss": 0.8755, "mean_token_accuracy": 0.870113730430603, "num_tokens": 461436728.0, "step": 12152 }, { "epoch": 1.5459865157104695, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.620075702667236, "learning_rate": 5e-05, "loss": 0.819, "mean_token_accuracy": 0.8825809955596924, "num_tokens": 461475418.0, "step": 12153 }, { "epoch": 1.54611372598906, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.171319961547852, "learning_rate": 5e-05, "loss": 0.8561, "mean_token_accuracy": 0.8631155490875244, "num_tokens": 461512094.0, "step": 12154 }, { "epoch": 1.5462409362676506, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.552969455718994, "learning_rate": 5e-05, "loss": 0.8807, "mean_token_accuracy": 0.8552223443984985, "num_tokens": 461553454.0, "step": 12155 }, { "epoch": 1.5463681465462409, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.025369644165039, "learning_rate": 5e-05, "loss": 0.8823, "mean_token_accuracy": 0.8520591855049133, "num_tokens": 461590818.0, "step": 12156 }, { "epoch": 1.5464953568248314, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.488604545593262, "learning_rate": 5e-05, "loss": 0.745, "mean_token_accuracy": 0.8849425911903381, "num_tokens": 461638692.0, "step": 12157 }, { "epoch": 1.546622567103422, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.871635437011719, "learning_rate": 5e-05, "loss": 0.7767, "mean_token_accuracy": 0.8735494613647461, "num_tokens": 461681040.0, "step": 12158 }, { "epoch": 1.5467497773820125, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.525400638580322, "learning_rate": 5e-05, "loss": 0.7654, "mean_token_accuracy": 0.8711295127868652, "num_tokens": 461716999.0, "step": 12159 }, { "epoch": 1.546876987660603, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.787937164306641, "learning_rate": 5e-05, "loss": 0.7909, "mean_token_accuracy": 0.8646949529647827, "num_tokens": 461755053.0, "step": 12160 }, { "epoch": 1.5470041979391935, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.492668151855469, "learning_rate": 5e-05, "loss": 0.74, "mean_token_accuracy": 0.8764256238937378, "num_tokens": 461795384.0, "step": 12161 }, { "epoch": 1.5471314082177838, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.774764060974121, "learning_rate": 5e-05, "loss": 0.7363, "mean_token_accuracy": 0.8748945593833923, "num_tokens": 461843951.0, "step": 12162 }, { "epoch": 1.5472586184963744, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.464390277862549, "learning_rate": 5e-05, "loss": 0.7885, "mean_token_accuracy": 0.8558986783027649, "num_tokens": 461884172.0, "step": 12163 }, { "epoch": 1.5473858287749649, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.732517242431641, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8684653043746948, "num_tokens": 461928493.0, "step": 12164 }, { "epoch": 1.5475130390535554, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.467133522033691, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.8755335807800293, "num_tokens": 461964462.0, "step": 12165 }, { "epoch": 1.547640249332146, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.6941142082214355, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.8784690499305725, "num_tokens": 462004123.0, "step": 12166 }, { "epoch": 1.5477674596107365, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.450244426727295, "learning_rate": 5e-05, "loss": 0.7817, "mean_token_accuracy": 0.8543870449066162, "num_tokens": 462042906.0, "step": 12167 }, { "epoch": 1.547894669889327, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.709625720977783, "learning_rate": 5e-05, "loss": 0.7016, "mean_token_accuracy": 0.883550763130188, "num_tokens": 462083802.0, "step": 12168 }, { "epoch": 1.5480218801679175, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.405560493469238, "learning_rate": 5e-05, "loss": 0.7322, "mean_token_accuracy": 0.8677601218223572, "num_tokens": 462119861.0, "step": 12169 }, { "epoch": 1.548149090446508, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.7407097816467285, "learning_rate": 5e-05, "loss": 0.7586, "mean_token_accuracy": 0.8644475936889648, "num_tokens": 462162445.0, "step": 12170 }, { "epoch": 1.5482763007250986, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.338316440582275, "learning_rate": 5e-05, "loss": 0.7164, "mean_token_accuracy": 0.8687498569488525, "num_tokens": 462195176.0, "step": 12171 }, { "epoch": 1.5484035110036891, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.814027309417725, "learning_rate": 5e-05, "loss": 0.7635, "mean_token_accuracy": 0.8635095357894897, "num_tokens": 462236155.0, "step": 12172 }, { "epoch": 1.5485307212822796, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.219207763671875, "learning_rate": 5e-05, "loss": 0.7036, "mean_token_accuracy": 0.8722565174102783, "num_tokens": 462267099.0, "step": 12173 }, { "epoch": 1.5486579315608702, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.976251125335693, "learning_rate": 5e-05, "loss": 0.805, "mean_token_accuracy": 0.8595677614212036, "num_tokens": 462311485.0, "step": 12174 }, { "epoch": 1.5487851418394607, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.984647274017334, "learning_rate": 5e-05, "loss": 0.6874, "mean_token_accuracy": 0.8727338314056396, "num_tokens": 462349078.0, "step": 12175 }, { "epoch": 1.5489123521180512, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.300201416015625, "learning_rate": 5e-05, "loss": 0.7735, "mean_token_accuracy": 0.871868371963501, "num_tokens": 462381477.0, "step": 12176 }, { "epoch": 1.5490395623966418, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.594629764556885, "learning_rate": 5e-05, "loss": 0.6705, "mean_token_accuracy": 0.8709595203399658, "num_tokens": 462415996.0, "step": 12177 }, { "epoch": 1.5491667726752323, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.906278610229492, "learning_rate": 5e-05, "loss": 0.8244, "mean_token_accuracy": 0.8754682540893555, "num_tokens": 462446737.0, "step": 12178 }, { "epoch": 1.5492939829538228, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000225067138671875, "grad_norm": 5.98144006729126, "learning_rate": 5e-05, "loss": 0.6581, "mean_token_accuracy": 0.8648340702056885, "num_tokens": 462487742.0, "step": 12179 }, { "epoch": 1.5494211932324133, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.947246551513672, "learning_rate": 5e-05, "loss": 0.9481, "mean_token_accuracy": 0.8750650882720947, "num_tokens": 462529231.0, "step": 12180 }, { "epoch": 1.5495484035110036, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 5.6558308601379395, "learning_rate": 5e-05, "loss": 0.6704, "mean_token_accuracy": 0.8738322854042053, "num_tokens": 462565608.0, "step": 12181 }, { "epoch": 1.5496756137895942, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 10.99563217163086, "learning_rate": 5e-05, "loss": 1.1342, "mean_token_accuracy": 0.8724005222320557, "num_tokens": 462600919.0, "step": 12182 }, { "epoch": 1.5498028240681847, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 7.287554740905762, "learning_rate": 5e-05, "loss": 0.9471, "mean_token_accuracy": 0.865728497505188, "num_tokens": 462635594.0, "step": 12183 }, { "epoch": 1.5499300343467752, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.510930061340332, "learning_rate": 5e-05, "loss": 1.1286, "mean_token_accuracy": 0.8581146001815796, "num_tokens": 462672349.0, "step": 12184 }, { "epoch": 1.5500572446253658, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 9.185901641845703, "learning_rate": 5e-05, "loss": 1.163, "mean_token_accuracy": 0.8753371238708496, "num_tokens": 462710286.0, "step": 12185 }, { "epoch": 1.5501844549039563, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.207137107849121, "learning_rate": 5e-05, "loss": 0.8568, "mean_token_accuracy": 0.8711332082748413, "num_tokens": 462744721.0, "step": 12186 }, { "epoch": 1.5503116651825466, "ewc_loss": 0.84375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 10.555985450744629, "learning_rate": 5e-05, "loss": 1.1789, "mean_token_accuracy": 0.8836854696273804, "num_tokens": 462782812.0, "step": 12187 }, { "epoch": 1.5504388754611371, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.2721967697143555, "learning_rate": 5e-05, "loss": 0.8211, "mean_token_accuracy": 0.8794527649879456, "num_tokens": 462827919.0, "step": 12188 }, { "epoch": 1.5505660857397277, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000843048095703125, "grad_norm": 11.862861633300781, "learning_rate": 5e-05, "loss": 1.2803, "mean_token_accuracy": 0.8674504160881042, "num_tokens": 462864878.0, "step": 12189 }, { "epoch": 1.5506932960183182, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 9.294931411743164, "learning_rate": 5e-05, "loss": 1.1854, "mean_token_accuracy": 0.8787184953689575, "num_tokens": 462906213.0, "step": 12190 }, { "epoch": 1.5508205062969087, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.901803970336914, "learning_rate": 5e-05, "loss": 1.129, "mean_token_accuracy": 0.8509430885314941, "num_tokens": 462943078.0, "step": 12191 }, { "epoch": 1.5509477165754992, "ewc_loss": 0.796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000743865966796875, "grad_norm": 9.265810012817383, "learning_rate": 5e-05, "loss": 1.1416, "mean_token_accuracy": 0.8799477815628052, "num_tokens": 462982444.0, "step": 12192 }, { "epoch": 1.5510749268540898, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.15665340423584, "learning_rate": 5e-05, "loss": 1.0049, "mean_token_accuracy": 0.8805400133132935, "num_tokens": 463024335.0, "step": 12193 }, { "epoch": 1.5512021371326803, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.533109664916992, "learning_rate": 5e-05, "loss": 0.958, "mean_token_accuracy": 0.8731400966644287, "num_tokens": 463062278.0, "step": 12194 }, { "epoch": 1.5513293474112708, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.657831192016602, "learning_rate": 5e-05, "loss": 1.0749, "mean_token_accuracy": 0.8496804237365723, "num_tokens": 463100257.0, "step": 12195 }, { "epoch": 1.5514565576898613, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.803285598754883, "learning_rate": 5e-05, "loss": 0.873, "mean_token_accuracy": 0.8737296462059021, "num_tokens": 463139468.0, "step": 12196 }, { "epoch": 1.5515837679684519, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.802279472351074, "learning_rate": 5e-05, "loss": 1.0382, "mean_token_accuracy": 0.8665468096733093, "num_tokens": 463176145.0, "step": 12197 }, { "epoch": 1.5517109782470424, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.365822792053223, "learning_rate": 5e-05, "loss": 0.8112, "mean_token_accuracy": 0.8788068294525146, "num_tokens": 463208899.0, "step": 12198 }, { "epoch": 1.551838188525633, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.406400680541992, "learning_rate": 5e-05, "loss": 1.0451, "mean_token_accuracy": 0.874544620513916, "num_tokens": 463250163.0, "step": 12199 }, { "epoch": 1.5519653988042235, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.704402923583984, "learning_rate": 5e-05, "loss": 0.9016, "mean_token_accuracy": 0.8685726523399353, "num_tokens": 463285447.0, "step": 12200 }, { "epoch": 1.552092609082814, "ewc_loss": 0.71875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.618895530700684, "learning_rate": 5e-05, "loss": 1.0453, "mean_token_accuracy": 0.8855767846107483, "num_tokens": 463322601.0, "step": 12201 }, { "epoch": 1.5522198193614045, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.538498878479004, "learning_rate": 5e-05, "loss": 0.9848, "mean_token_accuracy": 0.8812459707260132, "num_tokens": 463360624.0, "step": 12202 }, { "epoch": 1.552347029639995, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.990856170654297, "learning_rate": 5e-05, "loss": 1.0273, "mean_token_accuracy": 0.874273955821991, "num_tokens": 463401662.0, "step": 12203 }, { "epoch": 1.5524742399185856, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.81146240234375, "learning_rate": 5e-05, "loss": 1.0091, "mean_token_accuracy": 0.867958664894104, "num_tokens": 463433999.0, "step": 12204 }, { "epoch": 1.5526014501971759, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.04819393157959, "learning_rate": 5e-05, "loss": 1.0088, "mean_token_accuracy": 0.8552235960960388, "num_tokens": 463473673.0, "step": 12205 }, { "epoch": 1.5527286604757664, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.613462448120117, "learning_rate": 5e-05, "loss": 0.9706, "mean_token_accuracy": 0.8595410585403442, "num_tokens": 463507197.0, "step": 12206 }, { "epoch": 1.552855870754357, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.410262107849121, "learning_rate": 5e-05, "loss": 0.8914, "mean_token_accuracy": 0.8714967966079712, "num_tokens": 463540777.0, "step": 12207 }, { "epoch": 1.5529830810329475, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.050822257995605, "learning_rate": 5e-05, "loss": 0.7834, "mean_token_accuracy": 0.8886243104934692, "num_tokens": 463573335.0, "step": 12208 }, { "epoch": 1.553110291311538, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.50979232788086, "learning_rate": 5e-05, "loss": 0.8761, "mean_token_accuracy": 0.8751981258392334, "num_tokens": 463613898.0, "step": 12209 }, { "epoch": 1.5532375015901285, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.276161193847656, "learning_rate": 5e-05, "loss": 0.7156, "mean_token_accuracy": 0.883109450340271, "num_tokens": 463648948.0, "step": 12210 }, { "epoch": 1.5533647118687188, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.079462051391602, "learning_rate": 5e-05, "loss": 0.9101, "mean_token_accuracy": 0.8811228275299072, "num_tokens": 463683417.0, "step": 12211 }, { "epoch": 1.5534919221473094, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.90035343170166, "learning_rate": 5e-05, "loss": 0.7768, "mean_token_accuracy": 0.8655698299407959, "num_tokens": 463719533.0, "step": 12212 }, { "epoch": 1.5536191324258999, "ewc_loss": 0.625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.565180778503418, "learning_rate": 5e-05, "loss": 0.9583, "mean_token_accuracy": 0.882520854473114, "num_tokens": 463757740.0, "step": 12213 }, { "epoch": 1.5537463427044904, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.20737361907959, "learning_rate": 5e-05, "loss": 0.7918, "mean_token_accuracy": 0.871578574180603, "num_tokens": 463795619.0, "step": 12214 }, { "epoch": 1.553873552983081, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.581281661987305, "learning_rate": 5e-05, "loss": 0.9999, "mean_token_accuracy": 0.8737585544586182, "num_tokens": 463829934.0, "step": 12215 }, { "epoch": 1.5540007632616715, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.977329730987549, "learning_rate": 5e-05, "loss": 0.9531, "mean_token_accuracy": 0.8508470058441162, "num_tokens": 463869584.0, "step": 12216 }, { "epoch": 1.554127973540262, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.8125, "learning_rate": 5e-05, "loss": 0.9815, "mean_token_accuracy": 0.8572852611541748, "num_tokens": 463909220.0, "step": 12217 }, { "epoch": 1.5542551838188525, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.364860534667969, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.8791574239730835, "num_tokens": 463949367.0, "step": 12218 }, { "epoch": 1.554382394097443, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.865031719207764, "learning_rate": 5e-05, "loss": 0.8735, "mean_token_accuracy": 0.8660308122634888, "num_tokens": 463986037.0, "step": 12219 }, { "epoch": 1.5545096043760336, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.329054832458496, "learning_rate": 5e-05, "loss": 0.836, "mean_token_accuracy": 0.8809893131256104, "num_tokens": 464021903.0, "step": 12220 }, { "epoch": 1.554636814654624, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.5893235206604, "learning_rate": 5e-05, "loss": 0.7718, "mean_token_accuracy": 0.8776523470878601, "num_tokens": 464058738.0, "step": 12221 }, { "epoch": 1.5547640249332146, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.978172302246094, "learning_rate": 5e-05, "loss": 0.8537, "mean_token_accuracy": 0.8552310466766357, "num_tokens": 464090791.0, "step": 12222 }, { "epoch": 1.5548912352118052, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.710897445678711, "learning_rate": 5e-05, "loss": 0.7876, "mean_token_accuracy": 0.872517466545105, "num_tokens": 464133897.0, "step": 12223 }, { "epoch": 1.5550184454903957, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.657487869262695, "learning_rate": 5e-05, "loss": 0.731, "mean_token_accuracy": 0.8792657852172852, "num_tokens": 464173808.0, "step": 12224 }, { "epoch": 1.5551456557689862, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.845632076263428, "learning_rate": 5e-05, "loss": 0.7561, "mean_token_accuracy": 0.8772320747375488, "num_tokens": 464209456.0, "step": 12225 }, { "epoch": 1.5552728660475768, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.388331413269043, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8623901605606079, "num_tokens": 464250192.0, "step": 12226 }, { "epoch": 1.5554000763261673, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.120882034301758, "learning_rate": 5e-05, "loss": 0.858, "mean_token_accuracy": 0.8546417951583862, "num_tokens": 464291485.0, "step": 12227 }, { "epoch": 1.5555272866047578, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.997446060180664, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8627152442932129, "num_tokens": 464329963.0, "step": 12228 }, { "epoch": 1.5556544968833483, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.598433494567871, "learning_rate": 5e-05, "loss": 0.826, "mean_token_accuracy": 0.8749397993087769, "num_tokens": 464366944.0, "step": 12229 }, { "epoch": 1.5557817071619386, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.604172229766846, "learning_rate": 5e-05, "loss": 0.7162, "mean_token_accuracy": 0.8650306463241577, "num_tokens": 464409070.0, "step": 12230 }, { "epoch": 1.5559089174405292, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 9.081815719604492, "learning_rate": 5e-05, "loss": 0.8909, "mean_token_accuracy": 0.8787600994110107, "num_tokens": 464448305.0, "step": 12231 }, { "epoch": 1.5560361277191197, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.571413040161133, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.850804328918457, "num_tokens": 464482453.0, "step": 12232 }, { "epoch": 1.5561633379977102, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.224096298217773, "learning_rate": 5e-05, "loss": 0.9772, "mean_token_accuracy": 0.858592689037323, "num_tokens": 464518692.0, "step": 12233 }, { "epoch": 1.5562905482763008, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.0145721435546875, "learning_rate": 5e-05, "loss": 0.7741, "mean_token_accuracy": 0.8666353225708008, "num_tokens": 464554502.0, "step": 12234 }, { "epoch": 1.5564177585548913, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.857443809509277, "learning_rate": 5e-05, "loss": 0.9117, "mean_token_accuracy": 0.8794493675231934, "num_tokens": 464591414.0, "step": 12235 }, { "epoch": 1.5565449688334816, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.628354549407959, "learning_rate": 5e-05, "loss": 0.8602, "mean_token_accuracy": 0.8662330508232117, "num_tokens": 464627116.0, "step": 12236 }, { "epoch": 1.5566721791120721, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.11851978302002, "learning_rate": 5e-05, "loss": 0.8639, "mean_token_accuracy": 0.8713268041610718, "num_tokens": 464661783.0, "step": 12237 }, { "epoch": 1.5567993893906626, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.999826431274414, "learning_rate": 5e-05, "loss": 0.8929, "mean_token_accuracy": 0.8605179786682129, "num_tokens": 464704851.0, "step": 12238 }, { "epoch": 1.5569265996692532, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.560349464416504, "learning_rate": 5e-05, "loss": 0.8195, "mean_token_accuracy": 0.8629496693611145, "num_tokens": 464746583.0, "step": 12239 }, { "epoch": 1.5570538099478437, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.056528091430664, "learning_rate": 5e-05, "loss": 0.8352, "mean_token_accuracy": 0.8666824102401733, "num_tokens": 464783275.0, "step": 12240 }, { "epoch": 1.5571810202264342, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.447967529296875, "learning_rate": 5e-05, "loss": 0.746, "mean_token_accuracy": 0.8765708208084106, "num_tokens": 464818436.0, "step": 12241 }, { "epoch": 1.5573082305050248, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.960088729858398, "learning_rate": 5e-05, "loss": 0.7928, "mean_token_accuracy": 0.870619535446167, "num_tokens": 464859708.0, "step": 12242 }, { "epoch": 1.5574354407836153, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.385123252868652, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.8728652000427246, "num_tokens": 464894339.0, "step": 12243 }, { "epoch": 1.5575626510622058, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.019585609436035, "learning_rate": 5e-05, "loss": 0.7845, "mean_token_accuracy": 0.8792606592178345, "num_tokens": 464931517.0, "step": 12244 }, { "epoch": 1.5576898613407963, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.182783126831055, "learning_rate": 5e-05, "loss": 0.7493, "mean_token_accuracy": 0.867816150188446, "num_tokens": 464971616.0, "step": 12245 }, { "epoch": 1.5578170716193869, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.275959014892578, "learning_rate": 5e-05, "loss": 0.794, "mean_token_accuracy": 0.8784524202346802, "num_tokens": 465012551.0, "step": 12246 }, { "epoch": 1.5579442818979774, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.862506866455078, "learning_rate": 5e-05, "loss": 0.6853, "mean_token_accuracy": 0.8816810250282288, "num_tokens": 465054266.0, "step": 12247 }, { "epoch": 1.558071492176568, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.677179336547852, "learning_rate": 5e-05, "loss": 0.8143, "mean_token_accuracy": 0.8768896460533142, "num_tokens": 465090057.0, "step": 12248 }, { "epoch": 1.5581987024551585, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.575903415679932, "learning_rate": 5e-05, "loss": 0.6437, "mean_token_accuracy": 0.8888838887214661, "num_tokens": 465132169.0, "step": 12249 }, { "epoch": 1.558325912733749, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 9.032584190368652, "learning_rate": 5e-05, "loss": 0.9187, "mean_token_accuracy": 0.869470477104187, "num_tokens": 465169922.0, "step": 12250 }, { "epoch": 1.5584531230123395, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.60145378112793, "learning_rate": 5e-05, "loss": 0.6684, "mean_token_accuracy": 0.8847467303276062, "num_tokens": 465207792.0, "step": 12251 }, { "epoch": 1.55858033329093, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.054228782653809, "learning_rate": 5e-05, "loss": 0.9501, "mean_token_accuracy": 0.8628143072128296, "num_tokens": 465250498.0, "step": 12252 }, { "epoch": 1.5587075435695206, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.06094217300415, "learning_rate": 5e-05, "loss": 0.7976, "mean_token_accuracy": 0.8666415214538574, "num_tokens": 465288321.0, "step": 12253 }, { "epoch": 1.5588347538481109, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.509452819824219, "learning_rate": 5e-05, "loss": 0.8109, "mean_token_accuracy": 0.8954041004180908, "num_tokens": 465322241.0, "step": 12254 }, { "epoch": 1.5589619641267014, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.649209499359131, "learning_rate": 5e-05, "loss": 0.8013, "mean_token_accuracy": 0.8854568004608154, "num_tokens": 465359680.0, "step": 12255 }, { "epoch": 1.559089174405292, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.815621852874756, "learning_rate": 5e-05, "loss": 0.8119, "mean_token_accuracy": 0.8783527612686157, "num_tokens": 465396951.0, "step": 12256 }, { "epoch": 1.5592163846838825, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.933901786804199, "learning_rate": 5e-05, "loss": 0.8424, "mean_token_accuracy": 0.8690329790115356, "num_tokens": 465433087.0, "step": 12257 }, { "epoch": 1.559343594962473, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.533347129821777, "learning_rate": 5e-05, "loss": 0.7656, "mean_token_accuracy": 0.8765355944633484, "num_tokens": 465472810.0, "step": 12258 }, { "epoch": 1.5594708052410635, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.761030673980713, "learning_rate": 5e-05, "loss": 0.8267, "mean_token_accuracy": 0.8651265501976013, "num_tokens": 465512714.0, "step": 12259 }, { "epoch": 1.5595980155196538, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.768064498901367, "learning_rate": 5e-05, "loss": 0.8239, "mean_token_accuracy": 0.8567315936088562, "num_tokens": 465547629.0, "step": 12260 }, { "epoch": 1.5597252257982444, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.300899028778076, "learning_rate": 5e-05, "loss": 0.7508, "mean_token_accuracy": 0.8672570586204529, "num_tokens": 465584226.0, "step": 12261 }, { "epoch": 1.5598524360768349, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.211653709411621, "learning_rate": 5e-05, "loss": 0.8655, "mean_token_accuracy": 0.8585830926895142, "num_tokens": 465625854.0, "step": 12262 }, { "epoch": 1.5599796463554254, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.839301109313965, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8535558581352234, "num_tokens": 465660670.0, "step": 12263 }, { "epoch": 1.560106856634016, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.697067260742188, "learning_rate": 5e-05, "loss": 0.8565, "mean_token_accuracy": 0.8697972297668457, "num_tokens": 465700359.0, "step": 12264 }, { "epoch": 1.5602340669126065, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.493072032928467, "learning_rate": 5e-05, "loss": 0.6918, "mean_token_accuracy": 0.8745675086975098, "num_tokens": 465733648.0, "step": 12265 }, { "epoch": 1.560361277191197, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.161568641662598, "learning_rate": 5e-05, "loss": 0.9305, "mean_token_accuracy": 0.8646911978721619, "num_tokens": 465769814.0, "step": 12266 }, { "epoch": 1.5604884874697875, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.4930830001831055, "learning_rate": 5e-05, "loss": 0.6891, "mean_token_accuracy": 0.8771061301231384, "num_tokens": 465804637.0, "step": 12267 }, { "epoch": 1.560615697748378, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.220550537109375, "learning_rate": 5e-05, "loss": 0.9131, "mean_token_accuracy": 0.8843736052513123, "num_tokens": 465838649.0, "step": 12268 }, { "epoch": 1.5607429080269686, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.149336814880371, "learning_rate": 5e-05, "loss": 0.7745, "mean_token_accuracy": 0.882714033126831, "num_tokens": 465870628.0, "step": 12269 }, { "epoch": 1.560870118305559, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.41844367980957, "learning_rate": 5e-05, "loss": 0.9042, "mean_token_accuracy": 0.8682227730751038, "num_tokens": 465905612.0, "step": 12270 }, { "epoch": 1.5609973285841496, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.040332794189453, "learning_rate": 5e-05, "loss": 0.8751, "mean_token_accuracy": 0.8712046146392822, "num_tokens": 465945964.0, "step": 12271 }, { "epoch": 1.5611245388627402, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.247249603271484, "learning_rate": 5e-05, "loss": 0.7245, "mean_token_accuracy": 0.8861918449401855, "num_tokens": 465979600.0, "step": 12272 }, { "epoch": 1.5612517491413307, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.743293762207031, "learning_rate": 5e-05, "loss": 0.907, "mean_token_accuracy": 0.8662214279174805, "num_tokens": 466017406.0, "step": 12273 }, { "epoch": 1.5613789594199212, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.584481239318848, "learning_rate": 5e-05, "loss": 0.7307, "mean_token_accuracy": 0.8677082657814026, "num_tokens": 466059178.0, "step": 12274 }, { "epoch": 1.5615061696985117, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.207778930664062, "learning_rate": 5e-05, "loss": 0.9514, "mean_token_accuracy": 0.8686114549636841, "num_tokens": 466096385.0, "step": 12275 }, { "epoch": 1.5616333799771023, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.528282642364502, "learning_rate": 5e-05, "loss": 0.7304, "mean_token_accuracy": 0.8711444139480591, "num_tokens": 466131160.0, "step": 12276 }, { "epoch": 1.5617605902556928, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.35197639465332, "learning_rate": 5e-05, "loss": 0.9722, "mean_token_accuracy": 0.8723770380020142, "num_tokens": 466174842.0, "step": 12277 }, { "epoch": 1.5618878005342833, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.0287275314331055, "learning_rate": 5e-05, "loss": 0.8018, "mean_token_accuracy": 0.8763200044631958, "num_tokens": 466212943.0, "step": 12278 }, { "epoch": 1.5620150108128736, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.961112976074219, "learning_rate": 5e-05, "loss": 0.9297, "mean_token_accuracy": 0.8858003616333008, "num_tokens": 466251560.0, "step": 12279 }, { "epoch": 1.5621422210914642, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.667455673217773, "learning_rate": 5e-05, "loss": 0.9229, "mean_token_accuracy": 0.860731840133667, "num_tokens": 466287150.0, "step": 12280 }, { "epoch": 1.5622694313700547, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.274393081665039, "learning_rate": 5e-05, "loss": 0.9268, "mean_token_accuracy": 0.86905837059021, "num_tokens": 466328508.0, "step": 12281 }, { "epoch": 1.5623966416486452, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.8663434982299805, "learning_rate": 5e-05, "loss": 0.8862, "mean_token_accuracy": 0.8654193878173828, "num_tokens": 466362552.0, "step": 12282 }, { "epoch": 1.5625238519272358, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.030322074890137, "learning_rate": 5e-05, "loss": 0.8832, "mean_token_accuracy": 0.866620659828186, "num_tokens": 466404253.0, "step": 12283 }, { "epoch": 1.5626510622058263, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.447779655456543, "learning_rate": 5e-05, "loss": 0.8393, "mean_token_accuracy": 0.8626288175582886, "num_tokens": 466443258.0, "step": 12284 }, { "epoch": 1.5627782724844166, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.435993194580078, "learning_rate": 5e-05, "loss": 0.8996, "mean_token_accuracy": 0.8709424734115601, "num_tokens": 466477575.0, "step": 12285 }, { "epoch": 1.5629054827630071, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.877904891967773, "learning_rate": 5e-05, "loss": 0.7504, "mean_token_accuracy": 0.8661407232284546, "num_tokens": 466516301.0, "step": 12286 }, { "epoch": 1.5630326930415976, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.931172370910645, "learning_rate": 5e-05, "loss": 0.9595, "mean_token_accuracy": 0.8613433837890625, "num_tokens": 466553904.0, "step": 12287 }, { "epoch": 1.5631599033201882, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.737597465515137, "learning_rate": 5e-05, "loss": 0.7624, "mean_token_accuracy": 0.8665039539337158, "num_tokens": 466591668.0, "step": 12288 }, { "epoch": 1.5632871135987787, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.040830612182617, "learning_rate": 5e-05, "loss": 0.9481, "mean_token_accuracy": 0.8705507516860962, "num_tokens": 466626484.0, "step": 12289 }, { "epoch": 1.5634143238773692, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.108266830444336, "learning_rate": 5e-05, "loss": 0.7855, "mean_token_accuracy": 0.8800438642501831, "num_tokens": 466663590.0, "step": 12290 }, { "epoch": 1.5635415341559598, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.663537979125977, "learning_rate": 5e-05, "loss": 0.9085, "mean_token_accuracy": 0.8839943408966064, "num_tokens": 466699856.0, "step": 12291 }, { "epoch": 1.5636687444345503, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.611050128936768, "learning_rate": 5e-05, "loss": 0.83, "mean_token_accuracy": 0.8846598863601685, "num_tokens": 466738703.0, "step": 12292 }, { "epoch": 1.5637959547131408, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.056740760803223, "learning_rate": 5e-05, "loss": 0.8591, "mean_token_accuracy": 0.8765021562576294, "num_tokens": 466770948.0, "step": 12293 }, { "epoch": 1.5639231649917313, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.876402378082275, "learning_rate": 5e-05, "loss": 0.8522, "mean_token_accuracy": 0.8728281259536743, "num_tokens": 466804655.0, "step": 12294 }, { "epoch": 1.5640503752703219, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.624005317687988, "learning_rate": 5e-05, "loss": 0.7955, "mean_token_accuracy": 0.8771159052848816, "num_tokens": 466847592.0, "step": 12295 }, { "epoch": 1.5641775855489124, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.868348121643066, "learning_rate": 5e-05, "loss": 0.811, "mean_token_accuracy": 0.8792630434036255, "num_tokens": 466887855.0, "step": 12296 }, { "epoch": 1.564304795827503, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.581478595733643, "learning_rate": 5e-05, "loss": 0.7963, "mean_token_accuracy": 0.8665128946304321, "num_tokens": 466925749.0, "step": 12297 }, { "epoch": 1.5644320061060935, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.689338684082031, "learning_rate": 5e-05, "loss": 0.7239, "mean_token_accuracy": 0.8929990530014038, "num_tokens": 466966487.0, "step": 12298 }, { "epoch": 1.564559216384684, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.655362606048584, "learning_rate": 5e-05, "loss": 0.8557, "mean_token_accuracy": 0.8493570685386658, "num_tokens": 467008673.0, "step": 12299 }, { "epoch": 1.5646864266632745, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.565938949584961, "learning_rate": 5e-05, "loss": 0.7607, "mean_token_accuracy": 0.8780243396759033, "num_tokens": 467048389.0, "step": 12300 }, { "epoch": 1.564813636941865, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.6659932136535645, "learning_rate": 5e-05, "loss": 0.756, "mean_token_accuracy": 0.8761993050575256, "num_tokens": 467087059.0, "step": 12301 }, { "epoch": 1.5649408472204556, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.528717994689941, "learning_rate": 5e-05, "loss": 0.7003, "mean_token_accuracy": 0.8901010155677795, "num_tokens": 467123003.0, "step": 12302 }, { "epoch": 1.5650680574990459, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.667404651641846, "learning_rate": 5e-05, "loss": 0.7954, "mean_token_accuracy": 0.8611936569213867, "num_tokens": 467160526.0, "step": 12303 }, { "epoch": 1.5651952677776364, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.454519748687744, "learning_rate": 5e-05, "loss": 0.7384, "mean_token_accuracy": 0.8692503571510315, "num_tokens": 467195846.0, "step": 12304 }, { "epoch": 1.565322478056227, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.714249610900879, "learning_rate": 5e-05, "loss": 0.7905, "mean_token_accuracy": 0.860501766204834, "num_tokens": 467232412.0, "step": 12305 }, { "epoch": 1.5654496883348175, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.3547682762146, "learning_rate": 5e-05, "loss": 0.7468, "mean_token_accuracy": 0.8644522428512573, "num_tokens": 467269695.0, "step": 12306 }, { "epoch": 1.565576898613408, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.836596488952637, "learning_rate": 5e-05, "loss": 0.8015, "mean_token_accuracy": 0.8583266735076904, "num_tokens": 467304395.0, "step": 12307 }, { "epoch": 1.5657041088919985, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.156505107879639, "learning_rate": 5e-05, "loss": 0.7121, "mean_token_accuracy": 0.871319055557251, "num_tokens": 467348606.0, "step": 12308 }, { "epoch": 1.5658313191705888, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.073928833007812, "learning_rate": 5e-05, "loss": 0.7773, "mean_token_accuracy": 0.8721858263015747, "num_tokens": 467382673.0, "step": 12309 }, { "epoch": 1.5659585294491793, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.865799903869629, "learning_rate": 5e-05, "loss": 0.6768, "mean_token_accuracy": 0.8786762952804565, "num_tokens": 467426291.0, "step": 12310 }, { "epoch": 1.5660857397277699, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.4686861038208, "learning_rate": 5e-05, "loss": 0.8212, "mean_token_accuracy": 0.8686947226524353, "num_tokens": 467469299.0, "step": 12311 }, { "epoch": 1.5662129500063604, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.472639083862305, "learning_rate": 5e-05, "loss": 0.6826, "mean_token_accuracy": 0.8661824464797974, "num_tokens": 467510876.0, "step": 12312 }, { "epoch": 1.566340160284951, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 9.06621265411377, "learning_rate": 5e-05, "loss": 0.8779, "mean_token_accuracy": 0.8708841800689697, "num_tokens": 467546112.0, "step": 12313 }, { "epoch": 1.5664673705635415, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.078235626220703, "learning_rate": 5e-05, "loss": 0.6584, "mean_token_accuracy": 0.8735058307647705, "num_tokens": 467580009.0, "step": 12314 }, { "epoch": 1.566594580842132, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.752151489257812, "learning_rate": 5e-05, "loss": 0.9987, "mean_token_accuracy": 0.8643964529037476, "num_tokens": 467621009.0, "step": 12315 }, { "epoch": 1.5667217911207225, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.299474239349365, "learning_rate": 5e-05, "loss": 0.6982, "mean_token_accuracy": 0.8832119703292847, "num_tokens": 467662304.0, "step": 12316 }, { "epoch": 1.566849001399313, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.784770965576172, "learning_rate": 5e-05, "loss": 1.0705, "mean_token_accuracy": 0.8571285009384155, "num_tokens": 467697608.0, "step": 12317 }, { "epoch": 1.5669762116779036, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.630979061126709, "learning_rate": 5e-05, "loss": 0.9334, "mean_token_accuracy": 0.8729180097579956, "num_tokens": 467739978.0, "step": 12318 }, { "epoch": 1.567103421956494, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.323800086975098, "learning_rate": 5e-05, "loss": 0.9266, "mean_token_accuracy": 0.8608899116516113, "num_tokens": 467775055.0, "step": 12319 }, { "epoch": 1.5672306322350846, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.046005249023438, "learning_rate": 5e-05, "loss": 0.9881, "mean_token_accuracy": 0.8769557476043701, "num_tokens": 467812385.0, "step": 12320 }, { "epoch": 1.5673578425136752, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.337174892425537, "learning_rate": 5e-05, "loss": 0.7575, "mean_token_accuracy": 0.855431079864502, "num_tokens": 467846784.0, "step": 12321 }, { "epoch": 1.5674850527922657, "ewc_loss": 0.75, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 10.815418243408203, "learning_rate": 5e-05, "loss": 1.1401, "mean_token_accuracy": 0.8658835291862488, "num_tokens": 467880259.0, "step": 12322 }, { "epoch": 1.5676122630708562, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.034184455871582, "learning_rate": 5e-05, "loss": 0.9193, "mean_token_accuracy": 0.8656461834907532, "num_tokens": 467914613.0, "step": 12323 }, { "epoch": 1.5677394733494467, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 11.09048843383789, "learning_rate": 5e-05, "loss": 1.224, "mean_token_accuracy": 0.8478571176528931, "num_tokens": 467953070.0, "step": 12324 }, { "epoch": 1.5678666836280373, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.00054168701171875, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 8.575018882751465, "learning_rate": 5e-05, "loss": 1.1412, "mean_token_accuracy": 0.8575351238250732, "num_tokens": 467989404.0, "step": 12325 }, { "epoch": 1.5679938939066278, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.732030868530273, "learning_rate": 5e-05, "loss": 1.0061, "mean_token_accuracy": 0.8690617084503174, "num_tokens": 468030670.0, "step": 12326 }, { "epoch": 1.5681211041852183, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 9.167535781860352, "learning_rate": 5e-05, "loss": 1.115, "mean_token_accuracy": 0.8696779012680054, "num_tokens": 468064130.0, "step": 12327 }, { "epoch": 1.5682483144638086, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.372249603271484, "learning_rate": 5e-05, "loss": 0.9269, "mean_token_accuracy": 0.8708457946777344, "num_tokens": 468098393.0, "step": 12328 }, { "epoch": 1.5683755247423992, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.183985710144043, "learning_rate": 5e-05, "loss": 0.9764, "mean_token_accuracy": 0.8825547099113464, "num_tokens": 468134961.0, "step": 12329 }, { "epoch": 1.5685027350209897, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 7.970648765563965, "learning_rate": 5e-05, "loss": 0.8782, "mean_token_accuracy": 0.8806183934211731, "num_tokens": 468169389.0, "step": 12330 }, { "epoch": 1.5686299452995802, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.541315078735352, "learning_rate": 5e-05, "loss": 0.9414, "mean_token_accuracy": 0.8623059988021851, "num_tokens": 468211504.0, "step": 12331 }, { "epoch": 1.5687571555781707, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.233235359191895, "learning_rate": 5e-05, "loss": 0.9649, "mean_token_accuracy": 0.8602546453475952, "num_tokens": 468245204.0, "step": 12332 }, { "epoch": 1.5688843658567613, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.557374954223633, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.8730431199073792, "num_tokens": 468277054.0, "step": 12333 }, { "epoch": 1.5690115761353516, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.738235473632812, "learning_rate": 5e-05, "loss": 0.9068, "mean_token_accuracy": 0.8851882219314575, "num_tokens": 468316399.0, "step": 12334 }, { "epoch": 1.569138786413942, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.901695251464844, "learning_rate": 5e-05, "loss": 0.7282, "mean_token_accuracy": 0.8785775303840637, "num_tokens": 468351943.0, "step": 12335 }, { "epoch": 1.5692659966925326, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.398456573486328, "learning_rate": 5e-05, "loss": 0.981, "mean_token_accuracy": 0.869590699672699, "num_tokens": 468387443.0, "step": 12336 }, { "epoch": 1.5693932069711232, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.740589141845703, "learning_rate": 5e-05, "loss": 0.765, "mean_token_accuracy": 0.8751983642578125, "num_tokens": 468420503.0, "step": 12337 }, { "epoch": 1.5695204172497137, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.873918533325195, "learning_rate": 5e-05, "loss": 1.0474, "mean_token_accuracy": 0.8650360107421875, "num_tokens": 468456171.0, "step": 12338 }, { "epoch": 1.5696476275283042, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.274771213531494, "learning_rate": 5e-05, "loss": 0.8533, "mean_token_accuracy": 0.882896363735199, "num_tokens": 468494121.0, "step": 12339 }, { "epoch": 1.5697748378068948, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.751246452331543, "learning_rate": 5e-05, "loss": 1.0358, "mean_token_accuracy": 0.8721846342086792, "num_tokens": 468540960.0, "step": 12340 }, { "epoch": 1.5699020480854853, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 7.881167888641357, "learning_rate": 5e-05, "loss": 0.9943, "mean_token_accuracy": 0.869209885597229, "num_tokens": 468577851.0, "step": 12341 }, { "epoch": 1.5700292583640758, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.190814018249512, "learning_rate": 5e-05, "loss": 0.9943, "mean_token_accuracy": 0.8700095415115356, "num_tokens": 468613210.0, "step": 12342 }, { "epoch": 1.5701564686426663, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.899580955505371, "learning_rate": 5e-05, "loss": 0.9503, "mean_token_accuracy": 0.8668221831321716, "num_tokens": 468655871.0, "step": 12343 }, { "epoch": 1.5702836789212569, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.896147727966309, "learning_rate": 5e-05, "loss": 0.9399, "mean_token_accuracy": 0.8798763751983643, "num_tokens": 468692757.0, "step": 12344 }, { "epoch": 1.5704108891998474, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000576019287109375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.525323390960693, "learning_rate": 5e-05, "loss": 0.8528, "mean_token_accuracy": 0.8778901100158691, "num_tokens": 468730163.0, "step": 12345 }, { "epoch": 1.570538099478438, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.064387321472168, "learning_rate": 5e-05, "loss": 0.9799, "mean_token_accuracy": 0.8722935914993286, "num_tokens": 468768578.0, "step": 12346 }, { "epoch": 1.5706653097570284, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.282334804534912, "learning_rate": 5e-05, "loss": 0.7967, "mean_token_accuracy": 0.8836498856544495, "num_tokens": 468802027.0, "step": 12347 }, { "epoch": 1.570792520035619, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.23910903930664, "learning_rate": 5e-05, "loss": 1.0348, "mean_token_accuracy": 0.8609731197357178, "num_tokens": 468845266.0, "step": 12348 }, { "epoch": 1.5709197303142095, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.499347686767578, "learning_rate": 5e-05, "loss": 0.894, "mean_token_accuracy": 0.8738902807235718, "num_tokens": 468879780.0, "step": 12349 }, { "epoch": 1.5710469405928, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.9666109085083, "learning_rate": 5e-05, "loss": 0.9773, "mean_token_accuracy": 0.8748172521591187, "num_tokens": 468916846.0, "step": 12350 }, { "epoch": 1.5711741508713906, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.879683494567871, "learning_rate": 5e-05, "loss": 0.8827, "mean_token_accuracy": 0.8830039501190186, "num_tokens": 468947382.0, "step": 12351 }, { "epoch": 1.5713013611499809, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.337909698486328, "learning_rate": 5e-05, "loss": 0.8699, "mean_token_accuracy": 0.8840100765228271, "num_tokens": 468987258.0, "step": 12352 }, { "epoch": 1.5714285714285714, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.981245040893555, "learning_rate": 5e-05, "loss": 0.8737, "mean_token_accuracy": 0.8768395781517029, "num_tokens": 469024878.0, "step": 12353 }, { "epoch": 1.571555781707162, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.92700719833374, "learning_rate": 5e-05, "loss": 0.8544, "mean_token_accuracy": 0.8744882941246033, "num_tokens": 469065636.0, "step": 12354 }, { "epoch": 1.5716829919857525, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.775182247161865, "learning_rate": 5e-05, "loss": 0.8366, "mean_token_accuracy": 0.87111496925354, "num_tokens": 469099817.0, "step": 12355 }, { "epoch": 1.571810202264343, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.970666885375977, "learning_rate": 5e-05, "loss": 0.846, "mean_token_accuracy": 0.8696966767311096, "num_tokens": 469135185.0, "step": 12356 }, { "epoch": 1.5719374125429335, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.4672017097473145, "learning_rate": 5e-05, "loss": 0.7302, "mean_token_accuracy": 0.8879868984222412, "num_tokens": 469173980.0, "step": 12357 }, { "epoch": 1.5720646228215238, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.12324333190918, "learning_rate": 5e-05, "loss": 0.8338, "mean_token_accuracy": 0.8762059807777405, "num_tokens": 469206939.0, "step": 12358 }, { "epoch": 1.5721918331001143, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.307064056396484, "learning_rate": 5e-05, "loss": 0.7359, "mean_token_accuracy": 0.8780435919761658, "num_tokens": 469245261.0, "step": 12359 }, { "epoch": 1.5723190433787049, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.180524826049805, "learning_rate": 5e-05, "loss": 0.8346, "mean_token_accuracy": 0.8699608445167542, "num_tokens": 469285569.0, "step": 12360 }, { "epoch": 1.5724462536572954, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.206271648406982, "learning_rate": 5e-05, "loss": 0.7528, "mean_token_accuracy": 0.8674414157867432, "num_tokens": 469325249.0, "step": 12361 }, { "epoch": 1.572573463935886, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.278464317321777, "learning_rate": 5e-05, "loss": 0.8496, "mean_token_accuracy": 0.865510106086731, "num_tokens": 469360187.0, "step": 12362 }, { "epoch": 1.5727006742144765, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.09977912902832, "learning_rate": 5e-05, "loss": 0.7532, "mean_token_accuracy": 0.8675699830055237, "num_tokens": 469399363.0, "step": 12363 }, { "epoch": 1.572827884493067, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.386113166809082, "learning_rate": 5e-05, "loss": 0.8754, "mean_token_accuracy": 0.8707088828086853, "num_tokens": 469441523.0, "step": 12364 }, { "epoch": 1.5729550947716575, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.077075481414795, "learning_rate": 5e-05, "loss": 0.7105, "mean_token_accuracy": 0.8796550035476685, "num_tokens": 469475169.0, "step": 12365 }, { "epoch": 1.573082305050248, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.369237899780273, "learning_rate": 5e-05, "loss": 0.8807, "mean_token_accuracy": 0.8654420375823975, "num_tokens": 469509794.0, "step": 12366 }, { "epoch": 1.5732095153288386, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.182315826416016, "learning_rate": 5e-05, "loss": 0.7624, "mean_token_accuracy": 0.8672344088554382, "num_tokens": 469549007.0, "step": 12367 }, { "epoch": 1.573336725607429, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.214425086975098, "learning_rate": 5e-05, "loss": 0.8324, "mean_token_accuracy": 0.875637412071228, "num_tokens": 469586959.0, "step": 12368 }, { "epoch": 1.5734639358860196, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.364239692687988, "learning_rate": 5e-05, "loss": 0.7371, "mean_token_accuracy": 0.8791232109069824, "num_tokens": 469617767.0, "step": 12369 }, { "epoch": 1.5735911461646102, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.945088863372803, "learning_rate": 5e-05, "loss": 0.814, "mean_token_accuracy": 0.8682369589805603, "num_tokens": 469655300.0, "step": 12370 }, { "epoch": 1.5737183564432007, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.565069675445557, "learning_rate": 5e-05, "loss": 0.7639, "mean_token_accuracy": 0.8751771450042725, "num_tokens": 469693666.0, "step": 12371 }, { "epoch": 1.5738455667217912, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.624523162841797, "learning_rate": 5e-05, "loss": 0.7673, "mean_token_accuracy": 0.8746740818023682, "num_tokens": 469731081.0, "step": 12372 }, { "epoch": 1.5739727770003817, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.7782111167907715, "learning_rate": 5e-05, "loss": 0.7528, "mean_token_accuracy": 0.8755759596824646, "num_tokens": 469768081.0, "step": 12373 }, { "epoch": 1.5740999872789723, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.342739105224609, "learning_rate": 5e-05, "loss": 0.7473, "mean_token_accuracy": 0.8706779479980469, "num_tokens": 469806992.0, "step": 12374 }, { "epoch": 1.5742271975575628, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.950471878051758, "learning_rate": 5e-05, "loss": 0.8038, "mean_token_accuracy": 0.8683919906616211, "num_tokens": 469844969.0, "step": 12375 }, { "epoch": 1.5743544078361533, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.145970821380615, "learning_rate": 5e-05, "loss": 0.6981, "mean_token_accuracy": 0.8786141276359558, "num_tokens": 469881558.0, "step": 12376 }, { "epoch": 1.5744816181147436, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.100214004516602, "learning_rate": 5e-05, "loss": 0.794, "mean_token_accuracy": 0.8702676892280579, "num_tokens": 469914659.0, "step": 12377 }, { "epoch": 1.5746088283933342, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.982283592224121, "learning_rate": 5e-05, "loss": 0.7398, "mean_token_accuracy": 0.8668374419212341, "num_tokens": 469956012.0, "step": 12378 }, { "epoch": 1.5747360386719247, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.286083221435547, "learning_rate": 5e-05, "loss": 0.8172, "mean_token_accuracy": 0.867308497428894, "num_tokens": 469993840.0, "step": 12379 }, { "epoch": 1.5748632489505152, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.807463645935059, "learning_rate": 5e-05, "loss": 0.7254, "mean_token_accuracy": 0.8621668815612793, "num_tokens": 470032741.0, "step": 12380 }, { "epoch": 1.5749904592291057, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.496561050415039, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8838541507720947, "num_tokens": 470068969.0, "step": 12381 }, { "epoch": 1.5751176695076963, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.659599304199219, "learning_rate": 5e-05, "loss": 0.6816, "mean_token_accuracy": 0.8775849342346191, "num_tokens": 470108972.0, "step": 12382 }, { "epoch": 1.5752448797862866, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.716024398803711, "learning_rate": 5e-05, "loss": 0.8057, "mean_token_accuracy": 0.8862878084182739, "num_tokens": 470146655.0, "step": 12383 }, { "epoch": 1.575372090064877, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.627610683441162, "learning_rate": 5e-05, "loss": 0.7211, "mean_token_accuracy": 0.869017481803894, "num_tokens": 470184068.0, "step": 12384 }, { "epoch": 1.5754993003434676, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.746496200561523, "learning_rate": 5e-05, "loss": 0.8802, "mean_token_accuracy": 0.8689576387405396, "num_tokens": 470221328.0, "step": 12385 }, { "epoch": 1.5756265106220582, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.856045246124268, "learning_rate": 5e-05, "loss": 0.7336, "mean_token_accuracy": 0.8735921382904053, "num_tokens": 470261393.0, "step": 12386 }, { "epoch": 1.5757537209006487, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.420641899108887, "learning_rate": 5e-05, "loss": 0.8552, "mean_token_accuracy": 0.8702917695045471, "num_tokens": 470293705.0, "step": 12387 }, { "epoch": 1.5758809311792392, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.336174964904785, "learning_rate": 5e-05, "loss": 0.782, "mean_token_accuracy": 0.8700414896011353, "num_tokens": 470329891.0, "step": 12388 }, { "epoch": 1.5760081414578297, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.782614707946777, "learning_rate": 5e-05, "loss": 0.7891, "mean_token_accuracy": 0.8753063082695007, "num_tokens": 470368763.0, "step": 12389 }, { "epoch": 1.5761353517364203, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.827907085418701, "learning_rate": 5e-05, "loss": 0.7953, "mean_token_accuracy": 0.8711403608322144, "num_tokens": 470407267.0, "step": 12390 }, { "epoch": 1.5762625620150108, "ewc_loss": 0.375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.23671293258667, "learning_rate": 5e-05, "loss": 0.7423, "mean_token_accuracy": 0.8695217370986938, "num_tokens": 470448872.0, "step": 12391 }, { "epoch": 1.5763897722936013, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.149791717529297, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.8787761926651001, "num_tokens": 470486460.0, "step": 12392 }, { "epoch": 1.5765169825721919, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.984405517578125, "learning_rate": 5e-05, "loss": 0.6723, "mean_token_accuracy": 0.8839005827903748, "num_tokens": 470528800.0, "step": 12393 }, { "epoch": 1.5766441928507824, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.207884788513184, "learning_rate": 5e-05, "loss": 0.7883, "mean_token_accuracy": 0.8752226233482361, "num_tokens": 470568185.0, "step": 12394 }, { "epoch": 1.576771403129373, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.039252758026123, "learning_rate": 5e-05, "loss": 0.7269, "mean_token_accuracy": 0.8719382286071777, "num_tokens": 470607785.0, "step": 12395 }, { "epoch": 1.5768986134079634, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.036237716674805, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.8670840263366699, "num_tokens": 470648536.0, "step": 12396 }, { "epoch": 1.577025823686554, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.211109638214111, "learning_rate": 5e-05, "loss": 0.7017, "mean_token_accuracy": 0.8789374828338623, "num_tokens": 470686071.0, "step": 12397 }, { "epoch": 1.5771530339651445, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.804671764373779, "learning_rate": 5e-05, "loss": 0.7407, "mean_token_accuracy": 0.8804285526275635, "num_tokens": 470715165.0, "step": 12398 }, { "epoch": 1.577280244243735, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.457976341247559, "learning_rate": 5e-05, "loss": 0.712, "mean_token_accuracy": 0.8768013715744019, "num_tokens": 470748950.0, "step": 12399 }, { "epoch": 1.5774074545223256, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.488076686859131, "learning_rate": 5e-05, "loss": 0.6725, "mean_token_accuracy": 0.8886561989784241, "num_tokens": 470783856.0, "step": 12400 }, { "epoch": 1.5775346648009159, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.7293291091918945, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.8623743653297424, "num_tokens": 470823103.0, "step": 12401 }, { "epoch": 1.5776618750795064, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.189539432525635, "learning_rate": 5e-05, "loss": 0.7662, "mean_token_accuracy": 0.8552368879318237, "num_tokens": 470865582.0, "step": 12402 }, { "epoch": 1.577789085358097, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.990077018737793, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8621156811714172, "num_tokens": 470901710.0, "step": 12403 }, { "epoch": 1.5779162956366874, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.938262462615967, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.8541077375411987, "num_tokens": 470944850.0, "step": 12404 }, { "epoch": 1.578043505915278, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.276534080505371, "learning_rate": 5e-05, "loss": 0.8411, "mean_token_accuracy": 0.8616908192634583, "num_tokens": 470985092.0, "step": 12405 }, { "epoch": 1.5781707161938685, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.637301921844482, "learning_rate": 5e-05, "loss": 0.6624, "mean_token_accuracy": 0.8788856863975525, "num_tokens": 471026018.0, "step": 12406 }, { "epoch": 1.5782979264724588, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.721904754638672, "learning_rate": 5e-05, "loss": 0.8435, "mean_token_accuracy": 0.8678892254829407, "num_tokens": 471061365.0, "step": 12407 }, { "epoch": 1.5784251367510493, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.230747222900391, "learning_rate": 5e-05, "loss": 0.6606, "mean_token_accuracy": 0.8754116296768188, "num_tokens": 471098142.0, "step": 12408 }, { "epoch": 1.5785523470296399, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.357529640197754, "learning_rate": 5e-05, "loss": 0.9578, "mean_token_accuracy": 0.8588542342185974, "num_tokens": 471138001.0, "step": 12409 }, { "epoch": 1.5786795573082304, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 5.971316337585449, "learning_rate": 5e-05, "loss": 0.6728, "mean_token_accuracy": 0.8754619359970093, "num_tokens": 471170392.0, "step": 12410 }, { "epoch": 1.578806767586821, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.949091911315918, "learning_rate": 5e-05, "loss": 1.0251, "mean_token_accuracy": 0.8624187707901001, "num_tokens": 471215704.0, "step": 12411 }, { "epoch": 1.5789339778654115, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.5734639167785645, "learning_rate": 5e-05, "loss": 0.7523, "mean_token_accuracy": 0.8866933584213257, "num_tokens": 471252084.0, "step": 12412 }, { "epoch": 1.579061188144002, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.69119644165039, "learning_rate": 5e-05, "loss": 1.0395, "mean_token_accuracy": 0.866984486579895, "num_tokens": 471287096.0, "step": 12413 }, { "epoch": 1.5791883984225925, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.984968185424805, "learning_rate": 5e-05, "loss": 0.9317, "mean_token_accuracy": 0.8882694244384766, "num_tokens": 471321732.0, "step": 12414 }, { "epoch": 1.579315608701183, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.04463005065918, "learning_rate": 5e-05, "loss": 0.906, "mean_token_accuracy": 0.871971607208252, "num_tokens": 471353680.0, "step": 12415 }, { "epoch": 1.5794428189797736, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.15552043914795, "learning_rate": 5e-05, "loss": 1.0659, "mean_token_accuracy": 0.8583918809890747, "num_tokens": 471395286.0, "step": 12416 }, { "epoch": 1.579570029258364, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.4927754402160645, "learning_rate": 5e-05, "loss": 0.7129, "mean_token_accuracy": 0.8810921311378479, "num_tokens": 471433708.0, "step": 12417 }, { "epoch": 1.5796972395369546, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.227948188781738, "learning_rate": 5e-05, "loss": 1.057, "mean_token_accuracy": 0.8702238202095032, "num_tokens": 471472655.0, "step": 12418 }, { "epoch": 1.5798244498155452, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 6.776937007904053, "learning_rate": 5e-05, "loss": 0.8368, "mean_token_accuracy": 0.8709561824798584, "num_tokens": 471507013.0, "step": 12419 }, { "epoch": 1.5799516600941357, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.252049446105957, "learning_rate": 5e-05, "loss": 1.0653, "mean_token_accuracy": 0.8747021555900574, "num_tokens": 471539951.0, "step": 12420 }, { "epoch": 1.5800788703727262, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 8.202025413513184, "learning_rate": 5e-05, "loss": 1.0482, "mean_token_accuracy": 0.867546558380127, "num_tokens": 471580505.0, "step": 12421 }, { "epoch": 1.5802060806513167, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.945837020874023, "learning_rate": 5e-05, "loss": 0.9216, "mean_token_accuracy": 0.8799784183502197, "num_tokens": 471610408.0, "step": 12422 }, { "epoch": 1.5803332909299073, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 8.760760307312012, "learning_rate": 5e-05, "loss": 1.0475, "mean_token_accuracy": 0.8681021332740784, "num_tokens": 471648529.0, "step": 12423 }, { "epoch": 1.5804605012084978, "ewc_loss": 0.5, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.88352632522583, "learning_rate": 5e-05, "loss": 0.8696, "mean_token_accuracy": 0.8725014925003052, "num_tokens": 471682182.0, "step": 12424 }, { "epoch": 1.5805877114870883, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.621631622314453, "learning_rate": 5e-05, "loss": 0.9533, "mean_token_accuracy": 0.8743690252304077, "num_tokens": 471719180.0, "step": 12425 }, { "epoch": 1.5807149217656786, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.852344036102295, "learning_rate": 5e-05, "loss": 0.9042, "mean_token_accuracy": 0.8670085668563843, "num_tokens": 471759965.0, "step": 12426 }, { "epoch": 1.5808421320442692, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.856795787811279, "learning_rate": 5e-05, "loss": 0.8237, "mean_token_accuracy": 0.8821502923965454, "num_tokens": 471795405.0, "step": 12427 }, { "epoch": 1.5809693423228597, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.218615531921387, "learning_rate": 5e-05, "loss": 0.8867, "mean_token_accuracy": 0.8789277076721191, "num_tokens": 471837957.0, "step": 12428 }, { "epoch": 1.5810965526014502, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.179633617401123, "learning_rate": 5e-05, "loss": 0.7752, "mean_token_accuracy": 0.8712438941001892, "num_tokens": 471875422.0, "step": 12429 }, { "epoch": 1.5812237628800407, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.563118934631348, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8781030178070068, "num_tokens": 471913752.0, "step": 12430 }, { "epoch": 1.5813509731586313, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.900412082672119, "learning_rate": 5e-05, "loss": 0.7669, "mean_token_accuracy": 0.8653144836425781, "num_tokens": 471952362.0, "step": 12431 }, { "epoch": 1.5814781834372216, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.813091278076172, "learning_rate": 5e-05, "loss": 0.8978, "mean_token_accuracy": 0.879682183265686, "num_tokens": 471990798.0, "step": 12432 }, { "epoch": 1.581605393715812, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.792140007019043, "learning_rate": 5e-05, "loss": 0.764, "mean_token_accuracy": 0.8697459697723389, "num_tokens": 472027398.0, "step": 12433 }, { "epoch": 1.5817326039944026, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.016162872314453, "learning_rate": 5e-05, "loss": 0.9352, "mean_token_accuracy": 0.8760221600532532, "num_tokens": 472071725.0, "step": 12434 }, { "epoch": 1.5818598142729932, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 6.857702255249023, "learning_rate": 5e-05, "loss": 0.7533, "mean_token_accuracy": 0.8826931118965149, "num_tokens": 472106913.0, "step": 12435 }, { "epoch": 1.5819870245515837, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.99416732788086, "learning_rate": 5e-05, "loss": 0.9515, "mean_token_accuracy": 0.8706071376800537, "num_tokens": 472138136.0, "step": 12436 }, { "epoch": 1.5821142348301742, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.1105852127075195, "learning_rate": 5e-05, "loss": 0.8505, "mean_token_accuracy": 0.8694773316383362, "num_tokens": 472174884.0, "step": 12437 }, { "epoch": 1.5822414451087647, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.75589656829834, "learning_rate": 5e-05, "loss": 0.9545, "mean_token_accuracy": 0.8684750199317932, "num_tokens": 472215862.0, "step": 12438 }, { "epoch": 1.5823686553873553, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.294605731964111, "learning_rate": 5e-05, "loss": 0.8314, "mean_token_accuracy": 0.8786808252334595, "num_tokens": 472255770.0, "step": 12439 }, { "epoch": 1.5824958656659458, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.541053771972656, "learning_rate": 5e-05, "loss": 0.9625, "mean_token_accuracy": 0.8637309074401855, "num_tokens": 472290815.0, "step": 12440 }, { "epoch": 1.5826230759445363, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.317139625549316, "learning_rate": 5e-05, "loss": 0.8763, "mean_token_accuracy": 0.8635390996932983, "num_tokens": 472328180.0, "step": 12441 }, { "epoch": 1.5827502862231269, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.439502716064453, "learning_rate": 5e-05, "loss": 0.9319, "mean_token_accuracy": 0.8627026677131653, "num_tokens": 472366519.0, "step": 12442 }, { "epoch": 1.5828774965017174, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.309962749481201, "learning_rate": 5e-05, "loss": 0.8729, "mean_token_accuracy": 0.8578678965568542, "num_tokens": 472405270.0, "step": 12443 }, { "epoch": 1.583004706780308, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.316471099853516, "learning_rate": 5e-05, "loss": 0.8878, "mean_token_accuracy": 0.8770840167999268, "num_tokens": 472447572.0, "step": 12444 }, { "epoch": 1.5831319170588984, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.32232141494751, "learning_rate": 5e-05, "loss": 0.807, "mean_token_accuracy": 0.8751509189605713, "num_tokens": 472487869.0, "step": 12445 }, { "epoch": 1.583259127337489, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.213040351867676, "learning_rate": 5e-05, "loss": 0.8763, "mean_token_accuracy": 0.8748779892921448, "num_tokens": 472521856.0, "step": 12446 }, { "epoch": 1.5833863376160795, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.314740180969238, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.8730523586273193, "num_tokens": 472566550.0, "step": 12447 }, { "epoch": 1.58351354789467, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.134077072143555, "learning_rate": 5e-05, "loss": 0.8679, "mean_token_accuracy": 0.8688915371894836, "num_tokens": 472605054.0, "step": 12448 }, { "epoch": 1.5836407581732606, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.323944091796875, "learning_rate": 5e-05, "loss": 0.7861, "mean_token_accuracy": 0.8742316365242004, "num_tokens": 472649938.0, "step": 12449 }, { "epoch": 1.5837679684518509, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.024621963500977, "learning_rate": 5e-05, "loss": 0.8161, "mean_token_accuracy": 0.875531792640686, "num_tokens": 472688399.0, "step": 12450 }, { "epoch": 1.5838951787304414, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.39210319519043, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.865896463394165, "num_tokens": 472726383.0, "step": 12451 }, { "epoch": 1.584022389009032, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.8790812492370605, "learning_rate": 5e-05, "loss": 0.8151, "mean_token_accuracy": 0.866568922996521, "num_tokens": 472759692.0, "step": 12452 }, { "epoch": 1.5841495992876224, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.418868541717529, "learning_rate": 5e-05, "loss": 0.7628, "mean_token_accuracy": 0.877574622631073, "num_tokens": 472799679.0, "step": 12453 }, { "epoch": 1.584276809566213, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.792364597320557, "learning_rate": 5e-05, "loss": 0.8288, "mean_token_accuracy": 0.8595074415206909, "num_tokens": 472842043.0, "step": 12454 }, { "epoch": 1.5844040198448035, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.432529449462891, "learning_rate": 5e-05, "loss": 0.7869, "mean_token_accuracy": 0.8658322095870972, "num_tokens": 472877594.0, "step": 12455 }, { "epoch": 1.5845312301233938, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.732759475708008, "learning_rate": 5e-05, "loss": 0.7757, "mean_token_accuracy": 0.8741318583488464, "num_tokens": 472914052.0, "step": 12456 }, { "epoch": 1.5846584404019843, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.444468975067139, "learning_rate": 5e-05, "loss": 0.7455, "mean_token_accuracy": 0.8742150068283081, "num_tokens": 472953359.0, "step": 12457 }, { "epoch": 1.5847856506805749, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.665038108825684, "learning_rate": 5e-05, "loss": 0.771, "mean_token_accuracy": 0.8674829006195068, "num_tokens": 472993898.0, "step": 12458 }, { "epoch": 1.5849128609591654, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.438231468200684, "learning_rate": 5e-05, "loss": 0.7314, "mean_token_accuracy": 0.8748782873153687, "num_tokens": 473032026.0, "step": 12459 }, { "epoch": 1.585040071237756, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.652982711791992, "learning_rate": 5e-05, "loss": 0.7336, "mean_token_accuracy": 0.877299427986145, "num_tokens": 473075282.0, "step": 12460 }, { "epoch": 1.5851672815163464, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.410223484039307, "learning_rate": 5e-05, "loss": 0.715, "mean_token_accuracy": 0.875791609287262, "num_tokens": 473112834.0, "step": 12461 }, { "epoch": 1.585294491794937, "ewc_loss": 0.375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.660145282745361, "learning_rate": 5e-05, "loss": 0.7531, "mean_token_accuracy": 0.8705836534500122, "num_tokens": 473150331.0, "step": 12462 }, { "epoch": 1.5854217020735275, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.35822057723999, "learning_rate": 5e-05, "loss": 0.7139, "mean_token_accuracy": 0.8775238394737244, "num_tokens": 473185175.0, "step": 12463 }, { "epoch": 1.585548912352118, "ewc_loss": 0.375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.72420072555542, "learning_rate": 5e-05, "loss": 0.7608, "mean_token_accuracy": 0.8629839420318604, "num_tokens": 473223546.0, "step": 12464 }, { "epoch": 1.5856761226307086, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.251276969909668, "learning_rate": 5e-05, "loss": 0.7766, "mean_token_accuracy": 0.8579857349395752, "num_tokens": 473260098.0, "step": 12465 }, { "epoch": 1.585803332909299, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.8350911140441895, "learning_rate": 5e-05, "loss": 0.7393, "mean_token_accuracy": 0.8745949268341064, "num_tokens": 473299283.0, "step": 12466 }, { "epoch": 1.5859305431878896, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.1055755615234375, "learning_rate": 5e-05, "loss": 0.7161, "mean_token_accuracy": 0.8665838241577148, "num_tokens": 473332511.0, "step": 12467 }, { "epoch": 1.5860577534664801, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 8.021258354187012, "learning_rate": 5e-05, "loss": 0.7486, "mean_token_accuracy": 0.8776112198829651, "num_tokens": 473367259.0, "step": 12468 }, { "epoch": 1.5861849637450707, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.883113861083984, "learning_rate": 5e-05, "loss": 0.6862, "mean_token_accuracy": 0.8710386753082275, "num_tokens": 473412421.0, "step": 12469 }, { "epoch": 1.5863121740236612, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.350573539733887, "learning_rate": 5e-05, "loss": 0.8636, "mean_token_accuracy": 0.8495156764984131, "num_tokens": 473450444.0, "step": 12470 }, { "epoch": 1.5864393843022517, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.5199432373046875, "learning_rate": 5e-05, "loss": 0.6686, "mean_token_accuracy": 0.8704708814620972, "num_tokens": 473485093.0, "step": 12471 }, { "epoch": 1.5865665945808423, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.877472877502441, "learning_rate": 5e-05, "loss": 0.8103, "mean_token_accuracy": 0.8841066360473633, "num_tokens": 473522461.0, "step": 12472 }, { "epoch": 1.5866938048594328, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.114963531494141, "learning_rate": 5e-05, "loss": 0.6878, "mean_token_accuracy": 0.8671098351478577, "num_tokens": 473559860.0, "step": 12473 }, { "epoch": 1.5868210151380233, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.529687881469727, "learning_rate": 5e-05, "loss": 0.9205, "mean_token_accuracy": 0.880886435508728, "num_tokens": 473590144.0, "step": 12474 }, { "epoch": 1.5869482254166136, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.145697593688965, "learning_rate": 5e-05, "loss": 0.6919, "mean_token_accuracy": 0.8776195049285889, "num_tokens": 473631384.0, "step": 12475 }, { "epoch": 1.5870754356952042, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.686257362365723, "learning_rate": 5e-05, "loss": 1.0149, "mean_token_accuracy": 0.8702135682106018, "num_tokens": 473667396.0, "step": 12476 }, { "epoch": 1.5872026459737947, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.238959789276123, "learning_rate": 5e-05, "loss": 0.8742, "mean_token_accuracy": 0.8736197352409363, "num_tokens": 473699856.0, "step": 12477 }, { "epoch": 1.5873298562523852, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.461078643798828, "learning_rate": 5e-05, "loss": 0.9668, "mean_token_accuracy": 0.8523601293563843, "num_tokens": 473731164.0, "step": 12478 }, { "epoch": 1.5874570665309757, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.677565574645996, "learning_rate": 5e-05, "loss": 0.9794, "mean_token_accuracy": 0.869601845741272, "num_tokens": 473767005.0, "step": 12479 }, { "epoch": 1.5875842768095663, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.398380756378174, "learning_rate": 5e-05, "loss": 0.7157, "mean_token_accuracy": 0.8682535886764526, "num_tokens": 473805753.0, "step": 12480 }, { "epoch": 1.5877114870881566, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.617377281188965, "learning_rate": 5e-05, "loss": 1.0904, "mean_token_accuracy": 0.870566189289093, "num_tokens": 473845375.0, "step": 12481 }, { "epoch": 1.587838697366747, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.024387836456299, "learning_rate": 5e-05, "loss": 0.7612, "mean_token_accuracy": 0.8614953756332397, "num_tokens": 473889184.0, "step": 12482 }, { "epoch": 1.5879659076453376, "ewc_loss": 1.046875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0009918212890625, "grad_norm": 12.498242378234863, "learning_rate": 5e-05, "loss": 1.4034, "mean_token_accuracy": 0.8796555995941162, "num_tokens": 473927917.0, "step": 12483 }, { "epoch": 1.5880931179239282, "ewc_loss": 0.9609375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00090789794921875, "grad_norm": 9.604862213134766, "learning_rate": 5e-05, "loss": 1.3, "mean_token_accuracy": 0.8822667002677917, "num_tokens": 473964670.0, "step": 12484 }, { "epoch": 1.5882203282025187, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 11.131354331970215, "learning_rate": 5e-05, "loss": 1.0506, "mean_token_accuracy": 0.8873919248580933, "num_tokens": 473998455.0, "step": 12485 }, { "epoch": 1.5883475384811092, "ewc_loss": 0.921875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000865936279296875, "grad_norm": 9.829793930053711, "learning_rate": 5e-05, "loss": 1.3039, "mean_token_accuracy": 0.8713249564170837, "num_tokens": 474041623.0, "step": 12486 }, { "epoch": 1.5884747487596997, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.206416130065918, "learning_rate": 5e-05, "loss": 1.0627, "mean_token_accuracy": 0.8694034814834595, "num_tokens": 474083294.0, "step": 12487 }, { "epoch": 1.5886019590382903, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.176778793334961, "learning_rate": 5e-05, "loss": 1.041, "mean_token_accuracy": 0.8713613748550415, "num_tokens": 474117737.0, "step": 12488 }, { "epoch": 1.5887291693168808, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.212719917297363, "learning_rate": 5e-05, "loss": 1.0294, "mean_token_accuracy": 0.869435727596283, "num_tokens": 474154467.0, "step": 12489 }, { "epoch": 1.5888563795954713, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.711639404296875, "learning_rate": 5e-05, "loss": 0.9511, "mean_token_accuracy": 0.8688082695007324, "num_tokens": 474191845.0, "step": 12490 }, { "epoch": 1.5889835898740619, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.723819732666016, "learning_rate": 5e-05, "loss": 0.9327, "mean_token_accuracy": 0.876108705997467, "num_tokens": 474227787.0, "step": 12491 }, { "epoch": 1.5891108001526524, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.065923690795898, "learning_rate": 5e-05, "loss": 0.9005, "mean_token_accuracy": 0.8698809742927551, "num_tokens": 474271890.0, "step": 12492 }, { "epoch": 1.589238010431243, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.654176712036133, "learning_rate": 5e-05, "loss": 0.957, "mean_token_accuracy": 0.8591206669807434, "num_tokens": 474314493.0, "step": 12493 }, { "epoch": 1.5893652207098334, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.597811698913574, "learning_rate": 5e-05, "loss": 0.7875, "mean_token_accuracy": 0.8804558515548706, "num_tokens": 474349348.0, "step": 12494 }, { "epoch": 1.589492430988424, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.650223731994629, "learning_rate": 5e-05, "loss": 0.9258, "mean_token_accuracy": 0.8658365607261658, "num_tokens": 474384145.0, "step": 12495 }, { "epoch": 1.5896196412670145, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.276386260986328, "learning_rate": 5e-05, "loss": 0.756, "mean_token_accuracy": 0.8827000856399536, "num_tokens": 474419064.0, "step": 12496 }, { "epoch": 1.589746851545605, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.832437515258789, "learning_rate": 5e-05, "loss": 0.9281, "mean_token_accuracy": 0.869937539100647, "num_tokens": 474458716.0, "step": 12497 }, { "epoch": 1.5898740618241956, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.938922882080078, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.8776988983154297, "num_tokens": 474497672.0, "step": 12498 }, { "epoch": 1.5900012721027859, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.324005126953125, "learning_rate": 5e-05, "loss": 0.9539, "mean_token_accuracy": 0.8759604096412659, "num_tokens": 474530442.0, "step": 12499 }, { "epoch": 1.5901284823813764, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.812869071960449, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.8534338474273682, "num_tokens": 474573103.0, "step": 12500 }, { "epoch": 1.590255692659967, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.863643646240234, "learning_rate": 5e-05, "loss": 1.0084, "mean_token_accuracy": 0.8794828653335571, "num_tokens": 474610687.0, "step": 12501 }, { "epoch": 1.5903829029385574, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.369076251983643, "learning_rate": 5e-05, "loss": 0.9061, "mean_token_accuracy": 0.8654941320419312, "num_tokens": 474650893.0, "step": 12502 }, { "epoch": 1.590510113217148, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.756904602050781, "learning_rate": 5e-05, "loss": 1.0552, "mean_token_accuracy": 0.8712303638458252, "num_tokens": 474690808.0, "step": 12503 }, { "epoch": 1.5906373234957385, "ewc_loss": 0.625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.206870079040527, "learning_rate": 5e-05, "loss": 0.9631, "mean_token_accuracy": 0.8812766075134277, "num_tokens": 474726814.0, "step": 12504 }, { "epoch": 1.5907645337743288, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.817364692687988, "learning_rate": 5e-05, "loss": 0.9734, "mean_token_accuracy": 0.8693428039550781, "num_tokens": 474765170.0, "step": 12505 }, { "epoch": 1.5908917440529193, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.50039291381836, "learning_rate": 5e-05, "loss": 0.9932, "mean_token_accuracy": 0.8689787983894348, "num_tokens": 474808466.0, "step": 12506 }, { "epoch": 1.5910189543315099, "ewc_loss": 0.5, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.806739330291748, "learning_rate": 5e-05, "loss": 0.8529, "mean_token_accuracy": 0.8773670792579651, "num_tokens": 474847334.0, "step": 12507 }, { "epoch": 1.5911461646101004, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.439777374267578, "learning_rate": 5e-05, "loss": 0.9098, "mean_token_accuracy": 0.873069167137146, "num_tokens": 474885937.0, "step": 12508 }, { "epoch": 1.591273374888691, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.538663387298584, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.8769345283508301, "num_tokens": 474921089.0, "step": 12509 }, { "epoch": 1.5914005851672814, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.918057918548584, "learning_rate": 5e-05, "loss": 0.8126, "mean_token_accuracy": 0.8782128095626831, "num_tokens": 474961660.0, "step": 12510 }, { "epoch": 1.591527795445872, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.852179050445557, "learning_rate": 5e-05, "loss": 0.8224, "mean_token_accuracy": 0.8761445879936218, "num_tokens": 474997815.0, "step": 12511 }, { "epoch": 1.5916550057244625, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.375839710235596, "learning_rate": 5e-05, "loss": 0.8263, "mean_token_accuracy": 0.8468818068504333, "num_tokens": 475038334.0, "step": 12512 }, { "epoch": 1.591782216003053, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.125097274780273, "learning_rate": 5e-05, "loss": 0.8218, "mean_token_accuracy": 0.8737457990646362, "num_tokens": 475073522.0, "step": 12513 }, { "epoch": 1.5919094262816436, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.077091693878174, "learning_rate": 5e-05, "loss": 0.6993, "mean_token_accuracy": 0.8777740001678467, "num_tokens": 475111811.0, "step": 12514 }, { "epoch": 1.592036636560234, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.416133880615234, "learning_rate": 5e-05, "loss": 0.8035, "mean_token_accuracy": 0.8889309167861938, "num_tokens": 475146496.0, "step": 12515 }, { "epoch": 1.5921638468388246, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.720503807067871, "learning_rate": 5e-05, "loss": 0.7173, "mean_token_accuracy": 0.871584415435791, "num_tokens": 475187947.0, "step": 12516 }, { "epoch": 1.5922910571174151, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.001886367797852, "learning_rate": 5e-05, "loss": 0.9161, "mean_token_accuracy": 0.8726565837860107, "num_tokens": 475225420.0, "step": 12517 }, { "epoch": 1.5924182673960057, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.478118419647217, "learning_rate": 5e-05, "loss": 0.7429, "mean_token_accuracy": 0.864052414894104, "num_tokens": 475263913.0, "step": 12518 }, { "epoch": 1.5925454776745962, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.450081825256348, "learning_rate": 5e-05, "loss": 0.9491, "mean_token_accuracy": 0.8839775323867798, "num_tokens": 475300446.0, "step": 12519 }, { "epoch": 1.5926726879531867, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.976324558258057, "learning_rate": 5e-05, "loss": 0.8465, "mean_token_accuracy": 0.8596187829971313, "num_tokens": 475338728.0, "step": 12520 }, { "epoch": 1.5927998982317773, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.128314971923828, "learning_rate": 5e-05, "loss": 0.9757, "mean_token_accuracy": 0.869533896446228, "num_tokens": 475369453.0, "step": 12521 }, { "epoch": 1.5929271085103678, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 7.835928916931152, "learning_rate": 5e-05, "loss": 0.8714, "mean_token_accuracy": 0.8857625722885132, "num_tokens": 475407994.0, "step": 12522 }, { "epoch": 1.5930543187889583, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.111987113952637, "learning_rate": 5e-05, "loss": 0.8916, "mean_token_accuracy": 0.8681840300559998, "num_tokens": 475445480.0, "step": 12523 }, { "epoch": 1.5931815290675486, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.290035247802734, "learning_rate": 5e-05, "loss": 0.954, "mean_token_accuracy": 0.8578320741653442, "num_tokens": 475483971.0, "step": 12524 }, { "epoch": 1.5933087393461391, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.324004173278809, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8739161491394043, "num_tokens": 475521332.0, "step": 12525 }, { "epoch": 1.5934359496247297, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.425442695617676, "learning_rate": 5e-05, "loss": 0.8922, "mean_token_accuracy": 0.8667659759521484, "num_tokens": 475556085.0, "step": 12526 }, { "epoch": 1.5935631599033202, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.081579685211182, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.8698484897613525, "num_tokens": 475590966.0, "step": 12527 }, { "epoch": 1.5936903701819107, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.451446533203125, "learning_rate": 5e-05, "loss": 0.8922, "mean_token_accuracy": 0.864692211151123, "num_tokens": 475627430.0, "step": 12528 }, { "epoch": 1.5938175804605013, "ewc_loss": 0.375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.01786994934082, "learning_rate": 5e-05, "loss": 0.7087, "mean_token_accuracy": 0.8807033896446228, "num_tokens": 475662346.0, "step": 12529 }, { "epoch": 1.5939447907390916, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.55684757232666, "learning_rate": 5e-05, "loss": 0.9007, "mean_token_accuracy": 0.8642206192016602, "num_tokens": 475702312.0, "step": 12530 }, { "epoch": 1.594072001017682, "ewc_loss": 0.375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.9273223876953125, "learning_rate": 5e-05, "loss": 0.7534, "mean_token_accuracy": 0.8725922107696533, "num_tokens": 475743997.0, "step": 12531 }, { "epoch": 1.5941992112962726, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.71652603149414, "learning_rate": 5e-05, "loss": 0.9318, "mean_token_accuracy": 0.8612744808197021, "num_tokens": 475783369.0, "step": 12532 }, { "epoch": 1.5943264215748632, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.93461799621582, "learning_rate": 5e-05, "loss": 0.7194, "mean_token_accuracy": 0.8872787356376648, "num_tokens": 475815225.0, "step": 12533 }, { "epoch": 1.5944536318534537, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.723949432373047, "learning_rate": 5e-05, "loss": 0.9161, "mean_token_accuracy": 0.8663284778594971, "num_tokens": 475847917.0, "step": 12534 }, { "epoch": 1.5945808421320442, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.136948108673096, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8768691420555115, "num_tokens": 475889308.0, "step": 12535 }, { "epoch": 1.5947080524106347, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.46324348449707, "learning_rate": 5e-05, "loss": 0.8683, "mean_token_accuracy": 0.8807528614997864, "num_tokens": 475930603.0, "step": 12536 }, { "epoch": 1.5948352626892253, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.362967014312744, "learning_rate": 5e-05, "loss": 0.7847, "mean_token_accuracy": 0.879409909248352, "num_tokens": 475968288.0, "step": 12537 }, { "epoch": 1.5949624729678158, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.206273078918457, "learning_rate": 5e-05, "loss": 0.8463, "mean_token_accuracy": 0.8778485655784607, "num_tokens": 476002967.0, "step": 12538 }, { "epoch": 1.5950896832464063, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.388091564178467, "learning_rate": 5e-05, "loss": 0.8014, "mean_token_accuracy": 0.872969388961792, "num_tokens": 476037958.0, "step": 12539 }, { "epoch": 1.5952168935249968, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.13783073425293, "learning_rate": 5e-05, "loss": 0.9071, "mean_token_accuracy": 0.8515419960021973, "num_tokens": 476072036.0, "step": 12540 }, { "epoch": 1.5953441038035874, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.2320709228515625, "learning_rate": 5e-05, "loss": 0.7598, "mean_token_accuracy": 0.874420702457428, "num_tokens": 476115665.0, "step": 12541 }, { "epoch": 1.595471314082178, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.298579216003418, "learning_rate": 5e-05, "loss": 0.8556, "mean_token_accuracy": 0.8674532771110535, "num_tokens": 476155380.0, "step": 12542 }, { "epoch": 1.5955985243607684, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.987415790557861, "learning_rate": 5e-05, "loss": 0.7475, "mean_token_accuracy": 0.8661975860595703, "num_tokens": 476194833.0, "step": 12543 }, { "epoch": 1.595725734639359, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.550930976867676, "learning_rate": 5e-05, "loss": 0.866, "mean_token_accuracy": 0.8736026287078857, "num_tokens": 476233617.0, "step": 12544 }, { "epoch": 1.5958529449179495, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.898061275482178, "learning_rate": 5e-05, "loss": 0.7391, "mean_token_accuracy": 0.869290292263031, "num_tokens": 476270312.0, "step": 12545 }, { "epoch": 1.59598015519654, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.5984525680542, "learning_rate": 5e-05, "loss": 0.8085, "mean_token_accuracy": 0.8965535163879395, "num_tokens": 476304466.0, "step": 12546 }, { "epoch": 1.5961073654751305, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.0894269943237305, "learning_rate": 5e-05, "loss": 0.7462, "mean_token_accuracy": 0.8741940259933472, "num_tokens": 476345136.0, "step": 12547 }, { "epoch": 1.5962345757537209, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.26599407196045, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.8761695027351379, "num_tokens": 476381546.0, "step": 12548 }, { "epoch": 1.5963617860323114, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.471311569213867, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8712244033813477, "num_tokens": 476413990.0, "step": 12549 }, { "epoch": 1.596488996310902, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.752212047576904, "learning_rate": 5e-05, "loss": 0.758, "mean_token_accuracy": 0.884442925453186, "num_tokens": 476451487.0, "step": 12550 }, { "epoch": 1.5966162065894924, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.781430244445801, "learning_rate": 5e-05, "loss": 0.8404, "mean_token_accuracy": 0.8601679801940918, "num_tokens": 476493071.0, "step": 12551 }, { "epoch": 1.596743416868083, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.348332405090332, "learning_rate": 5e-05, "loss": 0.7206, "mean_token_accuracy": 0.879854679107666, "num_tokens": 476531390.0, "step": 12552 }, { "epoch": 1.5968706271466735, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.977443695068359, "learning_rate": 5e-05, "loss": 0.803, "mean_token_accuracy": 0.8701285719871521, "num_tokens": 476572491.0, "step": 12553 }, { "epoch": 1.5969978374252638, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.152479648590088, "learning_rate": 5e-05, "loss": 0.7571, "mean_token_accuracy": 0.8632800579071045, "num_tokens": 476613897.0, "step": 12554 }, { "epoch": 1.5971250477038543, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.016185760498047, "learning_rate": 5e-05, "loss": 0.8073, "mean_token_accuracy": 0.8673136234283447, "num_tokens": 476650773.0, "step": 12555 }, { "epoch": 1.5972522579824449, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.147763252258301, "learning_rate": 5e-05, "loss": 0.732, "mean_token_accuracy": 0.8734840154647827, "num_tokens": 476685546.0, "step": 12556 }, { "epoch": 1.5973794682610354, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.944949626922607, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8546047210693359, "num_tokens": 476720569.0, "step": 12557 }, { "epoch": 1.597506678539626, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.198009490966797, "learning_rate": 5e-05, "loss": 0.7099, "mean_token_accuracy": 0.8739774823188782, "num_tokens": 476760162.0, "step": 12558 }, { "epoch": 1.5976338888182164, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.861668586730957, "learning_rate": 5e-05, "loss": 0.7651, "mean_token_accuracy": 0.8692930936813354, "num_tokens": 476793575.0, "step": 12559 }, { "epoch": 1.597761099096807, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.240599632263184, "learning_rate": 5e-05, "loss": 0.7233, "mean_token_accuracy": 0.8746806979179382, "num_tokens": 476826714.0, "step": 12560 }, { "epoch": 1.5978883093753975, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.796701431274414, "learning_rate": 5e-05, "loss": 0.7143, "mean_token_accuracy": 0.8845925331115723, "num_tokens": 476866715.0, "step": 12561 }, { "epoch": 1.598015519653988, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.29325008392334, "learning_rate": 5e-05, "loss": 0.7306, "mean_token_accuracy": 0.866705060005188, "num_tokens": 476906036.0, "step": 12562 }, { "epoch": 1.5981427299325786, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.738431930541992, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.8690091371536255, "num_tokens": 476945005.0, "step": 12563 }, { "epoch": 1.598269940211169, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.311425685882568, "learning_rate": 5e-05, "loss": 0.7058, "mean_token_accuracy": 0.8770065307617188, "num_tokens": 476989491.0, "step": 12564 }, { "epoch": 1.5983971504897596, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.724514484405518, "learning_rate": 5e-05, "loss": 0.7604, "mean_token_accuracy": 0.8689453601837158, "num_tokens": 477026062.0, "step": 12565 }, { "epoch": 1.5985243607683501, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.2755303382873535, "learning_rate": 5e-05, "loss": 0.7182, "mean_token_accuracy": 0.8706048727035522, "num_tokens": 477060045.0, "step": 12566 }, { "epoch": 1.5986515710469407, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.767025470733643, "learning_rate": 5e-05, "loss": 0.7133, "mean_token_accuracy": 0.8817319869995117, "num_tokens": 477102343.0, "step": 12567 }, { "epoch": 1.5987787813255312, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.163422107696533, "learning_rate": 5e-05, "loss": 0.6783, "mean_token_accuracy": 0.8798673152923584, "num_tokens": 477136239.0, "step": 12568 }, { "epoch": 1.5989059916041217, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.937904357910156, "learning_rate": 5e-05, "loss": 0.7616, "mean_token_accuracy": 0.8708268404006958, "num_tokens": 477178291.0, "step": 12569 }, { "epoch": 1.5990332018827123, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.932039737701416, "learning_rate": 5e-05, "loss": 0.7429, "mean_token_accuracy": 0.854590654373169, "num_tokens": 477213993.0, "step": 12570 }, { "epoch": 1.5991604121613028, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.250819206237793, "learning_rate": 5e-05, "loss": 0.8228, "mean_token_accuracy": 0.8600722551345825, "num_tokens": 477258384.0, "step": 12571 }, { "epoch": 1.5992876224398933, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.548844814300537, "learning_rate": 5e-05, "loss": 0.6722, "mean_token_accuracy": 0.8720880150794983, "num_tokens": 477297386.0, "step": 12572 }, { "epoch": 1.5994148327184836, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.831109046936035, "learning_rate": 5e-05, "loss": 0.8149, "mean_token_accuracy": 0.8812313675880432, "num_tokens": 477334214.0, "step": 12573 }, { "epoch": 1.5995420429970741, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002307891845703125, "grad_norm": 5.99227237701416, "learning_rate": 5e-05, "loss": 0.6899, "mean_token_accuracy": 0.8608844876289368, "num_tokens": 477374777.0, "step": 12574 }, { "epoch": 1.5996692532756647, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.76416301727295, "learning_rate": 5e-05, "loss": 0.9545, "mean_token_accuracy": 0.8706815242767334, "num_tokens": 477415206.0, "step": 12575 }, { "epoch": 1.5997964635542552, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.74029016494751, "learning_rate": 5e-05, "loss": 0.6858, "mean_token_accuracy": 0.8680005073547363, "num_tokens": 477453864.0, "step": 12576 }, { "epoch": 1.5999236738328457, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.575430870056152, "learning_rate": 5e-05, "loss": 1.1364, "mean_token_accuracy": 0.8575602173805237, "num_tokens": 477489236.0, "step": 12577 }, { "epoch": 1.6000508841114363, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.134275913238525, "learning_rate": 5e-05, "loss": 0.8873, "mean_token_accuracy": 0.8817591667175293, "num_tokens": 477524459.0, "step": 12578 }, { "epoch": 1.6001780943900266, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.843939781188965, "learning_rate": 5e-05, "loss": 1.0498, "mean_token_accuracy": 0.8675507307052612, "num_tokens": 477560045.0, "step": 12579 }, { "epoch": 1.600305304668617, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 8.930535316467285, "learning_rate": 5e-05, "loss": 1.1542, "mean_token_accuracy": 0.8593392968177795, "num_tokens": 477596438.0, "step": 12580 }, { "epoch": 1.6004325149472076, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.382099151611328, "learning_rate": 5e-05, "loss": 0.7726, "mean_token_accuracy": 0.8738404512405396, "num_tokens": 477634370.0, "step": 12581 }, { "epoch": 1.6005597252257981, "ewc_loss": 0.8515625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 10.950739860534668, "learning_rate": 5e-05, "loss": 1.2431, "mean_token_accuracy": 0.8654235601425171, "num_tokens": 477669798.0, "step": 12582 }, { "epoch": 1.6006869355043887, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 6.829873085021973, "learning_rate": 5e-05, "loss": 0.8355, "mean_token_accuracy": 0.8703819513320923, "num_tokens": 477714127.0, "step": 12583 }, { "epoch": 1.6008141457829792, "ewc_loss": 1.078125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0010223388671875, "grad_norm": 13.07341194152832, "learning_rate": 5e-05, "loss": 1.441, "mean_token_accuracy": 0.8752651214599609, "num_tokens": 477753176.0, "step": 12584 }, { "epoch": 1.6009413560615697, "ewc_loss": 1.1015625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00104522705078125, "grad_norm": 10.571358680725098, "learning_rate": 5e-05, "loss": 1.4615, "mean_token_accuracy": 0.8729361295700073, "num_tokens": 477789372.0, "step": 12585 }, { "epoch": 1.6010685663401603, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 11.54393196105957, "learning_rate": 5e-05, "loss": 1.1116, "mean_token_accuracy": 0.8676344752311707, "num_tokens": 477820891.0, "step": 12586 }, { "epoch": 1.6011957766187508, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.751823425292969, "learning_rate": 5e-05, "loss": 1.2231, "mean_token_accuracy": 0.8700724840164185, "num_tokens": 477853952.0, "step": 12587 }, { "epoch": 1.6013229868973413, "ewc_loss": 0.84375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 11.646336555480957, "learning_rate": 5e-05, "loss": 1.1743, "mean_token_accuracy": 0.885463297367096, "num_tokens": 477888692.0, "step": 12588 }, { "epoch": 1.6014501971759318, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.714290618896484, "learning_rate": 5e-05, "loss": 0.9369, "mean_token_accuracy": 0.8833807706832886, "num_tokens": 477923769.0, "step": 12589 }, { "epoch": 1.6015774074545224, "ewc_loss": 0.90625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 11.96154499053955, "learning_rate": 5e-05, "loss": 1.2252, "mean_token_accuracy": 0.8878971338272095, "num_tokens": 477959997.0, "step": 12590 }, { "epoch": 1.601704617733113, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 9.424986839294434, "learning_rate": 5e-05, "loss": 1.1051, "mean_token_accuracy": 0.8555723428726196, "num_tokens": 477996211.0, "step": 12591 }, { "epoch": 1.6018318280117034, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 11.73514461517334, "learning_rate": 5e-05, "loss": 1.1796, "mean_token_accuracy": 0.874086320400238, "num_tokens": 478029817.0, "step": 12592 }, { "epoch": 1.601959038290294, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000732421875, "grad_norm": 10.095869064331055, "learning_rate": 5e-05, "loss": 1.1314, "mean_token_accuracy": 0.8792445659637451, "num_tokens": 478071216.0, "step": 12593 }, { "epoch": 1.6020862485688845, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.898981094360352, "learning_rate": 5e-05, "loss": 1.0528, "mean_token_accuracy": 0.878478467464447, "num_tokens": 478112581.0, "step": 12594 }, { "epoch": 1.602213458847475, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.587809562683105, "learning_rate": 5e-05, "loss": 1.0642, "mean_token_accuracy": 0.861324667930603, "num_tokens": 478148887.0, "step": 12595 }, { "epoch": 1.6023406691260655, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.46297836303711, "learning_rate": 5e-05, "loss": 1.089, "mean_token_accuracy": 0.8612308502197266, "num_tokens": 478187580.0, "step": 12596 }, { "epoch": 1.6024678794046558, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.987018585205078, "learning_rate": 5e-05, "loss": 0.9404, "mean_token_accuracy": 0.8757859468460083, "num_tokens": 478222679.0, "step": 12597 }, { "epoch": 1.6025950896832464, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.287408828735352, "learning_rate": 5e-05, "loss": 1.0238, "mean_token_accuracy": 0.8775589466094971, "num_tokens": 478259464.0, "step": 12598 }, { "epoch": 1.602722299961837, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.566457748413086, "learning_rate": 5e-05, "loss": 0.8715, "mean_token_accuracy": 0.8848533630371094, "num_tokens": 478293790.0, "step": 12599 }, { "epoch": 1.6028495102404274, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.413265228271484, "learning_rate": 5e-05, "loss": 1.0429, "mean_token_accuracy": 0.8756883144378662, "num_tokens": 478329311.0, "step": 12600 }, { "epoch": 1.602976720519018, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.493332862854004, "learning_rate": 5e-05, "loss": 0.9103, "mean_token_accuracy": 0.8718763589859009, "num_tokens": 478362818.0, "step": 12601 }, { "epoch": 1.6031039307976085, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.581097602844238, "learning_rate": 5e-05, "loss": 1.0933, "mean_token_accuracy": 0.8676891922950745, "num_tokens": 478402783.0, "step": 12602 }, { "epoch": 1.6032311410761988, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.753558158874512, "learning_rate": 5e-05, "loss": 0.9591, "mean_token_accuracy": 0.8594094514846802, "num_tokens": 478438141.0, "step": 12603 }, { "epoch": 1.6033583513547893, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.478084564208984, "learning_rate": 5e-05, "loss": 1.0442, "mean_token_accuracy": 0.8800020217895508, "num_tokens": 478473176.0, "step": 12604 }, { "epoch": 1.6034855616333799, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.003393173217773, "learning_rate": 5e-05, "loss": 0.9384, "mean_token_accuracy": 0.8785620927810669, "num_tokens": 478511899.0, "step": 12605 }, { "epoch": 1.6036127719119704, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.968087196350098, "learning_rate": 5e-05, "loss": 0.9736, "mean_token_accuracy": 0.8869079947471619, "num_tokens": 478545018.0, "step": 12606 }, { "epoch": 1.603739982190561, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.881550788879395, "learning_rate": 5e-05, "loss": 0.934, "mean_token_accuracy": 0.868252694606781, "num_tokens": 478582384.0, "step": 12607 }, { "epoch": 1.6038671924691514, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.450191497802734, "learning_rate": 5e-05, "loss": 0.9945, "mean_token_accuracy": 0.8695930242538452, "num_tokens": 478621491.0, "step": 12608 }, { "epoch": 1.603994402747742, "ewc_loss": 0.5, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.445195198059082, "learning_rate": 5e-05, "loss": 0.8099, "mean_token_accuracy": 0.8935394287109375, "num_tokens": 478662018.0, "step": 12609 }, { "epoch": 1.6041216130263325, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.296156883239746, "learning_rate": 5e-05, "loss": 0.963, "mean_token_accuracy": 0.8712446093559265, "num_tokens": 478702730.0, "step": 12610 }, { "epoch": 1.604248823304923, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.962305545806885, "learning_rate": 5e-05, "loss": 0.8191, "mean_token_accuracy": 0.8700594305992126, "num_tokens": 478736696.0, "step": 12611 }, { "epoch": 1.6043760335835135, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.536538124084473, "learning_rate": 5e-05, "loss": 1.0442, "mean_token_accuracy": 0.8542395234107971, "num_tokens": 478775873.0, "step": 12612 }, { "epoch": 1.604503243862104, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.819929599761963, "learning_rate": 5e-05, "loss": 0.7715, "mean_token_accuracy": 0.880280613899231, "num_tokens": 478810571.0, "step": 12613 }, { "epoch": 1.6046304541406946, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.76044750213623, "learning_rate": 5e-05, "loss": 1.0387, "mean_token_accuracy": 0.8624231815338135, "num_tokens": 478845696.0, "step": 12614 }, { "epoch": 1.6047576644192851, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.168286323547363, "learning_rate": 5e-05, "loss": 0.8569, "mean_token_accuracy": 0.8677349090576172, "num_tokens": 478884561.0, "step": 12615 }, { "epoch": 1.6048848746978757, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.45179557800293, "learning_rate": 5e-05, "loss": 1.0574, "mean_token_accuracy": 0.8501183390617371, "num_tokens": 478921989.0, "step": 12616 }, { "epoch": 1.6050120849764662, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.423933029174805, "learning_rate": 5e-05, "loss": 0.8666, "mean_token_accuracy": 0.8710892200469971, "num_tokens": 478958346.0, "step": 12617 }, { "epoch": 1.6051392952550567, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.784368515014648, "learning_rate": 5e-05, "loss": 0.9152, "mean_token_accuracy": 0.8754997253417969, "num_tokens": 479000141.0, "step": 12618 }, { "epoch": 1.6052665055336472, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.21489429473877, "learning_rate": 5e-05, "loss": 0.8164, "mean_token_accuracy": 0.8795813322067261, "num_tokens": 479038785.0, "step": 12619 }, { "epoch": 1.6053937158122378, "ewc_loss": 0.5, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.392292022705078, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8640540838241577, "num_tokens": 479074679.0, "step": 12620 }, { "epoch": 1.6055209260908283, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.801867961883545, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8767374753952026, "num_tokens": 479106745.0, "step": 12621 }, { "epoch": 1.6056481363694186, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.348517417907715, "learning_rate": 5e-05, "loss": 0.8428, "mean_token_accuracy": 0.8736001253128052, "num_tokens": 479145159.0, "step": 12622 }, { "epoch": 1.6057753466480091, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.451450347900391, "learning_rate": 5e-05, "loss": 0.7094, "mean_token_accuracy": 0.8854323625564575, "num_tokens": 479181033.0, "step": 12623 }, { "epoch": 1.6059025569265997, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.459230422973633, "learning_rate": 5e-05, "loss": 0.8742, "mean_token_accuracy": 0.8660898208618164, "num_tokens": 479223954.0, "step": 12624 }, { "epoch": 1.6060297672051902, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.291313171386719, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8596922159194946, "num_tokens": 479268676.0, "step": 12625 }, { "epoch": 1.6061569774837807, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.482170104980469, "learning_rate": 5e-05, "loss": 0.8161, "mean_token_accuracy": 0.8832310438156128, "num_tokens": 479305203.0, "step": 12626 }, { "epoch": 1.6062841877623713, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.25142765045166, "learning_rate": 5e-05, "loss": 0.8111, "mean_token_accuracy": 0.8472514748573303, "num_tokens": 479347358.0, "step": 12627 }, { "epoch": 1.6064113980409616, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.488202095031738, "learning_rate": 5e-05, "loss": 0.8787, "mean_token_accuracy": 0.8645021915435791, "num_tokens": 479392973.0, "step": 12628 }, { "epoch": 1.606538608319552, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.29227876663208, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8603709936141968, "num_tokens": 479430832.0, "step": 12629 }, { "epoch": 1.6066658185981426, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.389350891113281, "learning_rate": 5e-05, "loss": 0.8716, "mean_token_accuracy": 0.8606384992599487, "num_tokens": 479471363.0, "step": 12630 }, { "epoch": 1.6067930288767331, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.4307637214660645, "learning_rate": 5e-05, "loss": 0.7201, "mean_token_accuracy": 0.8778873682022095, "num_tokens": 479506041.0, "step": 12631 }, { "epoch": 1.6069202391553237, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.075784683227539, "learning_rate": 5e-05, "loss": 0.8101, "mean_token_accuracy": 0.8719207048416138, "num_tokens": 479537290.0, "step": 12632 }, { "epoch": 1.6070474494339142, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.634040832519531, "learning_rate": 5e-05, "loss": 0.7924, "mean_token_accuracy": 0.8564929962158203, "num_tokens": 479573191.0, "step": 12633 }, { "epoch": 1.6071746597125047, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.654812335968018, "learning_rate": 5e-05, "loss": 0.7081, "mean_token_accuracy": 0.8845292925834656, "num_tokens": 479606210.0, "step": 12634 }, { "epoch": 1.6073018699910953, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.848968505859375, "learning_rate": 5e-05, "loss": 0.7095, "mean_token_accuracy": 0.8857160806655884, "num_tokens": 479645198.0, "step": 12635 }, { "epoch": 1.6074290802696858, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.332388877868652, "learning_rate": 5e-05, "loss": 0.7321, "mean_token_accuracy": 0.8688303828239441, "num_tokens": 479686319.0, "step": 12636 }, { "epoch": 1.6075562905482763, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.066757202148438, "learning_rate": 5e-05, "loss": 0.7153, "mean_token_accuracy": 0.8925223350524902, "num_tokens": 479720170.0, "step": 12637 }, { "epoch": 1.6076835008268668, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.039441108703613, "learning_rate": 5e-05, "loss": 0.6887, "mean_token_accuracy": 0.8758673667907715, "num_tokens": 479757263.0, "step": 12638 }, { "epoch": 1.6078107111054574, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.351417541503906, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.8704026341438293, "num_tokens": 479793050.0, "step": 12639 }, { "epoch": 1.607937921384048, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.816744804382324, "learning_rate": 5e-05, "loss": 0.7083, "mean_token_accuracy": 0.8676730990409851, "num_tokens": 479831705.0, "step": 12640 }, { "epoch": 1.6080651316626384, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.533852577209473, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.885330855846405, "num_tokens": 479863888.0, "step": 12641 }, { "epoch": 1.608192341941229, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.844803333282471, "learning_rate": 5e-05, "loss": 0.6764, "mean_token_accuracy": 0.8822394609451294, "num_tokens": 479902280.0, "step": 12642 }, { "epoch": 1.6083195522198195, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.425050735473633, "learning_rate": 5e-05, "loss": 0.8009, "mean_token_accuracy": 0.8746025562286377, "num_tokens": 479935649.0, "step": 12643 }, { "epoch": 1.60844676249841, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.136824131011963, "learning_rate": 5e-05, "loss": 0.7287, "mean_token_accuracy": 0.8687863349914551, "num_tokens": 479971834.0, "step": 12644 }, { "epoch": 1.6085739727770005, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.997918605804443, "learning_rate": 5e-05, "loss": 0.8308, "mean_token_accuracy": 0.8579697608947754, "num_tokens": 480011985.0, "step": 12645 }, { "epoch": 1.6087011830555908, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.587055683135986, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8589242100715637, "num_tokens": 480047833.0, "step": 12646 }, { "epoch": 1.6088283933341814, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.441455364227295, "learning_rate": 5e-05, "loss": 0.7505, "mean_token_accuracy": 0.866157054901123, "num_tokens": 480080774.0, "step": 12647 }, { "epoch": 1.608955603612772, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.010224342346191, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8728804588317871, "num_tokens": 480119114.0, "step": 12648 }, { "epoch": 1.6090828138913624, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.039926052093506, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.8656529784202576, "num_tokens": 480156624.0, "step": 12649 }, { "epoch": 1.609210024169953, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.266292572021484, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8740139007568359, "num_tokens": 480192573.0, "step": 12650 }, { "epoch": 1.6093372344485435, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.893845558166504, "learning_rate": 5e-05, "loss": 0.7437, "mean_token_accuracy": 0.8592308163642883, "num_tokens": 480232139.0, "step": 12651 }, { "epoch": 1.6094644447271338, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.33349609375, "learning_rate": 5e-05, "loss": 0.8176, "mean_token_accuracy": 0.8650825023651123, "num_tokens": 480269811.0, "step": 12652 }, { "epoch": 1.6095916550057243, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.949675559997559, "learning_rate": 5e-05, "loss": 0.7256, "mean_token_accuracy": 0.8643169403076172, "num_tokens": 480307785.0, "step": 12653 }, { "epoch": 1.6097188652843148, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.205982208251953, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8680415749549866, "num_tokens": 480347028.0, "step": 12654 }, { "epoch": 1.6098460755629054, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.145565032958984, "learning_rate": 5e-05, "loss": 0.686, "mean_token_accuracy": 0.8786609172821045, "num_tokens": 480383915.0, "step": 12655 }, { "epoch": 1.609973285841496, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.918366432189941, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8728969097137451, "num_tokens": 480424226.0, "step": 12656 }, { "epoch": 1.6101004961200864, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.444732666015625, "learning_rate": 5e-05, "loss": 0.6943, "mean_token_accuracy": 0.8814990520477295, "num_tokens": 480461920.0, "step": 12657 }, { "epoch": 1.610227706398677, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.586262226104736, "learning_rate": 5e-05, "loss": 0.7183, "mean_token_accuracy": 0.8787316083908081, "num_tokens": 480498509.0, "step": 12658 }, { "epoch": 1.6103549166772675, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.652647495269775, "learning_rate": 5e-05, "loss": 0.76, "mean_token_accuracy": 0.8623348474502563, "num_tokens": 480541912.0, "step": 12659 }, { "epoch": 1.610482126955858, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.429931640625, "learning_rate": 5e-05, "loss": 0.7059, "mean_token_accuracy": 0.8763371706008911, "num_tokens": 480576034.0, "step": 12660 }, { "epoch": 1.6106093372344485, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.665055274963379, "learning_rate": 5e-05, "loss": 0.7516, "mean_token_accuracy": 0.8646080493927002, "num_tokens": 480612265.0, "step": 12661 }, { "epoch": 1.610736547513039, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.488828182220459, "learning_rate": 5e-05, "loss": 0.7805, "mean_token_accuracy": 0.8548015356063843, "num_tokens": 480650667.0, "step": 12662 }, { "epoch": 1.6108637577916296, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.461387634277344, "learning_rate": 5e-05, "loss": 0.7304, "mean_token_accuracy": 0.8688052892684937, "num_tokens": 480688931.0, "step": 12663 }, { "epoch": 1.6109909680702201, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.776828289031982, "learning_rate": 5e-05, "loss": 0.7642, "mean_token_accuracy": 0.8627290725708008, "num_tokens": 480719525.0, "step": 12664 }, { "epoch": 1.6111181783488107, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 7.067305088043213, "learning_rate": 5e-05, "loss": 0.6524, "mean_token_accuracy": 0.8833947777748108, "num_tokens": 480755872.0, "step": 12665 }, { "epoch": 1.6112453886274012, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.28757095336914, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.873580276966095, "num_tokens": 480794994.0, "step": 12666 }, { "epoch": 1.6113725989059917, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 6.448929309844971, "learning_rate": 5e-05, "loss": 0.6376, "mean_token_accuracy": 0.8775620460510254, "num_tokens": 480828421.0, "step": 12667 }, { "epoch": 1.6114998091845822, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 9.14247989654541, "learning_rate": 5e-05, "loss": 0.818, "mean_token_accuracy": 0.8836603164672852, "num_tokens": 480868571.0, "step": 12668 }, { "epoch": 1.6116270194631728, "ewc_loss": 0.26171875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00020599365234375, "grad_norm": 5.694305419921875, "learning_rate": 5e-05, "loss": 0.594, "mean_token_accuracy": 0.8829102516174316, "num_tokens": 480910897.0, "step": 12669 }, { "epoch": 1.6117542297417633, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 10.450544357299805, "learning_rate": 5e-05, "loss": 1.0084, "mean_token_accuracy": 0.8736632466316223, "num_tokens": 480942243.0, "step": 12670 }, { "epoch": 1.6118814400203536, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 5.671065330505371, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.8602551221847534, "num_tokens": 480980510.0, "step": 12671 }, { "epoch": 1.6120086502989441, "ewc_loss": 0.88671875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 11.566926956176758, "learning_rate": 5e-05, "loss": 1.2444, "mean_token_accuracy": 0.8767770528793335, "num_tokens": 481016837.0, "step": 12672 }, { "epoch": 1.6121358605775347, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 8.206838607788086, "learning_rate": 5e-05, "loss": 1.0425, "mean_token_accuracy": 0.8710330724716187, "num_tokens": 481055862.0, "step": 12673 }, { "epoch": 1.6122630708561252, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.538450241088867, "learning_rate": 5e-05, "loss": 1.0913, "mean_token_accuracy": 0.8626927137374878, "num_tokens": 481094632.0, "step": 12674 }, { "epoch": 1.6123902811347157, "ewc_loss": 0.85546875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 9.602417945861816, "learning_rate": 5e-05, "loss": 1.2318, "mean_token_accuracy": 0.871147871017456, "num_tokens": 481138468.0, "step": 12675 }, { "epoch": 1.6125174914133062, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.352450370788574, "learning_rate": 5e-05, "loss": 0.8825, "mean_token_accuracy": 0.867321252822876, "num_tokens": 481182983.0, "step": 12676 }, { "epoch": 1.6126447016918966, "ewc_loss": 0.86328125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0008087158203125, "grad_norm": 10.880577087402344, "learning_rate": 5e-05, "loss": 1.2272, "mean_token_accuracy": 0.8719614148139954, "num_tokens": 481225805.0, "step": 12677 }, { "epoch": 1.612771911970487, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 7.750175952911377, "learning_rate": 5e-05, "loss": 0.9076, "mean_token_accuracy": 0.8750401735305786, "num_tokens": 481260532.0, "step": 12678 }, { "epoch": 1.6128991222490776, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 12.040631294250488, "learning_rate": 5e-05, "loss": 1.2841, "mean_token_accuracy": 0.8663495779037476, "num_tokens": 481300494.0, "step": 12679 }, { "epoch": 1.6130263325276681, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.50789737701416, "learning_rate": 5e-05, "loss": 1.1648, "mean_token_accuracy": 0.8858919739723206, "num_tokens": 481335244.0, "step": 12680 }, { "epoch": 1.6131535428062587, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 11.266913414001465, "learning_rate": 5e-05, "loss": 1.0637, "mean_token_accuracy": 0.8761351108551025, "num_tokens": 481370894.0, "step": 12681 }, { "epoch": 1.6132807530848492, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 9.376534461975098, "learning_rate": 5e-05, "loss": 1.1506, "mean_token_accuracy": 0.8657751679420471, "num_tokens": 481413570.0, "step": 12682 }, { "epoch": 1.6134079633634397, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.807052612304688, "learning_rate": 5e-05, "loss": 1.0724, "mean_token_accuracy": 0.8739496469497681, "num_tokens": 481461380.0, "step": 12683 }, { "epoch": 1.6135351736420303, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.480377197265625, "learning_rate": 5e-05, "loss": 1.0019, "mean_token_accuracy": 0.8695748448371887, "num_tokens": 481497738.0, "step": 12684 }, { "epoch": 1.6136623839206208, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.559965133666992, "learning_rate": 5e-05, "loss": 1.0312, "mean_token_accuracy": 0.8822066783905029, "num_tokens": 481541000.0, "step": 12685 }, { "epoch": 1.6137895941992113, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.902707099914551, "learning_rate": 5e-05, "loss": 0.8427, "mean_token_accuracy": 0.8907606601715088, "num_tokens": 481578012.0, "step": 12686 }, { "epoch": 1.6139168044778018, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 10.940305709838867, "learning_rate": 5e-05, "loss": 1.1332, "mean_token_accuracy": 0.863680362701416, "num_tokens": 481612996.0, "step": 12687 }, { "epoch": 1.6140440147563924, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.113226890563965, "learning_rate": 5e-05, "loss": 0.9439, "mean_token_accuracy": 0.8743169903755188, "num_tokens": 481645570.0, "step": 12688 }, { "epoch": 1.614171225034983, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 11.371818542480469, "learning_rate": 5e-05, "loss": 1.1367, "mean_token_accuracy": 0.8788821697235107, "num_tokens": 481687873.0, "step": 12689 }, { "epoch": 1.6142984353135734, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.070292472839355, "learning_rate": 5e-05, "loss": 1.083, "mean_token_accuracy": 0.8628150820732117, "num_tokens": 481728105.0, "step": 12690 }, { "epoch": 1.614425645592164, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 10.841850280761719, "learning_rate": 5e-05, "loss": 1.0595, "mean_token_accuracy": 0.8809782862663269, "num_tokens": 481763558.0, "step": 12691 }, { "epoch": 1.6145528558707545, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.385127067565918, "learning_rate": 5e-05, "loss": 1.093, "mean_token_accuracy": 0.8588319420814514, "num_tokens": 481801153.0, "step": 12692 }, { "epoch": 1.614680066149345, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.9421968460083, "learning_rate": 5e-05, "loss": 0.9624, "mean_token_accuracy": 0.8900743722915649, "num_tokens": 481830523.0, "step": 12693 }, { "epoch": 1.6148072764279355, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.056333541870117, "learning_rate": 5e-05, "loss": 0.9859, "mean_token_accuracy": 0.8696836233139038, "num_tokens": 481870814.0, "step": 12694 }, { "epoch": 1.6149344867065258, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.16652774810791, "learning_rate": 5e-05, "loss": 0.9555, "mean_token_accuracy": 0.8739213943481445, "num_tokens": 481912279.0, "step": 12695 }, { "epoch": 1.6150616969851164, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.453447341918945, "learning_rate": 5e-05, "loss": 0.9016, "mean_token_accuracy": 0.8693748712539673, "num_tokens": 481957506.0, "step": 12696 }, { "epoch": 1.615188907263707, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.958843231201172, "learning_rate": 5e-05, "loss": 0.8981, "mean_token_accuracy": 0.8796198964118958, "num_tokens": 481996718.0, "step": 12697 }, { "epoch": 1.6153161175422974, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.6717329025268555, "learning_rate": 5e-05, "loss": 0.8047, "mean_token_accuracy": 0.8714129328727722, "num_tokens": 482035891.0, "step": 12698 }, { "epoch": 1.615443327820888, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.324060440063477, "learning_rate": 5e-05, "loss": 0.9486, "mean_token_accuracy": 0.8758696913719177, "num_tokens": 482070381.0, "step": 12699 }, { "epoch": 1.6155705380994785, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.253585338592529, "learning_rate": 5e-05, "loss": 0.7699, "mean_token_accuracy": 0.8743869662284851, "num_tokens": 482114484.0, "step": 12700 }, { "epoch": 1.6156977483780688, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.797515869140625, "learning_rate": 5e-05, "loss": 1.0083, "mean_token_accuracy": 0.8721072673797607, "num_tokens": 482147978.0, "step": 12701 }, { "epoch": 1.6158249586566593, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.526533603668213, "learning_rate": 5e-05, "loss": 0.7931, "mean_token_accuracy": 0.8772727251052856, "num_tokens": 482193017.0, "step": 12702 }, { "epoch": 1.6159521689352498, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.97607421875, "learning_rate": 5e-05, "loss": 1.0947, "mean_token_accuracy": 0.8576678037643433, "num_tokens": 482230856.0, "step": 12703 }, { "epoch": 1.6160793792138404, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.286931037902832, "learning_rate": 5e-05, "loss": 0.8773, "mean_token_accuracy": 0.883507490158081, "num_tokens": 482267540.0, "step": 12704 }, { "epoch": 1.616206589492431, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.323664665222168, "learning_rate": 5e-05, "loss": 0.9434, "mean_token_accuracy": 0.8850022554397583, "num_tokens": 482302256.0, "step": 12705 }, { "epoch": 1.6163337997710214, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.609281539916992, "learning_rate": 5e-05, "loss": 0.8987, "mean_token_accuracy": 0.8791705965995789, "num_tokens": 482340706.0, "step": 12706 }, { "epoch": 1.616461010049612, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.30338191986084, "learning_rate": 5e-05, "loss": 0.8477, "mean_token_accuracy": 0.8816767334938049, "num_tokens": 482377313.0, "step": 12707 }, { "epoch": 1.6165882203282025, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.424944877624512, "learning_rate": 5e-05, "loss": 0.8286, "mean_token_accuracy": 0.884202778339386, "num_tokens": 482415510.0, "step": 12708 }, { "epoch": 1.616715430606793, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.874922275543213, "learning_rate": 5e-05, "loss": 0.8144, "mean_token_accuracy": 0.8722946643829346, "num_tokens": 482453568.0, "step": 12709 }, { "epoch": 1.6168426408853835, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.9232964515686035, "learning_rate": 5e-05, "loss": 0.7865, "mean_token_accuracy": 0.8732408881187439, "num_tokens": 482487024.0, "step": 12710 }, { "epoch": 1.616969851163974, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.990488052368164, "learning_rate": 5e-05, "loss": 0.7841, "mean_token_accuracy": 0.8813307285308838, "num_tokens": 482524328.0, "step": 12711 }, { "epoch": 1.6170970614425646, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.452215671539307, "learning_rate": 5e-05, "loss": 0.7154, "mean_token_accuracy": 0.8813869953155518, "num_tokens": 482557042.0, "step": 12712 }, { "epoch": 1.6172242717211551, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.22593879699707, "learning_rate": 5e-05, "loss": 0.856, "mean_token_accuracy": 0.8513087034225464, "num_tokens": 482594486.0, "step": 12713 }, { "epoch": 1.6173514819997457, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.070371150970459, "learning_rate": 5e-05, "loss": 0.7151, "mean_token_accuracy": 0.8722845911979675, "num_tokens": 482635335.0, "step": 12714 }, { "epoch": 1.6174786922783362, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.691071510314941, "learning_rate": 5e-05, "loss": 0.8243, "mean_token_accuracy": 0.8836633563041687, "num_tokens": 482678945.0, "step": 12715 }, { "epoch": 1.6176059025569267, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.631284713745117, "learning_rate": 5e-05, "loss": 0.6864, "mean_token_accuracy": 0.8732099533081055, "num_tokens": 482717864.0, "step": 12716 }, { "epoch": 1.6177331128355172, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.329184532165527, "learning_rate": 5e-05, "loss": 0.9421, "mean_token_accuracy": 0.8701497316360474, "num_tokens": 482756361.0, "step": 12717 }, { "epoch": 1.6178603231141078, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.711328506469727, "learning_rate": 5e-05, "loss": 0.7073, "mean_token_accuracy": 0.8780309557914734, "num_tokens": 482794658.0, "step": 12718 }, { "epoch": 1.6179875333926983, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.420270919799805, "learning_rate": 5e-05, "loss": 0.9906, "mean_token_accuracy": 0.8690476417541504, "num_tokens": 482829426.0, "step": 12719 }, { "epoch": 1.6181147436712886, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.5801568031311035, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.8779916167259216, "num_tokens": 482867891.0, "step": 12720 }, { "epoch": 1.6182419539498791, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.569235801696777, "learning_rate": 5e-05, "loss": 0.8824, "mean_token_accuracy": 0.8772069215774536, "num_tokens": 482902307.0, "step": 12721 }, { "epoch": 1.6183691642284697, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.271429061889648, "learning_rate": 5e-05, "loss": 0.8724, "mean_token_accuracy": 0.868499219417572, "num_tokens": 482939751.0, "step": 12722 }, { "epoch": 1.6184963745070602, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.516170024871826, "learning_rate": 5e-05, "loss": 0.8035, "mean_token_accuracy": 0.8652829527854919, "num_tokens": 482978899.0, "step": 12723 }, { "epoch": 1.6186235847856507, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.521292686462402, "learning_rate": 5e-05, "loss": 0.8327, "mean_token_accuracy": 0.8806189298629761, "num_tokens": 483011086.0, "step": 12724 }, { "epoch": 1.6187507950642412, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.127920627593994, "learning_rate": 5e-05, "loss": 0.687, "mean_token_accuracy": 0.8857659101486206, "num_tokens": 483054235.0, "step": 12725 }, { "epoch": 1.6188780053428315, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.608430862426758, "learning_rate": 5e-05, "loss": 0.829, "mean_token_accuracy": 0.8772148489952087, "num_tokens": 483086843.0, "step": 12726 }, { "epoch": 1.619005215621422, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.059382915496826, "learning_rate": 5e-05, "loss": 0.7042, "mean_token_accuracy": 0.8767862319946289, "num_tokens": 483119198.0, "step": 12727 }, { "epoch": 1.6191324259000126, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.601723670959473, "learning_rate": 5e-05, "loss": 0.8313, "mean_token_accuracy": 0.8817746043205261, "num_tokens": 483160835.0, "step": 12728 }, { "epoch": 1.6192596361786031, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.17941427230835, "learning_rate": 5e-05, "loss": 0.7649, "mean_token_accuracy": 0.8651406764984131, "num_tokens": 483194906.0, "step": 12729 }, { "epoch": 1.6193868464571937, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.413130760192871, "learning_rate": 5e-05, "loss": 0.86, "mean_token_accuracy": 0.866233229637146, "num_tokens": 483231731.0, "step": 12730 }, { "epoch": 1.6195140567357842, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.321018695831299, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.8742311000823975, "num_tokens": 483270058.0, "step": 12731 }, { "epoch": 1.6196412670143747, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.195524215698242, "learning_rate": 5e-05, "loss": 0.7736, "mean_token_accuracy": 0.886246383190155, "num_tokens": 483303600.0, "step": 12732 }, { "epoch": 1.6197684772929652, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.405306339263916, "learning_rate": 5e-05, "loss": 0.7351, "mean_token_accuracy": 0.8764095902442932, "num_tokens": 483339933.0, "step": 12733 }, { "epoch": 1.6198956875715558, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.03469467163086, "learning_rate": 5e-05, "loss": 0.7831, "mean_token_accuracy": 0.8800947666168213, "num_tokens": 483377153.0, "step": 12734 }, { "epoch": 1.6200228978501463, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.414540767669678, "learning_rate": 5e-05, "loss": 0.6977, "mean_token_accuracy": 0.8856656551361084, "num_tokens": 483415089.0, "step": 12735 }, { "epoch": 1.6201501081287368, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.00892448425293, "learning_rate": 5e-05, "loss": 0.7862, "mean_token_accuracy": 0.8731113076210022, "num_tokens": 483454023.0, "step": 12736 }, { "epoch": 1.6202773184073274, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.279211521148682, "learning_rate": 5e-05, "loss": 0.7204, "mean_token_accuracy": 0.8732216954231262, "num_tokens": 483493572.0, "step": 12737 }, { "epoch": 1.6204045286859179, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.138437271118164, "learning_rate": 5e-05, "loss": 0.8068, "mean_token_accuracy": 0.8711814880371094, "num_tokens": 483529817.0, "step": 12738 }, { "epoch": 1.6205317389645084, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.045356750488281, "learning_rate": 5e-05, "loss": 0.7298, "mean_token_accuracy": 0.8664737939834595, "num_tokens": 483570905.0, "step": 12739 }, { "epoch": 1.620658949243099, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.392255783081055, "learning_rate": 5e-05, "loss": 0.8319, "mean_token_accuracy": 0.8689950704574585, "num_tokens": 483618177.0, "step": 12740 }, { "epoch": 1.6207861595216895, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.799245357513428, "learning_rate": 5e-05, "loss": 0.6984, "mean_token_accuracy": 0.8774974346160889, "num_tokens": 483651227.0, "step": 12741 }, { "epoch": 1.62091336980028, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000545501708984375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.707258224487305, "learning_rate": 5e-05, "loss": 0.8298, "mean_token_accuracy": 0.8770716190338135, "num_tokens": 483685577.0, "step": 12742 }, { "epoch": 1.6210405800788705, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.593393802642822, "learning_rate": 5e-05, "loss": 0.7334, "mean_token_accuracy": 0.8600605726242065, "num_tokens": 483721842.0, "step": 12743 }, { "epoch": 1.6211677903574608, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 9.019644737243652, "learning_rate": 5e-05, "loss": 0.912, "mean_token_accuracy": 0.8657624125480652, "num_tokens": 483758646.0, "step": 12744 }, { "epoch": 1.6212950006360514, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.637333869934082, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8595765829086304, "num_tokens": 483789216.0, "step": 12745 }, { "epoch": 1.621422210914642, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.963663101196289, "learning_rate": 5e-05, "loss": 0.8852, "mean_token_accuracy": 0.8819255232810974, "num_tokens": 483832901.0, "step": 12746 }, { "epoch": 1.6215494211932324, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.125312805175781, "learning_rate": 5e-05, "loss": 0.7944, "mean_token_accuracy": 0.8673653602600098, "num_tokens": 483874331.0, "step": 12747 }, { "epoch": 1.621676631471823, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.370285034179688, "learning_rate": 5e-05, "loss": 0.872, "mean_token_accuracy": 0.876849889755249, "num_tokens": 483911608.0, "step": 12748 }, { "epoch": 1.6218038417504135, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.727710723876953, "learning_rate": 5e-05, "loss": 0.8263, "mean_token_accuracy": 0.8722901940345764, "num_tokens": 483946419.0, "step": 12749 }, { "epoch": 1.6219310520290038, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.6077423095703125, "learning_rate": 5e-05, "loss": 0.8125, "mean_token_accuracy": 0.8627269864082336, "num_tokens": 483989740.0, "step": 12750 }, { "epoch": 1.6220582623075943, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.129265785217285, "learning_rate": 5e-05, "loss": 0.8443, "mean_token_accuracy": 0.8640666604042053, "num_tokens": 484025672.0, "step": 12751 }, { "epoch": 1.6221854725861848, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.209536075592041, "learning_rate": 5e-05, "loss": 0.7355, "mean_token_accuracy": 0.8745281100273132, "num_tokens": 484060421.0, "step": 12752 }, { "epoch": 1.6223126828647754, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.21179485321045, "learning_rate": 5e-05, "loss": 0.8078, "mean_token_accuracy": 0.8747975826263428, "num_tokens": 484102238.0, "step": 12753 }, { "epoch": 1.622439893143366, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.1514201164245605, "learning_rate": 5e-05, "loss": 0.7056, "mean_token_accuracy": 0.8778964877128601, "num_tokens": 484138314.0, "step": 12754 }, { "epoch": 1.6225671034219564, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.177730560302734, "learning_rate": 5e-05, "loss": 0.8546, "mean_token_accuracy": 0.8544209003448486, "num_tokens": 484175541.0, "step": 12755 }, { "epoch": 1.622694313700547, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.122974872589111, "learning_rate": 5e-05, "loss": 0.6975, "mean_token_accuracy": 0.8864299058914185, "num_tokens": 484211276.0, "step": 12756 }, { "epoch": 1.6228215239791375, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.241022109985352, "learning_rate": 5e-05, "loss": 0.875, "mean_token_accuracy": 0.8539063930511475, "num_tokens": 484250192.0, "step": 12757 }, { "epoch": 1.622948734257728, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.005060195922852, "learning_rate": 5e-05, "loss": 0.6915, "mean_token_accuracy": 0.8796512484550476, "num_tokens": 484287330.0, "step": 12758 }, { "epoch": 1.6230759445363185, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.374249458312988, "learning_rate": 5e-05, "loss": 0.8382, "mean_token_accuracy": 0.8686748743057251, "num_tokens": 484319937.0, "step": 12759 }, { "epoch": 1.623203154814909, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.909985542297363, "learning_rate": 5e-05, "loss": 0.7218, "mean_token_accuracy": 0.8678334355354309, "num_tokens": 484353305.0, "step": 12760 }, { "epoch": 1.6233303650934996, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.494721412658691, "learning_rate": 5e-05, "loss": 0.8835, "mean_token_accuracy": 0.8590402007102966, "num_tokens": 484391970.0, "step": 12761 }, { "epoch": 1.6234575753720901, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.85702657699585, "learning_rate": 5e-05, "loss": 0.6908, "mean_token_accuracy": 0.8807830214500427, "num_tokens": 484433448.0, "step": 12762 }, { "epoch": 1.6235847856506807, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.527111053466797, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.8782728314399719, "num_tokens": 484469729.0, "step": 12763 }, { "epoch": 1.6237119959292712, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.918820858001709, "learning_rate": 5e-05, "loss": 0.7568, "mean_token_accuracy": 0.8596900105476379, "num_tokens": 484509474.0, "step": 12764 }, { "epoch": 1.6238392062078617, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.445992469787598, "learning_rate": 5e-05, "loss": 0.7963, "mean_token_accuracy": 0.8900868892669678, "num_tokens": 484548999.0, "step": 12765 }, { "epoch": 1.6239664164864522, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.0983428955078125, "learning_rate": 5e-05, "loss": 0.7896, "mean_token_accuracy": 0.8556441068649292, "num_tokens": 484580534.0, "step": 12766 }, { "epoch": 1.6240936267650428, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.212762832641602, "learning_rate": 5e-05, "loss": 0.7952, "mean_token_accuracy": 0.8857673406600952, "num_tokens": 484616183.0, "step": 12767 }, { "epoch": 1.6242208370436333, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.300455093383789, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8718130588531494, "num_tokens": 484658290.0, "step": 12768 }, { "epoch": 1.6243480473222236, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.001744270324707, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8832096457481384, "num_tokens": 484690226.0, "step": 12769 }, { "epoch": 1.6244752576008141, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.38433313369751, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8706141114234924, "num_tokens": 484731911.0, "step": 12770 }, { "epoch": 1.6246024678794047, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.911894798278809, "learning_rate": 5e-05, "loss": 0.7817, "mean_token_accuracy": 0.8746487498283386, "num_tokens": 484768648.0, "step": 12771 }, { "epoch": 1.6247296781579952, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.343369483947754, "learning_rate": 5e-05, "loss": 0.7774, "mean_token_accuracy": 0.8625056147575378, "num_tokens": 484807266.0, "step": 12772 }, { "epoch": 1.6248568884365857, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.937658786773682, "learning_rate": 5e-05, "loss": 0.8214, "mean_token_accuracy": 0.8623064160346985, "num_tokens": 484850877.0, "step": 12773 }, { "epoch": 1.6249840987151762, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.22944450378418, "learning_rate": 5e-05, "loss": 0.7127, "mean_token_accuracy": 0.8763058185577393, "num_tokens": 484891223.0, "step": 12774 }, { "epoch": 1.6251113089937665, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.08577823638916, "learning_rate": 5e-05, "loss": 0.7725, "mean_token_accuracy": 0.8758188486099243, "num_tokens": 484922081.0, "step": 12775 }, { "epoch": 1.625238519272357, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.0480780601501465, "learning_rate": 5e-05, "loss": 0.7494, "mean_token_accuracy": 0.8602020740509033, "num_tokens": 484959016.0, "step": 12776 }, { "epoch": 1.6253657295509476, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.28960132598877, "learning_rate": 5e-05, "loss": 0.8032, "mean_token_accuracy": 0.8717384338378906, "num_tokens": 484998493.0, "step": 12777 }, { "epoch": 1.6254929398295381, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.848124027252197, "learning_rate": 5e-05, "loss": 0.7119, "mean_token_accuracy": 0.872520387172699, "num_tokens": 485035164.0, "step": 12778 }, { "epoch": 1.6256201501081287, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.520683288574219, "learning_rate": 5e-05, "loss": 0.8124, "mean_token_accuracy": 0.8834367990493774, "num_tokens": 485076347.0, "step": 12779 }, { "epoch": 1.6257473603867192, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.737915515899658, "learning_rate": 5e-05, "loss": 0.6921, "mean_token_accuracy": 0.8750667572021484, "num_tokens": 485107940.0, "step": 12780 }, { "epoch": 1.6258745706653097, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.632160186767578, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.8704836964607239, "num_tokens": 485145440.0, "step": 12781 }, { "epoch": 1.6260017809439002, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.863437652587891, "learning_rate": 5e-05, "loss": 0.7056, "mean_token_accuracy": 0.8778564929962158, "num_tokens": 485182948.0, "step": 12782 }, { "epoch": 1.6261289912224908, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.418749809265137, "learning_rate": 5e-05, "loss": 0.8278, "mean_token_accuracy": 0.8800057172775269, "num_tokens": 485220397.0, "step": 12783 }, { "epoch": 1.6262562015010813, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.255817413330078, "learning_rate": 5e-05, "loss": 0.7585, "mean_token_accuracy": 0.8738548755645752, "num_tokens": 485265382.0, "step": 12784 }, { "epoch": 1.6263834117796718, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.843769073486328, "learning_rate": 5e-05, "loss": 0.7973, "mean_token_accuracy": 0.8727700710296631, "num_tokens": 485306558.0, "step": 12785 }, { "epoch": 1.6265106220582624, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.84132719039917, "learning_rate": 5e-05, "loss": 0.82, "mean_token_accuracy": 0.8686820268630981, "num_tokens": 485341278.0, "step": 12786 }, { "epoch": 1.6266378323368529, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.091768264770508, "learning_rate": 5e-05, "loss": 0.6956, "mean_token_accuracy": 0.877954363822937, "num_tokens": 485376803.0, "step": 12787 }, { "epoch": 1.6267650426154434, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.491579055786133, "learning_rate": 5e-05, "loss": 0.8637, "mean_token_accuracy": 0.8607386350631714, "num_tokens": 485411168.0, "step": 12788 }, { "epoch": 1.626892252894034, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.429360389709473, "learning_rate": 5e-05, "loss": 0.7112, "mean_token_accuracy": 0.8643411993980408, "num_tokens": 485456320.0, "step": 12789 }, { "epoch": 1.6270194631726245, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 9.21326732635498, "learning_rate": 5e-05, "loss": 0.9119, "mean_token_accuracy": 0.8645961284637451, "num_tokens": 485491894.0, "step": 12790 }, { "epoch": 1.627146673451215, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.007828235626221, "learning_rate": 5e-05, "loss": 0.7312, "mean_token_accuracy": 0.8602519035339355, "num_tokens": 485526232.0, "step": 12791 }, { "epoch": 1.6272738837298055, "ewc_loss": 0.625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.884953498840332, "learning_rate": 5e-05, "loss": 1.0006, "mean_token_accuracy": 0.8684405088424683, "num_tokens": 485564752.0, "step": 12792 }, { "epoch": 1.6274010940083958, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.384095668792725, "learning_rate": 5e-05, "loss": 0.7262, "mean_token_accuracy": 0.8806508779525757, "num_tokens": 485600023.0, "step": 12793 }, { "epoch": 1.6275283042869864, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.839719772338867, "learning_rate": 5e-05, "loss": 1.0255, "mean_token_accuracy": 0.8740330934524536, "num_tokens": 485639195.0, "step": 12794 }, { "epoch": 1.6276555145655769, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 7.809960842132568, "learning_rate": 5e-05, "loss": 0.9485, "mean_token_accuracy": 0.8796283602714539, "num_tokens": 485676619.0, "step": 12795 }, { "epoch": 1.6277827248441674, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.32943344116211, "learning_rate": 5e-05, "loss": 0.9616, "mean_token_accuracy": 0.8585046529769897, "num_tokens": 485713739.0, "step": 12796 }, { "epoch": 1.627909935122758, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.05580997467041, "learning_rate": 5e-05, "loss": 1.015, "mean_token_accuracy": 0.8700680136680603, "num_tokens": 485750010.0, "step": 12797 }, { "epoch": 1.6280371454013485, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.63701057434082, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.8558589816093445, "num_tokens": 485790754.0, "step": 12798 }, { "epoch": 1.6281643556799388, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.111626625061035, "learning_rate": 5e-05, "loss": 1.0185, "mean_token_accuracy": 0.8767346143722534, "num_tokens": 485825472.0, "step": 12799 }, { "epoch": 1.6282915659585293, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 6.807180404663086, "learning_rate": 5e-05, "loss": 0.8237, "mean_token_accuracy": 0.8738200068473816, "num_tokens": 485866349.0, "step": 12800 }, { "epoch": 1.6284187762371198, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.006115913391113, "learning_rate": 5e-05, "loss": 1.0473, "mean_token_accuracy": 0.874183177947998, "num_tokens": 485908323.0, "step": 12801 }, { "epoch": 1.6285459865157104, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.171319961547852, "learning_rate": 5e-05, "loss": 0.9865, "mean_token_accuracy": 0.8818767070770264, "num_tokens": 485948637.0, "step": 12802 }, { "epoch": 1.628673196794301, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.556951522827148, "learning_rate": 5e-05, "loss": 0.932, "mean_token_accuracy": 0.8704150319099426, "num_tokens": 485988824.0, "step": 12803 }, { "epoch": 1.6288004070728914, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.87984848022461, "learning_rate": 5e-05, "loss": 1.0001, "mean_token_accuracy": 0.8756433725357056, "num_tokens": 486026364.0, "step": 12804 }, { "epoch": 1.628927617351482, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.515505790710449, "learning_rate": 5e-05, "loss": 0.8472, "mean_token_accuracy": 0.8672439455986023, "num_tokens": 486069578.0, "step": 12805 }, { "epoch": 1.6290548276300725, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.898409843444824, "learning_rate": 5e-05, "loss": 0.9839, "mean_token_accuracy": 0.8678609728813171, "num_tokens": 486105847.0, "step": 12806 }, { "epoch": 1.629182037908663, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.615634441375732, "learning_rate": 5e-05, "loss": 0.8382, "mean_token_accuracy": 0.884842038154602, "num_tokens": 486146680.0, "step": 12807 }, { "epoch": 1.6293092481872535, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.171021461486816, "learning_rate": 5e-05, "loss": 0.8509, "mean_token_accuracy": 0.8784105181694031, "num_tokens": 486184461.0, "step": 12808 }, { "epoch": 1.629436458465844, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.049304962158203, "learning_rate": 5e-05, "loss": 0.8725, "mean_token_accuracy": 0.8822693824768066, "num_tokens": 486214084.0, "step": 12809 }, { "epoch": 1.6295636687444346, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.4239606857299805, "learning_rate": 5e-05, "loss": 0.7842, "mean_token_accuracy": 0.8730238676071167, "num_tokens": 486250932.0, "step": 12810 }, { "epoch": 1.6296908790230251, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.406907081604004, "learning_rate": 5e-05, "loss": 0.8626, "mean_token_accuracy": 0.8805549144744873, "num_tokens": 486287964.0, "step": 12811 }, { "epoch": 1.6298180893016156, "ewc_loss": 0.375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.003034591674805, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8691025972366333, "num_tokens": 486321312.0, "step": 12812 }, { "epoch": 1.6299452995802062, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.756437301635742, "learning_rate": 5e-05, "loss": 0.9083, "mean_token_accuracy": 0.8728111386299133, "num_tokens": 486358992.0, "step": 12813 }, { "epoch": 1.6300725098587967, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.802032947540283, "learning_rate": 5e-05, "loss": 0.7972, "mean_token_accuracy": 0.860661506652832, "num_tokens": 486399632.0, "step": 12814 }, { "epoch": 1.6301997201373872, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.064508438110352, "learning_rate": 5e-05, "loss": 0.9345, "mean_token_accuracy": 0.8725982308387756, "num_tokens": 486438129.0, "step": 12815 }, { "epoch": 1.6303269304159778, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 6.796438217163086, "learning_rate": 5e-05, "loss": 0.7773, "mean_token_accuracy": 0.8707274198532104, "num_tokens": 486476167.0, "step": 12816 }, { "epoch": 1.630454140694568, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.142995834350586, "learning_rate": 5e-05, "loss": 0.922, "mean_token_accuracy": 0.8784473538398743, "num_tokens": 486512065.0, "step": 12817 }, { "epoch": 1.6305813509731586, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.00890588760376, "learning_rate": 5e-05, "loss": 0.874, "mean_token_accuracy": 0.8577334880828857, "num_tokens": 486555723.0, "step": 12818 }, { "epoch": 1.6307085612517491, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.010642051696777, "learning_rate": 5e-05, "loss": 0.9938, "mean_token_accuracy": 0.8603370189666748, "num_tokens": 486598598.0, "step": 12819 }, { "epoch": 1.6308357715303397, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.19197416305542, "learning_rate": 5e-05, "loss": 0.8497, "mean_token_accuracy": 0.8699887990951538, "num_tokens": 486633426.0, "step": 12820 }, { "epoch": 1.6309629818089302, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.879256248474121, "learning_rate": 5e-05, "loss": 0.9367, "mean_token_accuracy": 0.8753803968429565, "num_tokens": 486675584.0, "step": 12821 }, { "epoch": 1.6310901920875207, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.198901176452637, "learning_rate": 5e-05, "loss": 0.8299, "mean_token_accuracy": 0.8762431144714355, "num_tokens": 486722083.0, "step": 12822 }, { "epoch": 1.6312174023661112, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.904081344604492, "learning_rate": 5e-05, "loss": 0.9178, "mean_token_accuracy": 0.8824182748794556, "num_tokens": 486762658.0, "step": 12823 }, { "epoch": 1.6313446126447015, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.1848368644714355, "learning_rate": 5e-05, "loss": 0.8456, "mean_token_accuracy": 0.8691583275794983, "num_tokens": 486799441.0, "step": 12824 }, { "epoch": 1.631471822923292, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.900350570678711, "learning_rate": 5e-05, "loss": 0.9819, "mean_token_accuracy": 0.8623615503311157, "num_tokens": 486836742.0, "step": 12825 }, { "epoch": 1.6315990332018826, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.337717056274414, "learning_rate": 5e-05, "loss": 0.9274, "mean_token_accuracy": 0.8472275733947754, "num_tokens": 486873460.0, "step": 12826 }, { "epoch": 1.6317262434804731, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.664286613464355, "learning_rate": 5e-05, "loss": 0.931, "mean_token_accuracy": 0.8733394145965576, "num_tokens": 486909411.0, "step": 12827 }, { "epoch": 1.6318534537590637, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.567637920379639, "learning_rate": 5e-05, "loss": 0.8836, "mean_token_accuracy": 0.8635728359222412, "num_tokens": 486948427.0, "step": 12828 }, { "epoch": 1.6319806640376542, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.246809005737305, "learning_rate": 5e-05, "loss": 0.961, "mean_token_accuracy": 0.8557990789413452, "num_tokens": 486984594.0, "step": 12829 }, { "epoch": 1.6321078743162447, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.715317726135254, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.8786481022834778, "num_tokens": 487028122.0, "step": 12830 }, { "epoch": 1.6322350845948352, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.9040117263793945, "learning_rate": 5e-05, "loss": 0.8736, "mean_token_accuracy": 0.8630750775337219, "num_tokens": 487068731.0, "step": 12831 }, { "epoch": 1.6323622948734258, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.699807643890381, "learning_rate": 5e-05, "loss": 0.8259, "mean_token_accuracy": 0.8752156496047974, "num_tokens": 487103993.0, "step": 12832 }, { "epoch": 1.6324895051520163, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.812710285186768, "learning_rate": 5e-05, "loss": 0.8433, "mean_token_accuracy": 0.8713526129722595, "num_tokens": 487139206.0, "step": 12833 }, { "epoch": 1.6326167154306068, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.527071475982666, "learning_rate": 5e-05, "loss": 0.8887, "mean_token_accuracy": 0.842391312122345, "num_tokens": 487181553.0, "step": 12834 }, { "epoch": 1.6327439257091974, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.905738830566406, "learning_rate": 5e-05, "loss": 0.825, "mean_token_accuracy": 0.8666785359382629, "num_tokens": 487218931.0, "step": 12835 }, { "epoch": 1.6328711359877879, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.353308200836182, "learning_rate": 5e-05, "loss": 0.7754, "mean_token_accuracy": 0.865576982498169, "num_tokens": 487252829.0, "step": 12836 }, { "epoch": 1.6329983462663784, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.980612277984619, "learning_rate": 5e-05, "loss": 0.7976, "mean_token_accuracy": 0.8784437775611877, "num_tokens": 487292098.0, "step": 12837 }, { "epoch": 1.633125556544969, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.258822441101074, "learning_rate": 5e-05, "loss": 0.7491, "mean_token_accuracy": 0.8726825714111328, "num_tokens": 487330639.0, "step": 12838 }, { "epoch": 1.6332527668235595, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.042181968688965, "learning_rate": 5e-05, "loss": 0.7827, "mean_token_accuracy": 0.8800587058067322, "num_tokens": 487368499.0, "step": 12839 }, { "epoch": 1.63337997710215, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.190643787384033, "learning_rate": 5e-05, "loss": 0.7034, "mean_token_accuracy": 0.8821208477020264, "num_tokens": 487406626.0, "step": 12840 }, { "epoch": 1.6335071873807405, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.087467193603516, "learning_rate": 5e-05, "loss": 0.7959, "mean_token_accuracy": 0.8734025955200195, "num_tokens": 487441055.0, "step": 12841 }, { "epoch": 1.6336343976593308, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.124013900756836, "learning_rate": 5e-05, "loss": 0.6648, "mean_token_accuracy": 0.8937371373176575, "num_tokens": 487475973.0, "step": 12842 }, { "epoch": 1.6337616079379214, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.14937686920166, "learning_rate": 5e-05, "loss": 0.7974, "mean_token_accuracy": 0.8757752180099487, "num_tokens": 487513626.0, "step": 12843 }, { "epoch": 1.6338888182165119, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.059037208557129, "learning_rate": 5e-05, "loss": 0.7067, "mean_token_accuracy": 0.8803170919418335, "num_tokens": 487550053.0, "step": 12844 }, { "epoch": 1.6340160284951024, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.215784072875977, "learning_rate": 5e-05, "loss": 0.7889, "mean_token_accuracy": 0.8799076080322266, "num_tokens": 487586999.0, "step": 12845 }, { "epoch": 1.634143238773693, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.029234886169434, "learning_rate": 5e-05, "loss": 0.683, "mean_token_accuracy": 0.8832132816314697, "num_tokens": 487631433.0, "step": 12846 }, { "epoch": 1.6342704490522835, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.254229545593262, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.8741772174835205, "num_tokens": 487663259.0, "step": 12847 }, { "epoch": 1.6343976593308738, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.055665969848633, "learning_rate": 5e-05, "loss": 0.7539, "mean_token_accuracy": 0.8655295372009277, "num_tokens": 487697616.0, "step": 12848 }, { "epoch": 1.6345248696094643, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.15649127960205, "learning_rate": 5e-05, "loss": 0.7963, "mean_token_accuracy": 0.8818350434303284, "num_tokens": 487735693.0, "step": 12849 }, { "epoch": 1.6346520798880548, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.199823379516602, "learning_rate": 5e-05, "loss": 0.7341, "mean_token_accuracy": 0.874720573425293, "num_tokens": 487775519.0, "step": 12850 }, { "epoch": 1.6347792901666454, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.973718643188477, "learning_rate": 5e-05, "loss": 0.8221, "mean_token_accuracy": 0.8641088008880615, "num_tokens": 487814058.0, "step": 12851 }, { "epoch": 1.6349065004452359, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.378738880157471, "learning_rate": 5e-05, "loss": 0.7151, "mean_token_accuracy": 0.8817958235740662, "num_tokens": 487850128.0, "step": 12852 }, { "epoch": 1.6350337107238264, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.689396381378174, "learning_rate": 5e-05, "loss": 0.7235, "mean_token_accuracy": 0.885801374912262, "num_tokens": 487889465.0, "step": 12853 }, { "epoch": 1.635160921002417, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.628857135772705, "learning_rate": 5e-05, "loss": 0.7687, "mean_token_accuracy": 0.8671607971191406, "num_tokens": 487926264.0, "step": 12854 }, { "epoch": 1.6352881312810075, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.370264530181885, "learning_rate": 5e-05, "loss": 0.7289, "mean_token_accuracy": 0.8702052235603333, "num_tokens": 487959872.0, "step": 12855 }, { "epoch": 1.635415341559598, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.895619869232178, "learning_rate": 5e-05, "loss": 0.7998, "mean_token_accuracy": 0.8631411790847778, "num_tokens": 487997554.0, "step": 12856 }, { "epoch": 1.6355425518381885, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.029417991638184, "learning_rate": 5e-05, "loss": 0.6608, "mean_token_accuracy": 0.8861695528030396, "num_tokens": 488029172.0, "step": 12857 }, { "epoch": 1.635669762116779, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.260231971740723, "learning_rate": 5e-05, "loss": 0.8579, "mean_token_accuracy": 0.85393226146698, "num_tokens": 488066651.0, "step": 12858 }, { "epoch": 1.6357969723953696, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.635530948638916, "learning_rate": 5e-05, "loss": 0.7191, "mean_token_accuracy": 0.8620365858078003, "num_tokens": 488102207.0, "step": 12859 }, { "epoch": 1.6359241826739601, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.753921508789062, "learning_rate": 5e-05, "loss": 0.8083, "mean_token_accuracy": 0.882760763168335, "num_tokens": 488142502.0, "step": 12860 }, { "epoch": 1.6360513929525506, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.197114944458008, "learning_rate": 5e-05, "loss": 0.7136, "mean_token_accuracy": 0.8594374656677246, "num_tokens": 488182334.0, "step": 12861 }, { "epoch": 1.6361786032311412, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.44372272491455, "learning_rate": 5e-05, "loss": 0.9183, "mean_token_accuracy": 0.8749136328697205, "num_tokens": 488219434.0, "step": 12862 }, { "epoch": 1.6363058135097317, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 5.96572732925415, "learning_rate": 5e-05, "loss": 0.6705, "mean_token_accuracy": 0.8748242259025574, "num_tokens": 488256961.0, "step": 12863 }, { "epoch": 1.6364330237883222, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.981752395629883, "learning_rate": 5e-05, "loss": 1.0268, "mean_token_accuracy": 0.8652865886688232, "num_tokens": 488297069.0, "step": 12864 }, { "epoch": 1.6365602340669128, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 6.8123674392700195, "learning_rate": 5e-05, "loss": 0.8764, "mean_token_accuracy": 0.8560472130775452, "num_tokens": 488336436.0, "step": 12865 }, { "epoch": 1.636687444345503, "ewc_loss": 0.625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.35655403137207, "learning_rate": 5e-05, "loss": 0.9963, "mean_token_accuracy": 0.8670710921287537, "num_tokens": 488376145.0, "step": 12866 }, { "epoch": 1.6368146546240936, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.296605110168457, "learning_rate": 5e-05, "loss": 0.9836, "mean_token_accuracy": 0.8738301992416382, "num_tokens": 488418319.0, "step": 12867 }, { "epoch": 1.6369418649026841, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.370232105255127, "learning_rate": 5e-05, "loss": 0.7939, "mean_token_accuracy": 0.8721289038658142, "num_tokens": 488460377.0, "step": 12868 }, { "epoch": 1.6370690751812746, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.744105339050293, "learning_rate": 5e-05, "loss": 1.0622, "mean_token_accuracy": 0.8648726940155029, "num_tokens": 488491794.0, "step": 12869 }, { "epoch": 1.6371962854598652, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.145750045776367, "learning_rate": 5e-05, "loss": 0.7453, "mean_token_accuracy": 0.8642538785934448, "num_tokens": 488534106.0, "step": 12870 }, { "epoch": 1.6373234957384557, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00075531005859375, "grad_norm": 11.073783874511719, "learning_rate": 5e-05, "loss": 1.202, "mean_token_accuracy": 0.8648786544799805, "num_tokens": 488568198.0, "step": 12871 }, { "epoch": 1.6374507060170462, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 7.625831127166748, "learning_rate": 5e-05, "loss": 1.0037, "mean_token_accuracy": 0.8659614324569702, "num_tokens": 488606112.0, "step": 12872 }, { "epoch": 1.6375779162956365, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 10.738901138305664, "learning_rate": 5e-05, "loss": 1.1212, "mean_token_accuracy": 0.8695849776268005, "num_tokens": 488639585.0, "step": 12873 }, { "epoch": 1.637705126574227, "ewc_loss": 0.828125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000774383544921875, "grad_norm": 9.108389854431152, "learning_rate": 5e-05, "loss": 1.1505, "mean_token_accuracy": 0.8854178190231323, "num_tokens": 488670789.0, "step": 12874 }, { "epoch": 1.6378323368528176, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.229971885681152, "learning_rate": 5e-05, "loss": 0.9695, "mean_token_accuracy": 0.8711860775947571, "num_tokens": 488707837.0, "step": 12875 }, { "epoch": 1.6379595471314081, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.159393310546875, "learning_rate": 5e-05, "loss": 1.0608, "mean_token_accuracy": 0.8689148426055908, "num_tokens": 488747719.0, "step": 12876 }, { "epoch": 1.6380867574099987, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.350802421569824, "learning_rate": 5e-05, "loss": 0.9519, "mean_token_accuracy": 0.8672339916229248, "num_tokens": 488786493.0, "step": 12877 }, { "epoch": 1.6382139676885892, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.620288848876953, "learning_rate": 5e-05, "loss": 0.9497, "mean_token_accuracy": 0.866819441318512, "num_tokens": 488826390.0, "step": 12878 }, { "epoch": 1.6383411779671797, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.413558959960938, "learning_rate": 5e-05, "loss": 0.9592, "mean_token_accuracy": 0.8710184097290039, "num_tokens": 488861458.0, "step": 12879 }, { "epoch": 1.6384683882457702, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.569712162017822, "learning_rate": 5e-05, "loss": 0.8011, "mean_token_accuracy": 0.8728653192520142, "num_tokens": 488902762.0, "step": 12880 }, { "epoch": 1.6385955985243608, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.956915855407715, "learning_rate": 5e-05, "loss": 0.9148, "mean_token_accuracy": 0.883551836013794, "num_tokens": 488938139.0, "step": 12881 }, { "epoch": 1.6387228088029513, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.961919784545898, "learning_rate": 5e-05, "loss": 0.7492, "mean_token_accuracy": 0.8750783801078796, "num_tokens": 488976036.0, "step": 12882 }, { "epoch": 1.6388500190815418, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.421566009521484, "learning_rate": 5e-05, "loss": 1.0137, "mean_token_accuracy": 0.8610666394233704, "num_tokens": 489014333.0, "step": 12883 }, { "epoch": 1.6389772293601323, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.030707359313965, "learning_rate": 5e-05, "loss": 0.8137, "mean_token_accuracy": 0.874451756477356, "num_tokens": 489050060.0, "step": 12884 }, { "epoch": 1.6391044396387229, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.506282806396484, "learning_rate": 5e-05, "loss": 1.0146, "mean_token_accuracy": 0.8727232217788696, "num_tokens": 489090053.0, "step": 12885 }, { "epoch": 1.6392316499173134, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 7.411351203918457, "learning_rate": 5e-05, "loss": 0.9121, "mean_token_accuracy": 0.8701533675193787, "num_tokens": 489129081.0, "step": 12886 }, { "epoch": 1.639358860195904, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.289518356323242, "learning_rate": 5e-05, "loss": 1.0269, "mean_token_accuracy": 0.863381564617157, "num_tokens": 489174811.0, "step": 12887 }, { "epoch": 1.6394860704744945, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.675981521606445, "learning_rate": 5e-05, "loss": 0.9318, "mean_token_accuracy": 0.8712151050567627, "num_tokens": 489214694.0, "step": 12888 }, { "epoch": 1.639613280753085, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.874847412109375, "learning_rate": 5e-05, "loss": 0.9542, "mean_token_accuracy": 0.8781890273094177, "num_tokens": 489255299.0, "step": 12889 }, { "epoch": 1.6397404910316755, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.7181572914123535, "learning_rate": 5e-05, "loss": 0.8397, "mean_token_accuracy": 0.8895308971405029, "num_tokens": 489292409.0, "step": 12890 }, { "epoch": 1.6398677013102658, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.602557182312012, "learning_rate": 5e-05, "loss": 0.9756, "mean_token_accuracy": 0.8637787699699402, "num_tokens": 489335354.0, "step": 12891 }, { "epoch": 1.6399949115888564, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.508598327636719, "learning_rate": 5e-05, "loss": 0.925, "mean_token_accuracy": 0.8502360582351685, "num_tokens": 489369253.0, "step": 12892 }, { "epoch": 1.6401221218674469, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.667237281799316, "learning_rate": 5e-05, "loss": 0.9452, "mean_token_accuracy": 0.8721928596496582, "num_tokens": 489408453.0, "step": 12893 }, { "epoch": 1.6402493321460374, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.286658763885498, "learning_rate": 5e-05, "loss": 0.775, "mean_token_accuracy": 0.8813576698303223, "num_tokens": 489441609.0, "step": 12894 }, { "epoch": 1.640376542424628, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.77572250366211, "learning_rate": 5e-05, "loss": 0.9589, "mean_token_accuracy": 0.8715329170227051, "num_tokens": 489477228.0, "step": 12895 }, { "epoch": 1.6405037527032185, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.295727252960205, "learning_rate": 5e-05, "loss": 0.8609, "mean_token_accuracy": 0.859283447265625, "num_tokens": 489518622.0, "step": 12896 }, { "epoch": 1.6406309629818088, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.746752738952637, "learning_rate": 5e-05, "loss": 0.9951, "mean_token_accuracy": 0.8609534502029419, "num_tokens": 489557707.0, "step": 12897 }, { "epoch": 1.6407581732603993, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.410664081573486, "learning_rate": 5e-05, "loss": 0.8542, "mean_token_accuracy": 0.8668512105941772, "num_tokens": 489591180.0, "step": 12898 }, { "epoch": 1.6408853835389898, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.602495193481445, "learning_rate": 5e-05, "loss": 0.9532, "mean_token_accuracy": 0.8711197376251221, "num_tokens": 489631732.0, "step": 12899 }, { "epoch": 1.6410125938175804, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.562289237976074, "learning_rate": 5e-05, "loss": 0.8679, "mean_token_accuracy": 0.8669895529747009, "num_tokens": 489671484.0, "step": 12900 }, { "epoch": 1.6411398040961709, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.27266788482666, "learning_rate": 5e-05, "loss": 0.9133, "mean_token_accuracy": 0.8725473880767822, "num_tokens": 489706929.0, "step": 12901 }, { "epoch": 1.6412670143747614, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.719447612762451, "learning_rate": 5e-05, "loss": 0.8695, "mean_token_accuracy": 0.8704241514205933, "num_tokens": 489741830.0, "step": 12902 }, { "epoch": 1.641394224653352, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.887785911560059, "learning_rate": 5e-05, "loss": 0.8162, "mean_token_accuracy": 0.8831346035003662, "num_tokens": 489782808.0, "step": 12903 }, { "epoch": 1.6415214349319425, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.709345817565918, "learning_rate": 5e-05, "loss": 0.8504, "mean_token_accuracy": 0.8646406531333923, "num_tokens": 489823536.0, "step": 12904 }, { "epoch": 1.641648645210533, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.824732303619385, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.8771663308143616, "num_tokens": 489854730.0, "step": 12905 }, { "epoch": 1.6417758554891235, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.514655590057373, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8737218976020813, "num_tokens": 489886836.0, "step": 12906 }, { "epoch": 1.641903065767714, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.932628154754639, "learning_rate": 5e-05, "loss": 0.8109, "mean_token_accuracy": 0.8734762668609619, "num_tokens": 489928227.0, "step": 12907 }, { "epoch": 1.6420302760463046, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.351080894470215, "learning_rate": 5e-05, "loss": 0.7884, "mean_token_accuracy": 0.8597041964530945, "num_tokens": 489965697.0, "step": 12908 }, { "epoch": 1.642157486324895, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.010357856750488, "learning_rate": 5e-05, "loss": 0.7829, "mean_token_accuracy": 0.8778373003005981, "num_tokens": 490000344.0, "step": 12909 }, { "epoch": 1.6422846966034856, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.247176647186279, "learning_rate": 5e-05, "loss": 0.6854, "mean_token_accuracy": 0.893450140953064, "num_tokens": 490030965.0, "step": 12910 }, { "epoch": 1.6424119068820762, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.081770896911621, "learning_rate": 5e-05, "loss": 0.7936, "mean_token_accuracy": 0.8789653182029724, "num_tokens": 490069386.0, "step": 12911 }, { "epoch": 1.6425391171606667, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.133541107177734, "learning_rate": 5e-05, "loss": 0.7037, "mean_token_accuracy": 0.8797422051429749, "num_tokens": 490104040.0, "step": 12912 }, { "epoch": 1.6426663274392572, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.17767333984375, "learning_rate": 5e-05, "loss": 0.8452, "mean_token_accuracy": 0.8626338243484497, "num_tokens": 490147883.0, "step": 12913 }, { "epoch": 1.6427935377178478, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.086820125579834, "learning_rate": 5e-05, "loss": 0.7235, "mean_token_accuracy": 0.8736419081687927, "num_tokens": 490183889.0, "step": 12914 }, { "epoch": 1.642920747996438, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.194863319396973, "learning_rate": 5e-05, "loss": 0.8043, "mean_token_accuracy": 0.8759880065917969, "num_tokens": 490224132.0, "step": 12915 }, { "epoch": 1.6430479582750286, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.131701946258545, "learning_rate": 5e-05, "loss": 0.7352, "mean_token_accuracy": 0.8727275133132935, "num_tokens": 490262745.0, "step": 12916 }, { "epoch": 1.6431751685536191, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.128765106201172, "learning_rate": 5e-05, "loss": 0.8297, "mean_token_accuracy": 0.866599440574646, "num_tokens": 490301449.0, "step": 12917 }, { "epoch": 1.6433023788322096, "ewc_loss": 0.375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.221023082733154, "learning_rate": 5e-05, "loss": 0.7681, "mean_token_accuracy": 0.8594570159912109, "num_tokens": 490328873.0, "step": 12918 }, { "epoch": 1.6434295891108002, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.968400478363037, "learning_rate": 5e-05, "loss": 0.7869, "mean_token_accuracy": 0.8753857612609863, "num_tokens": 490361174.0, "step": 12919 }, { "epoch": 1.6435567993893907, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.386359214782715, "learning_rate": 5e-05, "loss": 0.7803, "mean_token_accuracy": 0.8646234273910522, "num_tokens": 490399708.0, "step": 12920 }, { "epoch": 1.6436840096679812, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.717514991760254, "learning_rate": 5e-05, "loss": 0.7306, "mean_token_accuracy": 0.8836117386817932, "num_tokens": 490433159.0, "step": 12921 }, { "epoch": 1.6438112199465715, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.593682765960693, "learning_rate": 5e-05, "loss": 0.7426, "mean_token_accuracy": 0.8737781643867493, "num_tokens": 490473335.0, "step": 12922 }, { "epoch": 1.643938430225162, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.429133892059326, "learning_rate": 5e-05, "loss": 0.7685, "mean_token_accuracy": 0.8641001582145691, "num_tokens": 490517091.0, "step": 12923 }, { "epoch": 1.6440656405037526, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.8141984939575195, "learning_rate": 5e-05, "loss": 0.7755, "mean_token_accuracy": 0.8688914775848389, "num_tokens": 490550336.0, "step": 12924 }, { "epoch": 1.6441928507823431, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.158626079559326, "learning_rate": 5e-05, "loss": 0.6644, "mean_token_accuracy": 0.8882846832275391, "num_tokens": 490593592.0, "step": 12925 }, { "epoch": 1.6443200610609336, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.065832138061523, "learning_rate": 5e-05, "loss": 0.7465, "mean_token_accuracy": 0.8817650079727173, "num_tokens": 490629354.0, "step": 12926 }, { "epoch": 1.6444472713395242, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.875221252441406, "learning_rate": 5e-05, "loss": 0.6264, "mean_token_accuracy": 0.8952792286872864, "num_tokens": 490668319.0, "step": 12927 }, { "epoch": 1.6445744816181147, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.400260925292969, "learning_rate": 5e-05, "loss": 0.7928, "mean_token_accuracy": 0.8812581300735474, "num_tokens": 490705978.0, "step": 12928 }, { "epoch": 1.6447016918967052, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.5268964767456055, "learning_rate": 5e-05, "loss": 0.6943, "mean_token_accuracy": 0.87281334400177, "num_tokens": 490742455.0, "step": 12929 }, { "epoch": 1.6448289021752958, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.903916358947754, "learning_rate": 5e-05, "loss": 0.906, "mean_token_accuracy": 0.8570565581321716, "num_tokens": 490781733.0, "step": 12930 }, { "epoch": 1.6449561124538863, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.176167964935303, "learning_rate": 5e-05, "loss": 0.6751, "mean_token_accuracy": 0.875307559967041, "num_tokens": 490821077.0, "step": 12931 }, { "epoch": 1.6450833227324768, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.471389770507812, "learning_rate": 5e-05, "loss": 0.9231, "mean_token_accuracy": 0.874954879283905, "num_tokens": 490862725.0, "step": 12932 }, { "epoch": 1.6452105330110673, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.224604606628418, "learning_rate": 5e-05, "loss": 0.7613, "mean_token_accuracy": 0.8581540584564209, "num_tokens": 490909612.0, "step": 12933 }, { "epoch": 1.6453377432896579, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.590934753417969, "learning_rate": 5e-05, "loss": 0.9706, "mean_token_accuracy": 0.8763399124145508, "num_tokens": 490943202.0, "step": 12934 }, { "epoch": 1.6454649535682484, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.220496654510498, "learning_rate": 5e-05, "loss": 0.8534, "mean_token_accuracy": 0.8715811967849731, "num_tokens": 490978481.0, "step": 12935 }, { "epoch": 1.645592163846839, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.438735008239746, "learning_rate": 5e-05, "loss": 0.8977, "mean_token_accuracy": 0.8751766085624695, "num_tokens": 491018672.0, "step": 12936 }, { "epoch": 1.6457193741254295, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.514545440673828, "learning_rate": 5e-05, "loss": 0.9396, "mean_token_accuracy": 0.8754817247390747, "num_tokens": 491055032.0, "step": 12937 }, { "epoch": 1.64584658440402, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.6671223640441895, "learning_rate": 5e-05, "loss": 0.7964, "mean_token_accuracy": 0.8475023508071899, "num_tokens": 491089780.0, "step": 12938 }, { "epoch": 1.6459737946826105, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.882949829101562, "learning_rate": 5e-05, "loss": 0.9843, "mean_token_accuracy": 0.8769159913063049, "num_tokens": 491124187.0, "step": 12939 }, { "epoch": 1.6461010049612008, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.0258989334106445, "learning_rate": 5e-05, "loss": 0.7145, "mean_token_accuracy": 0.8690994381904602, "num_tokens": 491167558.0, "step": 12940 }, { "epoch": 1.6462282152397913, "ewc_loss": 0.78125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.853813171386719, "learning_rate": 5e-05, "loss": 1.1751, "mean_token_accuracy": 0.8675891160964966, "num_tokens": 491210054.0, "step": 12941 }, { "epoch": 1.6463554255183819, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 7.672096252441406, "learning_rate": 5e-05, "loss": 0.9716, "mean_token_accuracy": 0.8783895969390869, "num_tokens": 491256535.0, "step": 12942 }, { "epoch": 1.6464826357969724, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.129495620727539, "learning_rate": 5e-05, "loss": 1.0259, "mean_token_accuracy": 0.8765116930007935, "num_tokens": 491295909.0, "step": 12943 }, { "epoch": 1.646609846075563, "ewc_loss": 0.765625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 9.01169204711914, "learning_rate": 5e-05, "loss": 1.1618, "mean_token_accuracy": 0.8620142936706543, "num_tokens": 491332547.0, "step": 12944 }, { "epoch": 1.6467370563541535, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.434497833251953, "learning_rate": 5e-05, "loss": 0.8576, "mean_token_accuracy": 0.8874062299728394, "num_tokens": 491373187.0, "step": 12945 }, { "epoch": 1.6468642666327438, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.338664054870605, "learning_rate": 5e-05, "loss": 1.0345, "mean_token_accuracy": 0.8726929426193237, "num_tokens": 491412476.0, "step": 12946 }, { "epoch": 1.6469914769113343, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.799703121185303, "learning_rate": 5e-05, "loss": 0.8856, "mean_token_accuracy": 0.8785843849182129, "num_tokens": 491456502.0, "step": 12947 }, { "epoch": 1.6471186871899248, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.830033302307129, "learning_rate": 5e-05, "loss": 0.93, "mean_token_accuracy": 0.8769164085388184, "num_tokens": 491492974.0, "step": 12948 }, { "epoch": 1.6472458974685154, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.2061185836792, "learning_rate": 5e-05, "loss": 0.9405, "mean_token_accuracy": 0.872873067855835, "num_tokens": 491523621.0, "step": 12949 }, { "epoch": 1.6473731077471059, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.596368312835693, "learning_rate": 5e-05, "loss": 0.8303, "mean_token_accuracy": 0.8615599870681763, "num_tokens": 491563366.0, "step": 12950 }, { "epoch": 1.6475003180256964, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.881709098815918, "learning_rate": 5e-05, "loss": 0.9609, "mean_token_accuracy": 0.8725626468658447, "num_tokens": 491594221.0, "step": 12951 }, { "epoch": 1.647627528304287, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.96419095993042, "learning_rate": 5e-05, "loss": 0.7822, "mean_token_accuracy": 0.8644731044769287, "num_tokens": 491632039.0, "step": 12952 }, { "epoch": 1.6477547385828775, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.229887962341309, "learning_rate": 5e-05, "loss": 0.9526, "mean_token_accuracy": 0.8787038922309875, "num_tokens": 491672282.0, "step": 12953 }, { "epoch": 1.647881948861468, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.084460258483887, "learning_rate": 5e-05, "loss": 0.8557, "mean_token_accuracy": 0.8583057522773743, "num_tokens": 491707969.0, "step": 12954 }, { "epoch": 1.6480091591400585, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.122867584228516, "learning_rate": 5e-05, "loss": 0.9297, "mean_token_accuracy": 0.8773066997528076, "num_tokens": 491743347.0, "step": 12955 }, { "epoch": 1.648136369418649, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.372739315032959, "learning_rate": 5e-05, "loss": 0.8839, "mean_token_accuracy": 0.8697026371955872, "num_tokens": 491777165.0, "step": 12956 }, { "epoch": 1.6482635796972396, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.798527717590332, "learning_rate": 5e-05, "loss": 0.8812, "mean_token_accuracy": 0.8864193558692932, "num_tokens": 491813859.0, "step": 12957 }, { "epoch": 1.64839078997583, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.4528961181640625, "learning_rate": 5e-05, "loss": 0.8829, "mean_token_accuracy": 0.8675075769424438, "num_tokens": 491853373.0, "step": 12958 }, { "epoch": 1.6485180002544206, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.51271915435791, "learning_rate": 5e-05, "loss": 0.8876, "mean_token_accuracy": 0.8768804669380188, "num_tokens": 491889401.0, "step": 12959 }, { "epoch": 1.6486452105330112, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.339045524597168, "learning_rate": 5e-05, "loss": 0.8336, "mean_token_accuracy": 0.8697057962417603, "num_tokens": 491933973.0, "step": 12960 }, { "epoch": 1.6487724208116017, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.502934455871582, "learning_rate": 5e-05, "loss": 0.9259, "mean_token_accuracy": 0.8624056577682495, "num_tokens": 491979198.0, "step": 12961 }, { "epoch": 1.6488996310901922, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.080379962921143, "learning_rate": 5e-05, "loss": 0.8165, "mean_token_accuracy": 0.8619611263275146, "num_tokens": 492009127.0, "step": 12962 }, { "epoch": 1.6490268413687827, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.781295776367188, "learning_rate": 5e-05, "loss": 0.9151, "mean_token_accuracy": 0.8743104934692383, "num_tokens": 492045022.0, "step": 12963 }, { "epoch": 1.649154051647373, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.8946099281311035, "learning_rate": 5e-05, "loss": 0.7618, "mean_token_accuracy": 0.8701884746551514, "num_tokens": 492092155.0, "step": 12964 }, { "epoch": 1.6492812619259636, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.00969409942627, "learning_rate": 5e-05, "loss": 1.0051, "mean_token_accuracy": 0.8554991483688354, "num_tokens": 492131807.0, "step": 12965 }, { "epoch": 1.649408472204554, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.066983222961426, "learning_rate": 5e-05, "loss": 0.7873, "mean_token_accuracy": 0.8729476928710938, "num_tokens": 492166794.0, "step": 12966 }, { "epoch": 1.6495356824831446, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.863462448120117, "learning_rate": 5e-05, "loss": 0.9526, "mean_token_accuracy": 0.8776812553405762, "num_tokens": 492208375.0, "step": 12967 }, { "epoch": 1.6496628927617352, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.463957786560059, "learning_rate": 5e-05, "loss": 0.8979, "mean_token_accuracy": 0.8606225252151489, "num_tokens": 492241166.0, "step": 12968 }, { "epoch": 1.6497901030403257, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.398754119873047, "learning_rate": 5e-05, "loss": 0.8742, "mean_token_accuracy": 0.8863949179649353, "num_tokens": 492278887.0, "step": 12969 }, { "epoch": 1.6499173133189162, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.759728908538818, "learning_rate": 5e-05, "loss": 0.8186, "mean_token_accuracy": 0.8865036368370056, "num_tokens": 492316523.0, "step": 12970 }, { "epoch": 1.6500445235975065, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.853785991668701, "learning_rate": 5e-05, "loss": 0.8634, "mean_token_accuracy": 0.8661791086196899, "num_tokens": 492354094.0, "step": 12971 }, { "epoch": 1.650171733876097, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.907738208770752, "learning_rate": 5e-05, "loss": 0.8742, "mean_token_accuracy": 0.8653762340545654, "num_tokens": 492387250.0, "step": 12972 }, { "epoch": 1.6502989441546876, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.483270645141602, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.8689227104187012, "num_tokens": 492423770.0, "step": 12973 }, { "epoch": 1.6504261544332781, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.935543537139893, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.8711041212081909, "num_tokens": 492465970.0, "step": 12974 }, { "epoch": 1.6505533647118686, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.392550945281982, "learning_rate": 5e-05, "loss": 0.7851, "mean_token_accuracy": 0.8662229776382446, "num_tokens": 492501668.0, "step": 12975 }, { "epoch": 1.6506805749904592, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.848527431488037, "learning_rate": 5e-05, "loss": 0.7984, "mean_token_accuracy": 0.8679554462432861, "num_tokens": 492541032.0, "step": 12976 }, { "epoch": 1.6508077852690497, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.432974338531494, "learning_rate": 5e-05, "loss": 0.7693, "mean_token_accuracy": 0.8683038353919983, "num_tokens": 492579639.0, "step": 12977 }, { "epoch": 1.6509349955476402, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.762263298034668, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.8702285289764404, "num_tokens": 492612324.0, "step": 12978 }, { "epoch": 1.6510622058262308, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.417377471923828, "learning_rate": 5e-05, "loss": 0.7744, "mean_token_accuracy": 0.8641676902770996, "num_tokens": 492648967.0, "step": 12979 }, { "epoch": 1.6511894161048213, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.7052178382873535, "learning_rate": 5e-05, "loss": 0.7301, "mean_token_accuracy": 0.8849149942398071, "num_tokens": 492687117.0, "step": 12980 }, { "epoch": 1.6513166263834118, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.43236780166626, "learning_rate": 5e-05, "loss": 0.7586, "mean_token_accuracy": 0.8727027177810669, "num_tokens": 492718760.0, "step": 12981 }, { "epoch": 1.6514438366620023, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.676554203033447, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8753775954246521, "num_tokens": 492760038.0, "step": 12982 }, { "epoch": 1.6515710469405929, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.425145149230957, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8622235059738159, "num_tokens": 492798450.0, "step": 12983 }, { "epoch": 1.6516982572191834, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.715277194976807, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8772431015968323, "num_tokens": 492835081.0, "step": 12984 }, { "epoch": 1.651825467497774, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.315140724182129, "learning_rate": 5e-05, "loss": 0.6791, "mean_token_accuracy": 0.8806237578392029, "num_tokens": 492878255.0, "step": 12985 }, { "epoch": 1.6519526777763645, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.848962306976318, "learning_rate": 5e-05, "loss": 0.7825, "mean_token_accuracy": 0.8614641427993774, "num_tokens": 492920800.0, "step": 12986 }, { "epoch": 1.652079888054955, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.128218650817871, "learning_rate": 5e-05, "loss": 0.7184, "mean_token_accuracy": 0.8700196146965027, "num_tokens": 492957170.0, "step": 12987 }, { "epoch": 1.6522070983335455, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.058268547058105, "learning_rate": 5e-05, "loss": 0.7674, "mean_token_accuracy": 0.8749166131019592, "num_tokens": 492995123.0, "step": 12988 }, { "epoch": 1.6523343086121358, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.86409330368042, "learning_rate": 5e-05, "loss": 0.677, "mean_token_accuracy": 0.8744210004806519, "num_tokens": 493034538.0, "step": 12989 }, { "epoch": 1.6524615188907263, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.469535827636719, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8589511513710022, "num_tokens": 493073218.0, "step": 12990 }, { "epoch": 1.6525887291693169, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.392744064331055, "learning_rate": 5e-05, "loss": 0.6632, "mean_token_accuracy": 0.8711528182029724, "num_tokens": 493110056.0, "step": 12991 }, { "epoch": 1.6527159394479074, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 9.199823379516602, "learning_rate": 5e-05, "loss": 0.8531, "mean_token_accuracy": 0.8796270489692688, "num_tokens": 493144324.0, "step": 12992 }, { "epoch": 1.652843149726498, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 5.872106075286865, "learning_rate": 5e-05, "loss": 0.6829, "mean_token_accuracy": 0.8619488477706909, "num_tokens": 493181685.0, "step": 12993 }, { "epoch": 1.6529703600050885, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 10.157705307006836, "learning_rate": 5e-05, "loss": 0.9786, "mean_token_accuracy": 0.8782153129577637, "num_tokens": 493217712.0, "step": 12994 }, { "epoch": 1.6530975702836788, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.154564380645752, "learning_rate": 5e-05, "loss": 0.7741, "mean_token_accuracy": 0.856968104839325, "num_tokens": 493253852.0, "step": 12995 }, { "epoch": 1.6532247805622693, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 10.511068344116211, "learning_rate": 5e-05, "loss": 1.1178, "mean_token_accuracy": 0.8725353479385376, "num_tokens": 493287797.0, "step": 12996 }, { "epoch": 1.6533519908408598, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.02935791015625, "learning_rate": 5e-05, "loss": 1.0083, "mean_token_accuracy": 0.8745531439781189, "num_tokens": 493326072.0, "step": 12997 }, { "epoch": 1.6534792011194503, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.9297456741333, "learning_rate": 5e-05, "loss": 0.995, "mean_token_accuracy": 0.8559208512306213, "num_tokens": 493365374.0, "step": 12998 }, { "epoch": 1.6536064113980409, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 9.465002059936523, "learning_rate": 5e-05, "loss": 1.0897, "mean_token_accuracy": 0.8875657320022583, "num_tokens": 493408526.0, "step": 12999 }, { "epoch": 1.6537336216766314, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.669994354248047, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8606337308883667, "num_tokens": 493445689.0, "step": 13000 }, { "epoch": 1.653860831955222, "ewc_loss": 0.91796875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00086212158203125, "grad_norm": 11.646618843078613, "learning_rate": 5e-05, "loss": 1.3133, "mean_token_accuracy": 0.8616419434547424, "num_tokens": 493480884.0, "step": 13001 }, { "epoch": 1.6539880422338125, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.763725757598877, "learning_rate": 5e-05, "loss": 0.979, "mean_token_accuracy": 0.8742929697036743, "num_tokens": 493523917.0, "step": 13002 }, { "epoch": 1.654115252512403, "ewc_loss": 0.9375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000881195068359375, "grad_norm": 12.467726707458496, "learning_rate": 5e-05, "loss": 1.2987, "mean_token_accuracy": 0.8708970546722412, "num_tokens": 493560438.0, "step": 13003 }, { "epoch": 1.6542424627909935, "ewc_loss": 1.046875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0009918212890625, "grad_norm": 10.328080177307129, "learning_rate": 5e-05, "loss": 1.4658, "mean_token_accuracy": 0.8610299825668335, "num_tokens": 493596609.0, "step": 13004 }, { "epoch": 1.654369673069584, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 11.16209888458252, "learning_rate": 5e-05, "loss": 1.1132, "mean_token_accuracy": 0.8658023476600647, "num_tokens": 493634369.0, "step": 13005 }, { "epoch": 1.6544968833481746, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 9.48966121673584, "learning_rate": 5e-05, "loss": 1.1529, "mean_token_accuracy": 0.8682328462600708, "num_tokens": 493669568.0, "step": 13006 }, { "epoch": 1.654624093626765, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 10.81971263885498, "learning_rate": 5e-05, "loss": 1.1147, "mean_token_accuracy": 0.8761962652206421, "num_tokens": 493710136.0, "step": 13007 }, { "epoch": 1.6547513039053556, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.615123748779297, "learning_rate": 5e-05, "loss": 0.9636, "mean_token_accuracy": 0.8674420714378357, "num_tokens": 493750028.0, "step": 13008 }, { "epoch": 1.6548785141839462, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.69931411743164, "learning_rate": 5e-05, "loss": 1.1702, "mean_token_accuracy": 0.859856128692627, "num_tokens": 493785167.0, "step": 13009 }, { "epoch": 1.6550057244625367, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.40542984008789, "learning_rate": 5e-05, "loss": 0.979, "mean_token_accuracy": 0.8599252104759216, "num_tokens": 493822724.0, "step": 13010 }, { "epoch": 1.6551329347411272, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.918317794799805, "learning_rate": 5e-05, "loss": 1.1538, "mean_token_accuracy": 0.863093376159668, "num_tokens": 493863910.0, "step": 13011 }, { "epoch": 1.6552601450197177, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.881470680236816, "learning_rate": 5e-05, "loss": 0.9899, "mean_token_accuracy": 0.8788175582885742, "num_tokens": 493901863.0, "step": 13012 }, { "epoch": 1.655387355298308, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.606133460998535, "learning_rate": 5e-05, "loss": 1.0814, "mean_token_accuracy": 0.8727884292602539, "num_tokens": 493938174.0, "step": 13013 }, { "epoch": 1.6555145655768986, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.035194396972656, "learning_rate": 5e-05, "loss": 1.0205, "mean_token_accuracy": 0.8710910081863403, "num_tokens": 493973483.0, "step": 13014 }, { "epoch": 1.655641775855489, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.071633338928223, "learning_rate": 5e-05, "loss": 1.003, "mean_token_accuracy": 0.8848079442977905, "num_tokens": 494016481.0, "step": 13015 }, { "epoch": 1.6557689861340796, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.735726356506348, "learning_rate": 5e-05, "loss": 0.9363, "mean_token_accuracy": 0.8737050294876099, "num_tokens": 494047543.0, "step": 13016 }, { "epoch": 1.6558961964126702, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.806411743164062, "learning_rate": 5e-05, "loss": 1.0401, "mean_token_accuracy": 0.8699522018432617, "num_tokens": 494082076.0, "step": 13017 }, { "epoch": 1.6560234066912607, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.328102111816406, "learning_rate": 5e-05, "loss": 0.9016, "mean_token_accuracy": 0.8682597875595093, "num_tokens": 494123822.0, "step": 13018 }, { "epoch": 1.6561506169698512, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.855489730834961, "learning_rate": 5e-05, "loss": 0.9761, "mean_token_accuracy": 0.8874545693397522, "num_tokens": 494152644.0, "step": 13019 }, { "epoch": 1.6562778272484415, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.168180465698242, "learning_rate": 5e-05, "loss": 0.9314, "mean_token_accuracy": 0.8575863838195801, "num_tokens": 494185667.0, "step": 13020 }, { "epoch": 1.656405037527032, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.979887008666992, "learning_rate": 5e-05, "loss": 1.0234, "mean_token_accuracy": 0.8814656734466553, "num_tokens": 494224277.0, "step": 13021 }, { "epoch": 1.6565322478056226, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.31728458404541, "learning_rate": 5e-05, "loss": 0.9222, "mean_token_accuracy": 0.8655497431755066, "num_tokens": 494254475.0, "step": 13022 }, { "epoch": 1.656659458084213, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.853179931640625, "learning_rate": 5e-05, "loss": 1.0311, "mean_token_accuracy": 0.8723232746124268, "num_tokens": 494291891.0, "step": 13023 }, { "epoch": 1.6567866683628036, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.510037422180176, "learning_rate": 5e-05, "loss": 0.9399, "mean_token_accuracy": 0.8648301959037781, "num_tokens": 494327659.0, "step": 13024 }, { "epoch": 1.6569138786413942, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.404610633850098, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.8856291174888611, "num_tokens": 494358149.0, "step": 13025 }, { "epoch": 1.6570410889199847, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.392524719238281, "learning_rate": 5e-05, "loss": 0.8794, "mean_token_accuracy": 0.8777374029159546, "num_tokens": 494393673.0, "step": 13026 }, { "epoch": 1.6571682991985752, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.08613395690918, "learning_rate": 5e-05, "loss": 0.9385, "mean_token_accuracy": 0.8740159869194031, "num_tokens": 494422811.0, "step": 13027 }, { "epoch": 1.6572955094771658, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.005806922912598, "learning_rate": 5e-05, "loss": 0.8397, "mean_token_accuracy": 0.8727613091468811, "num_tokens": 494467758.0, "step": 13028 }, { "epoch": 1.6574227197557563, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.143290519714355, "learning_rate": 5e-05, "loss": 0.9044, "mean_token_accuracy": 0.8869422078132629, "num_tokens": 494504698.0, "step": 13029 }, { "epoch": 1.6575499300343468, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.743739604949951, "learning_rate": 5e-05, "loss": 0.8366, "mean_token_accuracy": 0.8634894490242004, "num_tokens": 494541235.0, "step": 13030 }, { "epoch": 1.6576771403129373, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.271096229553223, "learning_rate": 5e-05, "loss": 0.9698, "mean_token_accuracy": 0.871202826499939, "num_tokens": 494584499.0, "step": 13031 }, { "epoch": 1.6578043505915279, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.777593612670898, "learning_rate": 5e-05, "loss": 0.8234, "mean_token_accuracy": 0.8695314526557922, "num_tokens": 494625845.0, "step": 13032 }, { "epoch": 1.6579315608701184, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.204561233520508, "learning_rate": 5e-05, "loss": 0.972, "mean_token_accuracy": 0.8765028119087219, "num_tokens": 494660820.0, "step": 13033 }, { "epoch": 1.658058771148709, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.953768730163574, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.885177493095398, "num_tokens": 494701215.0, "step": 13034 }, { "epoch": 1.6581859814272994, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.86677360534668, "learning_rate": 5e-05, "loss": 1.0329, "mean_token_accuracy": 0.8456694483757019, "num_tokens": 494739062.0, "step": 13035 }, { "epoch": 1.65831319170589, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.008134841918945, "learning_rate": 5e-05, "loss": 0.8378, "mean_token_accuracy": 0.8713233470916748, "num_tokens": 494779467.0, "step": 13036 }, { "epoch": 1.6584404019844805, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.476971626281738, "learning_rate": 5e-05, "loss": 0.9004, "mean_token_accuracy": 0.8709059953689575, "num_tokens": 494817237.0, "step": 13037 }, { "epoch": 1.6585676122630708, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.822084903717041, "learning_rate": 5e-05, "loss": 0.746, "mean_token_accuracy": 0.8879589438438416, "num_tokens": 494853485.0, "step": 13038 }, { "epoch": 1.6586948225416613, "ewc_loss": 0.5, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.347030639648438, "learning_rate": 5e-05, "loss": 0.8565, "mean_token_accuracy": 0.8754090070724487, "num_tokens": 494890456.0, "step": 13039 }, { "epoch": 1.6588220328202519, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.569187641143799, "learning_rate": 5e-05, "loss": 0.8041, "mean_token_accuracy": 0.8609880805015564, "num_tokens": 494929441.0, "step": 13040 }, { "epoch": 1.6589492430988424, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.373661994934082, "learning_rate": 5e-05, "loss": 0.8922, "mean_token_accuracy": 0.862459659576416, "num_tokens": 494963642.0, "step": 13041 }, { "epoch": 1.659076453377433, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.403505802154541, "learning_rate": 5e-05, "loss": 0.7765, "mean_token_accuracy": 0.8652522563934326, "num_tokens": 494998138.0, "step": 13042 }, { "epoch": 1.6592036636560235, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.396417617797852, "learning_rate": 5e-05, "loss": 0.8467, "mean_token_accuracy": 0.8730766177177429, "num_tokens": 495038080.0, "step": 13043 }, { "epoch": 1.6593308739346138, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.343811988830566, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8674478530883789, "num_tokens": 495079103.0, "step": 13044 }, { "epoch": 1.6594580842132043, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.35180377960205, "learning_rate": 5e-05, "loss": 0.8344, "mean_token_accuracy": 0.8725447654724121, "num_tokens": 495119385.0, "step": 13045 }, { "epoch": 1.6595852944917948, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.3583455085754395, "learning_rate": 5e-05, "loss": 0.7592, "mean_token_accuracy": 0.8643049597740173, "num_tokens": 495157829.0, "step": 13046 }, { "epoch": 1.6597125047703853, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.25301742553711, "learning_rate": 5e-05, "loss": 0.9101, "mean_token_accuracy": 0.8485444784164429, "num_tokens": 495198554.0, "step": 13047 }, { "epoch": 1.6598397150489759, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.4073686599731445, "learning_rate": 5e-05, "loss": 0.7264, "mean_token_accuracy": 0.8776910305023193, "num_tokens": 495241389.0, "step": 13048 }, { "epoch": 1.6599669253275664, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.1121187210083, "learning_rate": 5e-05, "loss": 0.7768, "mean_token_accuracy": 0.8816248178482056, "num_tokens": 495274216.0, "step": 13049 }, { "epoch": 1.660094135606157, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.505650520324707, "learning_rate": 5e-05, "loss": 0.7635, "mean_token_accuracy": 0.8634263277053833, "num_tokens": 495308941.0, "step": 13050 }, { "epoch": 1.6602213458847475, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.860509872436523, "learning_rate": 5e-05, "loss": 0.7614, "mean_token_accuracy": 0.8749279975891113, "num_tokens": 495351668.0, "step": 13051 }, { "epoch": 1.660348556163338, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.622092247009277, "learning_rate": 5e-05, "loss": 0.774, "mean_token_accuracy": 0.8638501763343811, "num_tokens": 495383402.0, "step": 13052 }, { "epoch": 1.6604757664419285, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.619869709014893, "learning_rate": 5e-05, "loss": 0.7129, "mean_token_accuracy": 0.882351279258728, "num_tokens": 495421371.0, "step": 13053 }, { "epoch": 1.660602976720519, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.738466262817383, "learning_rate": 5e-05, "loss": 0.7828, "mean_token_accuracy": 0.8651491403579712, "num_tokens": 495460692.0, "step": 13054 }, { "epoch": 1.6607301869991096, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.443014621734619, "learning_rate": 5e-05, "loss": 0.7208, "mean_token_accuracy": 0.8751732707023621, "num_tokens": 495493990.0, "step": 13055 }, { "epoch": 1.6608573972777, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.863940715789795, "learning_rate": 5e-05, "loss": 0.7755, "mean_token_accuracy": 0.8696908950805664, "num_tokens": 495531616.0, "step": 13056 }, { "epoch": 1.6609846075562906, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.280421257019043, "learning_rate": 5e-05, "loss": 0.7391, "mean_token_accuracy": 0.8623060584068298, "num_tokens": 495572065.0, "step": 13057 }, { "epoch": 1.6611118178348812, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.993941307067871, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8567671775817871, "num_tokens": 495612582.0, "step": 13058 }, { "epoch": 1.6612390281134717, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.158507347106934, "learning_rate": 5e-05, "loss": 0.7163, "mean_token_accuracy": 0.8688176870346069, "num_tokens": 495649211.0, "step": 13059 }, { "epoch": 1.6613662383920622, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.065399169921875, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8697259426116943, "num_tokens": 495689805.0, "step": 13060 }, { "epoch": 1.6614934486706527, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.108285903930664, "learning_rate": 5e-05, "loss": 0.7991, "mean_token_accuracy": 0.847669243812561, "num_tokens": 495727560.0, "step": 13061 }, { "epoch": 1.661620658949243, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.079344749450684, "learning_rate": 5e-05, "loss": 0.7628, "mean_token_accuracy": 0.8710057735443115, "num_tokens": 495759827.0, "step": 13062 }, { "epoch": 1.6617478692278336, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.118890285491943, "learning_rate": 5e-05, "loss": 0.7201, "mean_token_accuracy": 0.8725781440734863, "num_tokens": 495797106.0, "step": 13063 }, { "epoch": 1.661875079506424, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.065314292907715, "learning_rate": 5e-05, "loss": 0.7961, "mean_token_accuracy": 0.8661121129989624, "num_tokens": 495842331.0, "step": 13064 }, { "epoch": 1.6620022897850146, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.1525797843933105, "learning_rate": 5e-05, "loss": 0.7161, "mean_token_accuracy": 0.8667532205581665, "num_tokens": 495883491.0, "step": 13065 }, { "epoch": 1.6621295000636052, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 8.009905815124512, "learning_rate": 5e-05, "loss": 0.7873, "mean_token_accuracy": 0.8611612915992737, "num_tokens": 495924536.0, "step": 13066 }, { "epoch": 1.6622567103421957, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.214841365814209, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8603657484054565, "num_tokens": 495961783.0, "step": 13067 }, { "epoch": 1.662383920620786, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.904646396636963, "learning_rate": 5e-05, "loss": 0.7413, "mean_token_accuracy": 0.8745826482772827, "num_tokens": 495994131.0, "step": 13068 }, { "epoch": 1.6625111308993765, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.32012414932251, "learning_rate": 5e-05, "loss": 0.7307, "mean_token_accuracy": 0.8679383397102356, "num_tokens": 496035981.0, "step": 13069 }, { "epoch": 1.662638341177967, "ewc_loss": 0.375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.776023864746094, "learning_rate": 5e-05, "loss": 0.7798, "mean_token_accuracy": 0.8597770929336548, "num_tokens": 496075622.0, "step": 13070 }, { "epoch": 1.6627655514565576, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.412245273590088, "learning_rate": 5e-05, "loss": 0.7286, "mean_token_accuracy": 0.8700711727142334, "num_tokens": 496117130.0, "step": 13071 }, { "epoch": 1.662892761735148, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.679647445678711, "learning_rate": 5e-05, "loss": 0.6765, "mean_token_accuracy": 0.8916740417480469, "num_tokens": 496153309.0, "step": 13072 }, { "epoch": 1.6630199720137386, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.454807758331299, "learning_rate": 5e-05, "loss": 0.6822, "mean_token_accuracy": 0.8834083080291748, "num_tokens": 496194955.0, "step": 13073 }, { "epoch": 1.6631471822923292, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.6428351402282715, "learning_rate": 5e-05, "loss": 0.7399, "mean_token_accuracy": 0.8693951964378357, "num_tokens": 496238787.0, "step": 13074 }, { "epoch": 1.6632743925709197, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.448347091674805, "learning_rate": 5e-05, "loss": 0.7488, "mean_token_accuracy": 0.8606812953948975, "num_tokens": 496279606.0, "step": 13075 }, { "epoch": 1.6634016028495102, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.657771110534668, "learning_rate": 5e-05, "loss": 0.7059, "mean_token_accuracy": 0.8788212537765503, "num_tokens": 496315724.0, "step": 13076 }, { "epoch": 1.6635288131281007, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.367972373962402, "learning_rate": 5e-05, "loss": 0.6947, "mean_token_accuracy": 0.8740918040275574, "num_tokens": 496357254.0, "step": 13077 }, { "epoch": 1.6636560234066913, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.77858829498291, "learning_rate": 5e-05, "loss": 0.7303, "mean_token_accuracy": 0.8739979267120361, "num_tokens": 496394356.0, "step": 13078 }, { "epoch": 1.6637832336852818, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 7.175505638122559, "learning_rate": 5e-05, "loss": 0.7057, "mean_token_accuracy": 0.8677729964256287, "num_tokens": 496431569.0, "step": 13079 }, { "epoch": 1.6639104439638723, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 8.06236743927002, "learning_rate": 5e-05, "loss": 0.7468, "mean_token_accuracy": 0.8747061491012573, "num_tokens": 496465854.0, "step": 13080 }, { "epoch": 1.6640376542424629, "ewc_loss": 0.30078125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.786942005157471, "learning_rate": 5e-05, "loss": 0.6596, "mean_token_accuracy": 0.8733887076377869, "num_tokens": 496506023.0, "step": 13081 }, { "epoch": 1.6641648645210534, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.63827896118164, "learning_rate": 5e-05, "loss": 0.8416, "mean_token_accuracy": 0.859951913356781, "num_tokens": 496545950.0, "step": 13082 }, { "epoch": 1.664292074799644, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 6.139614582061768, "learning_rate": 5e-05, "loss": 0.6304, "mean_token_accuracy": 0.8748497366905212, "num_tokens": 496585728.0, "step": 13083 }, { "epoch": 1.6644192850782344, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 9.674690246582031, "learning_rate": 5e-05, "loss": 0.967, "mean_token_accuracy": 0.8530492782592773, "num_tokens": 496628491.0, "step": 13084 }, { "epoch": 1.664546495356825, "ewc_loss": 0.2578125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00020313262939453125, "grad_norm": 5.445988655090332, "learning_rate": 5e-05, "loss": 0.6551, "mean_token_accuracy": 0.8643304109573364, "num_tokens": 496665952.0, "step": 13085 }, { "epoch": 1.6646737056354155, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 11.08498764038086, "learning_rate": 5e-05, "loss": 1.0694, "mean_token_accuracy": 0.8807791471481323, "num_tokens": 496708188.0, "step": 13086 }, { "epoch": 1.6648009159140058, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.476210117340088, "learning_rate": 5e-05, "loss": 0.7737, "mean_token_accuracy": 0.878654420375824, "num_tokens": 496743079.0, "step": 13087 }, { "epoch": 1.6649281261925963, "ewc_loss": 0.87890625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000823974609375, "grad_norm": 11.587336540222168, "learning_rate": 5e-05, "loss": 1.2534, "mean_token_accuracy": 0.8726099729537964, "num_tokens": 496781364.0, "step": 13088 }, { "epoch": 1.6650553364711869, "ewc_loss": 0.859375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 9.253511428833008, "learning_rate": 5e-05, "loss": 1.2037, "mean_token_accuracy": 0.8827500939369202, "num_tokens": 496816419.0, "step": 13089 }, { "epoch": 1.6651825467497774, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.51961612701416, "learning_rate": 5e-05, "loss": 0.9307, "mean_token_accuracy": 0.8760097622871399, "num_tokens": 496851941.0, "step": 13090 }, { "epoch": 1.665309757028368, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 10.226908683776855, "learning_rate": 5e-05, "loss": 1.2174, "mean_token_accuracy": 0.8765606880187988, "num_tokens": 496884974.0, "step": 13091 }, { "epoch": 1.6654369673069584, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.805115222930908, "learning_rate": 5e-05, "loss": 0.8279, "mean_token_accuracy": 0.8768244981765747, "num_tokens": 496920331.0, "step": 13092 }, { "epoch": 1.6655641775855488, "ewc_loss": 0.921875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000865936279296875, "grad_norm": 11.898454666137695, "learning_rate": 5e-05, "loss": 1.2983, "mean_token_accuracy": 0.8669350147247314, "num_tokens": 496957599.0, "step": 13093 }, { "epoch": 1.6656913878641393, "ewc_loss": 0.78125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 9.367332458496094, "learning_rate": 5e-05, "loss": 1.1082, "mean_token_accuracy": 0.8819349408149719, "num_tokens": 496992804.0, "step": 13094 }, { "epoch": 1.6658185981427298, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 11.341832160949707, "learning_rate": 5e-05, "loss": 1.0455, "mean_token_accuracy": 0.8827922344207764, "num_tokens": 497025376.0, "step": 13095 }, { "epoch": 1.6659458084213203, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 9.440361022949219, "learning_rate": 5e-05, "loss": 1.1339, "mean_token_accuracy": 0.8786774277687073, "num_tokens": 497067923.0, "step": 13096 }, { "epoch": 1.6660730186999109, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.818349838256836, "learning_rate": 5e-05, "loss": 1.0994, "mean_token_accuracy": 0.8620349168777466, "num_tokens": 497113753.0, "step": 13097 }, { "epoch": 1.6662002289785014, "ewc_loss": 0.625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.451396942138672, "learning_rate": 5e-05, "loss": 0.9796, "mean_token_accuracy": 0.8745661377906799, "num_tokens": 497148001.0, "step": 13098 }, { "epoch": 1.666327439257092, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.540221214294434, "learning_rate": 5e-05, "loss": 1.1119, "mean_token_accuracy": 0.8518462181091309, "num_tokens": 497186702.0, "step": 13099 }, { "epoch": 1.6664546495356825, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.705865859985352, "learning_rate": 5e-05, "loss": 0.8946, "mean_token_accuracy": 0.8727160692214966, "num_tokens": 497224550.0, "step": 13100 }, { "epoch": 1.666581859814273, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.940272331237793, "learning_rate": 5e-05, "loss": 1.1326, "mean_token_accuracy": 0.8541467785835266, "num_tokens": 497261922.0, "step": 13101 }, { "epoch": 1.6667090700928635, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.867250442504883, "learning_rate": 5e-05, "loss": 0.9997, "mean_token_accuracy": 0.8485610485076904, "num_tokens": 497302297.0, "step": 13102 }, { "epoch": 1.666836280371454, "ewc_loss": 0.78125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 11.471717834472656, "learning_rate": 5e-05, "loss": 1.1632, "mean_token_accuracy": 0.8715804815292358, "num_tokens": 497337788.0, "step": 13103 }, { "epoch": 1.6669634906500446, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 8.960118293762207, "learning_rate": 5e-05, "loss": 1.0828, "mean_token_accuracy": 0.8586062788963318, "num_tokens": 497373455.0, "step": 13104 }, { "epoch": 1.667090700928635, "ewc_loss": 0.71875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 11.010845184326172, "learning_rate": 5e-05, "loss": 1.0914, "mean_token_accuracy": 0.8738175630569458, "num_tokens": 497414743.0, "step": 13105 }, { "epoch": 1.6672179112072256, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 9.300312995910645, "learning_rate": 5e-05, "loss": 1.0571, "mean_token_accuracy": 0.8747409582138062, "num_tokens": 497446769.0, "step": 13106 }, { "epoch": 1.6673451214858162, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.214590072631836, "learning_rate": 5e-05, "loss": 0.978, "mean_token_accuracy": 0.889353334903717, "num_tokens": 497488693.0, "step": 13107 }, { "epoch": 1.6674723317644067, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.03010082244873, "learning_rate": 5e-05, "loss": 0.9657, "mean_token_accuracy": 0.8786113858222961, "num_tokens": 497522258.0, "step": 13108 }, { "epoch": 1.6675995420429972, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.4586763381958, "learning_rate": 5e-05, "loss": 0.9858, "mean_token_accuracy": 0.8662412166595459, "num_tokens": 497557175.0, "step": 13109 }, { "epoch": 1.6677267523215877, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.541913986206055, "learning_rate": 5e-05, "loss": 0.8596, "mean_token_accuracy": 0.8848408460617065, "num_tokens": 497592880.0, "step": 13110 }, { "epoch": 1.667853962600178, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.137629508972168, "learning_rate": 5e-05, "loss": 0.9336, "mean_token_accuracy": 0.8747832179069519, "num_tokens": 497628416.0, "step": 13111 }, { "epoch": 1.6679811728787686, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.860378265380859, "learning_rate": 5e-05, "loss": 0.8585, "mean_token_accuracy": 0.8604059219360352, "num_tokens": 497666739.0, "step": 13112 }, { "epoch": 1.668108383157359, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.344348907470703, "learning_rate": 5e-05, "loss": 0.9427, "mean_token_accuracy": 0.871370255947113, "num_tokens": 497706307.0, "step": 13113 }, { "epoch": 1.6682355934359496, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.459517478942871, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.8578037023544312, "num_tokens": 497747623.0, "step": 13114 }, { "epoch": 1.6683628037145402, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.595398902893066, "learning_rate": 5e-05, "loss": 0.9834, "mean_token_accuracy": 0.8683995008468628, "num_tokens": 497783354.0, "step": 13115 }, { "epoch": 1.6684900139931307, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.57376766204834, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8739270567893982, "num_tokens": 497818178.0, "step": 13116 }, { "epoch": 1.668617224271721, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.633550643920898, "learning_rate": 5e-05, "loss": 0.9908, "mean_token_accuracy": 0.8732102513313293, "num_tokens": 497852899.0, "step": 13117 }, { "epoch": 1.6687444345503115, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.92239236831665, "learning_rate": 5e-05, "loss": 0.8472, "mean_token_accuracy": 0.8734462261199951, "num_tokens": 497896743.0, "step": 13118 }, { "epoch": 1.668871644828902, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.336216926574707, "learning_rate": 5e-05, "loss": 0.9619, "mean_token_accuracy": 0.877203106880188, "num_tokens": 497938837.0, "step": 13119 }, { "epoch": 1.6689988551074926, "ewc_loss": 0.5, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.187799453735352, "learning_rate": 5e-05, "loss": 0.8456, "mean_token_accuracy": 0.8793874382972717, "num_tokens": 497970335.0, "step": 13120 }, { "epoch": 1.669126065386083, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.696062088012695, "learning_rate": 5e-05, "loss": 0.9365, "mean_token_accuracy": 0.868382453918457, "num_tokens": 498003484.0, "step": 13121 }, { "epoch": 1.6692532756646736, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.131346702575684, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.8835272192955017, "num_tokens": 498040932.0, "step": 13122 }, { "epoch": 1.6693804859432642, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.277658462524414, "learning_rate": 5e-05, "loss": 0.9452, "mean_token_accuracy": 0.8469381332397461, "num_tokens": 498078911.0, "step": 13123 }, { "epoch": 1.6695076962218547, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.757335186004639, "learning_rate": 5e-05, "loss": 0.7592, "mean_token_accuracy": 0.8806953430175781, "num_tokens": 498119917.0, "step": 13124 }, { "epoch": 1.6696349065004452, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.346691131591797, "learning_rate": 5e-05, "loss": 0.8346, "mean_token_accuracy": 0.8774957060813904, "num_tokens": 498160359.0, "step": 13125 }, { "epoch": 1.6697621167790357, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.383354187011719, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8622438907623291, "num_tokens": 498197048.0, "step": 13126 }, { "epoch": 1.6698893270576263, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.497845649719238, "learning_rate": 5e-05, "loss": 0.9251, "mean_token_accuracy": 0.8549855947494507, "num_tokens": 498237007.0, "step": 13127 }, { "epoch": 1.6700165373362168, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.1292548179626465, "learning_rate": 5e-05, "loss": 0.7245, "mean_token_accuracy": 0.8702158331871033, "num_tokens": 498275324.0, "step": 13128 }, { "epoch": 1.6701437476148073, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.792597770690918, "learning_rate": 5e-05, "loss": 0.9464, "mean_token_accuracy": 0.855496883392334, "num_tokens": 498317018.0, "step": 13129 }, { "epoch": 1.6702709578933979, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.973060607910156, "learning_rate": 5e-05, "loss": 0.7109, "mean_token_accuracy": 0.872323751449585, "num_tokens": 498353903.0, "step": 13130 }, { "epoch": 1.6703981681719884, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.023987770080566, "learning_rate": 5e-05, "loss": 0.9566, "mean_token_accuracy": 0.8620250225067139, "num_tokens": 498390959.0, "step": 13131 }, { "epoch": 1.670525378450579, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.221658706665039, "learning_rate": 5e-05, "loss": 0.7317, "mean_token_accuracy": 0.880887508392334, "num_tokens": 498433067.0, "step": 13132 }, { "epoch": 1.6706525887291694, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.71020221710205, "learning_rate": 5e-05, "loss": 0.9481, "mean_token_accuracy": 0.8642334938049316, "num_tokens": 498476020.0, "step": 13133 }, { "epoch": 1.67077979900776, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.75554895401001, "learning_rate": 5e-05, "loss": 0.789, "mean_token_accuracy": 0.8744446635246277, "num_tokens": 498515644.0, "step": 13134 }, { "epoch": 1.6709070092863505, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.93177604675293, "learning_rate": 5e-05, "loss": 0.8164, "mean_token_accuracy": 0.8760109543800354, "num_tokens": 498551063.0, "step": 13135 }, { "epoch": 1.6710342195649408, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000553131103515625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.060927391052246, "learning_rate": 5e-05, "loss": 0.8204, "mean_token_accuracy": 0.8697399497032166, "num_tokens": 498588849.0, "step": 13136 }, { "epoch": 1.6711614298435313, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.483860015869141, "learning_rate": 5e-05, "loss": 0.7813, "mean_token_accuracy": 0.8647885322570801, "num_tokens": 498626073.0, "step": 13137 }, { "epoch": 1.6712886401221219, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.045116424560547, "learning_rate": 5e-05, "loss": 0.7763, "mean_token_accuracy": 0.8757508993148804, "num_tokens": 498661523.0, "step": 13138 }, { "epoch": 1.6714158504007124, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.4605889320373535, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.8628499507904053, "num_tokens": 498702075.0, "step": 13139 }, { "epoch": 1.671543060679303, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.8515801429748535, "learning_rate": 5e-05, "loss": 0.7917, "mean_token_accuracy": 0.8659799098968506, "num_tokens": 498737425.0, "step": 13140 }, { "epoch": 1.6716702709578934, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.599775314331055, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.8500752449035645, "num_tokens": 498779715.0, "step": 13141 }, { "epoch": 1.6717974812364838, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.566427707672119, "learning_rate": 5e-05, "loss": 0.7149, "mean_token_accuracy": 0.8809268474578857, "num_tokens": 498817312.0, "step": 13142 }, { "epoch": 1.6719246915150743, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.8844170570373535, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.8623238801956177, "num_tokens": 498857990.0, "step": 13143 }, { "epoch": 1.6720519017936648, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.175403118133545, "learning_rate": 5e-05, "loss": 0.7225, "mean_token_accuracy": 0.8689117431640625, "num_tokens": 498896275.0, "step": 13144 }, { "epoch": 1.6721791120722553, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.294816017150879, "learning_rate": 5e-05, "loss": 0.849, "mean_token_accuracy": 0.8618111610412598, "num_tokens": 498933662.0, "step": 13145 }, { "epoch": 1.6723063223508459, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.743690490722656, "learning_rate": 5e-05, "loss": 0.6836, "mean_token_accuracy": 0.8731866478919983, "num_tokens": 498974216.0, "step": 13146 }, { "epoch": 1.6724335326294364, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.850235939025879, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8700993061065674, "num_tokens": 499015889.0, "step": 13147 }, { "epoch": 1.672560742908027, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.400184631347656, "learning_rate": 5e-05, "loss": 0.6563, "mean_token_accuracy": 0.881986677646637, "num_tokens": 499056889.0, "step": 13148 }, { "epoch": 1.6726879531866174, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.274223327636719, "learning_rate": 5e-05, "loss": 0.9109, "mean_token_accuracy": 0.8714256286621094, "num_tokens": 499097440.0, "step": 13149 }, { "epoch": 1.672815163465208, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.580996513366699, "learning_rate": 5e-05, "loss": 0.6901, "mean_token_accuracy": 0.87955242395401, "num_tokens": 499135450.0, "step": 13150 }, { "epoch": 1.6729423737437985, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.037896156311035, "learning_rate": 5e-05, "loss": 0.9651, "mean_token_accuracy": 0.8616399765014648, "num_tokens": 499178487.0, "step": 13151 }, { "epoch": 1.673069584022389, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.363015651702881, "learning_rate": 5e-05, "loss": 0.7871, "mean_token_accuracy": 0.8669673204421997, "num_tokens": 499213824.0, "step": 13152 }, { "epoch": 1.6731967943009796, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.0985746383667, "learning_rate": 5e-05, "loss": 0.8665, "mean_token_accuracy": 0.8683974742889404, "num_tokens": 499252406.0, "step": 13153 }, { "epoch": 1.67332400457957, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.100133895874023, "learning_rate": 5e-05, "loss": 0.8684, "mean_token_accuracy": 0.8639373183250427, "num_tokens": 499295249.0, "step": 13154 }, { "epoch": 1.6734512148581606, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.235438346862793, "learning_rate": 5e-05, "loss": 0.7188, "mean_token_accuracy": 0.8800843954086304, "num_tokens": 499334115.0, "step": 13155 }, { "epoch": 1.6735784251367511, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.492688179016113, "learning_rate": 5e-05, "loss": 0.8675, "mean_token_accuracy": 0.8672809600830078, "num_tokens": 499372181.0, "step": 13156 }, { "epoch": 1.6737056354153417, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.985626220703125, "learning_rate": 5e-05, "loss": 0.7413, "mean_token_accuracy": 0.8624916076660156, "num_tokens": 499413617.0, "step": 13157 }, { "epoch": 1.6738328456939322, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.454007148742676, "learning_rate": 5e-05, "loss": 0.8408, "mean_token_accuracy": 0.8717778921127319, "num_tokens": 499443976.0, "step": 13158 }, { "epoch": 1.6739600559725227, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.113251686096191, "learning_rate": 5e-05, "loss": 0.7357, "mean_token_accuracy": 0.8650190830230713, "num_tokens": 499477973.0, "step": 13159 }, { "epoch": 1.674087266251113, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.246949195861816, "learning_rate": 5e-05, "loss": 0.8801, "mean_token_accuracy": 0.8559420108795166, "num_tokens": 499523163.0, "step": 13160 }, { "epoch": 1.6742144765297036, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.228675842285156, "learning_rate": 5e-05, "loss": 0.7791, "mean_token_accuracy": 0.8564743399620056, "num_tokens": 499567583.0, "step": 13161 }, { "epoch": 1.674341686808294, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.196039199829102, "learning_rate": 5e-05, "loss": 0.8339, "mean_token_accuracy": 0.8649829030036926, "num_tokens": 499603257.0, "step": 13162 }, { "epoch": 1.6744688970868846, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.129578113555908, "learning_rate": 5e-05, "loss": 0.7208, "mean_token_accuracy": 0.87538743019104, "num_tokens": 499648881.0, "step": 13163 }, { "epoch": 1.6745961073654752, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.361895561218262, "learning_rate": 5e-05, "loss": 0.8378, "mean_token_accuracy": 0.8700099587440491, "num_tokens": 499688632.0, "step": 13164 }, { "epoch": 1.6747233176440657, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.911160945892334, "learning_rate": 5e-05, "loss": 0.7135, "mean_token_accuracy": 0.8712610602378845, "num_tokens": 499723285.0, "step": 13165 }, { "epoch": 1.674850527922656, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.619029998779297, "learning_rate": 5e-05, "loss": 0.9146, "mean_token_accuracy": 0.8529379963874817, "num_tokens": 499761472.0, "step": 13166 }, { "epoch": 1.6749777382012465, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.728409767150879, "learning_rate": 5e-05, "loss": 0.7412, "mean_token_accuracy": 0.8610539436340332, "num_tokens": 499796997.0, "step": 13167 }, { "epoch": 1.675104948479837, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.851516723632812, "learning_rate": 5e-05, "loss": 0.8821, "mean_token_accuracy": 0.872223436832428, "num_tokens": 499834307.0, "step": 13168 }, { "epoch": 1.6752321587584276, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.745479106903076, "learning_rate": 5e-05, "loss": 0.7263, "mean_token_accuracy": 0.8714870810508728, "num_tokens": 499876497.0, "step": 13169 }, { "epoch": 1.675359369037018, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.82448959350586, "learning_rate": 5e-05, "loss": 0.9532, "mean_token_accuracy": 0.8599884510040283, "num_tokens": 499917620.0, "step": 13170 }, { "epoch": 1.6754865793156086, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.085946559906006, "learning_rate": 5e-05, "loss": 0.7562, "mean_token_accuracy": 0.8761197328567505, "num_tokens": 499957098.0, "step": 13171 }, { "epoch": 1.6756137895941992, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.399147987365723, "learning_rate": 5e-05, "loss": 0.897, "mean_token_accuracy": 0.8696990013122559, "num_tokens": 500001046.0, "step": 13172 }, { "epoch": 1.6757409998727897, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.537363529205322, "learning_rate": 5e-05, "loss": 0.8292, "mean_token_accuracy": 0.8662384748458862, "num_tokens": 500041443.0, "step": 13173 }, { "epoch": 1.6758682101513802, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.82274866104126, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8880447149276733, "num_tokens": 500076624.0, "step": 13174 }, { "epoch": 1.6759954204299707, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.8285017013549805, "learning_rate": 5e-05, "loss": 0.8635, "mean_token_accuracy": 0.8612784743309021, "num_tokens": 500118995.0, "step": 13175 }, { "epoch": 1.6761226307085613, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.540851593017578, "learning_rate": 5e-05, "loss": 0.7823, "mean_token_accuracy": 0.8661094903945923, "num_tokens": 500155134.0, "step": 13176 }, { "epoch": 1.6762498409871518, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.763818740844727, "learning_rate": 5e-05, "loss": 0.7558, "mean_token_accuracy": 0.8782680630683899, "num_tokens": 500189921.0, "step": 13177 }, { "epoch": 1.6763770512657423, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.656101703643799, "learning_rate": 5e-05, "loss": 0.8248, "mean_token_accuracy": 0.8534990549087524, "num_tokens": 500230655.0, "step": 13178 }, { "epoch": 1.6765042615443329, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.43308162689209, "learning_rate": 5e-05, "loss": 0.7565, "mean_token_accuracy": 0.8692922592163086, "num_tokens": 500270827.0, "step": 13179 }, { "epoch": 1.6766314718229234, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.977847576141357, "learning_rate": 5e-05, "loss": 0.8352, "mean_token_accuracy": 0.8594425916671753, "num_tokens": 500312810.0, "step": 13180 }, { "epoch": 1.676758682101514, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.059844970703125, "learning_rate": 5e-05, "loss": 0.743, "mean_token_accuracy": 0.8621429800987244, "num_tokens": 500345013.0, "step": 13181 }, { "epoch": 1.6768858923801044, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.373336791992188, "learning_rate": 5e-05, "loss": 0.8071, "mean_token_accuracy": 0.8723167777061462, "num_tokens": 500375628.0, "step": 13182 }, { "epoch": 1.677013102658695, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.706607818603516, "learning_rate": 5e-05, "loss": 0.6558, "mean_token_accuracy": 0.8860291838645935, "num_tokens": 500410084.0, "step": 13183 }, { "epoch": 1.6771403129372855, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.773701667785645, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.8881170749664307, "num_tokens": 500452221.0, "step": 13184 }, { "epoch": 1.6772675232158758, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.478843688964844, "learning_rate": 5e-05, "loss": 0.6993, "mean_token_accuracy": 0.8714851140975952, "num_tokens": 500491358.0, "step": 13185 }, { "epoch": 1.6773947334944663, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.094123840332031, "learning_rate": 5e-05, "loss": 0.9383, "mean_token_accuracy": 0.865256667137146, "num_tokens": 500530363.0, "step": 13186 }, { "epoch": 1.6775219437730569, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.6140265464782715, "learning_rate": 5e-05, "loss": 0.7144, "mean_token_accuracy": 0.8787509799003601, "num_tokens": 500568922.0, "step": 13187 }, { "epoch": 1.6776491540516474, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.917351722717285, "learning_rate": 5e-05, "loss": 0.9546, "mean_token_accuracy": 0.8620896935462952, "num_tokens": 500616445.0, "step": 13188 }, { "epoch": 1.677776364330238, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.307889938354492, "learning_rate": 5e-05, "loss": 0.7911, "mean_token_accuracy": 0.8796517848968506, "num_tokens": 500650692.0, "step": 13189 }, { "epoch": 1.6779035746088284, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.97088098526001, "learning_rate": 5e-05, "loss": 0.8887, "mean_token_accuracy": 0.8602076768875122, "num_tokens": 500692465.0, "step": 13190 }, { "epoch": 1.6780307848874187, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.203407287597656, "learning_rate": 5e-05, "loss": 0.8999, "mean_token_accuracy": 0.8604008555412292, "num_tokens": 500728715.0, "step": 13191 }, { "epoch": 1.6781579951660093, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.839087009429932, "learning_rate": 5e-05, "loss": 0.7351, "mean_token_accuracy": 0.8667445778846741, "num_tokens": 500768310.0, "step": 13192 }, { "epoch": 1.6782852054445998, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.05660343170166, "learning_rate": 5e-05, "loss": 0.9131, "mean_token_accuracy": 0.874500572681427, "num_tokens": 500809777.0, "step": 13193 }, { "epoch": 1.6784124157231903, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.225596904754639, "learning_rate": 5e-05, "loss": 0.7364, "mean_token_accuracy": 0.8607674241065979, "num_tokens": 500851412.0, "step": 13194 }, { "epoch": 1.6785396260017809, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.714436531066895, "learning_rate": 5e-05, "loss": 1.0438, "mean_token_accuracy": 0.8518244624137878, "num_tokens": 500890896.0, "step": 13195 }, { "epoch": 1.6786668362803714, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.419813632965088, "learning_rate": 5e-05, "loss": 0.7144, "mean_token_accuracy": 0.881213903427124, "num_tokens": 500926826.0, "step": 13196 }, { "epoch": 1.678794046558962, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.764656066894531, "learning_rate": 5e-05, "loss": 1.0553, "mean_token_accuracy": 0.8630236983299255, "num_tokens": 500960755.0, "step": 13197 }, { "epoch": 1.6789212568375524, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.587392807006836, "learning_rate": 5e-05, "loss": 0.9317, "mean_token_accuracy": 0.8716580271720886, "num_tokens": 500993177.0, "step": 13198 }, { "epoch": 1.679048467116143, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.685147285461426, "learning_rate": 5e-05, "loss": 0.9492, "mean_token_accuracy": 0.8693932890892029, "num_tokens": 501031391.0, "step": 13199 }, { "epoch": 1.6791756773947335, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.523524284362793, "learning_rate": 5e-05, "loss": 0.9497, "mean_token_accuracy": 0.8787778615951538, "num_tokens": 501072010.0, "step": 13200 }, { "epoch": 1.679302887673324, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.400962829589844, "learning_rate": 5e-05, "loss": 0.7987, "mean_token_accuracy": 0.8767189979553223, "num_tokens": 501112677.0, "step": 13201 }, { "epoch": 1.6794300979519146, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.792570114135742, "learning_rate": 5e-05, "loss": 0.9337, "mean_token_accuracy": 0.8725053668022156, "num_tokens": 501156634.0, "step": 13202 }, { "epoch": 1.679557308230505, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.3338704109191895, "learning_rate": 5e-05, "loss": 0.8142, "mean_token_accuracy": 0.8729585409164429, "num_tokens": 501188194.0, "step": 13203 }, { "epoch": 1.6796845185090956, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.057785034179688, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.8672731518745422, "num_tokens": 501216962.0, "step": 13204 }, { "epoch": 1.6798117287876861, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.030777931213379, "learning_rate": 5e-05, "loss": 0.8363, "mean_token_accuracy": 0.8835044503211975, "num_tokens": 501261964.0, "step": 13205 }, { "epoch": 1.6799389390662767, "ewc_loss": 0.375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.089478969573975, "learning_rate": 5e-05, "loss": 0.731, "mean_token_accuracy": 0.8737943172454834, "num_tokens": 501303701.0, "step": 13206 }, { "epoch": 1.6800661493448672, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.686944007873535, "learning_rate": 5e-05, "loss": 0.8754, "mean_token_accuracy": 0.879486083984375, "num_tokens": 501338531.0, "step": 13207 }, { "epoch": 1.6801933596234577, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.644796371459961, "learning_rate": 5e-05, "loss": 0.7246, "mean_token_accuracy": 0.8676049709320068, "num_tokens": 501375968.0, "step": 13208 }, { "epoch": 1.680320569902048, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.093167304992676, "learning_rate": 5e-05, "loss": 0.9462, "mean_token_accuracy": 0.8658273220062256, "num_tokens": 501414646.0, "step": 13209 }, { "epoch": 1.6804477801806386, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.6737751960754395, "learning_rate": 5e-05, "loss": 0.741, "mean_token_accuracy": 0.8683178424835205, "num_tokens": 501455035.0, "step": 13210 }, { "epoch": 1.680574990459229, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.180769920349121, "learning_rate": 5e-05, "loss": 0.9575, "mean_token_accuracy": 0.8714249730110168, "num_tokens": 501493198.0, "step": 13211 }, { "epoch": 1.6807022007378196, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.007169246673584, "learning_rate": 5e-05, "loss": 0.7993, "mean_token_accuracy": 0.8752087950706482, "num_tokens": 501534293.0, "step": 13212 }, { "epoch": 1.6808294110164101, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.031990051269531, "learning_rate": 5e-05, "loss": 0.9649, "mean_token_accuracy": 0.8696433305740356, "num_tokens": 501566614.0, "step": 13213 }, { "epoch": 1.6809566212950007, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.342153072357178, "learning_rate": 5e-05, "loss": 0.8988, "mean_token_accuracy": 0.8595374822616577, "num_tokens": 501607535.0, "step": 13214 }, { "epoch": 1.681083831573591, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.739797592163086, "learning_rate": 5e-05, "loss": 0.9669, "mean_token_accuracy": 0.8650586605072021, "num_tokens": 501644208.0, "step": 13215 }, { "epoch": 1.6812110418521815, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.432560920715332, "learning_rate": 5e-05, "loss": 0.9149, "mean_token_accuracy": 0.856934130191803, "num_tokens": 501684980.0, "step": 13216 }, { "epoch": 1.681338252130772, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.675699234008789, "learning_rate": 5e-05, "loss": 0.9291, "mean_token_accuracy": 0.8739749789237976, "num_tokens": 501722603.0, "step": 13217 }, { "epoch": 1.6814654624093626, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.304926872253418, "learning_rate": 5e-05, "loss": 0.8333, "mean_token_accuracy": 0.8761630058288574, "num_tokens": 501755650.0, "step": 13218 }, { "epoch": 1.681592672687953, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.751479148864746, "learning_rate": 5e-05, "loss": 0.9575, "mean_token_accuracy": 0.8699742555618286, "num_tokens": 501788848.0, "step": 13219 }, { "epoch": 1.6817198829665436, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.294014930725098, "learning_rate": 5e-05, "loss": 0.854, "mean_token_accuracy": 0.8682711124420166, "num_tokens": 501826396.0, "step": 13220 }, { "epoch": 1.6818470932451342, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.573301315307617, "learning_rate": 5e-05, "loss": 0.9709, "mean_token_accuracy": 0.8583143949508667, "num_tokens": 501863780.0, "step": 13221 }, { "epoch": 1.6819743035237247, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.532076358795166, "learning_rate": 5e-05, "loss": 0.8693, "mean_token_accuracy": 0.8685299158096313, "num_tokens": 501903275.0, "step": 13222 }, { "epoch": 1.6821015138023152, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.140419006347656, "learning_rate": 5e-05, "loss": 0.8115, "mean_token_accuracy": 0.8905484080314636, "num_tokens": 501940397.0, "step": 13223 }, { "epoch": 1.6822287240809057, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.7248711585998535, "learning_rate": 5e-05, "loss": 0.8722, "mean_token_accuracy": 0.8709596991539001, "num_tokens": 501979814.0, "step": 13224 }, { "epoch": 1.6823559343594963, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.804500579833984, "learning_rate": 5e-05, "loss": 0.8244, "mean_token_accuracy": 0.8737555742263794, "num_tokens": 502017349.0, "step": 13225 }, { "epoch": 1.6824831446380868, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.717226982116699, "learning_rate": 5e-05, "loss": 0.8653, "mean_token_accuracy": 0.8619451522827148, "num_tokens": 502062441.0, "step": 13226 }, { "epoch": 1.6826103549166773, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.696132659912109, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8690764904022217, "num_tokens": 502099897.0, "step": 13227 }, { "epoch": 1.6827375651952678, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.654160022735596, "learning_rate": 5e-05, "loss": 0.779, "mean_token_accuracy": 0.8757493495941162, "num_tokens": 502139682.0, "step": 13228 }, { "epoch": 1.6828647754738584, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.671924114227295, "learning_rate": 5e-05, "loss": 0.8084, "mean_token_accuracy": 0.8617027997970581, "num_tokens": 502179534.0, "step": 13229 }, { "epoch": 1.682991985752449, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.610775470733643, "learning_rate": 5e-05, "loss": 0.7706, "mean_token_accuracy": 0.875988781452179, "num_tokens": 502218031.0, "step": 13230 }, { "epoch": 1.6831191960310394, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.620571613311768, "learning_rate": 5e-05, "loss": 0.7525, "mean_token_accuracy": 0.8767595291137695, "num_tokens": 502251409.0, "step": 13231 }, { "epoch": 1.68324640630963, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.609918594360352, "learning_rate": 5e-05, "loss": 0.7586, "mean_token_accuracy": 0.874984860420227, "num_tokens": 502285757.0, "step": 13232 }, { "epoch": 1.6833736165882205, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.549015998840332, "learning_rate": 5e-05, "loss": 0.7313, "mean_token_accuracy": 0.8746674060821533, "num_tokens": 502325247.0, "step": 13233 }, { "epoch": 1.6835008268668108, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.642475128173828, "learning_rate": 5e-05, "loss": 0.772, "mean_token_accuracy": 0.8703285455703735, "num_tokens": 502367691.0, "step": 13234 }, { "epoch": 1.6836280371454013, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.47996711730957, "learning_rate": 5e-05, "loss": 0.7752, "mean_token_accuracy": 0.8601933717727661, "num_tokens": 502409084.0, "step": 13235 }, { "epoch": 1.6837552474239919, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.682537078857422, "learning_rate": 5e-05, "loss": 0.7455, "mean_token_accuracy": 0.8742542862892151, "num_tokens": 502448478.0, "step": 13236 }, { "epoch": 1.6838824577025824, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.4142560958862305, "learning_rate": 5e-05, "loss": 0.7494, "mean_token_accuracy": 0.8641194105148315, "num_tokens": 502481459.0, "step": 13237 }, { "epoch": 1.684009667981173, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.7300591468811035, "learning_rate": 5e-05, "loss": 0.82, "mean_token_accuracy": 0.8536416888237, "num_tokens": 502522300.0, "step": 13238 }, { "epoch": 1.6841368782597634, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.3408074378967285, "learning_rate": 5e-05, "loss": 0.7191, "mean_token_accuracy": 0.8753199577331543, "num_tokens": 502560257.0, "step": 13239 }, { "epoch": 1.6842640885383537, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.7967047691345215, "learning_rate": 5e-05, "loss": 0.7803, "mean_token_accuracy": 0.8616219162940979, "num_tokens": 502602815.0, "step": 13240 }, { "epoch": 1.6843912988169443, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.2174248695373535, "learning_rate": 5e-05, "loss": 0.6945, "mean_token_accuracy": 0.8735826015472412, "num_tokens": 502640067.0, "step": 13241 }, { "epoch": 1.6845185090955348, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.949059009552002, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8656660318374634, "num_tokens": 502680351.0, "step": 13242 }, { "epoch": 1.6846457193741253, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 7.006407737731934, "learning_rate": 5e-05, "loss": 0.6862, "mean_token_accuracy": 0.8783168792724609, "num_tokens": 502715279.0, "step": 13243 }, { "epoch": 1.6847729296527159, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.268794059753418, "learning_rate": 5e-05, "loss": 0.8271, "mean_token_accuracy": 0.8590008020401001, "num_tokens": 502762078.0, "step": 13244 }, { "epoch": 1.6849001399313064, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.613304615020752, "learning_rate": 5e-05, "loss": 0.6331, "mean_token_accuracy": 0.8851782083511353, "num_tokens": 502799886.0, "step": 13245 }, { "epoch": 1.685027350209897, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.886019706726074, "learning_rate": 5e-05, "loss": 0.8873, "mean_token_accuracy": 0.8601696491241455, "num_tokens": 502838414.0, "step": 13246 }, { "epoch": 1.6851545604884874, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 6.011701583862305, "learning_rate": 5e-05, "loss": 0.682, "mean_token_accuracy": 0.8622665405273438, "num_tokens": 502879504.0, "step": 13247 }, { "epoch": 1.685281770767078, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.8734712600708, "learning_rate": 5e-05, "loss": 0.8992, "mean_token_accuracy": 0.8898332118988037, "num_tokens": 502919167.0, "step": 13248 }, { "epoch": 1.6854089810456685, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 5.732934474945068, "learning_rate": 5e-05, "loss": 0.6221, "mean_token_accuracy": 0.8928097486495972, "num_tokens": 502956548.0, "step": 13249 }, { "epoch": 1.685536191324259, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 10.822990417480469, "learning_rate": 5e-05, "loss": 1.1212, "mean_token_accuracy": 0.8781346082687378, "num_tokens": 502993001.0, "step": 13250 }, { "epoch": 1.6856634016028496, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.304108619689941, "learning_rate": 5e-05, "loss": 0.9456, "mean_token_accuracy": 0.8737567663192749, "num_tokens": 503030690.0, "step": 13251 }, { "epoch": 1.68579061188144, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.135936737060547, "learning_rate": 5e-05, "loss": 1.0751, "mean_token_accuracy": 0.8651926517486572, "num_tokens": 503074010.0, "step": 13252 }, { "epoch": 1.6859178221600306, "ewc_loss": 0.828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000774383544921875, "grad_norm": 9.171984672546387, "learning_rate": 5e-05, "loss": 1.1928, "mean_token_accuracy": 0.8732763528823853, "num_tokens": 503111077.0, "step": 13253 }, { "epoch": 1.6860450324386211, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.604560852050781, "learning_rate": 5e-05, "loss": 0.7319, "mean_token_accuracy": 0.8925148844718933, "num_tokens": 503146594.0, "step": 13254 }, { "epoch": 1.6861722427172117, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 11.15752124786377, "learning_rate": 5e-05, "loss": 1.2168, "mean_token_accuracy": 0.8854897618293762, "num_tokens": 503181470.0, "step": 13255 }, { "epoch": 1.6862994529958022, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.125343322753906, "learning_rate": 5e-05, "loss": 0.8379, "mean_token_accuracy": 0.882955014705658, "num_tokens": 503215839.0, "step": 13256 }, { "epoch": 1.6864266632743927, "ewc_loss": 1.09375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00103759765625, "grad_norm": 13.195992469787598, "learning_rate": 5e-05, "loss": 1.4432, "mean_token_accuracy": 0.8790566921234131, "num_tokens": 503253131.0, "step": 13257 }, { "epoch": 1.686553873552983, "ewc_loss": 1.1171875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00106048583984375, "grad_norm": 10.791221618652344, "learning_rate": 5e-05, "loss": 1.5575, "mean_token_accuracy": 0.8487249612808228, "num_tokens": 503290331.0, "step": 13258 }, { "epoch": 1.6866810838315736, "ewc_loss": 0.75, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 11.628856658935547, "learning_rate": 5e-05, "loss": 1.1529, "mean_token_accuracy": 0.8582048416137695, "num_tokens": 503328551.0, "step": 13259 }, { "epoch": 1.686808294110164, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.801651954650879, "learning_rate": 5e-05, "loss": 1.3415, "mean_token_accuracy": 0.8378379344940186, "num_tokens": 503365288.0, "step": 13260 }, { "epoch": 1.6869355043887546, "ewc_loss": 0.875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 11.839437484741211, "learning_rate": 5e-05, "loss": 1.2189, "mean_token_accuracy": 0.8784496784210205, "num_tokens": 503402941.0, "step": 13261 }, { "epoch": 1.6870627146673451, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.715597152709961, "learning_rate": 5e-05, "loss": 1.0063, "mean_token_accuracy": 0.8612138032913208, "num_tokens": 503443174.0, "step": 13262 }, { "epoch": 1.6871899249459357, "ewc_loss": 0.9296875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000873565673828125, "grad_norm": 12.349740028381348, "learning_rate": 5e-05, "loss": 1.2938, "mean_token_accuracy": 0.8747649788856506, "num_tokens": 503481790.0, "step": 13263 }, { "epoch": 1.687317135224526, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 9.658920288085938, "learning_rate": 5e-05, "loss": 1.0756, "mean_token_accuracy": 0.8751303553581238, "num_tokens": 503518386.0, "step": 13264 }, { "epoch": 1.6874443455031165, "ewc_loss": 0.8515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 12.117027282714844, "learning_rate": 5e-05, "loss": 1.2105, "mean_token_accuracy": 0.8753623962402344, "num_tokens": 503558985.0, "step": 13265 }, { "epoch": 1.687571555781707, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00075531005859375, "grad_norm": 10.471952438354492, "learning_rate": 5e-05, "loss": 1.1608, "mean_token_accuracy": 0.8779247999191284, "num_tokens": 503602423.0, "step": 13266 }, { "epoch": 1.6876987660602976, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 11.208066940307617, "learning_rate": 5e-05, "loss": 1.0754, "mean_token_accuracy": 0.8792471885681152, "num_tokens": 503634300.0, "step": 13267 }, { "epoch": 1.687825976338888, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.874418258666992, "learning_rate": 5e-05, "loss": 1.0796, "mean_token_accuracy": 0.8627215027809143, "num_tokens": 503668225.0, "step": 13268 }, { "epoch": 1.6879531866174786, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.788281440734863, "learning_rate": 5e-05, "loss": 1.1158, "mean_token_accuracy": 0.8566061854362488, "num_tokens": 503705158.0, "step": 13269 }, { "epoch": 1.6880803968960691, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.266841888427734, "learning_rate": 5e-05, "loss": 0.9762, "mean_token_accuracy": 0.8727536201477051, "num_tokens": 503741322.0, "step": 13270 }, { "epoch": 1.6882076071746597, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.586748123168945, "learning_rate": 5e-05, "loss": 1.0112, "mean_token_accuracy": 0.8833892941474915, "num_tokens": 503777932.0, "step": 13271 }, { "epoch": 1.6883348174532502, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.834041595458984, "learning_rate": 5e-05, "loss": 0.8871, "mean_token_accuracy": 0.8821163177490234, "num_tokens": 503810175.0, "step": 13272 }, { "epoch": 1.6884620277318407, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.700749397277832, "learning_rate": 5e-05, "loss": 1.0769, "mean_token_accuracy": 0.8601778149604797, "num_tokens": 503844498.0, "step": 13273 }, { "epoch": 1.6885892380104313, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.730195999145508, "learning_rate": 5e-05, "loss": 0.8617, "mean_token_accuracy": 0.8903369903564453, "num_tokens": 503878898.0, "step": 13274 }, { "epoch": 1.6887164482890218, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.883512496948242, "learning_rate": 5e-05, "loss": 1.0466, "mean_token_accuracy": 0.8814606666564941, "num_tokens": 503915473.0, "step": 13275 }, { "epoch": 1.6888436585676123, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.993474960327148, "learning_rate": 5e-05, "loss": 0.9398, "mean_token_accuracy": 0.8729713559150696, "num_tokens": 503954234.0, "step": 13276 }, { "epoch": 1.6889708688462028, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.760936737060547, "learning_rate": 5e-05, "loss": 1.1437, "mean_token_accuracy": 0.851000189781189, "num_tokens": 503988740.0, "step": 13277 }, { "epoch": 1.6890980791247934, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.264241218566895, "learning_rate": 5e-05, "loss": 0.9558, "mean_token_accuracy": 0.873042643070221, "num_tokens": 504025755.0, "step": 13278 }, { "epoch": 1.689225289403384, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.206265449523926, "learning_rate": 5e-05, "loss": 1.0637, "mean_token_accuracy": 0.8578110933303833, "num_tokens": 504062639.0, "step": 13279 }, { "epoch": 1.6893524996819744, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.103572845458984, "learning_rate": 5e-05, "loss": 0.9826, "mean_token_accuracy": 0.852310299873352, "num_tokens": 504099645.0, "step": 13280 }, { "epoch": 1.689479709960565, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.655278205871582, "learning_rate": 5e-05, "loss": 0.9404, "mean_token_accuracy": 0.881361186504364, "num_tokens": 504137678.0, "step": 13281 }, { "epoch": 1.6896069202391555, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.604183197021484, "learning_rate": 5e-05, "loss": 0.8645, "mean_token_accuracy": 0.8732311725616455, "num_tokens": 504174037.0, "step": 13282 }, { "epoch": 1.6897341305177458, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.523177146911621, "learning_rate": 5e-05, "loss": 0.9951, "mean_token_accuracy": 0.8574228286743164, "num_tokens": 504220117.0, "step": 13283 }, { "epoch": 1.6898613407963363, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.081104278564453, "learning_rate": 5e-05, "loss": 0.7839, "mean_token_accuracy": 0.8805776238441467, "num_tokens": 504259104.0, "step": 13284 }, { "epoch": 1.6899885510749268, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.772743225097656, "learning_rate": 5e-05, "loss": 0.9887, "mean_token_accuracy": 0.8734562397003174, "num_tokens": 504302867.0, "step": 13285 }, { "epoch": 1.6901157613535174, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.975570201873779, "learning_rate": 5e-05, "loss": 0.771, "mean_token_accuracy": 0.8799247145652771, "num_tokens": 504337555.0, "step": 13286 }, { "epoch": 1.690242971632108, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.010220527648926, "learning_rate": 5e-05, "loss": 0.9702, "mean_token_accuracy": 0.8815089464187622, "num_tokens": 504372334.0, "step": 13287 }, { "epoch": 1.6903701819106984, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.361706733703613, "learning_rate": 5e-05, "loss": 0.8233, "mean_token_accuracy": 0.877159595489502, "num_tokens": 504412894.0, "step": 13288 }, { "epoch": 1.6904973921892887, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.695694923400879, "learning_rate": 5e-05, "loss": 0.9709, "mean_token_accuracy": 0.875045895576477, "num_tokens": 504452919.0, "step": 13289 }, { "epoch": 1.6906246024678793, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.632798194885254, "learning_rate": 5e-05, "loss": 0.8695, "mean_token_accuracy": 0.8740643262863159, "num_tokens": 504490694.0, "step": 13290 }, { "epoch": 1.6907518127464698, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.969245910644531, "learning_rate": 5e-05, "loss": 0.9135, "mean_token_accuracy": 0.8670114278793335, "num_tokens": 504526606.0, "step": 13291 }, { "epoch": 1.6908790230250603, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.373250007629395, "learning_rate": 5e-05, "loss": 0.81, "mean_token_accuracy": 0.8769958019256592, "num_tokens": 504568885.0, "step": 13292 }, { "epoch": 1.6910062333036509, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.549736022949219, "learning_rate": 5e-05, "loss": 0.8264, "mean_token_accuracy": 0.8862409591674805, "num_tokens": 504604947.0, "step": 13293 }, { "epoch": 1.6911334435822414, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.880093097686768, "learning_rate": 5e-05, "loss": 0.798, "mean_token_accuracy": 0.8638042211532593, "num_tokens": 504643937.0, "step": 13294 }, { "epoch": 1.691260653860832, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.52310562133789, "learning_rate": 5e-05, "loss": 0.8619, "mean_token_accuracy": 0.8708381056785583, "num_tokens": 504684298.0, "step": 13295 }, { "epoch": 1.6913878641394224, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.502783298492432, "learning_rate": 5e-05, "loss": 0.796, "mean_token_accuracy": 0.8601594567298889, "num_tokens": 504723578.0, "step": 13296 }, { "epoch": 1.691515074418013, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.615484237670898, "learning_rate": 5e-05, "loss": 0.8612, "mean_token_accuracy": 0.8635251522064209, "num_tokens": 504768190.0, "step": 13297 }, { "epoch": 1.6916422846966035, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.339745044708252, "learning_rate": 5e-05, "loss": 0.6943, "mean_token_accuracy": 0.8836672902107239, "num_tokens": 504807656.0, "step": 13298 }, { "epoch": 1.691769494975194, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.701027870178223, "learning_rate": 5e-05, "loss": 0.8489, "mean_token_accuracy": 0.868614673614502, "num_tokens": 504846631.0, "step": 13299 }, { "epoch": 1.6918967052537845, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.2942914962768555, "learning_rate": 5e-05, "loss": 0.7234, "mean_token_accuracy": 0.8748021125793457, "num_tokens": 504890451.0, "step": 13300 }, { "epoch": 1.692023915532375, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.698911666870117, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.8844252824783325, "num_tokens": 504931863.0, "step": 13301 }, { "epoch": 1.6921511258109656, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.40773868560791, "learning_rate": 5e-05, "loss": 0.7085, "mean_token_accuracy": 0.8814841508865356, "num_tokens": 504970247.0, "step": 13302 }, { "epoch": 1.6922783360895561, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.512177467346191, "learning_rate": 5e-05, "loss": 0.8121, "mean_token_accuracy": 0.878277063369751, "num_tokens": 505009987.0, "step": 13303 }, { "epoch": 1.6924055463681467, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.629278659820557, "learning_rate": 5e-05, "loss": 0.7321, "mean_token_accuracy": 0.8789196610450745, "num_tokens": 505051933.0, "step": 13304 }, { "epoch": 1.6925327566467372, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.047611236572266, "learning_rate": 5e-05, "loss": 0.7843, "mean_token_accuracy": 0.8749255537986755, "num_tokens": 505086216.0, "step": 13305 }, { "epoch": 1.6926599669253277, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.832759857177734, "learning_rate": 5e-05, "loss": 0.6844, "mean_token_accuracy": 0.8933138847351074, "num_tokens": 505123040.0, "step": 13306 }, { "epoch": 1.692787177203918, "ewc_loss": 0.375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.606637954711914, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.8751874566078186, "num_tokens": 505158546.0, "step": 13307 }, { "epoch": 1.6929143874825086, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 8.009296417236328, "learning_rate": 5e-05, "loss": 0.7735, "mean_token_accuracy": 0.871961236000061, "num_tokens": 505195433.0, "step": 13308 }, { "epoch": 1.693041597761099, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.292654991149902, "learning_rate": 5e-05, "loss": 0.688, "mean_token_accuracy": 0.8761311173439026, "num_tokens": 505232411.0, "step": 13309 }, { "epoch": 1.6931688080396896, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.22846794128418, "learning_rate": 5e-05, "loss": 0.8254, "mean_token_accuracy": 0.8587517142295837, "num_tokens": 505270590.0, "step": 13310 }, { "epoch": 1.6932960183182801, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.030837059020996, "learning_rate": 5e-05, "loss": 0.6931, "mean_token_accuracy": 0.8742054104804993, "num_tokens": 505307705.0, "step": 13311 }, { "epoch": 1.6934232285968707, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.476777076721191, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.8733528852462769, "num_tokens": 505342676.0, "step": 13312 }, { "epoch": 1.693550438875461, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.91331148147583, "learning_rate": 5e-05, "loss": 0.7446, "mean_token_accuracy": 0.8610774874687195, "num_tokens": 505387058.0, "step": 13313 }, { "epoch": 1.6936776491540515, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.4948148727417, "learning_rate": 5e-05, "loss": 0.835, "mean_token_accuracy": 0.8638980388641357, "num_tokens": 505429584.0, "step": 13314 }, { "epoch": 1.693804859432642, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.114588260650635, "learning_rate": 5e-05, "loss": 0.6836, "mean_token_accuracy": 0.8832998275756836, "num_tokens": 505464619.0, "step": 13315 }, { "epoch": 1.6939320697112326, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.175558090209961, "learning_rate": 5e-05, "loss": 0.8006, "mean_token_accuracy": 0.8674395084381104, "num_tokens": 505507390.0, "step": 13316 }, { "epoch": 1.694059279989823, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.4714202880859375, "learning_rate": 5e-05, "loss": 0.7271, "mean_token_accuracy": 0.8702191114425659, "num_tokens": 505543599.0, "step": 13317 }, { "epoch": 1.6941864902684136, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.727837085723877, "learning_rate": 5e-05, "loss": 0.7606, "mean_token_accuracy": 0.8711563348770142, "num_tokens": 505576283.0, "step": 13318 }, { "epoch": 1.6943137005470041, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.81960916519165, "learning_rate": 5e-05, "loss": 0.7626, "mean_token_accuracy": 0.86497962474823, "num_tokens": 505615244.0, "step": 13319 }, { "epoch": 1.6944409108255947, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.337684631347656, "learning_rate": 5e-05, "loss": 0.7447, "mean_token_accuracy": 0.8639036417007446, "num_tokens": 505653505.0, "step": 13320 }, { "epoch": 1.6945681211041852, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 8.096405982971191, "learning_rate": 5e-05, "loss": 0.7605, "mean_token_accuracy": 0.8736772537231445, "num_tokens": 505678666.0, "step": 13321 }, { "epoch": 1.6946953313827757, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.072688102722168, "learning_rate": 5e-05, "loss": 0.7351, "mean_token_accuracy": 0.8622450828552246, "num_tokens": 505718191.0, "step": 13322 }, { "epoch": 1.6948225416613663, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.245781898498535, "learning_rate": 5e-05, "loss": 0.7523, "mean_token_accuracy": 0.8878424167633057, "num_tokens": 505759174.0, "step": 13323 }, { "epoch": 1.6949497519399568, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.015743732452393, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.8583341836929321, "num_tokens": 505804483.0, "step": 13324 }, { "epoch": 1.6950769622185473, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00054931640625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.212549209594727, "learning_rate": 5e-05, "loss": 0.8036, "mean_token_accuracy": 0.8653353452682495, "num_tokens": 505839767.0, "step": 13325 }, { "epoch": 1.6952041724971378, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00055694580078125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.152317047119141, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8524488210678101, "num_tokens": 505880207.0, "step": 13326 }, { "epoch": 1.6953313827757284, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.9856343269348145, "learning_rate": 5e-05, "loss": 0.7542, "mean_token_accuracy": 0.8748107552528381, "num_tokens": 505914686.0, "step": 13327 }, { "epoch": 1.695458593054319, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.3927836418151855, "learning_rate": 5e-05, "loss": 0.7089, "mean_token_accuracy": 0.8736625909805298, "num_tokens": 505959029.0, "step": 13328 }, { "epoch": 1.6955858033329094, "ewc_loss": 0.375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.720958232879639, "learning_rate": 5e-05, "loss": 0.7757, "mean_token_accuracy": 0.8597390651702881, "num_tokens": 505997501.0, "step": 13329 }, { "epoch": 1.6957130136115, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.592918395996094, "learning_rate": 5e-05, "loss": 0.7546, "mean_token_accuracy": 0.8601411581039429, "num_tokens": 506029097.0, "step": 13330 }, { "epoch": 1.6958402238900905, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.5278849601745605, "learning_rate": 5e-05, "loss": 0.6938, "mean_token_accuracy": 0.8838592171669006, "num_tokens": 506065912.0, "step": 13331 }, { "epoch": 1.6959674341686808, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.6977996826171875, "learning_rate": 5e-05, "loss": 0.7296, "mean_token_accuracy": 0.877277135848999, "num_tokens": 506108843.0, "step": 13332 }, { "epoch": 1.6960946444472713, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.480427265167236, "learning_rate": 5e-05, "loss": 0.6785, "mean_token_accuracy": 0.8836730122566223, "num_tokens": 506147505.0, "step": 13333 }, { "epoch": 1.6962218547258618, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.617427349090576, "learning_rate": 5e-05, "loss": 0.7118, "mean_token_accuracy": 0.8779731392860413, "num_tokens": 506186491.0, "step": 13334 }, { "epoch": 1.6963490650044524, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.626937389373779, "learning_rate": 5e-05, "loss": 0.6985, "mean_token_accuracy": 0.8782474994659424, "num_tokens": 506223343.0, "step": 13335 }, { "epoch": 1.696476275283043, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.3739118576049805, "learning_rate": 5e-05, "loss": 0.7054, "mean_token_accuracy": 0.8755348324775696, "num_tokens": 506262328.0, "step": 13336 }, { "epoch": 1.6966034855616334, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.944091796875, "learning_rate": 5e-05, "loss": 0.7936, "mean_token_accuracy": 0.8604612350463867, "num_tokens": 506294758.0, "step": 13337 }, { "epoch": 1.6967306958402237, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.971681118011475, "learning_rate": 5e-05, "loss": 0.6517, "mean_token_accuracy": 0.8817315697669983, "num_tokens": 506325423.0, "step": 13338 }, { "epoch": 1.6968579061188143, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.472857475280762, "learning_rate": 5e-05, "loss": 0.8326, "mean_token_accuracy": 0.8600072860717773, "num_tokens": 506367722.0, "step": 13339 }, { "epoch": 1.6969851163974048, "ewc_loss": 0.28125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000225067138671875, "grad_norm": 6.3473801612854, "learning_rate": 5e-05, "loss": 0.6808, "mean_token_accuracy": 0.8607341051101685, "num_tokens": 506404638.0, "step": 13340 }, { "epoch": 1.6971123266759953, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 9.415602684020996, "learning_rate": 5e-05, "loss": 0.8713, "mean_token_accuracy": 0.8710138201713562, "num_tokens": 506440976.0, "step": 13341 }, { "epoch": 1.6972395369545858, "ewc_loss": 0.255859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002002716064453125, "grad_norm": 5.561528205871582, "learning_rate": 5e-05, "loss": 0.6329, "mean_token_accuracy": 0.8669235706329346, "num_tokens": 506475392.0, "step": 13342 }, { "epoch": 1.6973667472331764, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.830894470214844, "learning_rate": 5e-05, "loss": 1.0414, "mean_token_accuracy": 0.8716988563537598, "num_tokens": 506510933.0, "step": 13343 }, { "epoch": 1.697493957511767, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 5.86696720123291, "learning_rate": 5e-05, "loss": 0.7377, "mean_token_accuracy": 0.866270124912262, "num_tokens": 506550217.0, "step": 13344 }, { "epoch": 1.6976211677903574, "ewc_loss": 0.9375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000881195068359375, "grad_norm": 11.956426620483398, "learning_rate": 5e-05, "loss": 1.3074, "mean_token_accuracy": 0.8750585913658142, "num_tokens": 506583836.0, "step": 13345 }, { "epoch": 1.697748378068948, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 8.87486743927002, "learning_rate": 5e-05, "loss": 1.1576, "mean_token_accuracy": 0.8764932155609131, "num_tokens": 506616889.0, "step": 13346 }, { "epoch": 1.6978755883475385, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.513847351074219, "learning_rate": 5e-05, "loss": 1.0352, "mean_token_accuracy": 0.8770775198936462, "num_tokens": 506657198.0, "step": 13347 }, { "epoch": 1.698002798626129, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 9.881243705749512, "learning_rate": 5e-05, "loss": 1.2389, "mean_token_accuracy": 0.8804923295974731, "num_tokens": 506696439.0, "step": 13348 }, { "epoch": 1.6981300089047195, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.460737228393555, "learning_rate": 5e-05, "loss": 0.9159, "mean_token_accuracy": 0.8597246408462524, "num_tokens": 506736140.0, "step": 13349 }, { "epoch": 1.69825721918331, "ewc_loss": 0.875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 11.132806777954102, "learning_rate": 5e-05, "loss": 1.2207, "mean_token_accuracy": 0.880927562713623, "num_tokens": 506770976.0, "step": 13350 }, { "epoch": 1.6983844294619006, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.206404685974121, "learning_rate": 5e-05, "loss": 0.9692, "mean_token_accuracy": 0.8732059001922607, "num_tokens": 506804463.0, "step": 13351 }, { "epoch": 1.6985116397404911, "ewc_loss": 0.859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 11.901134490966797, "learning_rate": 5e-05, "loss": 1.1781, "mean_token_accuracy": 0.8897780179977417, "num_tokens": 506839498.0, "step": 13352 }, { "epoch": 1.6986388500190817, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.538206100463867, "learning_rate": 5e-05, "loss": 1.2154, "mean_token_accuracy": 0.8672876358032227, "num_tokens": 506875899.0, "step": 13353 }, { "epoch": 1.6987660602976722, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 11.119406700134277, "learning_rate": 5e-05, "loss": 1.0887, "mean_token_accuracy": 0.8639653921127319, "num_tokens": 506914760.0, "step": 13354 }, { "epoch": 1.6988932705762627, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 9.217842102050781, "learning_rate": 5e-05, "loss": 1.0933, "mean_token_accuracy": 0.8760676383972168, "num_tokens": 506950022.0, "step": 13355 }, { "epoch": 1.699020480854853, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.559006690979004, "learning_rate": 5e-05, "loss": 1.028, "mean_token_accuracy": 0.8764755129814148, "num_tokens": 506993349.0, "step": 13356 }, { "epoch": 1.6991476911334435, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.506516456604004, "learning_rate": 5e-05, "loss": 0.9989, "mean_token_accuracy": 0.8654686212539673, "num_tokens": 507037519.0, "step": 13357 }, { "epoch": 1.699274901412034, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 10.053918838500977, "learning_rate": 5e-05, "loss": 1.1047, "mean_token_accuracy": 0.8450579643249512, "num_tokens": 507071932.0, "step": 13358 }, { "epoch": 1.6994021116906246, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.949704170227051, "learning_rate": 5e-05, "loss": 0.8869, "mean_token_accuracy": 0.8712548613548279, "num_tokens": 507108140.0, "step": 13359 }, { "epoch": 1.6995293219692151, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.91879940032959, "learning_rate": 5e-05, "loss": 1.0211, "mean_token_accuracy": 0.8657808303833008, "num_tokens": 507148241.0, "step": 13360 }, { "epoch": 1.6996565322478057, "ewc_loss": 0.5, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.646024227142334, "learning_rate": 5e-05, "loss": 0.8673, "mean_token_accuracy": 0.8720482587814331, "num_tokens": 507183200.0, "step": 13361 }, { "epoch": 1.699783742526396, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.056244850158691, "learning_rate": 5e-05, "loss": 1.0062, "mean_token_accuracy": 0.8692031502723694, "num_tokens": 507219036.0, "step": 13362 }, { "epoch": 1.6999109528049865, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.736311912536621, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8839753866195679, "num_tokens": 507254767.0, "step": 13363 }, { "epoch": 1.700038163083577, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.978867530822754, "learning_rate": 5e-05, "loss": 0.9848, "mean_token_accuracy": 0.8875645399093628, "num_tokens": 507289945.0, "step": 13364 }, { "epoch": 1.7001653733621676, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.115200996398926, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.8709214925765991, "num_tokens": 507328350.0, "step": 13365 }, { "epoch": 1.700292583640758, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.452961921691895, "learning_rate": 5e-05, "loss": 1.0277, "mean_token_accuracy": 0.8591520190238953, "num_tokens": 507367151.0, "step": 13366 }, { "epoch": 1.7004197939193486, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.310498237609863, "learning_rate": 5e-05, "loss": 0.9205, "mean_token_accuracy": 0.8764681816101074, "num_tokens": 507404181.0, "step": 13367 }, { "epoch": 1.7005470041979391, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.700860023498535, "learning_rate": 5e-05, "loss": 0.8755, "mean_token_accuracy": 0.8836125135421753, "num_tokens": 507438402.0, "step": 13368 }, { "epoch": 1.7006742144765297, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.284629821777344, "learning_rate": 5e-05, "loss": 0.896, "mean_token_accuracy": 0.8738102912902832, "num_tokens": 507478091.0, "step": 13369 }, { "epoch": 1.7008014247551202, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.100336074829102, "learning_rate": 5e-05, "loss": 0.8564, "mean_token_accuracy": 0.8654743432998657, "num_tokens": 507518285.0, "step": 13370 }, { "epoch": 1.7009286350337107, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.06735610961914, "learning_rate": 5e-05, "loss": 0.8442, "mean_token_accuracy": 0.8723358511924744, "num_tokens": 507555250.0, "step": 13371 }, { "epoch": 1.7010558453123013, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.930160045623779, "learning_rate": 5e-05, "loss": 0.8022, "mean_token_accuracy": 0.8780384063720703, "num_tokens": 507591255.0, "step": 13372 }, { "epoch": 1.7011830555908918, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.7859697341918945, "learning_rate": 5e-05, "loss": 0.7805, "mean_token_accuracy": 0.8785833716392517, "num_tokens": 507623805.0, "step": 13373 }, { "epoch": 1.7013102658694823, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.87898063659668, "learning_rate": 5e-05, "loss": 0.8244, "mean_token_accuracy": 0.8622786998748779, "num_tokens": 507660312.0, "step": 13374 }, { "epoch": 1.7014374761480728, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.675848960876465, "learning_rate": 5e-05, "loss": 0.7716, "mean_token_accuracy": 0.8744038939476013, "num_tokens": 507700132.0, "step": 13375 }, { "epoch": 1.7015646864266634, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.788510799407959, "learning_rate": 5e-05, "loss": 0.7747, "mean_token_accuracy": 0.8726063370704651, "num_tokens": 507737521.0, "step": 13376 }, { "epoch": 1.701691896705254, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.63926362991333, "learning_rate": 5e-05, "loss": 0.694, "mean_token_accuracy": 0.8900055289268494, "num_tokens": 507779327.0, "step": 13377 }, { "epoch": 1.7018191069838444, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.798466205596924, "learning_rate": 5e-05, "loss": 0.7547, "mean_token_accuracy": 0.8745536804199219, "num_tokens": 507821899.0, "step": 13378 }, { "epoch": 1.701946317262435, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.408933639526367, "learning_rate": 5e-05, "loss": 0.6929, "mean_token_accuracy": 0.886932909488678, "num_tokens": 507862946.0, "step": 13379 }, { "epoch": 1.7020735275410255, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.082392692565918, "learning_rate": 5e-05, "loss": 0.8574, "mean_token_accuracy": 0.8521556854248047, "num_tokens": 507898409.0, "step": 13380 }, { "epoch": 1.7022007378196158, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.053531169891357, "learning_rate": 5e-05, "loss": 0.6872, "mean_token_accuracy": 0.8783861398696899, "num_tokens": 507936670.0, "step": 13381 }, { "epoch": 1.7023279480982063, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.47286319732666, "learning_rate": 5e-05, "loss": 0.8188, "mean_token_accuracy": 0.8729770183563232, "num_tokens": 507976160.0, "step": 13382 }, { "epoch": 1.7024551583767968, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.695693016052246, "learning_rate": 5e-05, "loss": 0.687, "mean_token_accuracy": 0.8760515451431274, "num_tokens": 508015500.0, "step": 13383 }, { "epoch": 1.7025823686553874, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 9.003711700439453, "learning_rate": 5e-05, "loss": 0.838, "mean_token_accuracy": 0.884677529335022, "num_tokens": 508050860.0, "step": 13384 }, { "epoch": 1.702709578933978, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.344964504241943, "learning_rate": 5e-05, "loss": 0.6959, "mean_token_accuracy": 0.869698703289032, "num_tokens": 508085794.0, "step": 13385 }, { "epoch": 1.7028367892125684, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.703170776367188, "learning_rate": 5e-05, "loss": 0.9245, "mean_token_accuracy": 0.885831892490387, "num_tokens": 508122731.0, "step": 13386 }, { "epoch": 1.7029639994911587, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.520938396453857, "learning_rate": 5e-05, "loss": 0.7283, "mean_token_accuracy": 0.8716034889221191, "num_tokens": 508168513.0, "step": 13387 }, { "epoch": 1.7030912097697493, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.905081748962402, "learning_rate": 5e-05, "loss": 0.9801, "mean_token_accuracy": 0.8885441422462463, "num_tokens": 508209758.0, "step": 13388 }, { "epoch": 1.7032184200483398, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.744075298309326, "learning_rate": 5e-05, "loss": 0.8693, "mean_token_accuracy": 0.882618248462677, "num_tokens": 508248172.0, "step": 13389 }, { "epoch": 1.7033456303269303, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.777400016784668, "learning_rate": 5e-05, "loss": 0.9582, "mean_token_accuracy": 0.8665884733200073, "num_tokens": 508288694.0, "step": 13390 }, { "epoch": 1.7034728406055208, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.77917194366455, "learning_rate": 5e-05, "loss": 0.9731, "mean_token_accuracy": 0.8676621913909912, "num_tokens": 508329021.0, "step": 13391 }, { "epoch": 1.7036000508841114, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.134023189544678, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.8579424619674683, "num_tokens": 508363706.0, "step": 13392 }, { "epoch": 1.703727261162702, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.514583587646484, "learning_rate": 5e-05, "loss": 0.9378, "mean_token_accuracy": 0.8823097348213196, "num_tokens": 508399879.0, "step": 13393 }, { "epoch": 1.7038544714412924, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.7028093338012695, "learning_rate": 5e-05, "loss": 0.7508, "mean_token_accuracy": 0.8663158416748047, "num_tokens": 508440172.0, "step": 13394 }, { "epoch": 1.703981681719883, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.674820899963379, "learning_rate": 5e-05, "loss": 1.0092, "mean_token_accuracy": 0.8685711622238159, "num_tokens": 508475531.0, "step": 13395 }, { "epoch": 1.7041088919984735, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.406157493591309, "learning_rate": 5e-05, "loss": 0.8132, "mean_token_accuracy": 0.8793137073516846, "num_tokens": 508509290.0, "step": 13396 }, { "epoch": 1.704236102277064, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.996133804321289, "learning_rate": 5e-05, "loss": 0.9425, "mean_token_accuracy": 0.8764146566390991, "num_tokens": 508545052.0, "step": 13397 }, { "epoch": 1.7043633125556545, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.049906730651855, "learning_rate": 5e-05, "loss": 0.8916, "mean_token_accuracy": 0.8758348226547241, "num_tokens": 508577922.0, "step": 13398 }, { "epoch": 1.704490522834245, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.183655738830566, "learning_rate": 5e-05, "loss": 0.8503, "mean_token_accuracy": 0.8771294355392456, "num_tokens": 508618135.0, "step": 13399 }, { "epoch": 1.7046177331128356, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.054966926574707, "learning_rate": 5e-05, "loss": 0.9115, "mean_token_accuracy": 0.8603346943855286, "num_tokens": 508656665.0, "step": 13400 }, { "epoch": 1.7047449433914261, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.980001926422119, "learning_rate": 5e-05, "loss": 0.8236, "mean_token_accuracy": 0.8712880611419678, "num_tokens": 508693898.0, "step": 13401 }, { "epoch": 1.7048721536700167, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.652381896972656, "learning_rate": 5e-05, "loss": 0.8431, "mean_token_accuracy": 0.8654006719589233, "num_tokens": 508726865.0, "step": 13402 }, { "epoch": 1.7049993639486072, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.1741304397583, "learning_rate": 5e-05, "loss": 0.8344, "mean_token_accuracy": 0.8749567270278931, "num_tokens": 508762040.0, "step": 13403 }, { "epoch": 1.7051265742271977, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.271780967712402, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8710336685180664, "num_tokens": 508803288.0, "step": 13404 }, { "epoch": 1.705253784505788, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.345966339111328, "learning_rate": 5e-05, "loss": 0.8169, "mean_token_accuracy": 0.8790627717971802, "num_tokens": 508840729.0, "step": 13405 }, { "epoch": 1.7053809947843785, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.14176082611084, "learning_rate": 5e-05, "loss": 0.6967, "mean_token_accuracy": 0.8813565969467163, "num_tokens": 508874219.0, "step": 13406 }, { "epoch": 1.705508205062969, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.358099937438965, "learning_rate": 5e-05, "loss": 0.7804, "mean_token_accuracy": 0.8917596936225891, "num_tokens": 508911746.0, "step": 13407 }, { "epoch": 1.7056354153415596, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.121579647064209, "learning_rate": 5e-05, "loss": 0.7061, "mean_token_accuracy": 0.8780619502067566, "num_tokens": 508944996.0, "step": 13408 }, { "epoch": 1.7057626256201501, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.394063949584961, "learning_rate": 5e-05, "loss": 0.8595, "mean_token_accuracy": 0.8640735149383545, "num_tokens": 508986416.0, "step": 13409 }, { "epoch": 1.7058898358987407, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.992459774017334, "learning_rate": 5e-05, "loss": 0.7239, "mean_token_accuracy": 0.871008574962616, "num_tokens": 509023690.0, "step": 13410 }, { "epoch": 1.706017046177331, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.592708587646484, "learning_rate": 5e-05, "loss": 0.8616, "mean_token_accuracy": 0.8743069171905518, "num_tokens": 509065405.0, "step": 13411 }, { "epoch": 1.7061442564559215, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.842748641967773, "learning_rate": 5e-05, "loss": 0.7385, "mean_token_accuracy": 0.8688011765480042, "num_tokens": 509104871.0, "step": 13412 }, { "epoch": 1.706271466734512, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.76876163482666, "learning_rate": 5e-05, "loss": 0.9163, "mean_token_accuracy": 0.8551981449127197, "num_tokens": 509140938.0, "step": 13413 }, { "epoch": 1.7063986770131025, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.897571086883545, "learning_rate": 5e-05, "loss": 0.7243, "mean_token_accuracy": 0.8752545118331909, "num_tokens": 509175324.0, "step": 13414 }, { "epoch": 1.706525887291693, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.67552661895752, "learning_rate": 5e-05, "loss": 0.8789, "mean_token_accuracy": 0.8744255304336548, "num_tokens": 509214983.0, "step": 13415 }, { "epoch": 1.7066530975702836, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.204421520233154, "learning_rate": 5e-05, "loss": 0.7905, "mean_token_accuracy": 0.8643522262573242, "num_tokens": 509249418.0, "step": 13416 }, { "epoch": 1.7067803078488741, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.28431224822998, "learning_rate": 5e-05, "loss": 0.8336, "mean_token_accuracy": 0.8797396421432495, "num_tokens": 509282884.0, "step": 13417 }, { "epoch": 1.7069075181274647, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.568266868591309, "learning_rate": 5e-05, "loss": 0.7879, "mean_token_accuracy": 0.8747279047966003, "num_tokens": 509317880.0, "step": 13418 }, { "epoch": 1.7070347284060552, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.7961297035217285, "learning_rate": 5e-05, "loss": 0.8212, "mean_token_accuracy": 0.866461992263794, "num_tokens": 509355760.0, "step": 13419 }, { "epoch": 1.7071619386846457, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.872115612030029, "learning_rate": 5e-05, "loss": 0.8, "mean_token_accuracy": 0.8748085498809814, "num_tokens": 509393088.0, "step": 13420 }, { "epoch": 1.7072891489632362, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.401339054107666, "learning_rate": 5e-05, "loss": 0.7741, "mean_token_accuracy": 0.8654703497886658, "num_tokens": 509431568.0, "step": 13421 }, { "epoch": 1.7074163592418268, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.06147289276123, "learning_rate": 5e-05, "loss": 0.8071, "mean_token_accuracy": 0.8701711893081665, "num_tokens": 509471044.0, "step": 13422 }, { "epoch": 1.7075435695204173, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.174272537231445, "learning_rate": 5e-05, "loss": 0.6964, "mean_token_accuracy": 0.8821572661399841, "num_tokens": 509507746.0, "step": 13423 }, { "epoch": 1.7076707797990078, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.209632873535156, "learning_rate": 5e-05, "loss": 0.7665, "mean_token_accuracy": 0.8832510709762573, "num_tokens": 509545753.0, "step": 13424 }, { "epoch": 1.7077979900775984, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.9957275390625, "learning_rate": 5e-05, "loss": 0.6984, "mean_token_accuracy": 0.8747022151947021, "num_tokens": 509581155.0, "step": 13425 }, { "epoch": 1.7079252003561889, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.431478500366211, "learning_rate": 5e-05, "loss": 0.8464, "mean_token_accuracy": 0.865654468536377, "num_tokens": 509615540.0, "step": 13426 }, { "epoch": 1.7080524106347794, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.761763572692871, "learning_rate": 5e-05, "loss": 0.7075, "mean_token_accuracy": 0.8693380951881409, "num_tokens": 509651257.0, "step": 13427 }, { "epoch": 1.70817962091337, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.749098777770996, "learning_rate": 5e-05, "loss": 0.9205, "mean_token_accuracy": 0.8540712594985962, "num_tokens": 509688088.0, "step": 13428 }, { "epoch": 1.7083068311919605, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.581982612609863, "learning_rate": 5e-05, "loss": 0.6911, "mean_token_accuracy": 0.8753613233566284, "num_tokens": 509721122.0, "step": 13429 }, { "epoch": 1.7084340414705508, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 9.024205207824707, "learning_rate": 5e-05, "loss": 0.9696, "mean_token_accuracy": 0.8524919748306274, "num_tokens": 509756711.0, "step": 13430 }, { "epoch": 1.7085612517491413, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.648082733154297, "learning_rate": 5e-05, "loss": 0.7264, "mean_token_accuracy": 0.8713855743408203, "num_tokens": 509793868.0, "step": 13431 }, { "epoch": 1.7086884620277318, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.97203254699707, "learning_rate": 5e-05, "loss": 0.8992, "mean_token_accuracy": 0.8807361125946045, "num_tokens": 509828741.0, "step": 13432 }, { "epoch": 1.7088156723063224, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.126976013183594, "learning_rate": 5e-05, "loss": 0.7766, "mean_token_accuracy": 0.880503237247467, "num_tokens": 509866534.0, "step": 13433 }, { "epoch": 1.708942882584913, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.390746116638184, "learning_rate": 5e-05, "loss": 0.8772, "mean_token_accuracy": 0.8706303834915161, "num_tokens": 509897664.0, "step": 13434 }, { "epoch": 1.7090700928635034, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.767508029937744, "learning_rate": 5e-05, "loss": 0.9392, "mean_token_accuracy": 0.846900463104248, "num_tokens": 509933349.0, "step": 13435 }, { "epoch": 1.7091973031420937, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.568723678588867, "learning_rate": 5e-05, "loss": 0.798, "mean_token_accuracy": 0.8695515394210815, "num_tokens": 509970912.0, "step": 13436 }, { "epoch": 1.7093245134206843, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.209271430969238, "learning_rate": 5e-05, "loss": 0.8869, "mean_token_accuracy": 0.8670488595962524, "num_tokens": 510007431.0, "step": 13437 }, { "epoch": 1.7094517236992748, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.148630619049072, "learning_rate": 5e-05, "loss": 0.7556, "mean_token_accuracy": 0.869766116142273, "num_tokens": 510044109.0, "step": 13438 }, { "epoch": 1.7095789339778653, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.265854835510254, "learning_rate": 5e-05, "loss": 0.8697, "mean_token_accuracy": 0.8557381629943848, "num_tokens": 510084665.0, "step": 13439 }, { "epoch": 1.7097061442564558, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.1897711753845215, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.8585262298583984, "num_tokens": 510123108.0, "step": 13440 }, { "epoch": 1.7098333545350464, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.014060974121094, "learning_rate": 5e-05, "loss": 0.8247, "mean_token_accuracy": 0.8666373491287231, "num_tokens": 510167918.0, "step": 13441 }, { "epoch": 1.709960564813637, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.442671298980713, "learning_rate": 5e-05, "loss": 0.724, "mean_token_accuracy": 0.880751371383667, "num_tokens": 510204555.0, "step": 13442 }, { "epoch": 1.7100877750922274, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.708515644073486, "learning_rate": 5e-05, "loss": 0.8217, "mean_token_accuracy": 0.8532588481903076, "num_tokens": 510239508.0, "step": 13443 }, { "epoch": 1.710214985370818, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.651210308074951, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8736885786056519, "num_tokens": 510275246.0, "step": 13444 }, { "epoch": 1.7103421956494085, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.451760292053223, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8718208074569702, "num_tokens": 510316558.0, "step": 13445 }, { "epoch": 1.710469405927999, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.840285301208496, "learning_rate": 5e-05, "loss": 0.7576, "mean_token_accuracy": 0.8766987323760986, "num_tokens": 510351784.0, "step": 13446 }, { "epoch": 1.7105966162065895, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.255373954772949, "learning_rate": 5e-05, "loss": 0.7013, "mean_token_accuracy": 0.8781615495681763, "num_tokens": 510391044.0, "step": 13447 }, { "epoch": 1.71072382648518, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.045716285705566, "learning_rate": 5e-05, "loss": 0.7854, "mean_token_accuracy": 0.8727155327796936, "num_tokens": 510431199.0, "step": 13448 }, { "epoch": 1.7108510367637706, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.00335693359375, "learning_rate": 5e-05, "loss": 0.6919, "mean_token_accuracy": 0.876422643661499, "num_tokens": 510471629.0, "step": 13449 }, { "epoch": 1.7109782470423611, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.342977523803711, "learning_rate": 5e-05, "loss": 0.7576, "mean_token_accuracy": 0.8851197957992554, "num_tokens": 510511546.0, "step": 13450 }, { "epoch": 1.7111054573209517, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.702675819396973, "learning_rate": 5e-05, "loss": 0.727, "mean_token_accuracy": 0.8629181385040283, "num_tokens": 510548942.0, "step": 13451 }, { "epoch": 1.7112326675995422, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.787335395812988, "learning_rate": 5e-05, "loss": 0.9287, "mean_token_accuracy": 0.8445513248443604, "num_tokens": 510589855.0, "step": 13452 }, { "epoch": 1.7113598778781327, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.290748596191406, "learning_rate": 5e-05, "loss": 0.6696, "mean_token_accuracy": 0.8740130066871643, "num_tokens": 510628960.0, "step": 13453 }, { "epoch": 1.711487088156723, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.45055103302002, "learning_rate": 5e-05, "loss": 0.9543, "mean_token_accuracy": 0.8668249249458313, "num_tokens": 510671107.0, "step": 13454 }, { "epoch": 1.7116142984353135, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.086975574493408, "learning_rate": 5e-05, "loss": 0.7125, "mean_token_accuracy": 0.868827760219574, "num_tokens": 510708957.0, "step": 13455 }, { "epoch": 1.711741508713904, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.94272518157959, "learning_rate": 5e-05, "loss": 1.0035, "mean_token_accuracy": 0.8722423315048218, "num_tokens": 510746150.0, "step": 13456 }, { "epoch": 1.7118687189924946, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 6.817056655883789, "learning_rate": 5e-05, "loss": 0.8646, "mean_token_accuracy": 0.8620123863220215, "num_tokens": 510775810.0, "step": 13457 }, { "epoch": 1.7119959292710851, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.430697441101074, "learning_rate": 5e-05, "loss": 1.0427, "mean_token_accuracy": 0.8580890893936157, "num_tokens": 510817438.0, "step": 13458 }, { "epoch": 1.7121231395496757, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.144489288330078, "learning_rate": 5e-05, "loss": 0.9899, "mean_token_accuracy": 0.8674507737159729, "num_tokens": 510854529.0, "step": 13459 }, { "epoch": 1.712250349828266, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.739021301269531, "learning_rate": 5e-05, "loss": 0.845, "mean_token_accuracy": 0.870060920715332, "num_tokens": 510891568.0, "step": 13460 }, { "epoch": 1.7123775601068565, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.228910446166992, "learning_rate": 5e-05, "loss": 0.984, "mean_token_accuracy": 0.8752167820930481, "num_tokens": 510928424.0, "step": 13461 }, { "epoch": 1.712504770385447, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.502485752105713, "learning_rate": 5e-05, "loss": 0.7561, "mean_token_accuracy": 0.8653647899627686, "num_tokens": 510971770.0, "step": 13462 }, { "epoch": 1.7126319806640375, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.952776908874512, "learning_rate": 5e-05, "loss": 1.0405, "mean_token_accuracy": 0.8630021810531616, "num_tokens": 511004087.0, "step": 13463 }, { "epoch": 1.712759190942628, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 6.872377872467041, "learning_rate": 5e-05, "loss": 0.886, "mean_token_accuracy": 0.8626444339752197, "num_tokens": 511046494.0, "step": 13464 }, { "epoch": 1.7128864012212186, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.603759765625, "learning_rate": 5e-05, "loss": 1.0433, "mean_token_accuracy": 0.8635903596878052, "num_tokens": 511088729.0, "step": 13465 }, { "epoch": 1.7130136114998091, "ewc_loss": 0.625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.028904914855957, "learning_rate": 5e-05, "loss": 0.9753, "mean_token_accuracy": 0.8736239075660706, "num_tokens": 511127581.0, "step": 13466 }, { "epoch": 1.7131408217783997, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.294002532958984, "learning_rate": 5e-05, "loss": 0.9122, "mean_token_accuracy": 0.8650360107421875, "num_tokens": 511166072.0, "step": 13467 }, { "epoch": 1.7132680320569902, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.638137817382812, "learning_rate": 5e-05, "loss": 0.9935, "mean_token_accuracy": 0.8687463998794556, "num_tokens": 511203529.0, "step": 13468 }, { "epoch": 1.7133952423355807, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.460684299468994, "learning_rate": 5e-05, "loss": 0.8205, "mean_token_accuracy": 0.8696457147598267, "num_tokens": 511245846.0, "step": 13469 }, { "epoch": 1.7135224526141712, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.617506980895996, "learning_rate": 5e-05, "loss": 0.9581, "mean_token_accuracy": 0.8649619817733765, "num_tokens": 511287280.0, "step": 13470 }, { "epoch": 1.7136496628927618, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.5009636878967285, "learning_rate": 5e-05, "loss": 0.8435, "mean_token_accuracy": 0.8727860450744629, "num_tokens": 511330905.0, "step": 13471 }, { "epoch": 1.7137768731713523, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.14923095703125, "learning_rate": 5e-05, "loss": 0.9158, "mean_token_accuracy": 0.8590478897094727, "num_tokens": 511369872.0, "step": 13472 }, { "epoch": 1.7139040834499428, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.740573883056641, "learning_rate": 5e-05, "loss": 0.8391, "mean_token_accuracy": 0.8770052194595337, "num_tokens": 511410638.0, "step": 13473 }, { "epoch": 1.7140312937285334, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.721707344055176, "learning_rate": 5e-05, "loss": 0.7892, "mean_token_accuracy": 0.8741253018379211, "num_tokens": 511445517.0, "step": 13474 }, { "epoch": 1.7141585040071239, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.853710651397705, "learning_rate": 5e-05, "loss": 0.8758, "mean_token_accuracy": 0.8571231961250305, "num_tokens": 511482749.0, "step": 13475 }, { "epoch": 1.7142857142857144, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.4813008308410645, "learning_rate": 5e-05, "loss": 0.8036, "mean_token_accuracy": 0.8616642951965332, "num_tokens": 511517001.0, "step": 13476 }, { "epoch": 1.714412924564305, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.9625983238220215, "learning_rate": 5e-05, "loss": 0.8505, "mean_token_accuracy": 0.8653349280357361, "num_tokens": 511559437.0, "step": 13477 }, { "epoch": 1.7145401348428955, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.257208824157715, "learning_rate": 5e-05, "loss": 0.7548, "mean_token_accuracy": 0.8728882670402527, "num_tokens": 511593713.0, "step": 13478 }, { "epoch": 1.7146673451214858, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.200915336608887, "learning_rate": 5e-05, "loss": 0.8659, "mean_token_accuracy": 0.8607503175735474, "num_tokens": 511633515.0, "step": 13479 }, { "epoch": 1.7147945554000763, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.987048625946045, "learning_rate": 5e-05, "loss": 0.7301, "mean_token_accuracy": 0.8687150478363037, "num_tokens": 511671539.0, "step": 13480 }, { "epoch": 1.7149217656786668, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.458630561828613, "learning_rate": 5e-05, "loss": 0.8124, "mean_token_accuracy": 0.879960298538208, "num_tokens": 511706972.0, "step": 13481 }, { "epoch": 1.7150489759572574, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.7797465324401855, "learning_rate": 5e-05, "loss": 0.7402, "mean_token_accuracy": 0.8660918474197388, "num_tokens": 511740186.0, "step": 13482 }, { "epoch": 1.7151761862358479, "ewc_loss": 0.5, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.718873023986816, "learning_rate": 5e-05, "loss": 0.8677, "mean_token_accuracy": 0.8679012060165405, "num_tokens": 511779755.0, "step": 13483 }, { "epoch": 1.7153033965144384, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.61067533493042, "learning_rate": 5e-05, "loss": 0.7501, "mean_token_accuracy": 0.8630624413490295, "num_tokens": 511822531.0, "step": 13484 }, { "epoch": 1.7154306067930287, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.013386726379395, "learning_rate": 5e-05, "loss": 0.92, "mean_token_accuracy": 0.8673415184020996, "num_tokens": 511852950.0, "step": 13485 }, { "epoch": 1.7155578170716193, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.52897834777832, "learning_rate": 5e-05, "loss": 0.7188, "mean_token_accuracy": 0.8782817125320435, "num_tokens": 511890251.0, "step": 13486 }, { "epoch": 1.7156850273502098, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.233994483947754, "learning_rate": 5e-05, "loss": 1.0098, "mean_token_accuracy": 0.8537307977676392, "num_tokens": 511933103.0, "step": 13487 }, { "epoch": 1.7158122376288003, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 6.77121639251709, "learning_rate": 5e-05, "loss": 0.776, "mean_token_accuracy": 0.8782459497451782, "num_tokens": 511975682.0, "step": 13488 }, { "epoch": 1.7159394479073908, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.063343048095703, "learning_rate": 5e-05, "loss": 0.9701, "mean_token_accuracy": 0.8687877655029297, "num_tokens": 512014899.0, "step": 13489 }, { "epoch": 1.7160666581859814, "ewc_loss": 0.5, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.327919960021973, "learning_rate": 5e-05, "loss": 0.8999, "mean_token_accuracy": 0.8603535890579224, "num_tokens": 512054951.0, "step": 13490 }, { "epoch": 1.716193868464572, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.47244930267334, "learning_rate": 5e-05, "loss": 0.9252, "mean_token_accuracy": 0.8697452545166016, "num_tokens": 512094273.0, "step": 13491 }, { "epoch": 1.7163210787431624, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.77290678024292, "learning_rate": 5e-05, "loss": 0.9032, "mean_token_accuracy": 0.8658503293991089, "num_tokens": 512133493.0, "step": 13492 }, { "epoch": 1.716448289021753, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.845758438110352, "learning_rate": 5e-05, "loss": 0.8513, "mean_token_accuracy": 0.8702397346496582, "num_tokens": 512171527.0, "step": 13493 }, { "epoch": 1.7165754993003435, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.907079219818115, "learning_rate": 5e-05, "loss": 0.8463, "mean_token_accuracy": 0.8755047917366028, "num_tokens": 512211106.0, "step": 13494 }, { "epoch": 1.716702709578934, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.649293422698975, "learning_rate": 5e-05, "loss": 0.7831, "mean_token_accuracy": 0.8768188953399658, "num_tokens": 512247444.0, "step": 13495 }, { "epoch": 1.7168299198575245, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.692020416259766, "learning_rate": 5e-05, "loss": 0.8391, "mean_token_accuracy": 0.8632255792617798, "num_tokens": 512293912.0, "step": 13496 }, { "epoch": 1.716957130136115, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.813242435455322, "learning_rate": 5e-05, "loss": 0.7751, "mean_token_accuracy": 0.881714940071106, "num_tokens": 512334149.0, "step": 13497 }, { "epoch": 1.7170843404147056, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.365631103515625, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8754165172576904, "num_tokens": 512371161.0, "step": 13498 }, { "epoch": 1.7172115506932961, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.022936820983887, "learning_rate": 5e-05, "loss": 0.8448, "mean_token_accuracy": 0.8598584532737732, "num_tokens": 512407217.0, "step": 13499 }, { "epoch": 1.7173387609718866, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.141977310180664, "learning_rate": 5e-05, "loss": 0.7033, "mean_token_accuracy": 0.8823884725570679, "num_tokens": 512441975.0, "step": 13500 }, { "epoch": 1.7174659712504772, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.187887191772461, "learning_rate": 5e-05, "loss": 0.8514, "mean_token_accuracy": 0.8634519577026367, "num_tokens": 512482956.0, "step": 13501 }, { "epoch": 1.7175931815290677, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.989389419555664, "learning_rate": 5e-05, "loss": 0.734, "mean_token_accuracy": 0.8691152930259705, "num_tokens": 512518514.0, "step": 13502 }, { "epoch": 1.717720391807658, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.358899116516113, "learning_rate": 5e-05, "loss": 0.8701, "mean_token_accuracy": 0.8592350482940674, "num_tokens": 512561620.0, "step": 13503 }, { "epoch": 1.7178476020862485, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.848099231719971, "learning_rate": 5e-05, "loss": 0.697, "mean_token_accuracy": 0.8729251027107239, "num_tokens": 512598234.0, "step": 13504 }, { "epoch": 1.717974812364839, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.537161827087402, "learning_rate": 5e-05, "loss": 0.8585, "mean_token_accuracy": 0.8696498870849609, "num_tokens": 512638219.0, "step": 13505 }, { "epoch": 1.7181020226434296, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.7985663414001465, "learning_rate": 5e-05, "loss": 0.7537, "mean_token_accuracy": 0.8604121804237366, "num_tokens": 512672277.0, "step": 13506 }, { "epoch": 1.7182292329220201, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.579297065734863, "learning_rate": 5e-05, "loss": 0.8431, "mean_token_accuracy": 0.8752032518386841, "num_tokens": 512709573.0, "step": 13507 }, { "epoch": 1.7183564432006107, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.930545330047607, "learning_rate": 5e-05, "loss": 0.7696, "mean_token_accuracy": 0.8643684387207031, "num_tokens": 512750243.0, "step": 13508 }, { "epoch": 1.718483653479201, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.415332794189453, "learning_rate": 5e-05, "loss": 0.9345, "mean_token_accuracy": 0.8480148315429688, "num_tokens": 512786001.0, "step": 13509 }, { "epoch": 1.7186108637577915, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.190042495727539, "learning_rate": 5e-05, "loss": 0.762, "mean_token_accuracy": 0.8759835958480835, "num_tokens": 512823844.0, "step": 13510 }, { "epoch": 1.718738074036382, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.047961235046387, "learning_rate": 5e-05, "loss": 0.8307, "mean_token_accuracy": 0.8724876642227173, "num_tokens": 512859474.0, "step": 13511 }, { "epoch": 1.7188652843149725, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.542303562164307, "learning_rate": 5e-05, "loss": 0.7569, "mean_token_accuracy": 0.8826833367347717, "num_tokens": 512889670.0, "step": 13512 }, { "epoch": 1.718992494593563, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.568868637084961, "learning_rate": 5e-05, "loss": 0.7541, "mean_token_accuracy": 0.8788223266601562, "num_tokens": 512919749.0, "step": 13513 }, { "epoch": 1.7191197048721536, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.871870517730713, "learning_rate": 5e-05, "loss": 0.8055, "mean_token_accuracy": 0.8695444464683533, "num_tokens": 512953206.0, "step": 13514 }, { "epoch": 1.7192469151507441, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.166613578796387, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.866205096244812, "num_tokens": 512989066.0, "step": 13515 }, { "epoch": 1.7193741254293347, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.163617134094238, "learning_rate": 5e-05, "loss": 0.7534, "mean_token_accuracy": 0.8883433938026428, "num_tokens": 513021975.0, "step": 13516 }, { "epoch": 1.7195013357079252, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.892592906951904, "learning_rate": 5e-05, "loss": 0.7156, "mean_token_accuracy": 0.8697414398193359, "num_tokens": 513061669.0, "step": 13517 }, { "epoch": 1.7196285459865157, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.429879188537598, "learning_rate": 5e-05, "loss": 0.8429, "mean_token_accuracy": 0.8669524192810059, "num_tokens": 513095714.0, "step": 13518 }, { "epoch": 1.7197557562651062, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.628384113311768, "learning_rate": 5e-05, "loss": 0.7123, "mean_token_accuracy": 0.8712681531906128, "num_tokens": 513137235.0, "step": 13519 }, { "epoch": 1.7198829665436968, "ewc_loss": 0.5, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.804255485534668, "learning_rate": 5e-05, "loss": 0.877, "mean_token_accuracy": 0.872721791267395, "num_tokens": 513176339.0, "step": 13520 }, { "epoch": 1.7200101768222873, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.402048110961914, "learning_rate": 5e-05, "loss": 0.6523, "mean_token_accuracy": 0.8879762887954712, "num_tokens": 513218984.0, "step": 13521 }, { "epoch": 1.7201373871008778, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.182780265808105, "learning_rate": 5e-05, "loss": 0.9701, "mean_token_accuracy": 0.8623402714729309, "num_tokens": 513259487.0, "step": 13522 }, { "epoch": 1.7202645973794684, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.508498668670654, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.86544269323349, "num_tokens": 513296270.0, "step": 13523 }, { "epoch": 1.7203918076580589, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.158294677734375, "learning_rate": 5e-05, "loss": 0.9863, "mean_token_accuracy": 0.8638721704483032, "num_tokens": 513333282.0, "step": 13524 }, { "epoch": 1.7205190179366494, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.16482400894165, "learning_rate": 5e-05, "loss": 0.814, "mean_token_accuracy": 0.874303936958313, "num_tokens": 513365679.0, "step": 13525 }, { "epoch": 1.72064622821524, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.381268501281738, "learning_rate": 5e-05, "loss": 0.9108, "mean_token_accuracy": 0.8707273006439209, "num_tokens": 513403387.0, "step": 13526 }, { "epoch": 1.7207734384938305, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.972200393676758, "learning_rate": 5e-05, "loss": 0.9833, "mean_token_accuracy": 0.8465501070022583, "num_tokens": 513442886.0, "step": 13527 }, { "epoch": 1.7209006487724208, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.351460933685303, "learning_rate": 5e-05, "loss": 0.7596, "mean_token_accuracy": 0.8778088688850403, "num_tokens": 513482401.0, "step": 13528 }, { "epoch": 1.7210278590510113, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.44742488861084, "learning_rate": 5e-05, "loss": 0.8767, "mean_token_accuracy": 0.8721818327903748, "num_tokens": 513517898.0, "step": 13529 }, { "epoch": 1.7211550693296018, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.968755722045898, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8726224303245544, "num_tokens": 513556761.0, "step": 13530 }, { "epoch": 1.7212822796081924, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.328618049621582, "learning_rate": 5e-05, "loss": 0.8877, "mean_token_accuracy": 0.8641378879547119, "num_tokens": 513590803.0, "step": 13531 }, { "epoch": 1.7214094898867829, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.24693489074707, "learning_rate": 5e-05, "loss": 0.7197, "mean_token_accuracy": 0.8918652534484863, "num_tokens": 513626718.0, "step": 13532 }, { "epoch": 1.7215367001653734, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.812227249145508, "learning_rate": 5e-05, "loss": 0.8258, "mean_token_accuracy": 0.8692978620529175, "num_tokens": 513663439.0, "step": 13533 }, { "epoch": 1.7216639104439637, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.673360824584961, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.8736045360565186, "num_tokens": 513699916.0, "step": 13534 }, { "epoch": 1.7217911207225542, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.3431549072265625, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.873889148235321, "num_tokens": 513739411.0, "step": 13535 }, { "epoch": 1.7219183310011448, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.975612163543701, "learning_rate": 5e-05, "loss": 0.8054, "mean_token_accuracy": 0.8731433153152466, "num_tokens": 513777497.0, "step": 13536 }, { "epoch": 1.7220455412797353, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.077512264251709, "learning_rate": 5e-05, "loss": 0.7482, "mean_token_accuracy": 0.865423321723938, "num_tokens": 513817658.0, "step": 13537 }, { "epoch": 1.7221727515583258, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.177861213684082, "learning_rate": 5e-05, "loss": 0.7764, "mean_token_accuracy": 0.8838924169540405, "num_tokens": 513853011.0, "step": 13538 }, { "epoch": 1.7222999618369164, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.881211280822754, "learning_rate": 5e-05, "loss": 0.654, "mean_token_accuracy": 0.8936747312545776, "num_tokens": 513892933.0, "step": 13539 }, { "epoch": 1.7224271721155069, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.409578323364258, "learning_rate": 5e-05, "loss": 0.8721, "mean_token_accuracy": 0.8582122921943665, "num_tokens": 513932284.0, "step": 13540 }, { "epoch": 1.7225543823940974, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.672512531280518, "learning_rate": 5e-05, "loss": 0.6809, "mean_token_accuracy": 0.878838062286377, "num_tokens": 513968151.0, "step": 13541 }, { "epoch": 1.722681592672688, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.720161437988281, "learning_rate": 5e-05, "loss": 0.8697, "mean_token_accuracy": 0.8709242343902588, "num_tokens": 514005323.0, "step": 13542 }, { "epoch": 1.7228088029512785, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.478888511657715, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.857873797416687, "num_tokens": 514044158.0, "step": 13543 }, { "epoch": 1.722936013229869, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.059869766235352, "learning_rate": 5e-05, "loss": 0.9186, "mean_token_accuracy": 0.8741409778594971, "num_tokens": 514081076.0, "step": 13544 }, { "epoch": 1.7230632235084595, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.4571661949157715, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8572688102722168, "num_tokens": 514122784.0, "step": 13545 }, { "epoch": 1.72319043378705, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.220760345458984, "learning_rate": 5e-05, "loss": 0.9674, "mean_token_accuracy": 0.8701397776603699, "num_tokens": 514158998.0, "step": 13546 }, { "epoch": 1.7233176440656406, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 6.8774518966674805, "learning_rate": 5e-05, "loss": 0.812, "mean_token_accuracy": 0.8668605089187622, "num_tokens": 514195815.0, "step": 13547 }, { "epoch": 1.7234448543442311, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.855374336242676, "learning_rate": 5e-05, "loss": 0.9654, "mean_token_accuracy": 0.8655887842178345, "num_tokens": 514228353.0, "step": 13548 }, { "epoch": 1.7235720646228216, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.54233980178833, "learning_rate": 5e-05, "loss": 0.8998, "mean_token_accuracy": 0.865814208984375, "num_tokens": 514267283.0, "step": 13549 }, { "epoch": 1.7236992749014122, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.038314819335938, "learning_rate": 5e-05, "loss": 0.8735, "mean_token_accuracy": 0.8765438199043274, "num_tokens": 514309950.0, "step": 13550 }, { "epoch": 1.7238264851800027, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.997989654541016, "learning_rate": 5e-05, "loss": 0.8371, "mean_token_accuracy": 0.885184109210968, "num_tokens": 514345043.0, "step": 13551 }, { "epoch": 1.723953695458593, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.451510429382324, "learning_rate": 5e-05, "loss": 0.8109, "mean_token_accuracy": 0.8721546530723572, "num_tokens": 514383692.0, "step": 13552 }, { "epoch": 1.7240809057371835, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.025761604309082, "learning_rate": 5e-05, "loss": 0.8924, "mean_token_accuracy": 0.857529878616333, "num_tokens": 514423552.0, "step": 13553 }, { "epoch": 1.724208116015774, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.45748233795166, "learning_rate": 5e-05, "loss": 0.809, "mean_token_accuracy": 0.8677866458892822, "num_tokens": 514460460.0, "step": 13554 }, { "epoch": 1.7243353262943646, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.6213860511779785, "learning_rate": 5e-05, "loss": 0.8062, "mean_token_accuracy": 0.8662863373756409, "num_tokens": 514504751.0, "step": 13555 }, { "epoch": 1.7244625365729551, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.837954521179199, "learning_rate": 5e-05, "loss": 0.818, "mean_token_accuracy": 0.8675001859664917, "num_tokens": 514539967.0, "step": 13556 }, { "epoch": 1.7245897468515456, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.155890941619873, "learning_rate": 5e-05, "loss": 0.738, "mean_token_accuracy": 0.87425297498703, "num_tokens": 514583191.0, "step": 13557 }, { "epoch": 1.724716957130136, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.146262168884277, "learning_rate": 5e-05, "loss": 0.8212, "mean_token_accuracy": 0.8738996982574463, "num_tokens": 514622832.0, "step": 13558 }, { "epoch": 1.7248441674087265, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.9104790687561035, "learning_rate": 5e-05, "loss": 0.7087, "mean_token_accuracy": 0.875864565372467, "num_tokens": 514663387.0, "step": 13559 }, { "epoch": 1.724971377687317, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.361722946166992, "learning_rate": 5e-05, "loss": 0.8597, "mean_token_accuracy": 0.8658192157745361, "num_tokens": 514702711.0, "step": 13560 }, { "epoch": 1.7250985879659075, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.777257442474365, "learning_rate": 5e-05, "loss": 0.7613, "mean_token_accuracy": 0.8564068675041199, "num_tokens": 514740341.0, "step": 13561 }, { "epoch": 1.725225798244498, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.551907539367676, "learning_rate": 5e-05, "loss": 0.9166, "mean_token_accuracy": 0.8552140593528748, "num_tokens": 514774834.0, "step": 13562 }, { "epoch": 1.7253530085230886, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.675938129425049, "learning_rate": 5e-05, "loss": 0.7597, "mean_token_accuracy": 0.8564151525497437, "num_tokens": 514814988.0, "step": 13563 }, { "epoch": 1.7254802188016791, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.723987579345703, "learning_rate": 5e-05, "loss": 0.8699, "mean_token_accuracy": 0.8771181106567383, "num_tokens": 514854936.0, "step": 13564 }, { "epoch": 1.7256074290802697, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.738764762878418, "learning_rate": 5e-05, "loss": 0.792, "mean_token_accuracy": 0.8527301549911499, "num_tokens": 514890528.0, "step": 13565 }, { "epoch": 1.7257346393588602, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.6557035446167, "learning_rate": 5e-05, "loss": 0.9147, "mean_token_accuracy": 0.8667299747467041, "num_tokens": 514935455.0, "step": 13566 }, { "epoch": 1.7258618496374507, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.06187105178833, "learning_rate": 5e-05, "loss": 0.8175, "mean_token_accuracy": 0.864082932472229, "num_tokens": 514975059.0, "step": 13567 }, { "epoch": 1.7259890599160412, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.250221252441406, "learning_rate": 5e-05, "loss": 0.8505, "mean_token_accuracy": 0.8779011964797974, "num_tokens": 515007820.0, "step": 13568 }, { "epoch": 1.7261162701946318, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.511015892028809, "learning_rate": 5e-05, "loss": 0.8662, "mean_token_accuracy": 0.8629812002182007, "num_tokens": 515051826.0, "step": 13569 }, { "epoch": 1.7262434804732223, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.623648643493652, "learning_rate": 5e-05, "loss": 0.8181, "mean_token_accuracy": 0.8658626079559326, "num_tokens": 515092316.0, "step": 13570 }, { "epoch": 1.7263706907518128, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.920777797698975, "learning_rate": 5e-05, "loss": 0.7936, "mean_token_accuracy": 0.8831051588058472, "num_tokens": 515124314.0, "step": 13571 }, { "epoch": 1.7264979010304033, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.127712726593018, "learning_rate": 5e-05, "loss": 0.7403, "mean_token_accuracy": 0.8698623180389404, "num_tokens": 515159725.0, "step": 13572 }, { "epoch": 1.7266251113089939, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.203617095947266, "learning_rate": 5e-05, "loss": 0.8747, "mean_token_accuracy": 0.8597795963287354, "num_tokens": 515201095.0, "step": 13573 }, { "epoch": 1.7267523215875844, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.839732646942139, "learning_rate": 5e-05, "loss": 0.7091, "mean_token_accuracy": 0.8729315996170044, "num_tokens": 515241761.0, "step": 13574 }, { "epoch": 1.726879531866175, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.473479270935059, "learning_rate": 5e-05, "loss": 0.8654, "mean_token_accuracy": 0.8691619634628296, "num_tokens": 515281983.0, "step": 13575 }, { "epoch": 1.7270067421447655, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.6138691902160645, "learning_rate": 5e-05, "loss": 0.736, "mean_token_accuracy": 0.8643680810928345, "num_tokens": 515315557.0, "step": 13576 }, { "epoch": 1.7271339524233558, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.81641674041748, "learning_rate": 5e-05, "loss": 0.9066, "mean_token_accuracy": 0.8634046912193298, "num_tokens": 515345969.0, "step": 13577 }, { "epoch": 1.7272611627019463, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.437748432159424, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.8646754622459412, "num_tokens": 515385520.0, "step": 13578 }, { "epoch": 1.7273883729805368, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.159965515136719, "learning_rate": 5e-05, "loss": 0.9955, "mean_token_accuracy": 0.8607670664787292, "num_tokens": 515424296.0, "step": 13579 }, { "epoch": 1.7275155832591274, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.638884544372559, "learning_rate": 5e-05, "loss": 0.7363, "mean_token_accuracy": 0.8804143667221069, "num_tokens": 515461602.0, "step": 13580 }, { "epoch": 1.7276427935377179, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.059635162353516, "learning_rate": 5e-05, "loss": 1.0075, "mean_token_accuracy": 0.8572525978088379, "num_tokens": 515504123.0, "step": 13581 }, { "epoch": 1.7277700038163084, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.329491138458252, "learning_rate": 5e-05, "loss": 0.868, "mean_token_accuracy": 0.8719724416732788, "num_tokens": 515543056.0, "step": 13582 }, { "epoch": 1.7278972140948987, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.247981071472168, "learning_rate": 5e-05, "loss": 0.9099, "mean_token_accuracy": 0.8667097687721252, "num_tokens": 515580194.0, "step": 13583 }, { "epoch": 1.7280244243734892, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.020318031311035, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.8745118379592896, "num_tokens": 515612493.0, "step": 13584 }, { "epoch": 1.7281516346520798, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.326785087585449, "learning_rate": 5e-05, "loss": 0.7629, "mean_token_accuracy": 0.8828715085983276, "num_tokens": 515656285.0, "step": 13585 }, { "epoch": 1.7282788449306703, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.322195053100586, "learning_rate": 5e-05, "loss": 0.8573, "mean_token_accuracy": 0.8760090470314026, "num_tokens": 515695326.0, "step": 13586 }, { "epoch": 1.7284060552092608, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.121548652648926, "learning_rate": 5e-05, "loss": 0.8222, "mean_token_accuracy": 0.8542960286140442, "num_tokens": 515730865.0, "step": 13587 }, { "epoch": 1.7285332654878514, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.012965202331543, "learning_rate": 5e-05, "loss": 0.8578, "mean_token_accuracy": 0.8622246980667114, "num_tokens": 515766725.0, "step": 13588 }, { "epoch": 1.7286604757664419, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.509552001953125, "learning_rate": 5e-05, "loss": 0.8171, "mean_token_accuracy": 0.8639466762542725, "num_tokens": 515807660.0, "step": 13589 }, { "epoch": 1.7287876860450324, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.404072284698486, "learning_rate": 5e-05, "loss": 0.7285, "mean_token_accuracy": 0.881263256072998, "num_tokens": 515844833.0, "step": 13590 }, { "epoch": 1.728914896323623, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.994395732879639, "learning_rate": 5e-05, "loss": 0.818, "mean_token_accuracy": 0.8702952265739441, "num_tokens": 515876538.0, "step": 13591 }, { "epoch": 1.7290421066022135, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.9123992919921875, "learning_rate": 5e-05, "loss": 0.7404, "mean_token_accuracy": 0.869177520275116, "num_tokens": 515916217.0, "step": 13592 }, { "epoch": 1.729169316880804, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.405452728271484, "learning_rate": 5e-05, "loss": 0.911, "mean_token_accuracy": 0.8501688838005066, "num_tokens": 515955129.0, "step": 13593 }, { "epoch": 1.7292965271593945, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.57855749130249, "learning_rate": 5e-05, "loss": 0.7235, "mean_token_accuracy": 0.8634567260742188, "num_tokens": 515993910.0, "step": 13594 }, { "epoch": 1.729423737437985, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.837780952453613, "learning_rate": 5e-05, "loss": 0.8457, "mean_token_accuracy": 0.885013222694397, "num_tokens": 516034376.0, "step": 13595 }, { "epoch": 1.7295509477165756, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.339752197265625, "learning_rate": 5e-05, "loss": 0.7371, "mean_token_accuracy": 0.8644859790802002, "num_tokens": 516070383.0, "step": 13596 }, { "epoch": 1.729678157995166, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.255321502685547, "learning_rate": 5e-05, "loss": 0.973, "mean_token_accuracy": 0.8714591860771179, "num_tokens": 516114527.0, "step": 13597 }, { "epoch": 1.7298053682737566, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.512181282043457, "learning_rate": 5e-05, "loss": 0.7297, "mean_token_accuracy": 0.8787487745285034, "num_tokens": 516156179.0, "step": 13598 }, { "epoch": 1.7299325785523472, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.266905784606934, "learning_rate": 5e-05, "loss": 0.9711, "mean_token_accuracy": 0.8752089142799377, "num_tokens": 516191000.0, "step": 13599 }, { "epoch": 1.7300597888309377, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.230852127075195, "learning_rate": 5e-05, "loss": 0.8592, "mean_token_accuracy": 0.8751747608184814, "num_tokens": 516224194.0, "step": 13600 }, { "epoch": 1.730186999109528, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.532488822937012, "learning_rate": 5e-05, "loss": 0.9326, "mean_token_accuracy": 0.8724462985992432, "num_tokens": 516262210.0, "step": 13601 }, { "epoch": 1.7303142093881185, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 7.96483039855957, "learning_rate": 5e-05, "loss": 0.945, "mean_token_accuracy": 0.8662671446800232, "num_tokens": 516298859.0, "step": 13602 }, { "epoch": 1.730441419666709, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.565589427947998, "learning_rate": 5e-05, "loss": 0.8593, "mean_token_accuracy": 0.8625502586364746, "num_tokens": 516344100.0, "step": 13603 }, { "epoch": 1.7305686299452996, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.240567207336426, "learning_rate": 5e-05, "loss": 0.9256, "mean_token_accuracy": 0.8628119230270386, "num_tokens": 516391313.0, "step": 13604 }, { "epoch": 1.7306958402238901, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.263956546783447, "learning_rate": 5e-05, "loss": 0.8328, "mean_token_accuracy": 0.8574002981185913, "num_tokens": 516430814.0, "step": 13605 }, { "epoch": 1.7308230505024806, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.015478134155273, "learning_rate": 5e-05, "loss": 0.8461, "mean_token_accuracy": 0.8710200786590576, "num_tokens": 516471380.0, "step": 13606 }, { "epoch": 1.730950260781071, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.451295852661133, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.8711248636245728, "num_tokens": 516518374.0, "step": 13607 }, { "epoch": 1.7310774710596615, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.635783672332764, "learning_rate": 5e-05, "loss": 0.8077, "mean_token_accuracy": 0.8672683238983154, "num_tokens": 516549917.0, "step": 13608 }, { "epoch": 1.731204681338252, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.607872009277344, "learning_rate": 5e-05, "loss": 0.8309, "mean_token_accuracy": 0.8560835719108582, "num_tokens": 516592869.0, "step": 13609 }, { "epoch": 1.7313318916168425, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.533214569091797, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8733848929405212, "num_tokens": 516630550.0, "step": 13610 }, { "epoch": 1.731459101895433, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.48988676071167, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.8778256177902222, "num_tokens": 516669285.0, "step": 13611 }, { "epoch": 1.7315863121740236, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.69780969619751, "learning_rate": 5e-05, "loss": 0.7442, "mean_token_accuracy": 0.8839308619499207, "num_tokens": 516704616.0, "step": 13612 }, { "epoch": 1.7317135224526141, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.212196350097656, "learning_rate": 5e-05, "loss": 0.7158, "mean_token_accuracy": 0.878254234790802, "num_tokens": 516749566.0, "step": 13613 }, { "epoch": 1.7318407327312046, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.002147674560547, "learning_rate": 5e-05, "loss": 0.7892, "mean_token_accuracy": 0.8739672899246216, "num_tokens": 516786264.0, "step": 13614 }, { "epoch": 1.7319679430097952, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.888418197631836, "learning_rate": 5e-05, "loss": 0.7297, "mean_token_accuracy": 0.8636918067932129, "num_tokens": 516822471.0, "step": 13615 }, { "epoch": 1.7320951532883857, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.353946685791016, "learning_rate": 5e-05, "loss": 0.8263, "mean_token_accuracy": 0.8688217997550964, "num_tokens": 516861840.0, "step": 13616 }, { "epoch": 1.7322223635669762, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.524770736694336, "learning_rate": 5e-05, "loss": 0.6997, "mean_token_accuracy": 0.86867356300354, "num_tokens": 516895451.0, "step": 13617 }, { "epoch": 1.7323495738455668, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.891814231872559, "learning_rate": 5e-05, "loss": 0.849, "mean_token_accuracy": 0.8789796829223633, "num_tokens": 516935926.0, "step": 13618 }, { "epoch": 1.7324767841241573, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.1569013595581055, "learning_rate": 5e-05, "loss": 0.6915, "mean_token_accuracy": 0.871425211429596, "num_tokens": 516976268.0, "step": 13619 }, { "epoch": 1.7326039944027478, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.54267692565918, "learning_rate": 5e-05, "loss": 0.9831, "mean_token_accuracy": 0.867617130279541, "num_tokens": 517021459.0, "step": 13620 }, { "epoch": 1.7327312046813383, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.262767314910889, "learning_rate": 5e-05, "loss": 0.7192, "mean_token_accuracy": 0.8767447471618652, "num_tokens": 517052441.0, "step": 13621 }, { "epoch": 1.7328584149599289, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.770224571228027, "learning_rate": 5e-05, "loss": 1.0687, "mean_token_accuracy": 0.8593109250068665, "num_tokens": 517094575.0, "step": 13622 }, { "epoch": 1.7329856252385194, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.3422417640686035, "learning_rate": 5e-05, "loss": 0.9165, "mean_token_accuracy": 0.8681106567382812, "num_tokens": 517135239.0, "step": 13623 }, { "epoch": 1.73311283551711, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.784968376159668, "learning_rate": 5e-05, "loss": 0.9624, "mean_token_accuracy": 0.8699089884757996, "num_tokens": 517174132.0, "step": 13624 }, { "epoch": 1.7332400457957005, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.517378807067871, "learning_rate": 5e-05, "loss": 0.9857, "mean_token_accuracy": 0.8738841414451599, "num_tokens": 517208693.0, "step": 13625 }, { "epoch": 1.7333672560742908, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.031937122344971, "learning_rate": 5e-05, "loss": 0.8412, "mean_token_accuracy": 0.8517606258392334, "num_tokens": 517249441.0, "step": 13626 }, { "epoch": 1.7334944663528813, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.595902442932129, "learning_rate": 5e-05, "loss": 1.0302, "mean_token_accuracy": 0.8664451837539673, "num_tokens": 517291368.0, "step": 13627 }, { "epoch": 1.7336216766314718, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.349000930786133, "learning_rate": 5e-05, "loss": 0.7799, "mean_token_accuracy": 0.8634771108627319, "num_tokens": 517328292.0, "step": 13628 }, { "epoch": 1.7337488869100623, "ewc_loss": 0.71875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.263263702392578, "learning_rate": 5e-05, "loss": 1.0753, "mean_token_accuracy": 0.8768205642700195, "num_tokens": 517361049.0, "step": 13629 }, { "epoch": 1.7338760971886529, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 7.4234185218811035, "learning_rate": 5e-05, "loss": 0.9695, "mean_token_accuracy": 0.8686903715133667, "num_tokens": 517395086.0, "step": 13630 }, { "epoch": 1.7340033074672434, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.626684188842773, "learning_rate": 5e-05, "loss": 1.028, "mean_token_accuracy": 0.8684777617454529, "num_tokens": 517432919.0, "step": 13631 }, { "epoch": 1.7341305177458337, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 8.563084602355957, "learning_rate": 5e-05, "loss": 1.0777, "mean_token_accuracy": 0.8682317733764648, "num_tokens": 517474676.0, "step": 13632 }, { "epoch": 1.7342577280244242, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.147619247436523, "learning_rate": 5e-05, "loss": 0.8576, "mean_token_accuracy": 0.8791160583496094, "num_tokens": 517514131.0, "step": 13633 }, { "epoch": 1.7343849383030148, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.918639183044434, "learning_rate": 5e-05, "loss": 1.001, "mean_token_accuracy": 0.873318076133728, "num_tokens": 517547704.0, "step": 13634 }, { "epoch": 1.7345121485816053, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.500447750091553, "learning_rate": 5e-05, "loss": 0.8545, "mean_token_accuracy": 0.8691270351409912, "num_tokens": 517586430.0, "step": 13635 }, { "epoch": 1.7346393588601958, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.642536163330078, "learning_rate": 5e-05, "loss": 1.0021, "mean_token_accuracy": 0.8567670583724976, "num_tokens": 517630303.0, "step": 13636 }, { "epoch": 1.7347665691387864, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.781364440917969, "learning_rate": 5e-05, "loss": 0.8396, "mean_token_accuracy": 0.8902626633644104, "num_tokens": 517663276.0, "step": 13637 }, { "epoch": 1.7348937794173769, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.7547221183776855, "learning_rate": 5e-05, "loss": 0.8363, "mean_token_accuracy": 0.8660346269607544, "num_tokens": 517697946.0, "step": 13638 }, { "epoch": 1.7350209896959674, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.215889930725098, "learning_rate": 5e-05, "loss": 0.9079, "mean_token_accuracy": 0.8693376183509827, "num_tokens": 517730820.0, "step": 13639 }, { "epoch": 1.735148199974558, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.109127521514893, "learning_rate": 5e-05, "loss": 0.7511, "mean_token_accuracy": 0.8747425079345703, "num_tokens": 517774576.0, "step": 13640 }, { "epoch": 1.7352754102531485, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.614293098449707, "learning_rate": 5e-05, "loss": 0.9409, "mean_token_accuracy": 0.8603616952896118, "num_tokens": 517814748.0, "step": 13641 }, { "epoch": 1.735402620531739, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.730474472045898, "learning_rate": 5e-05, "loss": 0.7418, "mean_token_accuracy": 0.8692721128463745, "num_tokens": 517846582.0, "step": 13642 }, { "epoch": 1.7355298308103295, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.1364164352417, "learning_rate": 5e-05, "loss": 1.0007, "mean_token_accuracy": 0.8551620244979858, "num_tokens": 517883354.0, "step": 13643 }, { "epoch": 1.73565704108892, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.649353981018066, "learning_rate": 5e-05, "loss": 0.821, "mean_token_accuracy": 0.8557467460632324, "num_tokens": 517920920.0, "step": 13644 }, { "epoch": 1.7357842513675106, "ewc_loss": 0.625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.466322898864746, "learning_rate": 5e-05, "loss": 1.0267, "mean_token_accuracy": 0.8579924702644348, "num_tokens": 517964844.0, "step": 13645 }, { "epoch": 1.735911461646101, "ewc_loss": 0.5, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.055407524108887, "learning_rate": 5e-05, "loss": 0.8804, "mean_token_accuracy": 0.8667945861816406, "num_tokens": 517998803.0, "step": 13646 }, { "epoch": 1.7360386719246916, "ewc_loss": 0.625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.28589153289795, "learning_rate": 5e-05, "loss": 1.0004, "mean_token_accuracy": 0.8676967024803162, "num_tokens": 518040012.0, "step": 13647 }, { "epoch": 1.7361658822032822, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.533376693725586, "learning_rate": 5e-05, "loss": 0.9525, "mean_token_accuracy": 0.8605939149856567, "num_tokens": 518082287.0, "step": 13648 }, { "epoch": 1.7362930924818727, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.877349853515625, "learning_rate": 5e-05, "loss": 0.9422, "mean_token_accuracy": 0.8716816902160645, "num_tokens": 518116898.0, "step": 13649 }, { "epoch": 1.736420302760463, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.539241313934326, "learning_rate": 5e-05, "loss": 0.8838, "mean_token_accuracy": 0.8763444423675537, "num_tokens": 518157559.0, "step": 13650 }, { "epoch": 1.7365475130390535, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.816370964050293, "learning_rate": 5e-05, "loss": 0.9966, "mean_token_accuracy": 0.862280011177063, "num_tokens": 518200691.0, "step": 13651 }, { "epoch": 1.736674723317644, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.253641605377197, "learning_rate": 5e-05, "loss": 0.9129, "mean_token_accuracy": 0.8552662134170532, "num_tokens": 518239188.0, "step": 13652 }, { "epoch": 1.7368019335962346, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.05903434753418, "learning_rate": 5e-05, "loss": 1.0017, "mean_token_accuracy": 0.8633549213409424, "num_tokens": 518274294.0, "step": 13653 }, { "epoch": 1.736929143874825, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.220297336578369, "learning_rate": 5e-05, "loss": 0.8192, "mean_token_accuracy": 0.8826084136962891, "num_tokens": 518310960.0, "step": 13654 }, { "epoch": 1.7370563541534156, "ewc_loss": 0.625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.055465698242188, "learning_rate": 5e-05, "loss": 1.0697, "mean_token_accuracy": 0.8498935699462891, "num_tokens": 518352133.0, "step": 13655 }, { "epoch": 1.737183564432006, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.602011680603027, "learning_rate": 5e-05, "loss": 0.9311, "mean_token_accuracy": 0.8601107597351074, "num_tokens": 518388793.0, "step": 13656 }, { "epoch": 1.7373107747105965, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.485946655273438, "learning_rate": 5e-05, "loss": 0.9234, "mean_token_accuracy": 0.8785416483879089, "num_tokens": 518425581.0, "step": 13657 }, { "epoch": 1.737437984989187, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 7.92971658706665, "learning_rate": 5e-05, "loss": 0.9118, "mean_token_accuracy": 0.8742703795433044, "num_tokens": 518465477.0, "step": 13658 }, { "epoch": 1.7375651952677775, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.821872711181641, "learning_rate": 5e-05, "loss": 0.8746, "mean_token_accuracy": 0.8612160682678223, "num_tokens": 518506790.0, "step": 13659 }, { "epoch": 1.737692405546368, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.944339275360107, "learning_rate": 5e-05, "loss": 0.8796, "mean_token_accuracy": 0.8702128529548645, "num_tokens": 518548240.0, "step": 13660 }, { "epoch": 1.7378196158249586, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.628061771392822, "learning_rate": 5e-05, "loss": 0.8635, "mean_token_accuracy": 0.8584340810775757, "num_tokens": 518583865.0, "step": 13661 }, { "epoch": 1.7379468261035491, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.674370765686035, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8791506886482239, "num_tokens": 518623303.0, "step": 13662 }, { "epoch": 1.7380740363821396, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.763339042663574, "learning_rate": 5e-05, "loss": 0.8289, "mean_token_accuracy": 0.8673628568649292, "num_tokens": 518662375.0, "step": 13663 }, { "epoch": 1.7382012466607302, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.409821510314941, "learning_rate": 5e-05, "loss": 0.8006, "mean_token_accuracy": 0.864293098449707, "num_tokens": 518702007.0, "step": 13664 }, { "epoch": 1.7383284569393207, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.856968879699707, "learning_rate": 5e-05, "loss": 0.7743, "mean_token_accuracy": 0.8805869817733765, "num_tokens": 518743676.0, "step": 13665 }, { "epoch": 1.7384556672179112, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.30925178527832, "learning_rate": 5e-05, "loss": 0.7369, "mean_token_accuracy": 0.8765914440155029, "num_tokens": 518783550.0, "step": 13666 }, { "epoch": 1.7385828774965018, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.831815242767334, "learning_rate": 5e-05, "loss": 0.7441, "mean_token_accuracy": 0.8859708309173584, "num_tokens": 518820788.0, "step": 13667 }, { "epoch": 1.7387100877750923, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.291223526000977, "learning_rate": 5e-05, "loss": 0.7829, "mean_token_accuracy": 0.8604097962379456, "num_tokens": 518865195.0, "step": 13668 }, { "epoch": 1.7388372980536828, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.847831726074219, "learning_rate": 5e-05, "loss": 0.7982, "mean_token_accuracy": 0.8683032989501953, "num_tokens": 518903683.0, "step": 13669 }, { "epoch": 1.7389645083322733, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.18098783493042, "learning_rate": 5e-05, "loss": 0.7038, "mean_token_accuracy": 0.8800880312919617, "num_tokens": 518939535.0, "step": 13670 }, { "epoch": 1.7390917186108639, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.003707885742188, "learning_rate": 5e-05, "loss": 0.8513, "mean_token_accuracy": 0.8576387763023376, "num_tokens": 518978753.0, "step": 13671 }, { "epoch": 1.7392189288894544, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.969526767730713, "learning_rate": 5e-05, "loss": 0.7488, "mean_token_accuracy": 0.8607203960418701, "num_tokens": 519014916.0, "step": 13672 }, { "epoch": 1.739346139168045, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.285897254943848, "learning_rate": 5e-05, "loss": 0.8373, "mean_token_accuracy": 0.8674542307853699, "num_tokens": 519043537.0, "step": 13673 }, { "epoch": 1.7394733494466355, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.696120738983154, "learning_rate": 5e-05, "loss": 0.6932, "mean_token_accuracy": 0.8732673525810242, "num_tokens": 519081388.0, "step": 13674 }, { "epoch": 1.7396005597252258, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.646349906921387, "learning_rate": 5e-05, "loss": 0.8344, "mean_token_accuracy": 0.879813015460968, "num_tokens": 519119531.0, "step": 13675 }, { "epoch": 1.7397277700038163, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.525300025939941, "learning_rate": 5e-05, "loss": 0.7613, "mean_token_accuracy": 0.8549587726593018, "num_tokens": 519157982.0, "step": 13676 }, { "epoch": 1.7398549802824068, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.902591705322266, "learning_rate": 5e-05, "loss": 0.9343, "mean_token_accuracy": 0.8646163940429688, "num_tokens": 519198928.0, "step": 13677 }, { "epoch": 1.7399821905609973, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.671205997467041, "learning_rate": 5e-05, "loss": 0.7549, "mean_token_accuracy": 0.8645392656326294, "num_tokens": 519237760.0, "step": 13678 }, { "epoch": 1.7401094008395879, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.7323637008667, "learning_rate": 5e-05, "loss": 0.9192, "mean_token_accuracy": 0.8712601661682129, "num_tokens": 519273557.0, "step": 13679 }, { "epoch": 1.7402366111181784, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000560760498046875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.239637851715088, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.8652006387710571, "num_tokens": 519316082.0, "step": 13680 }, { "epoch": 1.7403638213967687, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.992985248565674, "learning_rate": 5e-05, "loss": 0.8101, "mean_token_accuracy": 0.8838258385658264, "num_tokens": 519352528.0, "step": 13681 }, { "epoch": 1.7404910316753592, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.94480037689209, "learning_rate": 5e-05, "loss": 0.7994, "mean_token_accuracy": 0.8879866600036621, "num_tokens": 519384931.0, "step": 13682 }, { "epoch": 1.7406182419539498, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.035598278045654, "learning_rate": 5e-05, "loss": 0.7431, "mean_token_accuracy": 0.872352659702301, "num_tokens": 519421974.0, "step": 13683 }, { "epoch": 1.7407454522325403, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.5852689743042, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.8755994439125061, "num_tokens": 519460623.0, "step": 13684 }, { "epoch": 1.7408726625111308, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.4259867668151855, "learning_rate": 5e-05, "loss": 0.6702, "mean_token_accuracy": 0.8806865811347961, "num_tokens": 519497079.0, "step": 13685 }, { "epoch": 1.7409998727897213, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.166560173034668, "learning_rate": 5e-05, "loss": 0.9782, "mean_token_accuracy": 0.854155957698822, "num_tokens": 519542680.0, "step": 13686 }, { "epoch": 1.7411270830683119, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.266854763031006, "learning_rate": 5e-05, "loss": 0.7261, "mean_token_accuracy": 0.8672809600830078, "num_tokens": 519577025.0, "step": 13687 }, { "epoch": 1.7412542933469024, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.485603332519531, "learning_rate": 5e-05, "loss": 1.0033, "mean_token_accuracy": 0.8655688166618347, "num_tokens": 519618616.0, "step": 13688 }, { "epoch": 1.741381503625493, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 6.845794677734375, "learning_rate": 5e-05, "loss": 0.8214, "mean_token_accuracy": 0.8665705919265747, "num_tokens": 519658672.0, "step": 13689 }, { "epoch": 1.7415087139040835, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.048937797546387, "learning_rate": 5e-05, "loss": 0.9123, "mean_token_accuracy": 0.8919733166694641, "num_tokens": 519692830.0, "step": 13690 }, { "epoch": 1.741635924182674, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.820742607116699, "learning_rate": 5e-05, "loss": 0.9287, "mean_token_accuracy": 0.874148964881897, "num_tokens": 519729949.0, "step": 13691 }, { "epoch": 1.7417631344612645, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.885134220123291, "learning_rate": 5e-05, "loss": 0.8743, "mean_token_accuracy": 0.8677535057067871, "num_tokens": 519763556.0, "step": 13692 }, { "epoch": 1.741890344739855, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.39607048034668, "learning_rate": 5e-05, "loss": 0.9428, "mean_token_accuracy": 0.8612145781517029, "num_tokens": 519804556.0, "step": 13693 }, { "epoch": 1.7420175550184456, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.20607328414917, "learning_rate": 5e-05, "loss": 0.7687, "mean_token_accuracy": 0.8758825063705444, "num_tokens": 519845846.0, "step": 13694 }, { "epoch": 1.742144765297036, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.24441909790039, "learning_rate": 5e-05, "loss": 0.9077, "mean_token_accuracy": 0.8655971884727478, "num_tokens": 519883333.0, "step": 13695 }, { "epoch": 1.7422719755756266, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.4972991943359375, "learning_rate": 5e-05, "loss": 0.7525, "mean_token_accuracy": 0.8818180561065674, "num_tokens": 519915927.0, "step": 13696 }, { "epoch": 1.7423991858542172, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.451430320739746, "learning_rate": 5e-05, "loss": 0.7737, "mean_token_accuracy": 0.8713405132293701, "num_tokens": 519951555.0, "step": 13697 }, { "epoch": 1.7425263961328077, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.169411659240723, "learning_rate": 5e-05, "loss": 0.811, "mean_token_accuracy": 0.8806805610656738, "num_tokens": 519996616.0, "step": 13698 }, { "epoch": 1.742653606411398, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.7588934898376465, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8548246622085571, "num_tokens": 520032467.0, "step": 13699 }, { "epoch": 1.7427808166899885, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.738175392150879, "learning_rate": 5e-05, "loss": 0.8829, "mean_token_accuracy": 0.8731459379196167, "num_tokens": 520068893.0, "step": 13700 }, { "epoch": 1.742908026968579, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.454697132110596, "learning_rate": 5e-05, "loss": 0.7233, "mean_token_accuracy": 0.8706733584403992, "num_tokens": 520104745.0, "step": 13701 }, { "epoch": 1.7430352372471696, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.120091438293457, "learning_rate": 5e-05, "loss": 0.9905, "mean_token_accuracy": 0.8565904498100281, "num_tokens": 520146994.0, "step": 13702 }, { "epoch": 1.74316244752576, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.515339374542236, "learning_rate": 5e-05, "loss": 0.7053, "mean_token_accuracy": 0.880224347114563, "num_tokens": 520188220.0, "step": 13703 }, { "epoch": 1.7432896578043506, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.26262378692627, "learning_rate": 5e-05, "loss": 0.9297, "mean_token_accuracy": 0.8866493105888367, "num_tokens": 520223410.0, "step": 13704 }, { "epoch": 1.743416868082941, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.033039093017578, "learning_rate": 5e-05, "loss": 0.8371, "mean_token_accuracy": 0.869196891784668, "num_tokens": 520260017.0, "step": 13705 }, { "epoch": 1.7435440783615315, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.903800964355469, "learning_rate": 5e-05, "loss": 0.9434, "mean_token_accuracy": 0.8782477378845215, "num_tokens": 520296232.0, "step": 13706 }, { "epoch": 1.743671288640122, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.6091814041137695, "learning_rate": 5e-05, "loss": 0.93, "mean_token_accuracy": 0.8597925901412964, "num_tokens": 520330720.0, "step": 13707 }, { "epoch": 1.7437984989187125, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.245574951171875, "learning_rate": 5e-05, "loss": 0.8948, "mean_token_accuracy": 0.8786455392837524, "num_tokens": 520362438.0, "step": 13708 }, { "epoch": 1.743925709197303, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.796414852142334, "learning_rate": 5e-05, "loss": 0.8395, "mean_token_accuracy": 0.8839293718338013, "num_tokens": 520398170.0, "step": 13709 }, { "epoch": 1.7440529194758936, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.905751705169678, "learning_rate": 5e-05, "loss": 0.936, "mean_token_accuracy": 0.8542368412017822, "num_tokens": 520438219.0, "step": 13710 }, { "epoch": 1.744180129754484, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.576126575469971, "learning_rate": 5e-05, "loss": 0.8187, "mean_token_accuracy": 0.8748306035995483, "num_tokens": 520481642.0, "step": 13711 }, { "epoch": 1.7443073400330746, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.9836931228637695, "learning_rate": 5e-05, "loss": 0.8552, "mean_token_accuracy": 0.8713392615318298, "num_tokens": 520515945.0, "step": 13712 }, { "epoch": 1.7444345503116652, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.318361282348633, "learning_rate": 5e-05, "loss": 0.8081, "mean_token_accuracy": 0.8630452752113342, "num_tokens": 520553259.0, "step": 13713 }, { "epoch": 1.7445617605902557, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.993457794189453, "learning_rate": 5e-05, "loss": 0.8006, "mean_token_accuracy": 0.8834339380264282, "num_tokens": 520590371.0, "step": 13714 }, { "epoch": 1.7446889708688462, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.340854167938232, "learning_rate": 5e-05, "loss": 0.7744, "mean_token_accuracy": 0.8729815483093262, "num_tokens": 520624738.0, "step": 13715 }, { "epoch": 1.7448161811474368, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.757157802581787, "learning_rate": 5e-05, "loss": 0.7543, "mean_token_accuracy": 0.8838458061218262, "num_tokens": 520666631.0, "step": 13716 }, { "epoch": 1.7449433914260273, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.5372161865234375, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.877472996711731, "num_tokens": 520696567.0, "step": 13717 }, { "epoch": 1.7450706017046178, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.494505405426025, "learning_rate": 5e-05, "loss": 0.7726, "mean_token_accuracy": 0.8675377368927002, "num_tokens": 520737975.0, "step": 13718 }, { "epoch": 1.7451978119832083, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.664150714874268, "learning_rate": 5e-05, "loss": 0.7322, "mean_token_accuracy": 0.8841539025306702, "num_tokens": 520776884.0, "step": 13719 }, { "epoch": 1.7453250222617989, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.368168830871582, "learning_rate": 5e-05, "loss": 0.7931, "mean_token_accuracy": 0.8598169088363647, "num_tokens": 520817770.0, "step": 13720 }, { "epoch": 1.7454522325403894, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.699901103973389, "learning_rate": 5e-05, "loss": 0.7669, "mean_token_accuracy": 0.8698821663856506, "num_tokens": 520859381.0, "step": 13721 }, { "epoch": 1.74557944281898, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.303982257843018, "learning_rate": 5e-05, "loss": 0.6675, "mean_token_accuracy": 0.8930814862251282, "num_tokens": 520900007.0, "step": 13722 }, { "epoch": 1.7457066530975704, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.75222635269165, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.877453625202179, "num_tokens": 520940707.0, "step": 13723 }, { "epoch": 1.7458338633761608, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.182042121887207, "learning_rate": 5e-05, "loss": 0.7472, "mean_token_accuracy": 0.8640587329864502, "num_tokens": 520972565.0, "step": 13724 }, { "epoch": 1.7459610736547513, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.895679473876953, "learning_rate": 5e-05, "loss": 0.7773, "mean_token_accuracy": 0.8650292158126831, "num_tokens": 521012364.0, "step": 13725 }, { "epoch": 1.7460882839333418, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.996123790740967, "learning_rate": 5e-05, "loss": 0.6701, "mean_token_accuracy": 0.8832707405090332, "num_tokens": 521058761.0, "step": 13726 }, { "epoch": 1.7462154942119323, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.150130271911621, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.872821033000946, "num_tokens": 521096620.0, "step": 13727 }, { "epoch": 1.7463427044905229, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.6755242347717285, "learning_rate": 5e-05, "loss": 0.7131, "mean_token_accuracy": 0.8679516315460205, "num_tokens": 521133223.0, "step": 13728 }, { "epoch": 1.7464699147691134, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.609917640686035, "learning_rate": 5e-05, "loss": 0.8023, "mean_token_accuracy": 0.8845525979995728, "num_tokens": 521170334.0, "step": 13729 }, { "epoch": 1.7465971250477037, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.2327070236206055, "learning_rate": 5e-05, "loss": 0.6595, "mean_token_accuracy": 0.8776635527610779, "num_tokens": 521207458.0, "step": 13730 }, { "epoch": 1.7467243353262942, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.336161613464355, "learning_rate": 5e-05, "loss": 0.9116, "mean_token_accuracy": 0.8737343549728394, "num_tokens": 521245017.0, "step": 13731 }, { "epoch": 1.7468515456048848, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 5.930649280548096, "learning_rate": 5e-05, "loss": 0.6954, "mean_token_accuracy": 0.8664565682411194, "num_tokens": 521285792.0, "step": 13732 }, { "epoch": 1.7469787558834753, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.027633666992188, "learning_rate": 5e-05, "loss": 0.9812, "mean_token_accuracy": 0.8799062967300415, "num_tokens": 521327107.0, "step": 13733 }, { "epoch": 1.7471059661620658, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 6.728084087371826, "learning_rate": 5e-05, "loss": 0.8019, "mean_token_accuracy": 0.88198322057724, "num_tokens": 521364720.0, "step": 13734 }, { "epoch": 1.7472331764406563, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.655104637145996, "learning_rate": 5e-05, "loss": 0.9979, "mean_token_accuracy": 0.8804870843887329, "num_tokens": 521398183.0, "step": 13735 }, { "epoch": 1.7473603867192469, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.309812545776367, "learning_rate": 5e-05, "loss": 1.0133, "mean_token_accuracy": 0.8699016571044922, "num_tokens": 521436837.0, "step": 13736 }, { "epoch": 1.7474875969978374, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.640241622924805, "learning_rate": 5e-05, "loss": 0.8269, "mean_token_accuracy": 0.872207760810852, "num_tokens": 521477722.0, "step": 13737 }, { "epoch": 1.747614807276428, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.798500061035156, "learning_rate": 5e-05, "loss": 1.0329, "mean_token_accuracy": 0.8762473464012146, "num_tokens": 521513647.0, "step": 13738 }, { "epoch": 1.7477420175550185, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.0877556800842285, "learning_rate": 5e-05, "loss": 0.7237, "mean_token_accuracy": 0.8714672923088074, "num_tokens": 521545480.0, "step": 13739 }, { "epoch": 1.747869227833609, "ewc_loss": 0.9140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 11.750497817993164, "learning_rate": 5e-05, "loss": 1.3099, "mean_token_accuracy": 0.8605164289474487, "num_tokens": 521587363.0, "step": 13740 }, { "epoch": 1.7479964381121995, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 8.125738143920898, "learning_rate": 5e-05, "loss": 1.0737, "mean_token_accuracy": 0.8664061427116394, "num_tokens": 521629399.0, "step": 13741 }, { "epoch": 1.74812364839079, "ewc_loss": 0.828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000774383544921875, "grad_norm": 11.71078109741211, "learning_rate": 5e-05, "loss": 1.2071, "mean_token_accuracy": 0.869710385799408, "num_tokens": 521667065.0, "step": 13742 }, { "epoch": 1.7482508586693806, "ewc_loss": 0.9453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000888824462890625, "grad_norm": 9.816643714904785, "learning_rate": 5e-05, "loss": 1.2928, "mean_token_accuracy": 0.8770252466201782, "num_tokens": 521696317.0, "step": 13743 }, { "epoch": 1.748378068947971, "ewc_loss": 0.6875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.526683807373047, "learning_rate": 5e-05, "loss": 1.0422, "mean_token_accuracy": 0.8763970136642456, "num_tokens": 521736948.0, "step": 13744 }, { "epoch": 1.7485052792265616, "ewc_loss": 0.71875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 9.199396133422852, "learning_rate": 5e-05, "loss": 1.0508, "mean_token_accuracy": 0.8826876282691956, "num_tokens": 521770320.0, "step": 13745 }, { "epoch": 1.7486324895051522, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 9.762269973754883, "learning_rate": 5e-05, "loss": 1.0181, "mean_token_accuracy": 0.8845627307891846, "num_tokens": 521803836.0, "step": 13746 }, { "epoch": 1.7487596997837427, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.515667915344238, "learning_rate": 5e-05, "loss": 0.9848, "mean_token_accuracy": 0.8596508502960205, "num_tokens": 521838066.0, "step": 13747 }, { "epoch": 1.748886910062333, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.42579174041748, "learning_rate": 5e-05, "loss": 1.0021, "mean_token_accuracy": 0.8762021064758301, "num_tokens": 521876806.0, "step": 13748 }, { "epoch": 1.7490141203409235, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.837172031402588, "learning_rate": 5e-05, "loss": 0.8987, "mean_token_accuracy": 0.8627390265464783, "num_tokens": 521916241.0, "step": 13749 }, { "epoch": 1.749141330619514, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.63363265991211, "learning_rate": 5e-05, "loss": 0.998, "mean_token_accuracy": 0.8746793866157532, "num_tokens": 521949369.0, "step": 13750 }, { "epoch": 1.7492685408981046, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.649026870727539, "learning_rate": 5e-05, "loss": 0.8956, "mean_token_accuracy": 0.8622530698776245, "num_tokens": 521989251.0, "step": 13751 }, { "epoch": 1.749395751176695, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.795184135437012, "learning_rate": 5e-05, "loss": 1.0006, "mean_token_accuracy": 0.8765721321105957, "num_tokens": 522019221.0, "step": 13752 }, { "epoch": 1.7495229614552856, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 7.845631122589111, "learning_rate": 5e-05, "loss": 0.9206, "mean_token_accuracy": 0.8680682182312012, "num_tokens": 522058858.0, "step": 13753 }, { "epoch": 1.749650171733876, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.690568923950195, "learning_rate": 5e-05, "loss": 1.017, "mean_token_accuracy": 0.8722943067550659, "num_tokens": 522089834.0, "step": 13754 }, { "epoch": 1.7497773820124665, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.04481315612793, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8794305324554443, "num_tokens": 522124474.0, "step": 13755 }, { "epoch": 1.749904592291057, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.446246147155762, "learning_rate": 5e-05, "loss": 1.0332, "mean_token_accuracy": 0.8635212182998657, "num_tokens": 522159046.0, "step": 13756 }, { "epoch": 1.7500318025696475, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.016544342041016, "learning_rate": 5e-05, "loss": 0.9144, "mean_token_accuracy": 0.8697595596313477, "num_tokens": 522200698.0, "step": 13757 }, { "epoch": 1.750159012848238, "ewc_loss": 0.625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.182262420654297, "learning_rate": 5e-05, "loss": 1.0312, "mean_token_accuracy": 0.8558099865913391, "num_tokens": 522233533.0, "step": 13758 }, { "epoch": 1.7502862231268286, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.969101428985596, "learning_rate": 5e-05, "loss": 0.9424, "mean_token_accuracy": 0.859510064125061, "num_tokens": 522274234.0, "step": 13759 }, { "epoch": 1.750413433405419, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.005436897277832, "learning_rate": 5e-05, "loss": 0.9882, "mean_token_accuracy": 0.8738209009170532, "num_tokens": 522313055.0, "step": 13760 }, { "epoch": 1.7505406436840096, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.70521354675293, "learning_rate": 5e-05, "loss": 0.8386, "mean_token_accuracy": 0.8724913597106934, "num_tokens": 522346372.0, "step": 13761 }, { "epoch": 1.7506678539626002, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.1334867477417, "learning_rate": 5e-05, "loss": 0.9955, "mean_token_accuracy": 0.8718416690826416, "num_tokens": 522381199.0, "step": 13762 }, { "epoch": 1.7507950642411907, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.622091770172119, "learning_rate": 5e-05, "loss": 0.915, "mean_token_accuracy": 0.8542239665985107, "num_tokens": 522419509.0, "step": 13763 }, { "epoch": 1.7509222745197812, "ewc_loss": 0.625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.165263175964355, "learning_rate": 5e-05, "loss": 0.9896, "mean_token_accuracy": 0.8723657727241516, "num_tokens": 522458433.0, "step": 13764 }, { "epoch": 1.7510494847983717, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.795557975769043, "learning_rate": 5e-05, "loss": 0.8758, "mean_token_accuracy": 0.8726792931556702, "num_tokens": 522492010.0, "step": 13765 }, { "epoch": 1.7511766950769623, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.846863746643066, "learning_rate": 5e-05, "loss": 0.976, "mean_token_accuracy": 0.8729143738746643, "num_tokens": 522524190.0, "step": 13766 }, { "epoch": 1.7513039053555528, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.920314788818359, "learning_rate": 5e-05, "loss": 0.9032, "mean_token_accuracy": 0.8587591648101807, "num_tokens": 522558218.0, "step": 13767 }, { "epoch": 1.7514311156341433, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.447406768798828, "learning_rate": 5e-05, "loss": 0.9149, "mean_token_accuracy": 0.8762257099151611, "num_tokens": 522594683.0, "step": 13768 }, { "epoch": 1.7515583259127339, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.7795562744140625, "learning_rate": 5e-05, "loss": 0.8225, "mean_token_accuracy": 0.8792926669120789, "num_tokens": 522633409.0, "step": 13769 }, { "epoch": 1.7516855361913244, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.311905860900879, "learning_rate": 5e-05, "loss": 0.9055, "mean_token_accuracy": 0.8670961856842041, "num_tokens": 522672469.0, "step": 13770 }, { "epoch": 1.751812746469915, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.50906229019165, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.874945878982544, "num_tokens": 522714676.0, "step": 13771 }, { "epoch": 1.7519399567485054, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.439523696899414, "learning_rate": 5e-05, "loss": 0.8654, "mean_token_accuracy": 0.8827440142631531, "num_tokens": 522748726.0, "step": 13772 }, { "epoch": 1.7520671670270958, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.287364482879639, "learning_rate": 5e-05, "loss": 0.8551, "mean_token_accuracy": 0.8525967001914978, "num_tokens": 522783761.0, "step": 13773 }, { "epoch": 1.7521943773056863, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.588872909545898, "learning_rate": 5e-05, "loss": 0.9007, "mean_token_accuracy": 0.8786332607269287, "num_tokens": 522822027.0, "step": 13774 }, { "epoch": 1.7523215875842768, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.269282341003418, "learning_rate": 5e-05, "loss": 0.7616, "mean_token_accuracy": 0.8784180879592896, "num_tokens": 522856952.0, "step": 13775 }, { "epoch": 1.7524487978628673, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.52582836151123, "learning_rate": 5e-05, "loss": 0.9257, "mean_token_accuracy": 0.865339994430542, "num_tokens": 522897450.0, "step": 13776 }, { "epoch": 1.7525760081414579, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.449431896209717, "learning_rate": 5e-05, "loss": 0.7816, "mean_token_accuracy": 0.8759939074516296, "num_tokens": 522930547.0, "step": 13777 }, { "epoch": 1.7527032184200484, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.199968338012695, "learning_rate": 5e-05, "loss": 0.8869, "mean_token_accuracy": 0.8685014843940735, "num_tokens": 522970145.0, "step": 13778 }, { "epoch": 1.7528304286986387, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.575048923492432, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8854641914367676, "num_tokens": 523006492.0, "step": 13779 }, { "epoch": 1.7529576389772292, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.983557224273682, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8650064468383789, "num_tokens": 523045524.0, "step": 13780 }, { "epoch": 1.7530848492558198, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.535940647125244, "learning_rate": 5e-05, "loss": 0.7714, "mean_token_accuracy": 0.8754926323890686, "num_tokens": 523086091.0, "step": 13781 }, { "epoch": 1.7532120595344103, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.891207695007324, "learning_rate": 5e-05, "loss": 0.7905, "mean_token_accuracy": 0.8791884779930115, "num_tokens": 523127351.0, "step": 13782 }, { "epoch": 1.7533392698130008, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.49625301361084, "learning_rate": 5e-05, "loss": 0.7653, "mean_token_accuracy": 0.8690617084503174, "num_tokens": 523161775.0, "step": 13783 }, { "epoch": 1.7534664800915913, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.8036956787109375, "learning_rate": 5e-05, "loss": 0.8024, "mean_token_accuracy": 0.8662006855010986, "num_tokens": 523198977.0, "step": 13784 }, { "epoch": 1.7535936903701819, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.467260360717773, "learning_rate": 5e-05, "loss": 0.7656, "mean_token_accuracy": 0.8667502403259277, "num_tokens": 523233917.0, "step": 13785 }, { "epoch": 1.7537209006487724, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.777252674102783, "learning_rate": 5e-05, "loss": 0.7669, "mean_token_accuracy": 0.8759491443634033, "num_tokens": 523267679.0, "step": 13786 }, { "epoch": 1.753848110927363, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.4013671875, "learning_rate": 5e-05, "loss": 0.7769, "mean_token_accuracy": 0.8605592250823975, "num_tokens": 523301512.0, "step": 13787 }, { "epoch": 1.7539753212059535, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.790846824645996, "learning_rate": 5e-05, "loss": 0.752, "mean_token_accuracy": 0.8799649477005005, "num_tokens": 523342618.0, "step": 13788 }, { "epoch": 1.754102531484544, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.343205451965332, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.870802104473114, "num_tokens": 523379555.0, "step": 13789 }, { "epoch": 1.7542297417631345, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.795352458953857, "learning_rate": 5e-05, "loss": 0.7204, "mean_token_accuracy": 0.89092618227005, "num_tokens": 523414621.0, "step": 13790 }, { "epoch": 1.754356952041725, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.304683208465576, "learning_rate": 5e-05, "loss": 0.6655, "mean_token_accuracy": 0.8935567140579224, "num_tokens": 523459022.0, "step": 13791 }, { "epoch": 1.7544841623203156, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.842377662658691, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.8854647874832153, "num_tokens": 523498718.0, "step": 13792 }, { "epoch": 1.754611372598906, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.228488922119141, "learning_rate": 5e-05, "loss": 0.6941, "mean_token_accuracy": 0.8823028802871704, "num_tokens": 523531926.0, "step": 13793 }, { "epoch": 1.7547385828774966, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.881441116333008, "learning_rate": 5e-05, "loss": 0.7566, "mean_token_accuracy": 0.8734822273254395, "num_tokens": 523572457.0, "step": 13794 }, { "epoch": 1.7548657931560872, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.195115566253662, "learning_rate": 5e-05, "loss": 0.744, "mean_token_accuracy": 0.8644463419914246, "num_tokens": 523607317.0, "step": 13795 }, { "epoch": 1.7549930034346777, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.895340442657471, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8662564754486084, "num_tokens": 523638979.0, "step": 13796 }, { "epoch": 1.755120213713268, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.163504600524902, "learning_rate": 5e-05, "loss": 0.6986, "mean_token_accuracy": 0.8741223216056824, "num_tokens": 523681839.0, "step": 13797 }, { "epoch": 1.7552474239918585, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.928042888641357, "learning_rate": 5e-05, "loss": 0.7404, "mean_token_accuracy": 0.8782204389572144, "num_tokens": 523719797.0, "step": 13798 }, { "epoch": 1.755374634270449, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.117722988128662, "learning_rate": 5e-05, "loss": 0.6813, "mean_token_accuracy": 0.8777729272842407, "num_tokens": 523761774.0, "step": 13799 }, { "epoch": 1.7555018445490396, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.973660469055176, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.8736560344696045, "num_tokens": 523799522.0, "step": 13800 }, { "epoch": 1.75562905482763, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.082485675811768, "learning_rate": 5e-05, "loss": 0.6908, "mean_token_accuracy": 0.8745108842849731, "num_tokens": 523831927.0, "step": 13801 }, { "epoch": 1.7557562651062206, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.981068134307861, "learning_rate": 5e-05, "loss": 0.7952, "mean_token_accuracy": 0.8638086318969727, "num_tokens": 523863058.0, "step": 13802 }, { "epoch": 1.755883475384811, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.090875148773193, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8526721000671387, "num_tokens": 523907017.0, "step": 13803 }, { "epoch": 1.7560106856634015, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.935844421386719, "learning_rate": 5e-05, "loss": 0.786, "mean_token_accuracy": 0.8685925006866455, "num_tokens": 523943376.0, "step": 13804 }, { "epoch": 1.756137895941992, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.158414363861084, "learning_rate": 5e-05, "loss": 0.6918, "mean_token_accuracy": 0.8811538219451904, "num_tokens": 523982902.0, "step": 13805 }, { "epoch": 1.7562651062205825, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.8346123695373535, "learning_rate": 5e-05, "loss": 0.7686, "mean_token_accuracy": 0.8660356998443604, "num_tokens": 524016423.0, "step": 13806 }, { "epoch": 1.756392316499173, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.264782905578613, "learning_rate": 5e-05, "loss": 0.7779, "mean_token_accuracy": 0.856720507144928, "num_tokens": 524056686.0, "step": 13807 }, { "epoch": 1.7565195267777636, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.656258583068848, "learning_rate": 5e-05, "loss": 0.7416, "mean_token_accuracy": 0.871283769607544, "num_tokens": 524093731.0, "step": 13808 }, { "epoch": 1.756646737056354, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.469812870025635, "learning_rate": 5e-05, "loss": 0.7374, "mean_token_accuracy": 0.8713841438293457, "num_tokens": 524139398.0, "step": 13809 }, { "epoch": 1.7567739473349446, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.377600193023682, "learning_rate": 5e-05, "loss": 0.7393, "mean_token_accuracy": 0.8645087480545044, "num_tokens": 524177012.0, "step": 13810 }, { "epoch": 1.7569011576135352, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.77429723739624, "learning_rate": 5e-05, "loss": 0.6982, "mean_token_accuracy": 0.8857176303863525, "num_tokens": 524211785.0, "step": 13811 }, { "epoch": 1.7570283678921257, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.973294258117676, "learning_rate": 5e-05, "loss": 0.6559, "mean_token_accuracy": 0.8852150440216064, "num_tokens": 524252263.0, "step": 13812 }, { "epoch": 1.7571555781707162, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.274760246276855, "learning_rate": 5e-05, "loss": 0.7248, "mean_token_accuracy": 0.8904887437820435, "num_tokens": 524286561.0, "step": 13813 }, { "epoch": 1.7572827884493067, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002288818359375, "grad_norm": 6.388388156890869, "learning_rate": 5e-05, "loss": 0.6333, "mean_token_accuracy": 0.8784875869750977, "num_tokens": 524320910.0, "step": 13814 }, { "epoch": 1.7574099987278973, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 9.135597229003906, "learning_rate": 5e-05, "loss": 0.8358, "mean_token_accuracy": 0.881926417350769, "num_tokens": 524362416.0, "step": 13815 }, { "epoch": 1.7575372090064878, "ewc_loss": 0.25390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0001983642578125, "grad_norm": 5.579405307769775, "learning_rate": 5e-05, "loss": 0.5983, "mean_token_accuracy": 0.8784764409065247, "num_tokens": 524405432.0, "step": 13816 }, { "epoch": 1.7576644192850783, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.587388038635254, "learning_rate": 5e-05, "loss": 1.0146, "mean_token_accuracy": 0.8763340711593628, "num_tokens": 524440605.0, "step": 13817 }, { "epoch": 1.7577916295636689, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 5.534730434417725, "learning_rate": 5e-05, "loss": 0.6991, "mean_token_accuracy": 0.8697879314422607, "num_tokens": 524483126.0, "step": 13818 }, { "epoch": 1.7579188398422594, "ewc_loss": 0.96484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000911712646484375, "grad_norm": 12.13652515411377, "learning_rate": 5e-05, "loss": 1.3156, "mean_token_accuracy": 0.8802651166915894, "num_tokens": 524516675.0, "step": 13819 }, { "epoch": 1.75804605012085, "ewc_loss": 0.828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000774383544921875, "grad_norm": 8.69510269165039, "learning_rate": 5e-05, "loss": 1.1793, "mean_token_accuracy": 0.8757333755493164, "num_tokens": 524554394.0, "step": 13820 }, { "epoch": 1.7581732603994404, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 11.126352310180664, "learning_rate": 5e-05, "loss": 1.0945, "mean_token_accuracy": 0.8684961795806885, "num_tokens": 524591195.0, "step": 13821 }, { "epoch": 1.7583004706780307, "ewc_loss": 0.96484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000911712646484375, "grad_norm": 9.757049560546875, "learning_rate": 5e-05, "loss": 1.3502, "mean_token_accuracy": 0.8634389638900757, "num_tokens": 524629077.0, "step": 13822 }, { "epoch": 1.7584276809566213, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.551163673400879, "learning_rate": 5e-05, "loss": 1.0312, "mean_token_accuracy": 0.8563046455383301, "num_tokens": 524672072.0, "step": 13823 }, { "epoch": 1.7585548912352118, "ewc_loss": 0.78125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 9.896041870117188, "learning_rate": 5e-05, "loss": 1.1498, "mean_token_accuracy": 0.8730576038360596, "num_tokens": 524706735.0, "step": 13824 }, { "epoch": 1.7586821015138023, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.336737632751465, "learning_rate": 5e-05, "loss": 0.9374, "mean_token_accuracy": 0.8780063390731812, "num_tokens": 524741750.0, "step": 13825 }, { "epoch": 1.7588093117923929, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 10.367534637451172, "learning_rate": 5e-05, "loss": 1.0472, "mean_token_accuracy": 0.8868237733840942, "num_tokens": 524774565.0, "step": 13826 }, { "epoch": 1.7589365220709834, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.374018669128418, "learning_rate": 5e-05, "loss": 0.9789, "mean_token_accuracy": 0.8796614408493042, "num_tokens": 524803731.0, "step": 13827 }, { "epoch": 1.7590637323495737, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.872624397277832, "learning_rate": 5e-05, "loss": 1.0224, "mean_token_accuracy": 0.8698641061782837, "num_tokens": 524842570.0, "step": 13828 }, { "epoch": 1.7591909426281642, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.522910118103027, "learning_rate": 5e-05, "loss": 1.0207, "mean_token_accuracy": 0.8715089559555054, "num_tokens": 524880695.0, "step": 13829 }, { "epoch": 1.7593181529067548, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.798381805419922, "learning_rate": 5e-05, "loss": 0.9018, "mean_token_accuracy": 0.875444233417511, "num_tokens": 524921331.0, "step": 13830 }, { "epoch": 1.7594453631853453, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.755118370056152, "learning_rate": 5e-05, "loss": 0.9958, "mean_token_accuracy": 0.86644446849823, "num_tokens": 524955769.0, "step": 13831 }, { "epoch": 1.7595725734639358, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.564777851104736, "learning_rate": 5e-05, "loss": 0.8382, "mean_token_accuracy": 0.8591127991676331, "num_tokens": 524998299.0, "step": 13832 }, { "epoch": 1.7596997837425263, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.592048645019531, "learning_rate": 5e-05, "loss": 1.04, "mean_token_accuracy": 0.8604849576950073, "num_tokens": 525038590.0, "step": 13833 }, { "epoch": 1.7598269940211169, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.894528865814209, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.8705266714096069, "num_tokens": 525073688.0, "step": 13834 }, { "epoch": 1.7599542042997074, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.594732284545898, "learning_rate": 5e-05, "loss": 1.015, "mean_token_accuracy": 0.8916746377944946, "num_tokens": 525112109.0, "step": 13835 }, { "epoch": 1.760081414578298, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.711315155029297, "learning_rate": 5e-05, "loss": 0.9387, "mean_token_accuracy": 0.8764380216598511, "num_tokens": 525147554.0, "step": 13836 }, { "epoch": 1.7602086248568884, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.56654167175293, "learning_rate": 5e-05, "loss": 1.0921, "mean_token_accuracy": 0.8597500324249268, "num_tokens": 525187233.0, "step": 13837 }, { "epoch": 1.760335835135479, "ewc_loss": 0.703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 8.623520851135254, "learning_rate": 5e-05, "loss": 1.1074, "mean_token_accuracy": 0.864771842956543, "num_tokens": 525218104.0, "step": 13838 }, { "epoch": 1.7604630454140695, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.718344688415527, "learning_rate": 5e-05, "loss": 0.9658, "mean_token_accuracy": 0.877957284450531, "num_tokens": 525255690.0, "step": 13839 }, { "epoch": 1.76059025569266, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.48752498626709, "learning_rate": 5e-05, "loss": 1.0215, "mean_token_accuracy": 0.8725360035896301, "num_tokens": 525287972.0, "step": 13840 }, { "epoch": 1.7607174659712506, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.059550285339355, "learning_rate": 5e-05, "loss": 0.9522, "mean_token_accuracy": 0.8651039600372314, "num_tokens": 525328003.0, "step": 13841 }, { "epoch": 1.760844676249841, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.961995601654053, "learning_rate": 5e-05, "loss": 0.898, "mean_token_accuracy": 0.8789733648300171, "num_tokens": 525370642.0, "step": 13842 }, { "epoch": 1.7609718865284316, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.846287727355957, "learning_rate": 5e-05, "loss": 0.9788, "mean_token_accuracy": 0.8579193949699402, "num_tokens": 525408668.0, "step": 13843 }, { "epoch": 1.7610990968070221, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.339681148529053, "learning_rate": 5e-05, "loss": 0.807, "mean_token_accuracy": 0.8722084760665894, "num_tokens": 525447566.0, "step": 13844 }, { "epoch": 1.7612263070856127, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.064623832702637, "learning_rate": 5e-05, "loss": 0.9422, "mean_token_accuracy": 0.8660418391227722, "num_tokens": 525485456.0, "step": 13845 }, { "epoch": 1.761353517364203, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.9532389640808105, "learning_rate": 5e-05, "loss": 0.827, "mean_token_accuracy": 0.8574923276901245, "num_tokens": 525521190.0, "step": 13846 }, { "epoch": 1.7614807276427935, "ewc_loss": 0.625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.462870597839355, "learning_rate": 5e-05, "loss": 0.9924, "mean_token_accuracy": 0.8728355765342712, "num_tokens": 525559331.0, "step": 13847 }, { "epoch": 1.761607937921384, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.1342597007751465, "learning_rate": 5e-05, "loss": 0.8277, "mean_token_accuracy": 0.8653136491775513, "num_tokens": 525598380.0, "step": 13848 }, { "epoch": 1.7617351481999746, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.556453704833984, "learning_rate": 5e-05, "loss": 1.0293, "mean_token_accuracy": 0.8649182319641113, "num_tokens": 525637901.0, "step": 13849 }, { "epoch": 1.761862358478565, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 7.770690441131592, "learning_rate": 5e-05, "loss": 0.9294, "mean_token_accuracy": 0.8668402433395386, "num_tokens": 525676072.0, "step": 13850 }, { "epoch": 1.7619895687571556, "ewc_loss": 0.578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.952099800109863, "learning_rate": 5e-05, "loss": 0.8912, "mean_token_accuracy": 0.8897996544837952, "num_tokens": 525710655.0, "step": 13851 }, { "epoch": 1.762116779035746, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.252389907836914, "learning_rate": 5e-05, "loss": 0.916, "mean_token_accuracy": 0.8783140182495117, "num_tokens": 525745714.0, "step": 13852 }, { "epoch": 1.7622439893143365, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.9431352615356445, "learning_rate": 5e-05, "loss": 0.7865, "mean_token_accuracy": 0.8859166502952576, "num_tokens": 525782246.0, "step": 13853 }, { "epoch": 1.762371199592927, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.433570861816406, "learning_rate": 5e-05, "loss": 0.8606, "mean_token_accuracy": 0.8880786895751953, "num_tokens": 525814232.0, "step": 13854 }, { "epoch": 1.7624984098715175, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.303944110870361, "learning_rate": 5e-05, "loss": 0.775, "mean_token_accuracy": 0.8708490133285522, "num_tokens": 525847775.0, "step": 13855 }, { "epoch": 1.762625620150108, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.48217487335205, "learning_rate": 5e-05, "loss": 0.8703, "mean_token_accuracy": 0.8744019865989685, "num_tokens": 525890760.0, "step": 13856 }, { "epoch": 1.7627528304286986, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.193105697631836, "learning_rate": 5e-05, "loss": 0.7177, "mean_token_accuracy": 0.8843522071838379, "num_tokens": 525926893.0, "step": 13857 }, { "epoch": 1.762880040707289, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.314064979553223, "learning_rate": 5e-05, "loss": 0.8754, "mean_token_accuracy": 0.86506187915802, "num_tokens": 525968907.0, "step": 13858 }, { "epoch": 1.7630072509858796, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.3163065910339355, "learning_rate": 5e-05, "loss": 0.7349, "mean_token_accuracy": 0.8776732087135315, "num_tokens": 526003988.0, "step": 13859 }, { "epoch": 1.7631344612644702, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.047466278076172, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8756235837936401, "num_tokens": 526044575.0, "step": 13860 }, { "epoch": 1.7632616715430607, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.4309306144714355, "learning_rate": 5e-05, "loss": 0.7601, "mean_token_accuracy": 0.8734210729598999, "num_tokens": 526087803.0, "step": 13861 }, { "epoch": 1.7633888818216512, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.84588623046875, "learning_rate": 5e-05, "loss": 0.8189, "mean_token_accuracy": 0.8568159341812134, "num_tokens": 526123432.0, "step": 13862 }, { "epoch": 1.7635160921002417, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.470030784606934, "learning_rate": 5e-05, "loss": 0.7148, "mean_token_accuracy": 0.8825598359107971, "num_tokens": 526161285.0, "step": 13863 }, { "epoch": 1.7636433023788323, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.81467342376709, "learning_rate": 5e-05, "loss": 0.839, "mean_token_accuracy": 0.8568198680877686, "num_tokens": 526203096.0, "step": 13864 }, { "epoch": 1.7637705126574228, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.295078277587891, "learning_rate": 5e-05, "loss": 0.692, "mean_token_accuracy": 0.8837978839874268, "num_tokens": 526237741.0, "step": 13865 }, { "epoch": 1.7638977229360133, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.087209701538086, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.87429279088974, "num_tokens": 526279704.0, "step": 13866 }, { "epoch": 1.7640249332146039, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.907751083374023, "learning_rate": 5e-05, "loss": 0.7114, "mean_token_accuracy": 0.8724544048309326, "num_tokens": 526315708.0, "step": 13867 }, { "epoch": 1.7641521434931944, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.522513389587402, "learning_rate": 5e-05, "loss": 0.8327, "mean_token_accuracy": 0.8700257539749146, "num_tokens": 526360725.0, "step": 13868 }, { "epoch": 1.764279353771785, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.554507732391357, "learning_rate": 5e-05, "loss": 0.7158, "mean_token_accuracy": 0.8670558333396912, "num_tokens": 526398920.0, "step": 13869 }, { "epoch": 1.7644065640503754, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.960271835327148, "learning_rate": 5e-05, "loss": 0.8285, "mean_token_accuracy": 0.8864545822143555, "num_tokens": 526433453.0, "step": 13870 }, { "epoch": 1.7645337743289657, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.440391540527344, "learning_rate": 5e-05, "loss": 0.6798, "mean_token_accuracy": 0.8805314302444458, "num_tokens": 526476256.0, "step": 13871 }, { "epoch": 1.7646609846075563, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.231147766113281, "learning_rate": 5e-05, "loss": 0.9205, "mean_token_accuracy": 0.8769549131393433, "num_tokens": 526517707.0, "step": 13872 }, { "epoch": 1.7647881948861468, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.735161304473877, "learning_rate": 5e-05, "loss": 0.7486, "mean_token_accuracy": 0.8744753003120422, "num_tokens": 526551696.0, "step": 13873 }, { "epoch": 1.7649154051647373, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.020297050476074, "learning_rate": 5e-05, "loss": 0.9175, "mean_token_accuracy": 0.8830586671829224, "num_tokens": 526587884.0, "step": 13874 }, { "epoch": 1.7650426154433279, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.453721046447754, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.8691293597221375, "num_tokens": 526625398.0, "step": 13875 }, { "epoch": 1.7651698257219184, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.134453773498535, "learning_rate": 5e-05, "loss": 0.8745, "mean_token_accuracy": 0.8693776726722717, "num_tokens": 526662481.0, "step": 13876 }, { "epoch": 1.7652970360005087, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.132203102111816, "learning_rate": 5e-05, "loss": 0.9044, "mean_token_accuracy": 0.862912654876709, "num_tokens": 526700719.0, "step": 13877 }, { "epoch": 1.7654242462790992, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.174671649932861, "learning_rate": 5e-05, "loss": 0.7411, "mean_token_accuracy": 0.8736391067504883, "num_tokens": 526732941.0, "step": 13878 }, { "epoch": 1.7655514565576897, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.618653297424316, "learning_rate": 5e-05, "loss": 0.8813, "mean_token_accuracy": 0.8748985528945923, "num_tokens": 526777825.0, "step": 13879 }, { "epoch": 1.7656786668362803, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.678933143615723, "learning_rate": 5e-05, "loss": 0.7281, "mean_token_accuracy": 0.8661863803863525, "num_tokens": 526811331.0, "step": 13880 }, { "epoch": 1.7658058771148708, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.000946998596191, "learning_rate": 5e-05, "loss": 0.9149, "mean_token_accuracy": 0.8731927871704102, "num_tokens": 526846788.0, "step": 13881 }, { "epoch": 1.7659330873934613, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.624173164367676, "learning_rate": 5e-05, "loss": 0.6997, "mean_token_accuracy": 0.8801900148391724, "num_tokens": 526880127.0, "step": 13882 }, { "epoch": 1.7660602976720519, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.155988693237305, "learning_rate": 5e-05, "loss": 0.9773, "mean_token_accuracy": 0.8637882471084595, "num_tokens": 526914099.0, "step": 13883 }, { "epoch": 1.7661875079506424, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 6.989371299743652, "learning_rate": 5e-05, "loss": 0.8251, "mean_token_accuracy": 0.8645370006561279, "num_tokens": 526956409.0, "step": 13884 }, { "epoch": 1.766314718229233, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.866418838500977, "learning_rate": 5e-05, "loss": 0.947, "mean_token_accuracy": 0.8709291815757751, "num_tokens": 526995175.0, "step": 13885 }, { "epoch": 1.7664419285078234, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.587191104888916, "learning_rate": 5e-05, "loss": 0.8776, "mean_token_accuracy": 0.8713794946670532, "num_tokens": 527035833.0, "step": 13886 }, { "epoch": 1.766569138786414, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.139303207397461, "learning_rate": 5e-05, "loss": 0.8717, "mean_token_accuracy": 0.8730754852294922, "num_tokens": 527078780.0, "step": 13887 }, { "epoch": 1.7666963490650045, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.896963596343994, "learning_rate": 5e-05, "loss": 0.8533, "mean_token_accuracy": 0.8801578283309937, "num_tokens": 527119494.0, "step": 13888 }, { "epoch": 1.766823559343595, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.739907741546631, "learning_rate": 5e-05, "loss": 0.826, "mean_token_accuracy": 0.8698129653930664, "num_tokens": 527158112.0, "step": 13889 }, { "epoch": 1.7669507696221856, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.705363750457764, "learning_rate": 5e-05, "loss": 0.8079, "mean_token_accuracy": 0.8739264607429504, "num_tokens": 527195885.0, "step": 13890 }, { "epoch": 1.767077979900776, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.907192707061768, "learning_rate": 5e-05, "loss": 0.806, "mean_token_accuracy": 0.8735257387161255, "num_tokens": 527229681.0, "step": 13891 }, { "epoch": 1.7672051901793666, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.237703323364258, "learning_rate": 5e-05, "loss": 0.813, "mean_token_accuracy": 0.8493281006813049, "num_tokens": 527267248.0, "step": 13892 }, { "epoch": 1.7673324004579571, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.264348030090332, "learning_rate": 5e-05, "loss": 0.8249, "mean_token_accuracy": 0.8740394115447998, "num_tokens": 527299695.0, "step": 13893 }, { "epoch": 1.7674596107365477, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.8839921951293945, "learning_rate": 5e-05, "loss": 0.7266, "mean_token_accuracy": 0.8685365915298462, "num_tokens": 527334398.0, "step": 13894 }, { "epoch": 1.767586821015138, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.553542137145996, "learning_rate": 5e-05, "loss": 0.9117, "mean_token_accuracy": 0.859176754951477, "num_tokens": 527376954.0, "step": 13895 }, { "epoch": 1.7677140312937285, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.708066940307617, "learning_rate": 5e-05, "loss": 0.7214, "mean_token_accuracy": 0.8704800009727478, "num_tokens": 527419737.0, "step": 13896 }, { "epoch": 1.767841241572319, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.782102584838867, "learning_rate": 5e-05, "loss": 0.9435, "mean_token_accuracy": 0.8534521460533142, "num_tokens": 527455874.0, "step": 13897 }, { "epoch": 1.7679684518509096, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.732941627502441, "learning_rate": 5e-05, "loss": 0.7961, "mean_token_accuracy": 0.8458739519119263, "num_tokens": 527494796.0, "step": 13898 }, { "epoch": 1.7680956621295, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.815855026245117, "learning_rate": 5e-05, "loss": 0.9718, "mean_token_accuracy": 0.8538675904273987, "num_tokens": 527536955.0, "step": 13899 }, { "epoch": 1.7682228724080906, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.035048007965088, "learning_rate": 5e-05, "loss": 0.7763, "mean_token_accuracy": 0.868506908416748, "num_tokens": 527569472.0, "step": 13900 }, { "epoch": 1.768350082686681, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.459443092346191, "learning_rate": 5e-05, "loss": 0.8753, "mean_token_accuracy": 0.8742810487747192, "num_tokens": 527603641.0, "step": 13901 }, { "epoch": 1.7684772929652715, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.476959228515625, "learning_rate": 5e-05, "loss": 0.8515, "mean_token_accuracy": 0.862990140914917, "num_tokens": 527642942.0, "step": 13902 }, { "epoch": 1.768604503243862, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.8757219314575195, "learning_rate": 5e-05, "loss": 0.8508, "mean_token_accuracy": 0.8657349348068237, "num_tokens": 527682687.0, "step": 13903 }, { "epoch": 1.7687317135224525, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.81161642074585, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.8829469084739685, "num_tokens": 527721355.0, "step": 13904 }, { "epoch": 1.768858923801043, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.416805267333984, "learning_rate": 5e-05, "loss": 0.762, "mean_token_accuracy": 0.8725737929344177, "num_tokens": 527760887.0, "step": 13905 }, { "epoch": 1.7689861340796336, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.994920253753662, "learning_rate": 5e-05, "loss": 0.8534, "mean_token_accuracy": 0.8623993992805481, "num_tokens": 527808669.0, "step": 13906 }, { "epoch": 1.769113344358224, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.188188076019287, "learning_rate": 5e-05, "loss": 0.789, "mean_token_accuracy": 0.8591105937957764, "num_tokens": 527840262.0, "step": 13907 }, { "epoch": 1.7692405546368146, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.085845947265625, "learning_rate": 5e-05, "loss": 0.842, "mean_token_accuracy": 0.8613696694374084, "num_tokens": 527880822.0, "step": 13908 }, { "epoch": 1.7693677649154052, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.027519702911377, "learning_rate": 5e-05, "loss": 0.7198, "mean_token_accuracy": 0.8771666288375854, "num_tokens": 527923677.0, "step": 13909 }, { "epoch": 1.7694949751939957, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.252554893493652, "learning_rate": 5e-05, "loss": 0.8665, "mean_token_accuracy": 0.8586087226867676, "num_tokens": 527966004.0, "step": 13910 }, { "epoch": 1.7696221854725862, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.844232082366943, "learning_rate": 5e-05, "loss": 0.719, "mean_token_accuracy": 0.8727237582206726, "num_tokens": 528004725.0, "step": 13911 }, { "epoch": 1.7697493957511767, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.484946250915527, "learning_rate": 5e-05, "loss": 0.8431, "mean_token_accuracy": 0.8691242933273315, "num_tokens": 528043262.0, "step": 13912 }, { "epoch": 1.7698766060297673, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.6852617263793945, "learning_rate": 5e-05, "loss": 0.73, "mean_token_accuracy": 0.8655062317848206, "num_tokens": 528082810.0, "step": 13913 }, { "epoch": 1.7700038163083578, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.709175109863281, "learning_rate": 5e-05, "loss": 0.8874, "mean_token_accuracy": 0.866097092628479, "num_tokens": 528117672.0, "step": 13914 }, { "epoch": 1.7701310265869483, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.644705772399902, "learning_rate": 5e-05, "loss": 0.728, "mean_token_accuracy": 0.8712732195854187, "num_tokens": 528155756.0, "step": 13915 }, { "epoch": 1.7702582368655388, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.781867980957031, "learning_rate": 5e-05, "loss": 0.8724, "mean_token_accuracy": 0.8794522881507874, "num_tokens": 528194664.0, "step": 13916 }, { "epoch": 1.7703854471441294, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.883673191070557, "learning_rate": 5e-05, "loss": 0.7165, "mean_token_accuracy": 0.8825305104255676, "num_tokens": 528234334.0, "step": 13917 }, { "epoch": 1.77051265742272, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.44186782836914, "learning_rate": 5e-05, "loss": 0.8432, "mean_token_accuracy": 0.8871191143989563, "num_tokens": 528270543.0, "step": 13918 }, { "epoch": 1.7706398677013102, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.361520767211914, "learning_rate": 5e-05, "loss": 0.8302, "mean_token_accuracy": 0.865250825881958, "num_tokens": 528309698.0, "step": 13919 }, { "epoch": 1.7707670779799007, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.8727898597717285, "learning_rate": 5e-05, "loss": 0.83, "mean_token_accuracy": 0.8676609992980957, "num_tokens": 528356396.0, "step": 13920 }, { "epoch": 1.7708942882584913, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.744210720062256, "learning_rate": 5e-05, "loss": 0.8329, "mean_token_accuracy": 0.864065408706665, "num_tokens": 528394542.0, "step": 13921 }, { "epoch": 1.7710214985370818, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.45808744430542, "learning_rate": 5e-05, "loss": 0.7292, "mean_token_accuracy": 0.882881224155426, "num_tokens": 528435433.0, "step": 13922 }, { "epoch": 1.7711487088156723, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.830282688140869, "learning_rate": 5e-05, "loss": 0.7544, "mean_token_accuracy": 0.887604296207428, "num_tokens": 528471152.0, "step": 13923 }, { "epoch": 1.7712759190942629, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.416857719421387, "learning_rate": 5e-05, "loss": 0.7273, "mean_token_accuracy": 0.8815146684646606, "num_tokens": 528505084.0, "step": 13924 }, { "epoch": 1.7714031293728534, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.647706508636475, "learning_rate": 5e-05, "loss": 0.8529, "mean_token_accuracy": 0.844420313835144, "num_tokens": 528545234.0, "step": 13925 }, { "epoch": 1.7715303396514437, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.606180667877197, "learning_rate": 5e-05, "loss": 0.7415, "mean_token_accuracy": 0.8751566410064697, "num_tokens": 528584099.0, "step": 13926 }, { "epoch": 1.7716575499300342, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.330835342407227, "learning_rate": 5e-05, "loss": 0.7681, "mean_token_accuracy": 0.8627815246582031, "num_tokens": 528624143.0, "step": 13927 }, { "epoch": 1.7717847602086247, "ewc_loss": 0.40625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.913747787475586, "learning_rate": 5e-05, "loss": 0.7724, "mean_token_accuracy": 0.8715686798095703, "num_tokens": 528661521.0, "step": 13928 }, { "epoch": 1.7719119704872153, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.981326580047607, "learning_rate": 5e-05, "loss": 0.6816, "mean_token_accuracy": 0.8831489086151123, "num_tokens": 528699317.0, "step": 13929 }, { "epoch": 1.7720391807658058, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.292033195495605, "learning_rate": 5e-05, "loss": 0.842, "mean_token_accuracy": 0.858799934387207, "num_tokens": 528737277.0, "step": 13930 }, { "epoch": 1.7721663910443963, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.596353054046631, "learning_rate": 5e-05, "loss": 0.7303, "mean_token_accuracy": 0.8614996671676636, "num_tokens": 528777076.0, "step": 13931 }, { "epoch": 1.7722936013229869, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.785646438598633, "learning_rate": 5e-05, "loss": 0.8517, "mean_token_accuracy": 0.8730594515800476, "num_tokens": 528813940.0, "step": 13932 }, { "epoch": 1.7724208116015774, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.2411298751831055, "learning_rate": 5e-05, "loss": 0.6306, "mean_token_accuracy": 0.889320969581604, "num_tokens": 528852883.0, "step": 13933 }, { "epoch": 1.772548021880168, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.373517036437988, "learning_rate": 5e-05, "loss": 0.9623, "mean_token_accuracy": 0.8608628511428833, "num_tokens": 528891561.0, "step": 13934 }, { "epoch": 1.7726752321587584, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.223808765411377, "learning_rate": 5e-05, "loss": 0.7199, "mean_token_accuracy": 0.867606520652771, "num_tokens": 528930755.0, "step": 13935 }, { "epoch": 1.772802442437349, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.578961372375488, "learning_rate": 5e-05, "loss": 1.0006, "mean_token_accuracy": 0.8670092225074768, "num_tokens": 528969230.0, "step": 13936 }, { "epoch": 1.7729296527159395, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.098090648651123, "learning_rate": 5e-05, "loss": 0.8312, "mean_token_accuracy": 0.8776991963386536, "num_tokens": 529005157.0, "step": 13937 }, { "epoch": 1.77305686299453, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.687716484069824, "learning_rate": 5e-05, "loss": 0.9068, "mean_token_accuracy": 0.8776332139968872, "num_tokens": 529042934.0, "step": 13938 }, { "epoch": 1.7731840732731206, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.293946266174316, "learning_rate": 5e-05, "loss": 0.957, "mean_token_accuracy": 0.8751223087310791, "num_tokens": 529087811.0, "step": 13939 }, { "epoch": 1.773311283551711, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.961907863616943, "learning_rate": 5e-05, "loss": 0.7901, "mean_token_accuracy": 0.8621179461479187, "num_tokens": 529133143.0, "step": 13940 }, { "epoch": 1.7734384938303016, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.533706665039062, "learning_rate": 5e-05, "loss": 0.955, "mean_token_accuracy": 0.8836054801940918, "num_tokens": 529172737.0, "step": 13941 }, { "epoch": 1.7735657041088921, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.078503608703613, "learning_rate": 5e-05, "loss": 0.7076, "mean_token_accuracy": 0.8734422922134399, "num_tokens": 529208913.0, "step": 13942 }, { "epoch": 1.7736929143874827, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 10.527839660644531, "learning_rate": 5e-05, "loss": 1.1071, "mean_token_accuracy": 0.8698480725288391, "num_tokens": 529250905.0, "step": 13943 }, { "epoch": 1.773820124666073, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.160452365875244, "learning_rate": 5e-05, "loss": 0.8658, "mean_token_accuracy": 0.8883870840072632, "num_tokens": 529289542.0, "step": 13944 }, { "epoch": 1.7739473349446635, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.228936195373535, "learning_rate": 5e-05, "loss": 1.0491, "mean_token_accuracy": 0.8793580532073975, "num_tokens": 529324759.0, "step": 13945 }, { "epoch": 1.774074545223254, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 8.767969131469727, "learning_rate": 5e-05, "loss": 1.1223, "mean_token_accuracy": 0.8706302642822266, "num_tokens": 529361845.0, "step": 13946 }, { "epoch": 1.7742017555018446, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.505415916442871, "learning_rate": 5e-05, "loss": 0.9116, "mean_token_accuracy": 0.8721873760223389, "num_tokens": 529402220.0, "step": 13947 }, { "epoch": 1.774328965780435, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.330921173095703, "learning_rate": 5e-05, "loss": 1.0939, "mean_token_accuracy": 0.8583698272705078, "num_tokens": 529440532.0, "step": 13948 }, { "epoch": 1.7744561760590256, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.474070072174072, "learning_rate": 5e-05, "loss": 0.8365, "mean_token_accuracy": 0.8747847080230713, "num_tokens": 529480357.0, "step": 13949 }, { "epoch": 1.774583386337616, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.370509147644043, "learning_rate": 5e-05, "loss": 1.0234, "mean_token_accuracy": 0.8643456101417542, "num_tokens": 529514075.0, "step": 13950 }, { "epoch": 1.7747105966162064, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.7517499923706055, "learning_rate": 5e-05, "loss": 0.9806, "mean_token_accuracy": 0.8597934246063232, "num_tokens": 529550738.0, "step": 13951 }, { "epoch": 1.774837806894797, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.432352066040039, "learning_rate": 5e-05, "loss": 0.9315, "mean_token_accuracy": 0.8634316921234131, "num_tokens": 529594151.0, "step": 13952 }, { "epoch": 1.7749650171733875, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.2622652053833, "learning_rate": 5e-05, "loss": 0.9286, "mean_token_accuracy": 0.8755711913108826, "num_tokens": 529633631.0, "step": 13953 }, { "epoch": 1.775092227451978, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.470635414123535, "learning_rate": 5e-05, "loss": 0.7881, "mean_token_accuracy": 0.869279682636261, "num_tokens": 529674377.0, "step": 13954 }, { "epoch": 1.7752194377305686, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.624530792236328, "learning_rate": 5e-05, "loss": 0.9687, "mean_token_accuracy": 0.8598820567131042, "num_tokens": 529707209.0, "step": 13955 }, { "epoch": 1.775346648009159, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.078039646148682, "learning_rate": 5e-05, "loss": 0.7816, "mean_token_accuracy": 0.8707632422447205, "num_tokens": 529737546.0, "step": 13956 }, { "epoch": 1.7754738582877496, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.848484992980957, "learning_rate": 5e-05, "loss": 0.9295, "mean_token_accuracy": 0.8722752332687378, "num_tokens": 529775532.0, "step": 13957 }, { "epoch": 1.7756010685663401, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 6.9930877685546875, "learning_rate": 5e-05, "loss": 0.7692, "mean_token_accuracy": 0.8799792528152466, "num_tokens": 529814705.0, "step": 13958 }, { "epoch": 1.7757282788449307, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.953965187072754, "learning_rate": 5e-05, "loss": 0.9296, "mean_token_accuracy": 0.8704276084899902, "num_tokens": 529850912.0, "step": 13959 }, { "epoch": 1.7758554891235212, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.110220909118652, "learning_rate": 5e-05, "loss": 0.8009, "mean_token_accuracy": 0.8802800178527832, "num_tokens": 529889042.0, "step": 13960 }, { "epoch": 1.7759826994021117, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.761703491210938, "learning_rate": 5e-05, "loss": 0.896, "mean_token_accuracy": 0.875463604927063, "num_tokens": 529926749.0, "step": 13961 }, { "epoch": 1.7761099096807023, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.293242454528809, "learning_rate": 5e-05, "loss": 0.832, "mean_token_accuracy": 0.8765735626220703, "num_tokens": 529951775.0, "step": 13962 }, { "epoch": 1.7762371199592928, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.454176902770996, "learning_rate": 5e-05, "loss": 0.8444, "mean_token_accuracy": 0.8843924403190613, "num_tokens": 529996172.0, "step": 13963 }, { "epoch": 1.7763643302378833, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.3227314949035645, "learning_rate": 5e-05, "loss": 0.8818, "mean_token_accuracy": 0.8538830280303955, "num_tokens": 530039577.0, "step": 13964 }, { "epoch": 1.7764915405164738, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.330168724060059, "learning_rate": 5e-05, "loss": 0.848, "mean_token_accuracy": 0.8816049098968506, "num_tokens": 530076516.0, "step": 13965 }, { "epoch": 1.7766187507950644, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.172749996185303, "learning_rate": 5e-05, "loss": 0.759, "mean_token_accuracy": 0.8821736574172974, "num_tokens": 530118552.0, "step": 13966 }, { "epoch": 1.776745961073655, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.394962310791016, "learning_rate": 5e-05, "loss": 0.8799, "mean_token_accuracy": 0.8668491840362549, "num_tokens": 530154971.0, "step": 13967 }, { "epoch": 1.7768731713522452, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.012022495269775, "learning_rate": 5e-05, "loss": 0.7073, "mean_token_accuracy": 0.8833209276199341, "num_tokens": 530190509.0, "step": 13968 }, { "epoch": 1.7770003816308357, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.526321411132812, "learning_rate": 5e-05, "loss": 0.8567, "mean_token_accuracy": 0.876444399356842, "num_tokens": 530229814.0, "step": 13969 }, { "epoch": 1.7771275919094263, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.911118030548096, "learning_rate": 5e-05, "loss": 0.7622, "mean_token_accuracy": 0.8665348291397095, "num_tokens": 530269426.0, "step": 13970 }, { "epoch": 1.7772548021880168, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.636680603027344, "learning_rate": 5e-05, "loss": 0.9081, "mean_token_accuracy": 0.8652769923210144, "num_tokens": 530305036.0, "step": 13971 }, { "epoch": 1.7773820124666073, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.924229145050049, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8684511184692383, "num_tokens": 530344995.0, "step": 13972 }, { "epoch": 1.7775092227451978, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.656501770019531, "learning_rate": 5e-05, "loss": 0.8868, "mean_token_accuracy": 0.878216564655304, "num_tokens": 530378283.0, "step": 13973 }, { "epoch": 1.7776364330237884, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.1189398765563965, "learning_rate": 5e-05, "loss": 0.8461, "mean_token_accuracy": 0.8557909727096558, "num_tokens": 530416147.0, "step": 13974 }, { "epoch": 1.7777636433023787, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.382617950439453, "learning_rate": 5e-05, "loss": 0.8794, "mean_token_accuracy": 0.870315432548523, "num_tokens": 530452096.0, "step": 13975 }, { "epoch": 1.7778908535809692, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.4372782707214355, "learning_rate": 5e-05, "loss": 0.8611, "mean_token_accuracy": 0.8600530624389648, "num_tokens": 530489347.0, "step": 13976 }, { "epoch": 1.7780180638595597, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.904177665710449, "learning_rate": 5e-05, "loss": 0.8463, "mean_token_accuracy": 0.8663778901100159, "num_tokens": 530528265.0, "step": 13977 }, { "epoch": 1.7781452741381503, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.731874942779541, "learning_rate": 5e-05, "loss": 0.7956, "mean_token_accuracy": 0.8833099007606506, "num_tokens": 530567501.0, "step": 13978 }, { "epoch": 1.7782724844167408, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.481564998626709, "learning_rate": 5e-05, "loss": 0.7982, "mean_token_accuracy": 0.8701232671737671, "num_tokens": 530610038.0, "step": 13979 }, { "epoch": 1.7783996946953313, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.891406059265137, "learning_rate": 5e-05, "loss": 0.8198, "mean_token_accuracy": 0.8682827353477478, "num_tokens": 530649039.0, "step": 13980 }, { "epoch": 1.7785269049739219, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.275932788848877, "learning_rate": 5e-05, "loss": 0.7289, "mean_token_accuracy": 0.8755474090576172, "num_tokens": 530687142.0, "step": 13981 }, { "epoch": 1.7786541152525124, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.925929546356201, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8824108839035034, "num_tokens": 530724399.0, "step": 13982 }, { "epoch": 1.778781325531103, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.2356133460998535, "learning_rate": 5e-05, "loss": 0.7315, "mean_token_accuracy": 0.8717643022537231, "num_tokens": 530764552.0, "step": 13983 }, { "epoch": 1.7789085358096934, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.8732757568359375, "learning_rate": 5e-05, "loss": 0.7879, "mean_token_accuracy": 0.8654634952545166, "num_tokens": 530808587.0, "step": 13984 }, { "epoch": 1.779035746088284, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.258554458618164, "learning_rate": 5e-05, "loss": 0.7481, "mean_token_accuracy": 0.8671761751174927, "num_tokens": 530845887.0, "step": 13985 }, { "epoch": 1.7791629563668745, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.813729286193848, "learning_rate": 5e-05, "loss": 0.8022, "mean_token_accuracy": 0.8593610525131226, "num_tokens": 530887398.0, "step": 13986 }, { "epoch": 1.779290166645465, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.283308982849121, "learning_rate": 5e-05, "loss": 0.6987, "mean_token_accuracy": 0.8811085224151611, "num_tokens": 530916377.0, "step": 13987 }, { "epoch": 1.7794173769240555, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.78097677230835, "learning_rate": 5e-05, "loss": 0.7955, "mean_token_accuracy": 0.8618595600128174, "num_tokens": 530954651.0, "step": 13988 }, { "epoch": 1.779544587202646, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.263147830963135, "learning_rate": 5e-05, "loss": 0.7564, "mean_token_accuracy": 0.8612945675849915, "num_tokens": 530999406.0, "step": 13989 }, { "epoch": 1.7796717974812366, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.790044784545898, "learning_rate": 5e-05, "loss": 0.7415, "mean_token_accuracy": 0.8760011792182922, "num_tokens": 531036925.0, "step": 13990 }, { "epoch": 1.7797990077598271, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.208240032196045, "learning_rate": 5e-05, "loss": 0.7086, "mean_token_accuracy": 0.8735865950584412, "num_tokens": 531076168.0, "step": 13991 }, { "epoch": 1.7799262180384177, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.878023147583008, "learning_rate": 5e-05, "loss": 0.7556, "mean_token_accuracy": 0.874459981918335, "num_tokens": 531113217.0, "step": 13992 }, { "epoch": 1.780053428317008, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.058319091796875, "learning_rate": 5e-05, "loss": 0.7093, "mean_token_accuracy": 0.8703518509864807, "num_tokens": 531155938.0, "step": 13993 }, { "epoch": 1.7801806385955985, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 8.077269554138184, "learning_rate": 5e-05, "loss": 0.7737, "mean_token_accuracy": 0.8719921112060547, "num_tokens": 531194556.0, "step": 13994 }, { "epoch": 1.780307848874189, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.791543006896973, "learning_rate": 5e-05, "loss": 0.7304, "mean_token_accuracy": 0.8588786721229553, "num_tokens": 531230704.0, "step": 13995 }, { "epoch": 1.7804350591527796, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.485113143920898, "learning_rate": 5e-05, "loss": 0.7966, "mean_token_accuracy": 0.8762319087982178, "num_tokens": 531270672.0, "step": 13996 }, { "epoch": 1.78056226943137, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.392001152038574, "learning_rate": 5e-05, "loss": 0.6331, "mean_token_accuracy": 0.8826375007629395, "num_tokens": 531302043.0, "step": 13997 }, { "epoch": 1.7806894797099606, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 9.069623947143555, "learning_rate": 5e-05, "loss": 0.8318, "mean_token_accuracy": 0.8858746290206909, "num_tokens": 531337014.0, "step": 13998 }, { "epoch": 1.780816689988551, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.006855010986328, "learning_rate": 5e-05, "loss": 0.6603, "mean_token_accuracy": 0.8785935640335083, "num_tokens": 531371537.0, "step": 13999 }, { "epoch": 1.7809439002671414, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.750465393066406, "learning_rate": 5e-05, "loss": 0.9489, "mean_token_accuracy": 0.8807420134544373, "num_tokens": 531406171.0, "step": 14000 }, { "epoch": 1.781071110545732, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.260748386383057, "learning_rate": 5e-05, "loss": 0.7811, "mean_token_accuracy": 0.8608641624450684, "num_tokens": 531445625.0, "step": 14001 }, { "epoch": 1.7811983208243225, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.814589500427246, "learning_rate": 5e-05, "loss": 0.9922, "mean_token_accuracy": 0.8814171552658081, "num_tokens": 531490748.0, "step": 14002 }, { "epoch": 1.781325531102913, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.6045451164245605, "learning_rate": 5e-05, "loss": 0.9356, "mean_token_accuracy": 0.8739005327224731, "num_tokens": 531526142.0, "step": 14003 }, { "epoch": 1.7814527413815036, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.353114128112793, "learning_rate": 5e-05, "loss": 0.8993, "mean_token_accuracy": 0.8756797909736633, "num_tokens": 531568325.0, "step": 14004 }, { "epoch": 1.781579951660094, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.029642105102539, "learning_rate": 5e-05, "loss": 1.023, "mean_token_accuracy": 0.8690279126167297, "num_tokens": 531608898.0, "step": 14005 }, { "epoch": 1.7817071619386846, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.32059383392334, "learning_rate": 5e-05, "loss": 0.7013, "mean_token_accuracy": 0.8712881803512573, "num_tokens": 531640860.0, "step": 14006 }, { "epoch": 1.7818343722172751, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 10.978425025939941, "learning_rate": 5e-05, "loss": 1.1411, "mean_token_accuracy": 0.8742754459381104, "num_tokens": 531680121.0, "step": 14007 }, { "epoch": 1.7819615824958657, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 6.6332807540893555, "learning_rate": 5e-05, "loss": 0.8958, "mean_token_accuracy": 0.8551039099693298, "num_tokens": 531713496.0, "step": 14008 }, { "epoch": 1.7820887927744562, "ewc_loss": 0.97265625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000919342041015625, "grad_norm": 12.295099258422852, "learning_rate": 5e-05, "loss": 1.3752, "mean_token_accuracy": 0.8631341457366943, "num_tokens": 531749989.0, "step": 14009 }, { "epoch": 1.7822160030530467, "ewc_loss": 0.99609375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00093841552734375, "grad_norm": 9.809137344360352, "learning_rate": 5e-05, "loss": 1.3683, "mean_token_accuracy": 0.8690236806869507, "num_tokens": 531788032.0, "step": 14010 }, { "epoch": 1.7823432133316373, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.802613258361816, "learning_rate": 5e-05, "loss": 1.1204, "mean_token_accuracy": 0.8480392694473267, "num_tokens": 531822178.0, "step": 14011 }, { "epoch": 1.7824704236102278, "ewc_loss": 0.85546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 9.598766326904297, "learning_rate": 5e-05, "loss": 1.2529, "mean_token_accuracy": 0.8635445237159729, "num_tokens": 531855713.0, "step": 14012 }, { "epoch": 1.7825976338888183, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 9.953706741333008, "learning_rate": 5e-05, "loss": 1.1016, "mean_token_accuracy": 0.8602303266525269, "num_tokens": 531898714.0, "step": 14013 }, { "epoch": 1.7827248441674088, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.913249969482422, "learning_rate": 5e-05, "loss": 1.0379, "mean_token_accuracy": 0.8675421476364136, "num_tokens": 531933968.0, "step": 14014 }, { "epoch": 1.7828520544459994, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.22197437286377, "learning_rate": 5e-05, "loss": 1.03, "mean_token_accuracy": 0.8728346824645996, "num_tokens": 531980672.0, "step": 14015 }, { "epoch": 1.78297926472459, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.233287811279297, "learning_rate": 5e-05, "loss": 0.8976, "mean_token_accuracy": 0.8732839822769165, "num_tokens": 532022461.0, "step": 14016 }, { "epoch": 1.7831064750031802, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.131970405578613, "learning_rate": 5e-05, "loss": 0.9459, "mean_token_accuracy": 0.8888598084449768, "num_tokens": 532059090.0, "step": 14017 }, { "epoch": 1.7832336852817707, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.637260437011719, "learning_rate": 5e-05, "loss": 0.8784, "mean_token_accuracy": 0.8629517555236816, "num_tokens": 532095926.0, "step": 14018 }, { "epoch": 1.7833608955603613, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.38771915435791, "learning_rate": 5e-05, "loss": 0.9844, "mean_token_accuracy": 0.8740063309669495, "num_tokens": 532131623.0, "step": 14019 }, { "epoch": 1.7834881058389518, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.456010341644287, "learning_rate": 5e-05, "loss": 0.8919, "mean_token_accuracy": 0.8556435108184814, "num_tokens": 532171594.0, "step": 14020 }, { "epoch": 1.7836153161175423, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.504997253417969, "learning_rate": 5e-05, "loss": 0.9585, "mean_token_accuracy": 0.8796478509902954, "num_tokens": 532209884.0, "step": 14021 }, { "epoch": 1.7837425263961328, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.571652412414551, "learning_rate": 5e-05, "loss": 0.9238, "mean_token_accuracy": 0.8607597947120667, "num_tokens": 532246182.0, "step": 14022 }, { "epoch": 1.7838697366747234, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.491628646850586, "learning_rate": 5e-05, "loss": 1.0003, "mean_token_accuracy": 0.8725694417953491, "num_tokens": 532281133.0, "step": 14023 }, { "epoch": 1.7839969469533137, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.68146276473999, "learning_rate": 5e-05, "loss": 0.8901, "mean_token_accuracy": 0.8677911758422852, "num_tokens": 532318593.0, "step": 14024 }, { "epoch": 1.7841241572319042, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.448850631713867, "learning_rate": 5e-05, "loss": 1.0507, "mean_token_accuracy": 0.8621618747711182, "num_tokens": 532353473.0, "step": 14025 }, { "epoch": 1.7842513675104947, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.804774284362793, "learning_rate": 5e-05, "loss": 0.9429, "mean_token_accuracy": 0.858199417591095, "num_tokens": 532395462.0, "step": 14026 }, { "epoch": 1.7843785777890853, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.254756927490234, "learning_rate": 5e-05, "loss": 0.9811, "mean_token_accuracy": 0.8778654336929321, "num_tokens": 532434677.0, "step": 14027 }, { "epoch": 1.7845057880676758, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.893946170806885, "learning_rate": 5e-05, "loss": 0.9151, "mean_token_accuracy": 0.8681045770645142, "num_tokens": 532470585.0, "step": 14028 }, { "epoch": 1.7846329983462663, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.00507640838623, "learning_rate": 5e-05, "loss": 1.002, "mean_token_accuracy": 0.8642364144325256, "num_tokens": 532502657.0, "step": 14029 }, { "epoch": 1.7847602086248568, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.858227252960205, "learning_rate": 5e-05, "loss": 0.8973, "mean_token_accuracy": 0.86323082447052, "num_tokens": 532548372.0, "step": 14030 }, { "epoch": 1.7848874189034474, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.798749923706055, "learning_rate": 5e-05, "loss": 0.9879, "mean_token_accuracy": 0.8621240854263306, "num_tokens": 532584707.0, "step": 14031 }, { "epoch": 1.785014629182038, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.738424777984619, "learning_rate": 5e-05, "loss": 0.8509, "mean_token_accuracy": 0.8712437748908997, "num_tokens": 532623179.0, "step": 14032 }, { "epoch": 1.7851418394606284, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.684673309326172, "learning_rate": 5e-05, "loss": 0.9443, "mean_token_accuracy": 0.8738064169883728, "num_tokens": 532663054.0, "step": 14033 }, { "epoch": 1.785269049739219, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.591439247131348, "learning_rate": 5e-05, "loss": 0.8359, "mean_token_accuracy": 0.8703193068504333, "num_tokens": 532696456.0, "step": 14034 }, { "epoch": 1.7853962600178095, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.695588111877441, "learning_rate": 5e-05, "loss": 0.8917, "mean_token_accuracy": 0.8861140608787537, "num_tokens": 532730726.0, "step": 14035 }, { "epoch": 1.7855234702964, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.477099418640137, "learning_rate": 5e-05, "loss": 0.8163, "mean_token_accuracy": 0.8684982657432556, "num_tokens": 532765561.0, "step": 14036 }, { "epoch": 1.7856506805749905, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.702889442443848, "learning_rate": 5e-05, "loss": 0.9145, "mean_token_accuracy": 0.8785692453384399, "num_tokens": 532803225.0, "step": 14037 }, { "epoch": 1.785777890853581, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.520340919494629, "learning_rate": 5e-05, "loss": 0.8251, "mean_token_accuracy": 0.869882345199585, "num_tokens": 532843284.0, "step": 14038 }, { "epoch": 1.7859051011321716, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.508744239807129, "learning_rate": 5e-05, "loss": 0.9273, "mean_token_accuracy": 0.8684422969818115, "num_tokens": 532877637.0, "step": 14039 }, { "epoch": 1.7860323114107621, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.616796493530273, "learning_rate": 5e-05, "loss": 0.8103, "mean_token_accuracy": 0.8717055320739746, "num_tokens": 532914678.0, "step": 14040 }, { "epoch": 1.7861595216893527, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.24167537689209, "learning_rate": 5e-05, "loss": 0.8816, "mean_token_accuracy": 0.8708913922309875, "num_tokens": 532952309.0, "step": 14041 }, { "epoch": 1.786286731967943, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.59786319732666, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8773568868637085, "num_tokens": 532986440.0, "step": 14042 }, { "epoch": 1.7864139422465335, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.136509895324707, "learning_rate": 5e-05, "loss": 0.8747, "mean_token_accuracy": 0.8638485074043274, "num_tokens": 533020488.0, "step": 14043 }, { "epoch": 1.786541152525124, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.480092525482178, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8750698566436768, "num_tokens": 533052535.0, "step": 14044 }, { "epoch": 1.7866683628037145, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.080427169799805, "learning_rate": 5e-05, "loss": 0.8203, "mean_token_accuracy": 0.8807368278503418, "num_tokens": 533092924.0, "step": 14045 }, { "epoch": 1.786795573082305, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.4173784255981445, "learning_rate": 5e-05, "loss": 0.7882, "mean_token_accuracy": 0.8666994571685791, "num_tokens": 533130497.0, "step": 14046 }, { "epoch": 1.7869227833608956, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.068737983703613, "learning_rate": 5e-05, "loss": 0.8846, "mean_token_accuracy": 0.8513609170913696, "num_tokens": 533170012.0, "step": 14047 }, { "epoch": 1.787049993639486, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.310436725616455, "learning_rate": 5e-05, "loss": 0.7544, "mean_token_accuracy": 0.8707525730133057, "num_tokens": 533207433.0, "step": 14048 }, { "epoch": 1.7871772039180764, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.102760314941406, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8843849301338196, "num_tokens": 533248809.0, "step": 14049 }, { "epoch": 1.787304414196667, "ewc_loss": 0.375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.2816548347473145, "learning_rate": 5e-05, "loss": 0.7205, "mean_token_accuracy": 0.8813003301620483, "num_tokens": 533283381.0, "step": 14050 }, { "epoch": 1.7874316244752575, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.043532371520996, "learning_rate": 5e-05, "loss": 0.7829, "mean_token_accuracy": 0.8817279934883118, "num_tokens": 533320951.0, "step": 14051 }, { "epoch": 1.787558834753848, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.3177409172058105, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.8741126656532288, "num_tokens": 533361080.0, "step": 14052 }, { "epoch": 1.7876860450324386, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.951087951660156, "learning_rate": 5e-05, "loss": 0.7603, "mean_token_accuracy": 0.8844363689422607, "num_tokens": 533396448.0, "step": 14053 }, { "epoch": 1.787813255311029, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.3627142906188965, "learning_rate": 5e-05, "loss": 0.744, "mean_token_accuracy": 0.8730621933937073, "num_tokens": 533434924.0, "step": 14054 }, { "epoch": 1.7879404655896196, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.8702216148376465, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.8722809553146362, "num_tokens": 533471248.0, "step": 14055 }, { "epoch": 1.7880676758682101, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.382232666015625, "learning_rate": 5e-05, "loss": 0.7199, "mean_token_accuracy": 0.8800405263900757, "num_tokens": 533512521.0, "step": 14056 }, { "epoch": 1.7881948861468007, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.768219947814941, "learning_rate": 5e-05, "loss": 0.7506, "mean_token_accuracy": 0.8788051605224609, "num_tokens": 533550385.0, "step": 14057 }, { "epoch": 1.7883220964253912, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.4615020751953125, "learning_rate": 5e-05, "loss": 0.7342, "mean_token_accuracy": 0.8733863234519958, "num_tokens": 533580048.0, "step": 14058 }, { "epoch": 1.7884493067039817, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.630367755889893, "learning_rate": 5e-05, "loss": 0.7201, "mean_token_accuracy": 0.8821167945861816, "num_tokens": 533615115.0, "step": 14059 }, { "epoch": 1.7885765169825723, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.545877456665039, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.8736483454704285, "num_tokens": 533653774.0, "step": 14060 }, { "epoch": 1.7887037272611628, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.511202812194824, "learning_rate": 5e-05, "loss": 0.7055, "mean_token_accuracy": 0.8837598562240601, "num_tokens": 533692628.0, "step": 14061 }, { "epoch": 1.7888309375397533, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.606660842895508, "learning_rate": 5e-05, "loss": 0.6964, "mean_token_accuracy": 0.889339804649353, "num_tokens": 533734935.0, "step": 14062 }, { "epoch": 1.7889581478183438, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.4313530921936035, "learning_rate": 5e-05, "loss": 0.7361, "mean_token_accuracy": 0.8652467131614685, "num_tokens": 533772541.0, "step": 14063 }, { "epoch": 1.7890853580969344, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.656855583190918, "learning_rate": 5e-05, "loss": 0.7206, "mean_token_accuracy": 0.8777204155921936, "num_tokens": 533816036.0, "step": 14064 }, { "epoch": 1.789212568375525, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.3481903076171875, "learning_rate": 5e-05, "loss": 0.7265, "mean_token_accuracy": 0.8710796236991882, "num_tokens": 533848174.0, "step": 14065 }, { "epoch": 1.7893397786541152, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.704174518585205, "learning_rate": 5e-05, "loss": 0.7791, "mean_token_accuracy": 0.8599066734313965, "num_tokens": 533890011.0, "step": 14066 }, { "epoch": 1.7894669889327057, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.292147636413574, "learning_rate": 5e-05, "loss": 0.7378, "mean_token_accuracy": 0.8616666197776794, "num_tokens": 533927650.0, "step": 14067 }, { "epoch": 1.7895941992112963, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.763597011566162, "learning_rate": 5e-05, "loss": 0.7601, "mean_token_accuracy": 0.8658584952354431, "num_tokens": 533959117.0, "step": 14068 }, { "epoch": 1.7897214094898868, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.185450553894043, "learning_rate": 5e-05, "loss": 0.743, "mean_token_accuracy": 0.8592990636825562, "num_tokens": 533995396.0, "step": 14069 }, { "epoch": 1.7898486197684773, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.887992858886719, "learning_rate": 5e-05, "loss": 0.7223, "mean_token_accuracy": 0.8805995583534241, "num_tokens": 534032498.0, "step": 14070 }, { "epoch": 1.7899758300470678, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 7.021342754364014, "learning_rate": 5e-05, "loss": 0.736, "mean_token_accuracy": 0.8635451793670654, "num_tokens": 534071790.0, "step": 14071 }, { "epoch": 1.7901030403256584, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.091426849365234, "learning_rate": 5e-05, "loss": 0.7918, "mean_token_accuracy": 0.8624974489212036, "num_tokens": 534112778.0, "step": 14072 }, { "epoch": 1.7902302506042487, "ewc_loss": 0.3125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.774970531463623, "learning_rate": 5e-05, "loss": 0.666, "mean_token_accuracy": 0.8765261173248291, "num_tokens": 534150815.0, "step": 14073 }, { "epoch": 1.7903574608828392, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.468234062194824, "learning_rate": 5e-05, "loss": 0.8094, "mean_token_accuracy": 0.869498610496521, "num_tokens": 534186148.0, "step": 14074 }, { "epoch": 1.7904846711614297, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 6.358524799346924, "learning_rate": 5e-05, "loss": 0.6722, "mean_token_accuracy": 0.8676884174346924, "num_tokens": 534219496.0, "step": 14075 }, { "epoch": 1.7906118814400203, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 9.12795352935791, "learning_rate": 5e-05, "loss": 0.8517, "mean_token_accuracy": 0.8770315647125244, "num_tokens": 534263093.0, "step": 14076 }, { "epoch": 1.7907390917186108, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002288818359375, "grad_norm": 5.888726711273193, "learning_rate": 5e-05, "loss": 0.6205, "mean_token_accuracy": 0.8824484944343567, "num_tokens": 534298515.0, "step": 14077 }, { "epoch": 1.7908663019972013, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.955506324768066, "learning_rate": 5e-05, "loss": 0.9685, "mean_token_accuracy": 0.8747621774673462, "num_tokens": 534337210.0, "step": 14078 }, { "epoch": 1.7909935122757918, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.10266637802124, "learning_rate": 5e-05, "loss": 0.7595, "mean_token_accuracy": 0.863140344619751, "num_tokens": 534373440.0, "step": 14079 }, { "epoch": 1.7911207225543824, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.144196510314941, "learning_rate": 5e-05, "loss": 0.9767, "mean_token_accuracy": 0.8947728276252747, "num_tokens": 534405560.0, "step": 14080 }, { "epoch": 1.791247932832973, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.706247329711914, "learning_rate": 5e-05, "loss": 0.9608, "mean_token_accuracy": 0.8724139928817749, "num_tokens": 534442824.0, "step": 14081 }, { "epoch": 1.7913751431115634, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.50364875793457, "learning_rate": 5e-05, "loss": 0.9111, "mean_token_accuracy": 0.8723634481430054, "num_tokens": 534482569.0, "step": 14082 }, { "epoch": 1.791502353390154, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.370854377746582, "learning_rate": 5e-05, "loss": 1.0719, "mean_token_accuracy": 0.8633770942687988, "num_tokens": 534518481.0, "step": 14083 }, { "epoch": 1.7916295636687445, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.189510822296143, "learning_rate": 5e-05, "loss": 0.7604, "mean_token_accuracy": 0.8501428961753845, "num_tokens": 534553273.0, "step": 14084 }, { "epoch": 1.791756773947335, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 11.986026763916016, "learning_rate": 5e-05, "loss": 1.2897, "mean_token_accuracy": 0.8652604818344116, "num_tokens": 534587322.0, "step": 14085 }, { "epoch": 1.7918839842259255, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 7.712095260620117, "learning_rate": 5e-05, "loss": 0.9364, "mean_token_accuracy": 0.8800326585769653, "num_tokens": 534622999.0, "step": 14086 }, { "epoch": 1.792011194504516, "ewc_loss": 1.0, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000946044921875, "grad_norm": 13.127240180969238, "learning_rate": 5e-05, "loss": 1.4081, "mean_token_accuracy": 0.8619264960289001, "num_tokens": 534655953.0, "step": 14087 }, { "epoch": 1.7921384047831066, "ewc_loss": 1.078125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0010223388671875, "grad_norm": 10.909634590148926, "learning_rate": 5e-05, "loss": 1.4558, "mean_token_accuracy": 0.8659588694572449, "num_tokens": 534699495.0, "step": 14088 }, { "epoch": 1.7922656150616971, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 11.77569580078125, "learning_rate": 5e-05, "loss": 1.1454, "mean_token_accuracy": 0.8786253929138184, "num_tokens": 534737650.0, "step": 14089 }, { "epoch": 1.7923928253402877, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 9.751910209655762, "learning_rate": 5e-05, "loss": 1.1355, "mean_token_accuracy": 0.875403881072998, "num_tokens": 534771169.0, "step": 14090 }, { "epoch": 1.792520035618878, "ewc_loss": 0.890625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 11.849048614501953, "learning_rate": 5e-05, "loss": 1.2276, "mean_token_accuracy": 0.880467414855957, "num_tokens": 534802968.0, "step": 14091 }, { "epoch": 1.7926472458974685, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.273305892944336, "learning_rate": 5e-05, "loss": 1.0055, "mean_token_accuracy": 0.8741097450256348, "num_tokens": 534846975.0, "step": 14092 }, { "epoch": 1.792774456176059, "ewc_loss": 0.828125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00077056884765625, "grad_norm": 11.572090148925781, "learning_rate": 5e-05, "loss": 1.187, "mean_token_accuracy": 0.876280665397644, "num_tokens": 534880969.0, "step": 14093 }, { "epoch": 1.7929016664546495, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 9.78095531463623, "learning_rate": 5e-05, "loss": 1.1081, "mean_token_accuracy": 0.8653852939605713, "num_tokens": 534918423.0, "step": 14094 }, { "epoch": 1.79302887673324, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.776267051696777, "learning_rate": 5e-05, "loss": 1.0324, "mean_token_accuracy": 0.878399133682251, "num_tokens": 534952733.0, "step": 14095 }, { "epoch": 1.7931560870118306, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.438478469848633, "learning_rate": 5e-05, "loss": 1.0242, "mean_token_accuracy": 0.8774269819259644, "num_tokens": 534981868.0, "step": 14096 }, { "epoch": 1.793283297290421, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.301688194274902, "learning_rate": 5e-05, "loss": 1.0454, "mean_token_accuracy": 0.86933833360672, "num_tokens": 535020940.0, "step": 14097 }, { "epoch": 1.7934105075690114, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.73240852355957, "learning_rate": 5e-05, "loss": 0.9159, "mean_token_accuracy": 0.8761235475540161, "num_tokens": 535053073.0, "step": 14098 }, { "epoch": 1.793537717847602, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.281517028808594, "learning_rate": 5e-05, "loss": 1.0433, "mean_token_accuracy": 0.8697818517684937, "num_tokens": 535097731.0, "step": 14099 }, { "epoch": 1.7936649281261925, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.187268257141113, "learning_rate": 5e-05, "loss": 0.8951, "mean_token_accuracy": 0.8658811450004578, "num_tokens": 535145571.0, "step": 14100 }, { "epoch": 1.793792138404783, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.80312728881836, "learning_rate": 5e-05, "loss": 1.1161, "mean_token_accuracy": 0.8695494532585144, "num_tokens": 535184226.0, "step": 14101 }, { "epoch": 1.7939193486833735, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.503019332885742, "learning_rate": 5e-05, "loss": 0.9922, "mean_token_accuracy": 0.8519719839096069, "num_tokens": 535221990.0, "step": 14102 }, { "epoch": 1.794046558961964, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 11.044988632202148, "learning_rate": 5e-05, "loss": 1.14, "mean_token_accuracy": 0.864362895488739, "num_tokens": 535261588.0, "step": 14103 }, { "epoch": 1.7941737692405546, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.323288917541504, "learning_rate": 5e-05, "loss": 1.0581, "mean_token_accuracy": 0.8526129126548767, "num_tokens": 535299023.0, "step": 14104 }, { "epoch": 1.7943009795191451, "ewc_loss": 0.671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.288393020629883, "learning_rate": 5e-05, "loss": 1.0487, "mean_token_accuracy": 0.8680826425552368, "num_tokens": 535332174.0, "step": 14105 }, { "epoch": 1.7944281897977357, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.314253807067871, "learning_rate": 5e-05, "loss": 0.9731, "mean_token_accuracy": 0.8756341934204102, "num_tokens": 535367242.0, "step": 14106 }, { "epoch": 1.7945554000763262, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.420401573181152, "learning_rate": 5e-05, "loss": 0.9568, "mean_token_accuracy": 0.8739484548568726, "num_tokens": 535401962.0, "step": 14107 }, { "epoch": 1.7946826103549167, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.78621768951416, "learning_rate": 5e-05, "loss": 0.9297, "mean_token_accuracy": 0.8627519607543945, "num_tokens": 535441481.0, "step": 14108 }, { "epoch": 1.7948098206335072, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.957815170288086, "learning_rate": 5e-05, "loss": 0.892, "mean_token_accuracy": 0.8787329196929932, "num_tokens": 535473619.0, "step": 14109 }, { "epoch": 1.7949370309120978, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.057059288024902, "learning_rate": 5e-05, "loss": 0.7847, "mean_token_accuracy": 0.8870217800140381, "num_tokens": 535515340.0, "step": 14110 }, { "epoch": 1.7950642411906883, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.140398979187012, "learning_rate": 5e-05, "loss": 0.9761, "mean_token_accuracy": 0.8558331727981567, "num_tokens": 535555090.0, "step": 14111 }, { "epoch": 1.7951914514692788, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.415361404418945, "learning_rate": 5e-05, "loss": 0.7812, "mean_token_accuracy": 0.8714009523391724, "num_tokens": 535595484.0, "step": 14112 }, { "epoch": 1.7953186617478694, "ewc_loss": 0.625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.798150062561035, "learning_rate": 5e-05, "loss": 1.0062, "mean_token_accuracy": 0.8703352808952332, "num_tokens": 535634170.0, "step": 14113 }, { "epoch": 1.7954458720264599, "ewc_loss": 0.421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.498046398162842, "learning_rate": 5e-05, "loss": 0.774, "mean_token_accuracy": 0.8749498128890991, "num_tokens": 535671771.0, "step": 14114 }, { "epoch": 1.7955730823050502, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.249119758605957, "learning_rate": 5e-05, "loss": 1.0353, "mean_token_accuracy": 0.8793975114822388, "num_tokens": 535710280.0, "step": 14115 }, { "epoch": 1.7957002925836407, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.495555877685547, "learning_rate": 5e-05, "loss": 0.9155, "mean_token_accuracy": 0.8684971332550049, "num_tokens": 535749509.0, "step": 14116 }, { "epoch": 1.7958275028622313, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.462740898132324, "learning_rate": 5e-05, "loss": 0.9527, "mean_token_accuracy": 0.879203200340271, "num_tokens": 535794184.0, "step": 14117 }, { "epoch": 1.7959547131408218, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.901154518127441, "learning_rate": 5e-05, "loss": 0.9845, "mean_token_accuracy": 0.8527576327323914, "num_tokens": 535828818.0, "step": 14118 }, { "epoch": 1.7960819234194123, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.20500659942627, "learning_rate": 5e-05, "loss": 0.875, "mean_token_accuracy": 0.8665329217910767, "num_tokens": 535869461.0, "step": 14119 }, { "epoch": 1.7962091336980028, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.649178504943848, "learning_rate": 5e-05, "loss": 0.8525, "mean_token_accuracy": 0.8747227191925049, "num_tokens": 535907287.0, "step": 14120 }, { "epoch": 1.7963363439765934, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.612908840179443, "learning_rate": 5e-05, "loss": 0.7605, "mean_token_accuracy": 0.8816888332366943, "num_tokens": 535949063.0, "step": 14121 }, { "epoch": 1.7964635542551837, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.28342056274414, "learning_rate": 5e-05, "loss": 0.8109, "mean_token_accuracy": 0.8748389482498169, "num_tokens": 535987202.0, "step": 14122 }, { "epoch": 1.7965907645337742, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.520635604858398, "learning_rate": 5e-05, "loss": 0.7474, "mean_token_accuracy": 0.8803672790527344, "num_tokens": 536021790.0, "step": 14123 }, { "epoch": 1.7967179748123647, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.109010696411133, "learning_rate": 5e-05, "loss": 0.8905, "mean_token_accuracy": 0.8460572957992554, "num_tokens": 536069932.0, "step": 14124 }, { "epoch": 1.7968451850909553, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.330482006072998, "learning_rate": 5e-05, "loss": 0.6969, "mean_token_accuracy": 0.8812708854675293, "num_tokens": 536110428.0, "step": 14125 }, { "epoch": 1.7969723953695458, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.3392915725708, "learning_rate": 5e-05, "loss": 0.8341, "mean_token_accuracy": 0.8682844638824463, "num_tokens": 536147711.0, "step": 14126 }, { "epoch": 1.7970996056481363, "ewc_loss": 0.34375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.04525089263916, "learning_rate": 5e-05, "loss": 0.6978, "mean_token_accuracy": 0.8780500888824463, "num_tokens": 536184422.0, "step": 14127 }, { "epoch": 1.7972268159267268, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.624648094177246, "learning_rate": 5e-05, "loss": 0.7962, "mean_token_accuracy": 0.8851494789123535, "num_tokens": 536224802.0, "step": 14128 }, { "epoch": 1.7973540262053174, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.903750896453857, "learning_rate": 5e-05, "loss": 0.7191, "mean_token_accuracy": 0.8704540729522705, "num_tokens": 536267596.0, "step": 14129 }, { "epoch": 1.797481236483908, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.847476959228516, "learning_rate": 5e-05, "loss": 0.8559, "mean_token_accuracy": 0.8753267526626587, "num_tokens": 536306130.0, "step": 14130 }, { "epoch": 1.7976084467624984, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.978697776794434, "learning_rate": 5e-05, "loss": 0.7283, "mean_token_accuracy": 0.8700026273727417, "num_tokens": 536345451.0, "step": 14131 }, { "epoch": 1.797735657041089, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.87816047668457, "learning_rate": 5e-05, "loss": 0.8841, "mean_token_accuracy": 0.8661007881164551, "num_tokens": 536377953.0, "step": 14132 }, { "epoch": 1.7978628673196795, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.253385543823242, "learning_rate": 5e-05, "loss": 0.7028, "mean_token_accuracy": 0.8833783864974976, "num_tokens": 536410879.0, "step": 14133 }, { "epoch": 1.79799007759827, "ewc_loss": 0.5, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.615646362304688, "learning_rate": 5e-05, "loss": 0.9094, "mean_token_accuracy": 0.8583548069000244, "num_tokens": 536448351.0, "step": 14134 }, { "epoch": 1.7981172878768605, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.527379989624023, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8668102025985718, "num_tokens": 536488968.0, "step": 14135 }, { "epoch": 1.798244498155451, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.225253105163574, "learning_rate": 5e-05, "loss": 0.8117, "mean_token_accuracy": 0.8810907006263733, "num_tokens": 536529189.0, "step": 14136 }, { "epoch": 1.7983717084340416, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.574716567993164, "learning_rate": 5e-05, "loss": 0.7053, "mean_token_accuracy": 0.8873578310012817, "num_tokens": 536567778.0, "step": 14137 }, { "epoch": 1.7984989187126321, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.175954818725586, "learning_rate": 5e-05, "loss": 0.8178, "mean_token_accuracy": 0.871314287185669, "num_tokens": 536609193.0, "step": 14138 }, { "epoch": 1.7986261289912227, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.239438056945801, "learning_rate": 5e-05, "loss": 0.7758, "mean_token_accuracy": 0.8552193641662598, "num_tokens": 536646122.0, "step": 14139 }, { "epoch": 1.798753339269813, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.560622215270996, "learning_rate": 5e-05, "loss": 0.8402, "mean_token_accuracy": 0.8695859313011169, "num_tokens": 536682146.0, "step": 14140 }, { "epoch": 1.7988805495484035, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.864121437072754, "learning_rate": 5e-05, "loss": 0.7378, "mean_token_accuracy": 0.8636796474456787, "num_tokens": 536715449.0, "step": 14141 }, { "epoch": 1.799007759826994, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.924783706665039, "learning_rate": 5e-05, "loss": 0.8955, "mean_token_accuracy": 0.8659565448760986, "num_tokens": 536752430.0, "step": 14142 }, { "epoch": 1.7991349701055845, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.934336185455322, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.8585695028305054, "num_tokens": 536785577.0, "step": 14143 }, { "epoch": 1.799262180384175, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.778950691223145, "learning_rate": 5e-05, "loss": 0.8506, "mean_token_accuracy": 0.881202220916748, "num_tokens": 536824074.0, "step": 14144 }, { "epoch": 1.7993893906627656, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.453146934509277, "learning_rate": 5e-05, "loss": 0.7247, "mean_token_accuracy": 0.8877622485160828, "num_tokens": 536865819.0, "step": 14145 }, { "epoch": 1.799516600941356, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.065876007080078, "learning_rate": 5e-05, "loss": 0.82, "mean_token_accuracy": 0.8719395399093628, "num_tokens": 536905538.0, "step": 14146 }, { "epoch": 1.7996438112199464, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.986176490783691, "learning_rate": 5e-05, "loss": 0.7882, "mean_token_accuracy": 0.8794838190078735, "num_tokens": 536944895.0, "step": 14147 }, { "epoch": 1.799771021498537, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.279918193817139, "learning_rate": 5e-05, "loss": 0.7508, "mean_token_accuracy": 0.8702622652053833, "num_tokens": 536981094.0, "step": 14148 }, { "epoch": 1.7998982317771275, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.335256576538086, "learning_rate": 5e-05, "loss": 0.8647, "mean_token_accuracy": 0.8592468500137329, "num_tokens": 537020242.0, "step": 14149 }, { "epoch": 1.800025442055718, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.970188140869141, "learning_rate": 5e-05, "loss": 0.7493, "mean_token_accuracy": 0.8593780994415283, "num_tokens": 537062475.0, "step": 14150 }, { "epoch": 1.8001526523343085, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.392741203308105, "learning_rate": 5e-05, "loss": 0.823, "mean_token_accuracy": 0.8681697845458984, "num_tokens": 537100969.0, "step": 14151 }, { "epoch": 1.800279862612899, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.022890090942383, "learning_rate": 5e-05, "loss": 0.7113, "mean_token_accuracy": 0.8701473474502563, "num_tokens": 537141828.0, "step": 14152 }, { "epoch": 1.8004070728914896, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.2172269821167, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.8764546513557434, "num_tokens": 537181324.0, "step": 14153 }, { "epoch": 1.8005342831700801, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.246103763580322, "learning_rate": 5e-05, "loss": 0.7125, "mean_token_accuracy": 0.877593994140625, "num_tokens": 537224593.0, "step": 14154 }, { "epoch": 1.8006614934486707, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.8940324783325195, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.866165816783905, "num_tokens": 537262835.0, "step": 14155 }, { "epoch": 1.8007887037272612, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.534505367279053, "learning_rate": 5e-05, "loss": 0.7354, "mean_token_accuracy": 0.8685756325721741, "num_tokens": 537299722.0, "step": 14156 }, { "epoch": 1.8009159140058517, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.530407428741455, "learning_rate": 5e-05, "loss": 0.7274, "mean_token_accuracy": 0.875313401222229, "num_tokens": 537345445.0, "step": 14157 }, { "epoch": 1.8010431242844422, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.856926918029785, "learning_rate": 5e-05, "loss": 0.7496, "mean_token_accuracy": 0.8722779750823975, "num_tokens": 537382327.0, "step": 14158 }, { "epoch": 1.8011703345630328, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.175506591796875, "learning_rate": 5e-05, "loss": 0.7156, "mean_token_accuracy": 0.8709098696708679, "num_tokens": 537419013.0, "step": 14159 }, { "epoch": 1.8012975448416233, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.138676643371582, "learning_rate": 5e-05, "loss": 0.7847, "mean_token_accuracy": 0.8702192306518555, "num_tokens": 537459841.0, "step": 14160 }, { "epoch": 1.8014247551202138, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.913414001464844, "learning_rate": 5e-05, "loss": 0.6551, "mean_token_accuracy": 0.8842257261276245, "num_tokens": 537494124.0, "step": 14161 }, { "epoch": 1.8015519653988044, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.407271385192871, "learning_rate": 5e-05, "loss": 0.7823, "mean_token_accuracy": 0.8784832954406738, "num_tokens": 537532798.0, "step": 14162 }, { "epoch": 1.8016791756773949, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.745466709136963, "learning_rate": 5e-05, "loss": 0.6955, "mean_token_accuracy": 0.8699318766593933, "num_tokens": 537568548.0, "step": 14163 }, { "epoch": 1.8018063859559852, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.609038352966309, "learning_rate": 5e-05, "loss": 0.7892, "mean_token_accuracy": 0.8849451541900635, "num_tokens": 537604413.0, "step": 14164 }, { "epoch": 1.8019335962345757, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.654429912567139, "learning_rate": 5e-05, "loss": 0.6957, "mean_token_accuracy": 0.8728606700897217, "num_tokens": 537640588.0, "step": 14165 }, { "epoch": 1.8020608065131662, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.670857429504395, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.8635349869728088, "num_tokens": 537682664.0, "step": 14166 }, { "epoch": 1.8021880167917568, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.790985107421875, "learning_rate": 5e-05, "loss": 0.7452, "mean_token_accuracy": 0.8557518720626831, "num_tokens": 537716020.0, "step": 14167 }, { "epoch": 1.8023152270703473, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.511286735534668, "learning_rate": 5e-05, "loss": 0.8295, "mean_token_accuracy": 0.8750187158584595, "num_tokens": 537756895.0, "step": 14168 }, { "epoch": 1.8024424373489378, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.067400932312012, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.8672217726707458, "num_tokens": 537794130.0, "step": 14169 }, { "epoch": 1.8025696476275284, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.184383392333984, "learning_rate": 5e-05, "loss": 0.7782, "mean_token_accuracy": 0.8831061124801636, "num_tokens": 537822865.0, "step": 14170 }, { "epoch": 1.8026968579061187, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.30689001083374, "learning_rate": 5e-05, "loss": 0.7246, "mean_token_accuracy": 0.8763895034790039, "num_tokens": 537854594.0, "step": 14171 }, { "epoch": 1.8028240681847092, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.966864585876465, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8720511794090271, "num_tokens": 537889667.0, "step": 14172 }, { "epoch": 1.8029512784632997, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.35852575302124, "learning_rate": 5e-05, "loss": 0.7744, "mean_token_accuracy": 0.8582472205162048, "num_tokens": 537924322.0, "step": 14173 }, { "epoch": 1.8030784887418903, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.986422061920166, "learning_rate": 5e-05, "loss": 0.7615, "mean_token_accuracy": 0.8807587623596191, "num_tokens": 537962196.0, "step": 14174 }, { "epoch": 1.8032056990204808, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.129300117492676, "learning_rate": 5e-05, "loss": 0.7085, "mean_token_accuracy": 0.8789548277854919, "num_tokens": 537999916.0, "step": 14175 }, { "epoch": 1.8033329092990713, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.318466186523438, "learning_rate": 5e-05, "loss": 0.7853, "mean_token_accuracy": 0.8807210326194763, "num_tokens": 538035823.0, "step": 14176 }, { "epoch": 1.8034601195776618, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.737361907958984, "learning_rate": 5e-05, "loss": 0.7363, "mean_token_accuracy": 0.861553430557251, "num_tokens": 538080664.0, "step": 14177 }, { "epoch": 1.8035873298562524, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.807133674621582, "learning_rate": 5e-05, "loss": 0.8861, "mean_token_accuracy": 0.8649470210075378, "num_tokens": 538123178.0, "step": 14178 }, { "epoch": 1.803714540134843, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.464203834533691, "learning_rate": 5e-05, "loss": 0.6818, "mean_token_accuracy": 0.8761686086654663, "num_tokens": 538158757.0, "step": 14179 }, { "epoch": 1.8038417504134334, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.147470474243164, "learning_rate": 5e-05, "loss": 0.9354, "mean_token_accuracy": 0.8656489253044128, "num_tokens": 538199019.0, "step": 14180 }, { "epoch": 1.803968960692024, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.730195045471191, "learning_rate": 5e-05, "loss": 0.7325, "mean_token_accuracy": 0.869952917098999, "num_tokens": 538239839.0, "step": 14181 }, { "epoch": 1.8040961709706145, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.752450942993164, "learning_rate": 5e-05, "loss": 0.9039, "mean_token_accuracy": 0.8739409446716309, "num_tokens": 538287677.0, "step": 14182 }, { "epoch": 1.804223381249205, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.62205171585083, "learning_rate": 5e-05, "loss": 0.8241, "mean_token_accuracy": 0.8734560012817383, "num_tokens": 538322304.0, "step": 14183 }, { "epoch": 1.8043505915277955, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.48624849319458, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.8768196105957031, "num_tokens": 538364087.0, "step": 14184 }, { "epoch": 1.804477801806386, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.67101001739502, "learning_rate": 5e-05, "loss": 0.8838, "mean_token_accuracy": 0.869382381439209, "num_tokens": 538401507.0, "step": 14185 }, { "epoch": 1.8046050120849766, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.299993515014648, "learning_rate": 5e-05, "loss": 0.6759, "mean_token_accuracy": 0.8758276700973511, "num_tokens": 538440666.0, "step": 14186 }, { "epoch": 1.8047322223635671, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.738635063171387, "learning_rate": 5e-05, "loss": 1.0066, "mean_token_accuracy": 0.8628454208374023, "num_tokens": 538478915.0, "step": 14187 }, { "epoch": 1.8048594326421576, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 5.943839073181152, "learning_rate": 5e-05, "loss": 0.6856, "mean_token_accuracy": 0.874995231628418, "num_tokens": 538517770.0, "step": 14188 }, { "epoch": 1.804986642920748, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.4185791015625, "learning_rate": 5e-05, "loss": 1.1022, "mean_token_accuracy": 0.8607202768325806, "num_tokens": 538555812.0, "step": 14189 }, { "epoch": 1.8051138531993385, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.200074195861816, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.8875223398208618, "num_tokens": 538596023.0, "step": 14190 }, { "epoch": 1.805241063477929, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.713157653808594, "learning_rate": 5e-05, "loss": 0.9902, "mean_token_accuracy": 0.8861707448959351, "num_tokens": 538630840.0, "step": 14191 }, { "epoch": 1.8053682737565195, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 8.736466407775879, "learning_rate": 5e-05, "loss": 1.0877, "mean_token_accuracy": 0.8625883460044861, "num_tokens": 538668515.0, "step": 14192 }, { "epoch": 1.80549548403511, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.754075527191162, "learning_rate": 5e-05, "loss": 0.8712, "mean_token_accuracy": 0.8637754321098328, "num_tokens": 538703996.0, "step": 14193 }, { "epoch": 1.8056226943137006, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.709758758544922, "learning_rate": 5e-05, "loss": 1.0236, "mean_token_accuracy": 0.8708993196487427, "num_tokens": 538742819.0, "step": 14194 }, { "epoch": 1.805749904592291, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 6.927227973937988, "learning_rate": 5e-05, "loss": 0.7671, "mean_token_accuracy": 0.8811817169189453, "num_tokens": 538776218.0, "step": 14195 }, { "epoch": 1.8058771148708814, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.033482551574707, "learning_rate": 5e-05, "loss": 1.0452, "mean_token_accuracy": 0.8731248378753662, "num_tokens": 538813377.0, "step": 14196 }, { "epoch": 1.806004325149472, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.796849250793457, "learning_rate": 5e-05, "loss": 0.9622, "mean_token_accuracy": 0.86890709400177, "num_tokens": 538856755.0, "step": 14197 }, { "epoch": 1.8061315354280625, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.943902015686035, "learning_rate": 5e-05, "loss": 0.948, "mean_token_accuracy": 0.8718970417976379, "num_tokens": 538898175.0, "step": 14198 }, { "epoch": 1.806258745706653, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.553923606872559, "learning_rate": 5e-05, "loss": 1.011, "mean_token_accuracy": 0.8649278879165649, "num_tokens": 538935539.0, "step": 14199 }, { "epoch": 1.8063859559852435, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.728973388671875, "learning_rate": 5e-05, "loss": 0.8757, "mean_token_accuracy": 0.8577855825424194, "num_tokens": 538978999.0, "step": 14200 }, { "epoch": 1.806513166263834, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.806142807006836, "learning_rate": 5e-05, "loss": 0.9225, "mean_token_accuracy": 0.8803892135620117, "num_tokens": 539013480.0, "step": 14201 }, { "epoch": 1.8066403765424246, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.438181400299072, "learning_rate": 5e-05, "loss": 0.8582, "mean_token_accuracy": 0.8637934923171997, "num_tokens": 539047120.0, "step": 14202 }, { "epoch": 1.8067675868210151, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.555858612060547, "learning_rate": 5e-05, "loss": 0.8818, "mean_token_accuracy": 0.8824765682220459, "num_tokens": 539088243.0, "step": 14203 }, { "epoch": 1.8068947970996057, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.604351043701172, "learning_rate": 5e-05, "loss": 0.8802, "mean_token_accuracy": 0.8641213178634644, "num_tokens": 539128240.0, "step": 14204 }, { "epoch": 1.8070220073781962, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.034538269042969, "learning_rate": 5e-05, "loss": 0.8448, "mean_token_accuracy": 0.8688871264457703, "num_tokens": 539168792.0, "step": 14205 }, { "epoch": 1.8071492176567867, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.829165458679199, "learning_rate": 5e-05, "loss": 0.8231, "mean_token_accuracy": 0.8757367134094238, "num_tokens": 539210510.0, "step": 14206 }, { "epoch": 1.8072764279353772, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.564374923706055, "learning_rate": 5e-05, "loss": 0.7748, "mean_token_accuracy": 0.8732659816741943, "num_tokens": 539248926.0, "step": 14207 }, { "epoch": 1.8074036382139678, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.062955856323242, "learning_rate": 5e-05, "loss": 0.8456, "mean_token_accuracy": 0.8672924637794495, "num_tokens": 539285795.0, "step": 14208 }, { "epoch": 1.8075308484925583, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.218856334686279, "learning_rate": 5e-05, "loss": 0.7214, "mean_token_accuracy": 0.8766763210296631, "num_tokens": 539324891.0, "step": 14209 }, { "epoch": 1.8076580587711488, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.356566429138184, "learning_rate": 5e-05, "loss": 0.8604, "mean_token_accuracy": 0.861944317817688, "num_tokens": 539364684.0, "step": 14210 }, { "epoch": 1.8077852690497394, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.9079766273498535, "learning_rate": 5e-05, "loss": 0.7031, "mean_token_accuracy": 0.8769791722297668, "num_tokens": 539403312.0, "step": 14211 }, { "epoch": 1.8079124793283299, "ewc_loss": 0.5, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.701743125915527, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.8824804425239563, "num_tokens": 539443567.0, "step": 14212 }, { "epoch": 1.8080396896069202, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.711564540863037, "learning_rate": 5e-05, "loss": 0.7265, "mean_token_accuracy": 0.8663330674171448, "num_tokens": 539480969.0, "step": 14213 }, { "epoch": 1.8081668998855107, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.880797386169434, "learning_rate": 5e-05, "loss": 0.9041, "mean_token_accuracy": 0.8715108633041382, "num_tokens": 539518105.0, "step": 14214 }, { "epoch": 1.8082941101641012, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.724727153778076, "learning_rate": 5e-05, "loss": 0.7779, "mean_token_accuracy": 0.8571926355361938, "num_tokens": 539561364.0, "step": 14215 }, { "epoch": 1.8084213204426918, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.921709060668945, "learning_rate": 5e-05, "loss": 0.9334, "mean_token_accuracy": 0.865554928779602, "num_tokens": 539594602.0, "step": 14216 }, { "epoch": 1.8085485307212823, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.856815814971924, "learning_rate": 5e-05, "loss": 0.8092, "mean_token_accuracy": 0.8575900793075562, "num_tokens": 539635985.0, "step": 14217 }, { "epoch": 1.8086757409998728, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.8391695022583, "learning_rate": 5e-05, "loss": 0.9061, "mean_token_accuracy": 0.8735941052436829, "num_tokens": 539675201.0, "step": 14218 }, { "epoch": 1.8088029512784631, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.030954360961914, "learning_rate": 5e-05, "loss": 0.768, "mean_token_accuracy": 0.8818008899688721, "num_tokens": 539712661.0, "step": 14219 }, { "epoch": 1.8089301615570537, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.682720184326172, "learning_rate": 5e-05, "loss": 0.9461, "mean_token_accuracy": 0.8577154874801636, "num_tokens": 539749941.0, "step": 14220 }, { "epoch": 1.8090573718356442, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.190511226654053, "learning_rate": 5e-05, "loss": 0.8204, "mean_token_accuracy": 0.8667581081390381, "num_tokens": 539786906.0, "step": 14221 }, { "epoch": 1.8091845821142347, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.494291305541992, "learning_rate": 5e-05, "loss": 0.8843, "mean_token_accuracy": 0.8764333724975586, "num_tokens": 539824235.0, "step": 14222 }, { "epoch": 1.8093117923928252, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.276533603668213, "learning_rate": 5e-05, "loss": 0.8493, "mean_token_accuracy": 0.8598330616950989, "num_tokens": 539864326.0, "step": 14223 }, { "epoch": 1.8094390026714158, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.366361618041992, "learning_rate": 5e-05, "loss": 0.9271, "mean_token_accuracy": 0.8573824167251587, "num_tokens": 539909105.0, "step": 14224 }, { "epoch": 1.8095662129500063, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.30053186416626, "learning_rate": 5e-05, "loss": 0.7981, "mean_token_accuracy": 0.872214674949646, "num_tokens": 539944846.0, "step": 14225 }, { "epoch": 1.8096934232285968, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.24089241027832, "learning_rate": 5e-05, "loss": 0.8433, "mean_token_accuracy": 0.874426007270813, "num_tokens": 539984720.0, "step": 14226 }, { "epoch": 1.8098206335071874, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.330686092376709, "learning_rate": 5e-05, "loss": 0.8063, "mean_token_accuracy": 0.8641969561576843, "num_tokens": 540025002.0, "step": 14227 }, { "epoch": 1.8099478437857779, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.12602710723877, "learning_rate": 5e-05, "loss": 0.8777, "mean_token_accuracy": 0.8616607189178467, "num_tokens": 540060256.0, "step": 14228 }, { "epoch": 1.8100750540643684, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.363783359527588, "learning_rate": 5e-05, "loss": 0.7841, "mean_token_accuracy": 0.8695162534713745, "num_tokens": 540098073.0, "step": 14229 }, { "epoch": 1.810202264342959, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.009381294250488, "learning_rate": 5e-05, "loss": 0.8077, "mean_token_accuracy": 0.8731680512428284, "num_tokens": 540128704.0, "step": 14230 }, { "epoch": 1.8103294746215495, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.399701118469238, "learning_rate": 5e-05, "loss": 0.7296, "mean_token_accuracy": 0.8821545839309692, "num_tokens": 540161293.0, "step": 14231 }, { "epoch": 1.81045668490014, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.91337776184082, "learning_rate": 5e-05, "loss": 0.7689, "mean_token_accuracy": 0.8842486143112183, "num_tokens": 540193685.0, "step": 14232 }, { "epoch": 1.8105838951787305, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.418072700500488, "learning_rate": 5e-05, "loss": 0.7106, "mean_token_accuracy": 0.8834615349769592, "num_tokens": 540232217.0, "step": 14233 }, { "epoch": 1.810711105457321, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.831244945526123, "learning_rate": 5e-05, "loss": 0.794, "mean_token_accuracy": 0.8661760687828064, "num_tokens": 540267561.0, "step": 14234 }, { "epoch": 1.8108383157359116, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.414753437042236, "learning_rate": 5e-05, "loss": 0.7251, "mean_token_accuracy": 0.8794435858726501, "num_tokens": 540305788.0, "step": 14235 }, { "epoch": 1.8109655260145021, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.83900785446167, "learning_rate": 5e-05, "loss": 0.7918, "mean_token_accuracy": 0.8662261366844177, "num_tokens": 540341105.0, "step": 14236 }, { "epoch": 1.8110927362930926, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.345834732055664, "learning_rate": 5e-05, "loss": 0.7596, "mean_token_accuracy": 0.8677194714546204, "num_tokens": 540380502.0, "step": 14237 }, { "epoch": 1.811219946571683, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.867598533630371, "learning_rate": 5e-05, "loss": 0.8313, "mean_token_accuracy": 0.8526079654693604, "num_tokens": 540420561.0, "step": 14238 }, { "epoch": 1.8113471568502735, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.274519443511963, "learning_rate": 5e-05, "loss": 0.741, "mean_token_accuracy": 0.8651610016822815, "num_tokens": 540460561.0, "step": 14239 }, { "epoch": 1.811474367128864, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.928707122802734, "learning_rate": 5e-05, "loss": 0.783, "mean_token_accuracy": 0.8686757683753967, "num_tokens": 540498939.0, "step": 14240 }, { "epoch": 1.8116015774074545, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.180559158325195, "learning_rate": 5e-05, "loss": 0.7403, "mean_token_accuracy": 0.8612602949142456, "num_tokens": 540536127.0, "step": 14241 }, { "epoch": 1.811728787686045, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.038942337036133, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8842114210128784, "num_tokens": 540567812.0, "step": 14242 }, { "epoch": 1.8118559979646356, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.057912826538086, "learning_rate": 5e-05, "loss": 0.6733, "mean_token_accuracy": 0.8834707140922546, "num_tokens": 540604176.0, "step": 14243 }, { "epoch": 1.811983208243226, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.177334785461426, "learning_rate": 5e-05, "loss": 0.8403, "mean_token_accuracy": 0.8539067506790161, "num_tokens": 540640989.0, "step": 14244 }, { "epoch": 1.8121104185218164, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.90225887298584, "learning_rate": 5e-05, "loss": 0.69, "mean_token_accuracy": 0.876308798789978, "num_tokens": 540680590.0, "step": 14245 }, { "epoch": 1.812237628800407, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.369593620300293, "learning_rate": 5e-05, "loss": 0.8, "mean_token_accuracy": 0.8739741444587708, "num_tokens": 540723739.0, "step": 14246 }, { "epoch": 1.8123648390789975, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.779932022094727, "learning_rate": 5e-05, "loss": 0.7005, "mean_token_accuracy": 0.8706609606742859, "num_tokens": 540759686.0, "step": 14247 }, { "epoch": 1.812492049357588, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.516508102416992, "learning_rate": 5e-05, "loss": 0.7987, "mean_token_accuracy": 0.884486198425293, "num_tokens": 540795710.0, "step": 14248 }, { "epoch": 1.8126192596361785, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.7461724281311035, "learning_rate": 5e-05, "loss": 0.7068, "mean_token_accuracy": 0.8707116842269897, "num_tokens": 540838558.0, "step": 14249 }, { "epoch": 1.812746469914769, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.524138450622559, "learning_rate": 5e-05, "loss": 0.8145, "mean_token_accuracy": 0.8835351467132568, "num_tokens": 540874543.0, "step": 14250 }, { "epoch": 1.8128736801933596, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.9214019775390625, "learning_rate": 5e-05, "loss": 0.7323, "mean_token_accuracy": 0.873843252658844, "num_tokens": 540909903.0, "step": 14251 }, { "epoch": 1.8130008904719501, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.235428810119629, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.8709036111831665, "num_tokens": 540947977.0, "step": 14252 }, { "epoch": 1.8131281007505406, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.346977233886719, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8613504767417908, "num_tokens": 540986288.0, "step": 14253 }, { "epoch": 1.8132553110291312, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.618744850158691, "learning_rate": 5e-05, "loss": 0.7605, "mean_token_accuracy": 0.8721495866775513, "num_tokens": 541027607.0, "step": 14254 }, { "epoch": 1.8133825213077217, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.959672927856445, "learning_rate": 5e-05, "loss": 0.7784, "mean_token_accuracy": 0.8744111061096191, "num_tokens": 541064424.0, "step": 14255 }, { "epoch": 1.8135097315863122, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.861276626586914, "learning_rate": 5e-05, "loss": 0.7064, "mean_token_accuracy": 0.8659946918487549, "num_tokens": 541097815.0, "step": 14256 }, { "epoch": 1.8136369418649028, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.691790580749512, "learning_rate": 5e-05, "loss": 0.8793, "mean_token_accuracy": 0.8644164204597473, "num_tokens": 541136567.0, "step": 14257 }, { "epoch": 1.8137641521434933, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.169503688812256, "learning_rate": 5e-05, "loss": 0.6647, "mean_token_accuracy": 0.8743994235992432, "num_tokens": 541174305.0, "step": 14258 }, { "epoch": 1.8138913624220838, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.603190422058105, "learning_rate": 5e-05, "loss": 0.9542, "mean_token_accuracy": 0.8705074787139893, "num_tokens": 541207118.0, "step": 14259 }, { "epoch": 1.8140185727006743, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.783487319946289, "learning_rate": 5e-05, "loss": 0.6454, "mean_token_accuracy": 0.8821605443954468, "num_tokens": 541242208.0, "step": 14260 }, { "epoch": 1.8141457829792649, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.458009719848633, "learning_rate": 5e-05, "loss": 1.0756, "mean_token_accuracy": 0.8732355833053589, "num_tokens": 541278553.0, "step": 14261 }, { "epoch": 1.8142729932578552, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 6.841970443725586, "learning_rate": 5e-05, "loss": 0.8845, "mean_token_accuracy": 0.8670096397399902, "num_tokens": 541318329.0, "step": 14262 }, { "epoch": 1.8144002035364457, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.104334831237793, "learning_rate": 5e-05, "loss": 1.0575, "mean_token_accuracy": 0.8790837526321411, "num_tokens": 541353406.0, "step": 14263 }, { "epoch": 1.8145274138150362, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 8.648818016052246, "learning_rate": 5e-05, "loss": 1.1112, "mean_token_accuracy": 0.8670153617858887, "num_tokens": 541385504.0, "step": 14264 }, { "epoch": 1.8146546240936268, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.9927077293396, "learning_rate": 5e-05, "loss": 0.8818, "mean_token_accuracy": 0.8644512891769409, "num_tokens": 541426172.0, "step": 14265 }, { "epoch": 1.8147818343722173, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.022356986999512, "learning_rate": 5e-05, "loss": 1.1329, "mean_token_accuracy": 0.8741944432258606, "num_tokens": 541464785.0, "step": 14266 }, { "epoch": 1.8149090446508078, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.628671646118164, "learning_rate": 5e-05, "loss": 0.8084, "mean_token_accuracy": 0.8637180328369141, "num_tokens": 541500271.0, "step": 14267 }, { "epoch": 1.8150362549293981, "ewc_loss": 0.87109375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 11.480790138244629, "learning_rate": 5e-05, "loss": 1.2561, "mean_token_accuracy": 0.8668705224990845, "num_tokens": 541537482.0, "step": 14268 }, { "epoch": 1.8151634652079887, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 8.318218231201172, "learning_rate": 5e-05, "loss": 1.0638, "mean_token_accuracy": 0.8745726346969604, "num_tokens": 541576669.0, "step": 14269 }, { "epoch": 1.8152906754865792, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 11.030317306518555, "learning_rate": 5e-05, "loss": 1.1049, "mean_token_accuracy": 0.8710840940475464, "num_tokens": 541616040.0, "step": 14270 }, { "epoch": 1.8154178857651697, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0007781982421875, "grad_norm": 9.110759735107422, "learning_rate": 5e-05, "loss": 1.2516, "mean_token_accuracy": 0.8564473986625671, "num_tokens": 541649994.0, "step": 14271 }, { "epoch": 1.8155450960437602, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.114851951599121, "learning_rate": 5e-05, "loss": 1.0862, "mean_token_accuracy": 0.8579648733139038, "num_tokens": 541688468.0, "step": 14272 }, { "epoch": 1.8156723063223508, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.630032539367676, "learning_rate": 5e-05, "loss": 1.0294, "mean_token_accuracy": 0.8682016134262085, "num_tokens": 541725173.0, "step": 14273 }, { "epoch": 1.8157995166009413, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.518474578857422, "learning_rate": 5e-05, "loss": 1.0379, "mean_token_accuracy": 0.8686562776565552, "num_tokens": 541762523.0, "step": 14274 }, { "epoch": 1.8159267268795318, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.947839736938477, "learning_rate": 5e-05, "loss": 0.9023, "mean_token_accuracy": 0.8678334951400757, "num_tokens": 541799731.0, "step": 14275 }, { "epoch": 1.8160539371581224, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.55888557434082, "learning_rate": 5e-05, "loss": 0.997, "mean_token_accuracy": 0.8808706402778625, "num_tokens": 541844172.0, "step": 14276 }, { "epoch": 1.8161811474367129, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.472853660583496, "learning_rate": 5e-05, "loss": 0.8547, "mean_token_accuracy": 0.8727765083312988, "num_tokens": 541887526.0, "step": 14277 }, { "epoch": 1.8163083577153034, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.97492504119873, "learning_rate": 5e-05, "loss": 1.0383, "mean_token_accuracy": 0.8706601858139038, "num_tokens": 541932712.0, "step": 14278 }, { "epoch": 1.816435567993894, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.6220293045043945, "learning_rate": 5e-05, "loss": 0.9122, "mean_token_accuracy": 0.871351957321167, "num_tokens": 541968915.0, "step": 14279 }, { "epoch": 1.8165627782724845, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.087281227111816, "learning_rate": 5e-05, "loss": 1.0898, "mean_token_accuracy": 0.8651328682899475, "num_tokens": 542010296.0, "step": 14280 }, { "epoch": 1.816689988551075, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.129788398742676, "learning_rate": 5e-05, "loss": 0.9222, "mean_token_accuracy": 0.8866719007492065, "num_tokens": 542050096.0, "step": 14281 }, { "epoch": 1.8168171988296655, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.696579933166504, "learning_rate": 5e-05, "loss": 1.0479, "mean_token_accuracy": 0.867148756980896, "num_tokens": 542086746.0, "step": 14282 }, { "epoch": 1.816944409108256, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.35775089263916, "learning_rate": 5e-05, "loss": 0.9694, "mean_token_accuracy": 0.8760715126991272, "num_tokens": 542123095.0, "step": 14283 }, { "epoch": 1.8170716193868466, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.08786392211914, "learning_rate": 5e-05, "loss": 0.9702, "mean_token_accuracy": 0.8763226270675659, "num_tokens": 542157797.0, "step": 14284 }, { "epoch": 1.817198829665437, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.257979393005371, "learning_rate": 5e-05, "loss": 0.9385, "mean_token_accuracy": 0.8644567728042603, "num_tokens": 542196091.0, "step": 14285 }, { "epoch": 1.8173260399440276, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.638890266418457, "learning_rate": 5e-05, "loss": 0.9905, "mean_token_accuracy": 0.8578128814697266, "num_tokens": 542234242.0, "step": 14286 }, { "epoch": 1.817453250222618, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.849935054779053, "learning_rate": 5e-05, "loss": 0.8426, "mean_token_accuracy": 0.87773597240448, "num_tokens": 542273187.0, "step": 14287 }, { "epoch": 1.8175804605012085, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.749728202819824, "learning_rate": 5e-05, "loss": 0.9233, "mean_token_accuracy": 0.8785816431045532, "num_tokens": 542308116.0, "step": 14288 }, { "epoch": 1.817707670779799, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.33741569519043, "learning_rate": 5e-05, "loss": 0.7879, "mean_token_accuracy": 0.876394510269165, "num_tokens": 542339964.0, "step": 14289 }, { "epoch": 1.8178348810583895, "ewc_loss": 0.625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.224078178405762, "learning_rate": 5e-05, "loss": 1.0164, "mean_token_accuracy": 0.8648600578308105, "num_tokens": 542381923.0, "step": 14290 }, { "epoch": 1.81796209133698, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.29926872253418, "learning_rate": 5e-05, "loss": 0.7903, "mean_token_accuracy": 0.8803149461746216, "num_tokens": 542418468.0, "step": 14291 }, { "epoch": 1.8180893016155706, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.364702224731445, "learning_rate": 5e-05, "loss": 1.1022, "mean_token_accuracy": 0.8517570495605469, "num_tokens": 542448806.0, "step": 14292 }, { "epoch": 1.818216511894161, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.784493446350098, "learning_rate": 5e-05, "loss": 0.8882, "mean_token_accuracy": 0.8742346167564392, "num_tokens": 542483991.0, "step": 14293 }, { "epoch": 1.8183437221727514, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.849630355834961, "learning_rate": 5e-05, "loss": 0.9712, "mean_token_accuracy": 0.8757007122039795, "num_tokens": 542528334.0, "step": 14294 }, { "epoch": 1.818470932451342, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.12615966796875, "learning_rate": 5e-05, "loss": 0.8727, "mean_token_accuracy": 0.8843071460723877, "num_tokens": 542562903.0, "step": 14295 }, { "epoch": 1.8185981427299325, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.111449241638184, "learning_rate": 5e-05, "loss": 0.8823, "mean_token_accuracy": 0.8748717904090881, "num_tokens": 542600150.0, "step": 14296 }, { "epoch": 1.818725353008523, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.096248626708984, "learning_rate": 5e-05, "loss": 0.8318, "mean_token_accuracy": 0.8799241781234741, "num_tokens": 542637894.0, "step": 14297 }, { "epoch": 1.8188525632871135, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.789897441864014, "learning_rate": 5e-05, "loss": 0.8715, "mean_token_accuracy": 0.8579649925231934, "num_tokens": 542667729.0, "step": 14298 }, { "epoch": 1.818979773565704, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.719233989715576, "learning_rate": 5e-05, "loss": 0.7894, "mean_token_accuracy": 0.8751384019851685, "num_tokens": 542699994.0, "step": 14299 }, { "epoch": 1.8191069838442946, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.072256088256836, "learning_rate": 5e-05, "loss": 0.8184, "mean_token_accuracy": 0.8796402215957642, "num_tokens": 542743849.0, "step": 14300 }, { "epoch": 1.8192341941228851, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.220795631408691, "learning_rate": 5e-05, "loss": 0.7676, "mean_token_accuracy": 0.863259494304657, "num_tokens": 542784788.0, "step": 14301 }, { "epoch": 1.8193614044014756, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.475447654724121, "learning_rate": 5e-05, "loss": 0.8857, "mean_token_accuracy": 0.8703198432922363, "num_tokens": 542820942.0, "step": 14302 }, { "epoch": 1.8194886146800662, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.947782039642334, "learning_rate": 5e-05, "loss": 0.6982, "mean_token_accuracy": 0.882402241230011, "num_tokens": 542857345.0, "step": 14303 }, { "epoch": 1.8196158249586567, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.68427562713623, "learning_rate": 5e-05, "loss": 0.8765, "mean_token_accuracy": 0.8800917863845825, "num_tokens": 542890952.0, "step": 14304 }, { "epoch": 1.8197430352372472, "ewc_loss": 0.375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.9789719581604, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.8584026098251343, "num_tokens": 542927487.0, "step": 14305 }, { "epoch": 1.8198702455158378, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.707693099975586, "learning_rate": 5e-05, "loss": 0.9114, "mean_token_accuracy": 0.871399462223053, "num_tokens": 542967554.0, "step": 14306 }, { "epoch": 1.8199974557944283, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.206830978393555, "learning_rate": 5e-05, "loss": 0.8048, "mean_token_accuracy": 0.8590257167816162, "num_tokens": 543003768.0, "step": 14307 }, { "epoch": 1.8201246660730188, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.441864013671875, "learning_rate": 5e-05, "loss": 0.8802, "mean_token_accuracy": 0.8757036924362183, "num_tokens": 543041937.0, "step": 14308 }, { "epoch": 1.8202518763516093, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.5441389083862305, "learning_rate": 5e-05, "loss": 0.754, "mean_token_accuracy": 0.8879660367965698, "num_tokens": 543084358.0, "step": 14309 }, { "epoch": 1.8203790866301999, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.915838241577148, "learning_rate": 5e-05, "loss": 0.8301, "mean_token_accuracy": 0.8772540092468262, "num_tokens": 543121749.0, "step": 14310 }, { "epoch": 1.8205062969087902, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.837170124053955, "learning_rate": 5e-05, "loss": 0.8326, "mean_token_accuracy": 0.8619204759597778, "num_tokens": 543161160.0, "step": 14311 }, { "epoch": 1.8206335071873807, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.469377040863037, "learning_rate": 5e-05, "loss": 0.7801, "mean_token_accuracy": 0.8700549602508545, "num_tokens": 543197364.0, "step": 14312 }, { "epoch": 1.8207607174659712, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.9813079833984375, "learning_rate": 5e-05, "loss": 0.7584, "mean_token_accuracy": 0.8864079117774963, "num_tokens": 543233381.0, "step": 14313 }, { "epoch": 1.8208879277445618, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.287003993988037, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8641073107719421, "num_tokens": 543266139.0, "step": 14314 }, { "epoch": 1.8210151380231523, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.992284774780273, "learning_rate": 5e-05, "loss": 0.8488, "mean_token_accuracy": 0.858238935470581, "num_tokens": 543307681.0, "step": 14315 }, { "epoch": 1.8211423483017428, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.245065212249756, "learning_rate": 5e-05, "loss": 0.772, "mean_token_accuracy": 0.8637720346450806, "num_tokens": 543347916.0, "step": 14316 }, { "epoch": 1.8212695585803331, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.998958587646484, "learning_rate": 5e-05, "loss": 0.7612, "mean_token_accuracy": 0.8801382780075073, "num_tokens": 543386326.0, "step": 14317 }, { "epoch": 1.8213967688589237, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.196755886077881, "learning_rate": 5e-05, "loss": 0.743, "mean_token_accuracy": 0.8680765628814697, "num_tokens": 543423220.0, "step": 14318 }, { "epoch": 1.8215239791375142, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.016338348388672, "learning_rate": 5e-05, "loss": 0.763, "mean_token_accuracy": 0.8781347870826721, "num_tokens": 543461672.0, "step": 14319 }, { "epoch": 1.8216511894161047, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.147078990936279, "learning_rate": 5e-05, "loss": 0.7066, "mean_token_accuracy": 0.8755334615707397, "num_tokens": 543498775.0, "step": 14320 }, { "epoch": 1.8217783996946952, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.06589126586914, "learning_rate": 5e-05, "loss": 0.7481, "mean_token_accuracy": 0.8838146328926086, "num_tokens": 543539966.0, "step": 14321 }, { "epoch": 1.8219056099732858, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.114356994628906, "learning_rate": 5e-05, "loss": 0.7082, "mean_token_accuracy": 0.8722055554389954, "num_tokens": 543577176.0, "step": 14322 }, { "epoch": 1.8220328202518763, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.07359790802002, "learning_rate": 5e-05, "loss": 0.8261, "mean_token_accuracy": 0.8610002994537354, "num_tokens": 543621328.0, "step": 14323 }, { "epoch": 1.8221600305304668, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.099607944488525, "learning_rate": 5e-05, "loss": 0.7423, "mean_token_accuracy": 0.8600168228149414, "num_tokens": 543662446.0, "step": 14324 }, { "epoch": 1.8222872408090574, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.093951225280762, "learning_rate": 5e-05, "loss": 0.746, "mean_token_accuracy": 0.8826775550842285, "num_tokens": 543695739.0, "step": 14325 }, { "epoch": 1.8224144510876479, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.094860076904297, "learning_rate": 5e-05, "loss": 0.6968, "mean_token_accuracy": 0.8773041367530823, "num_tokens": 543735823.0, "step": 14326 }, { "epoch": 1.8225416613662384, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.117952346801758, "learning_rate": 5e-05, "loss": 0.8377, "mean_token_accuracy": 0.8546245694160461, "num_tokens": 543771800.0, "step": 14327 }, { "epoch": 1.822668871644829, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.032944679260254, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8628261089324951, "num_tokens": 543812425.0, "step": 14328 }, { "epoch": 1.8227960819234195, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.213462829589844, "learning_rate": 5e-05, "loss": 0.7858, "mean_token_accuracy": 0.8739476203918457, "num_tokens": 543849867.0, "step": 14329 }, { "epoch": 1.82292329220201, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.91382360458374, "learning_rate": 5e-05, "loss": 0.6812, "mean_token_accuracy": 0.8827970027923584, "num_tokens": 543887289.0, "step": 14330 }, { "epoch": 1.8230505024806005, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.377279281616211, "learning_rate": 5e-05, "loss": 0.7959, "mean_token_accuracy": 0.8764858245849609, "num_tokens": 543923432.0, "step": 14331 }, { "epoch": 1.823177712759191, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.777280330657959, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.8579477071762085, "num_tokens": 543966655.0, "step": 14332 }, { "epoch": 1.8233049230377816, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.573620796203613, "learning_rate": 5e-05, "loss": 0.8201, "mean_token_accuracy": 0.8784384727478027, "num_tokens": 544006920.0, "step": 14333 }, { "epoch": 1.823432133316372, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.688027858734131, "learning_rate": 5e-05, "loss": 0.7031, "mean_token_accuracy": 0.872307538986206, "num_tokens": 544044066.0, "step": 14334 }, { "epoch": 1.8235593435949626, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.683095932006836, "learning_rate": 5e-05, "loss": 0.8806, "mean_token_accuracy": 0.8649464845657349, "num_tokens": 544073795.0, "step": 14335 }, { "epoch": 1.823686553873553, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.783638000488281, "learning_rate": 5e-05, "loss": 0.7121, "mean_token_accuracy": 0.8758893609046936, "num_tokens": 544107378.0, "step": 14336 }, { "epoch": 1.8238137641521435, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.53262996673584, "learning_rate": 5e-05, "loss": 0.846, "mean_token_accuracy": 0.8753240704536438, "num_tokens": 544144377.0, "step": 14337 }, { "epoch": 1.823940974430734, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.159904479980469, "learning_rate": 5e-05, "loss": 0.7868, "mean_token_accuracy": 0.8675122261047363, "num_tokens": 544176484.0, "step": 14338 }, { "epoch": 1.8240681847093245, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.967294692993164, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8779863119125366, "num_tokens": 544209152.0, "step": 14339 }, { "epoch": 1.824195394987915, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.722550868988037, "learning_rate": 5e-05, "loss": 0.8329, "mean_token_accuracy": 0.8667256832122803, "num_tokens": 544249197.0, "step": 14340 }, { "epoch": 1.8243226052665056, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.224907398223877, "learning_rate": 5e-05, "loss": 0.6994, "mean_token_accuracy": 0.8850532174110413, "num_tokens": 544290225.0, "step": 14341 }, { "epoch": 1.8244498155450959, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.307571411132812, "learning_rate": 5e-05, "loss": 0.8295, "mean_token_accuracy": 0.8690045475959778, "num_tokens": 544333015.0, "step": 14342 }, { "epoch": 1.8245770258236864, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.648983001708984, "learning_rate": 5e-05, "loss": 0.7266, "mean_token_accuracy": 0.8647682666778564, "num_tokens": 544370165.0, "step": 14343 }, { "epoch": 1.824704236102277, "ewc_loss": 0.5, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.814784049987793, "learning_rate": 5e-05, "loss": 0.8695, "mean_token_accuracy": 0.871936023235321, "num_tokens": 544407284.0, "step": 14344 }, { "epoch": 1.8248314463808675, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.320753574371338, "learning_rate": 5e-05, "loss": 0.7442, "mean_token_accuracy": 0.8590939044952393, "num_tokens": 544441558.0, "step": 14345 }, { "epoch": 1.824958656659458, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.214868545532227, "learning_rate": 5e-05, "loss": 0.9978, "mean_token_accuracy": 0.8459444046020508, "num_tokens": 544478468.0, "step": 14346 }, { "epoch": 1.8250858669380485, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.361850261688232, "learning_rate": 5e-05, "loss": 0.6717, "mean_token_accuracy": 0.8866056203842163, "num_tokens": 544516271.0, "step": 14347 }, { "epoch": 1.825213077216639, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.229411125183105, "learning_rate": 5e-05, "loss": 0.9392, "mean_token_accuracy": 0.8813140392303467, "num_tokens": 544553198.0, "step": 14348 }, { "epoch": 1.8253402874952296, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.04902982711792, "learning_rate": 5e-05, "loss": 0.7887, "mean_token_accuracy": 0.8772482872009277, "num_tokens": 544586381.0, "step": 14349 }, { "epoch": 1.8254674977738201, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.427399635314941, "learning_rate": 5e-05, "loss": 0.913, "mean_token_accuracy": 0.8679390549659729, "num_tokens": 544628342.0, "step": 14350 }, { "epoch": 1.8255947080524106, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.967169761657715, "learning_rate": 5e-05, "loss": 0.903, "mean_token_accuracy": 0.8684539794921875, "num_tokens": 544667096.0, "step": 14351 }, { "epoch": 1.8257219183310012, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.237809181213379, "learning_rate": 5e-05, "loss": 0.7577, "mean_token_accuracy": 0.8771716356277466, "num_tokens": 544711467.0, "step": 14352 }, { "epoch": 1.8258491286095917, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.714094161987305, "learning_rate": 5e-05, "loss": 0.9058, "mean_token_accuracy": 0.8741537928581238, "num_tokens": 544751672.0, "step": 14353 }, { "epoch": 1.8259763388881822, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.647136211395264, "learning_rate": 5e-05, "loss": 0.6605, "mean_token_accuracy": 0.8943642377853394, "num_tokens": 544785479.0, "step": 14354 }, { "epoch": 1.8261035491667728, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.910265922546387, "learning_rate": 5e-05, "loss": 0.9201, "mean_token_accuracy": 0.8755682706832886, "num_tokens": 544823122.0, "step": 14355 }, { "epoch": 1.8262307594453633, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.905410289764404, "learning_rate": 5e-05, "loss": 0.7448, "mean_token_accuracy": 0.8826611042022705, "num_tokens": 544862544.0, "step": 14356 }, { "epoch": 1.8263579697239538, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.386902809143066, "learning_rate": 5e-05, "loss": 0.9019, "mean_token_accuracy": 0.8638330698013306, "num_tokens": 544899538.0, "step": 14357 }, { "epoch": 1.8264851800025443, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.558548927307129, "learning_rate": 5e-05, "loss": 0.8028, "mean_token_accuracy": 0.8790283203125, "num_tokens": 544934544.0, "step": 14358 }, { "epoch": 1.8266123902811349, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.57873010635376, "learning_rate": 5e-05, "loss": 0.8186, "mean_token_accuracy": 0.8639663457870483, "num_tokens": 544973304.0, "step": 14359 }, { "epoch": 1.8267396005597252, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.050410270690918, "learning_rate": 5e-05, "loss": 0.8052, "mean_token_accuracy": 0.8792791366577148, "num_tokens": 545010748.0, "step": 14360 }, { "epoch": 1.8268668108383157, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.147677898406982, "learning_rate": 5e-05, "loss": 0.7543, "mean_token_accuracy": 0.8665862083435059, "num_tokens": 545050742.0, "step": 14361 }, { "epoch": 1.8269940211169062, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.226994514465332, "learning_rate": 5e-05, "loss": 0.8435, "mean_token_accuracy": 0.8687390089035034, "num_tokens": 545083904.0, "step": 14362 }, { "epoch": 1.8271212313954968, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.091323375701904, "learning_rate": 5e-05, "loss": 0.7075, "mean_token_accuracy": 0.8810738325119019, "num_tokens": 545123885.0, "step": 14363 }, { "epoch": 1.8272484416740873, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.120368003845215, "learning_rate": 5e-05, "loss": 0.8301, "mean_token_accuracy": 0.8680227994918823, "num_tokens": 545160295.0, "step": 14364 }, { "epoch": 1.8273756519526778, "ewc_loss": 0.375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.213093280792236, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8598864674568176, "num_tokens": 545203469.0, "step": 14365 }, { "epoch": 1.8275028622312681, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.921505928039551, "learning_rate": 5e-05, "loss": 0.7771, "mean_token_accuracy": 0.8781000971794128, "num_tokens": 545242173.0, "step": 14366 }, { "epoch": 1.8276300725098586, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.349801540374756, "learning_rate": 5e-05, "loss": 0.7603, "mean_token_accuracy": 0.8682628273963928, "num_tokens": 545284968.0, "step": 14367 }, { "epoch": 1.8277572827884492, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.806038856506348, "learning_rate": 5e-05, "loss": 0.8353, "mean_token_accuracy": 0.8562632203102112, "num_tokens": 545324764.0, "step": 14368 }, { "epoch": 1.8278844930670397, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.297000885009766, "learning_rate": 5e-05, "loss": 0.7526, "mean_token_accuracy": 0.8666773438453674, "num_tokens": 545362783.0, "step": 14369 }, { "epoch": 1.8280117033456302, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.899832725524902, "learning_rate": 5e-05, "loss": 0.7809, "mean_token_accuracy": 0.8800380229949951, "num_tokens": 545395757.0, "step": 14370 }, { "epoch": 1.8281389136242208, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.09716272354126, "learning_rate": 5e-05, "loss": 0.6786, "mean_token_accuracy": 0.8865487575531006, "num_tokens": 545432194.0, "step": 14371 }, { "epoch": 1.8282661239028113, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.18385124206543, "learning_rate": 5e-05, "loss": 0.7854, "mean_token_accuracy": 0.8761518597602844, "num_tokens": 545469743.0, "step": 14372 }, { "epoch": 1.8283933341814018, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.70682430267334, "learning_rate": 5e-05, "loss": 0.7173, "mean_token_accuracy": 0.8664587140083313, "num_tokens": 545508976.0, "step": 14373 }, { "epoch": 1.8285205444599923, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.734585762023926, "learning_rate": 5e-05, "loss": 0.8457, "mean_token_accuracy": 0.8731120824813843, "num_tokens": 545548223.0, "step": 14374 }, { "epoch": 1.8286477547385829, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.163813591003418, "learning_rate": 5e-05, "loss": 0.7284, "mean_token_accuracy": 0.8556221723556519, "num_tokens": 545586110.0, "step": 14375 }, { "epoch": 1.8287749650171734, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.623259544372559, "learning_rate": 5e-05, "loss": 0.9579, "mean_token_accuracy": 0.8684194087982178, "num_tokens": 545625779.0, "step": 14376 }, { "epoch": 1.828902175295764, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 5.822300910949707, "learning_rate": 5e-05, "loss": 0.7475, "mean_token_accuracy": 0.8531568050384521, "num_tokens": 545669464.0, "step": 14377 }, { "epoch": 1.8290293855743545, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.504530906677246, "learning_rate": 5e-05, "loss": 1.044, "mean_token_accuracy": 0.881203293800354, "num_tokens": 545702674.0, "step": 14378 }, { "epoch": 1.829156595852945, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 6.912617206573486, "learning_rate": 5e-05, "loss": 0.8668, "mean_token_accuracy": 0.876361608505249, "num_tokens": 545742410.0, "step": 14379 }, { "epoch": 1.8292838061315355, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.229504585266113, "learning_rate": 5e-05, "loss": 1.1137, "mean_token_accuracy": 0.8650009632110596, "num_tokens": 545779748.0, "step": 14380 }, { "epoch": 1.829411016410126, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 8.704715728759766, "learning_rate": 5e-05, "loss": 1.1594, "mean_token_accuracy": 0.86518794298172, "num_tokens": 545811401.0, "step": 14381 }, { "epoch": 1.8295382266887166, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.148448944091797, "learning_rate": 5e-05, "loss": 0.8477, "mean_token_accuracy": 0.8777775764465332, "num_tokens": 545848565.0, "step": 14382 }, { "epoch": 1.829665436967307, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000732421875, "grad_norm": 9.969948768615723, "learning_rate": 5e-05, "loss": 1.1797, "mean_token_accuracy": 0.8697747588157654, "num_tokens": 545891245.0, "step": 14383 }, { "epoch": 1.8297926472458976, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 6.636096000671387, "learning_rate": 5e-05, "loss": 0.721, "mean_token_accuracy": 0.8900782465934753, "num_tokens": 545924860.0, "step": 14384 }, { "epoch": 1.829919857524488, "ewc_loss": 0.921875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00086212158203125, "grad_norm": 11.657127380371094, "learning_rate": 5e-05, "loss": 1.2608, "mean_token_accuracy": 0.8827379941940308, "num_tokens": 545958119.0, "step": 14385 }, { "epoch": 1.8300470678030785, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 8.403226852416992, "learning_rate": 5e-05, "loss": 1.144, "mean_token_accuracy": 0.8639646768569946, "num_tokens": 545994195.0, "step": 14386 }, { "epoch": 1.830174278081669, "ewc_loss": 0.78125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 11.375876426696777, "learning_rate": 5e-05, "loss": 1.1555, "mean_token_accuracy": 0.8727340698242188, "num_tokens": 546031998.0, "step": 14387 }, { "epoch": 1.8303014883602595, "ewc_loss": 0.921875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00086212158203125, "grad_norm": 9.504487037658691, "learning_rate": 5e-05, "loss": 1.2843, "mean_token_accuracy": 0.8704707026481628, "num_tokens": 546067152.0, "step": 14388 }, { "epoch": 1.83042869863885, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.333641052246094, "learning_rate": 5e-05, "loss": 1.0082, "mean_token_accuracy": 0.8854496479034424, "num_tokens": 546104207.0, "step": 14389 }, { "epoch": 1.8305559089174406, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 8.969100952148438, "learning_rate": 5e-05, "loss": 1.1093, "mean_token_accuracy": 0.860192060470581, "num_tokens": 546150659.0, "step": 14390 }, { "epoch": 1.8306831191960309, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.620918273925781, "learning_rate": 5e-05, "loss": 1.0305, "mean_token_accuracy": 0.8786600828170776, "num_tokens": 546188427.0, "step": 14391 }, { "epoch": 1.8308103294746214, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.235721588134766, "learning_rate": 5e-05, "loss": 0.9545, "mean_token_accuracy": 0.8634907603263855, "num_tokens": 546227249.0, "step": 14392 }, { "epoch": 1.830937539753212, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.464264869689941, "learning_rate": 5e-05, "loss": 1.0572, "mean_token_accuracy": 0.8619582653045654, "num_tokens": 546263469.0, "step": 14393 }, { "epoch": 1.8310647500318025, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.5445027351379395, "learning_rate": 5e-05, "loss": 0.8635, "mean_token_accuracy": 0.8687440156936646, "num_tokens": 546300820.0, "step": 14394 }, { "epoch": 1.831191960310393, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.870015144348145, "learning_rate": 5e-05, "loss": 1.0136, "mean_token_accuracy": 0.876419186592102, "num_tokens": 546336639.0, "step": 14395 }, { "epoch": 1.8313191705889835, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.414499282836914, "learning_rate": 5e-05, "loss": 0.8956, "mean_token_accuracy": 0.8672312498092651, "num_tokens": 546377769.0, "step": 14396 }, { "epoch": 1.831446380867574, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.27372932434082, "learning_rate": 5e-05, "loss": 1.0494, "mean_token_accuracy": 0.8811432719230652, "num_tokens": 546419286.0, "step": 14397 }, { "epoch": 1.8315735911461646, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.019929885864258, "learning_rate": 5e-05, "loss": 1.0057, "mean_token_accuracy": 0.8615784645080566, "num_tokens": 546458907.0, "step": 14398 }, { "epoch": 1.831700801424755, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.030664443969727, "learning_rate": 5e-05, "loss": 0.9822, "mean_token_accuracy": 0.8913160562515259, "num_tokens": 546491682.0, "step": 14399 }, { "epoch": 1.8318280117033456, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.50690746307373, "learning_rate": 5e-05, "loss": 1.0621, "mean_token_accuracy": 0.8530869483947754, "num_tokens": 546537141.0, "step": 14400 }, { "epoch": 1.8319552219819362, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.273406028747559, "learning_rate": 5e-05, "loss": 1.0039, "mean_token_accuracy": 0.8695374131202698, "num_tokens": 546574318.0, "step": 14401 }, { "epoch": 1.8320824322605267, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.500320434570312, "learning_rate": 5e-05, "loss": 0.9968, "mean_token_accuracy": 0.8623837828636169, "num_tokens": 546615938.0, "step": 14402 }, { "epoch": 1.8322096425391172, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.541875839233398, "learning_rate": 5e-05, "loss": 0.9239, "mean_token_accuracy": 0.8748130202293396, "num_tokens": 546652608.0, "step": 14403 }, { "epoch": 1.8323368528177078, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.120229721069336, "learning_rate": 5e-05, "loss": 0.8318, "mean_token_accuracy": 0.8829288482666016, "num_tokens": 546695363.0, "step": 14404 }, { "epoch": 1.8324640630962983, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.442741394042969, "learning_rate": 5e-05, "loss": 0.9474, "mean_token_accuracy": 0.8589448928833008, "num_tokens": 546741174.0, "step": 14405 }, { "epoch": 1.8325912733748888, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.416642665863037, "learning_rate": 5e-05, "loss": 0.7624, "mean_token_accuracy": 0.879031777381897, "num_tokens": 546777827.0, "step": 14406 }, { "epoch": 1.8327184836534793, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.008183479309082, "learning_rate": 5e-05, "loss": 0.9284, "mean_token_accuracy": 0.880039393901825, "num_tokens": 546821199.0, "step": 14407 }, { "epoch": 1.8328456939320699, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.95205545425415, "learning_rate": 5e-05, "loss": 0.7578, "mean_token_accuracy": 0.8733291625976562, "num_tokens": 546860151.0, "step": 14408 }, { "epoch": 1.8329729042106602, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.62640380859375, "learning_rate": 5e-05, "loss": 0.9841, "mean_token_accuracy": 0.8821700215339661, "num_tokens": 546892456.0, "step": 14409 }, { "epoch": 1.8331001144892507, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.313895225524902, "learning_rate": 5e-05, "loss": 0.8592, "mean_token_accuracy": 0.8662627935409546, "num_tokens": 546931611.0, "step": 14410 }, { "epoch": 1.8332273247678412, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.570449829101562, "learning_rate": 5e-05, "loss": 1.0211, "mean_token_accuracy": 0.8765326738357544, "num_tokens": 546968408.0, "step": 14411 }, { "epoch": 1.8333545350464318, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.180190086364746, "learning_rate": 5e-05, "loss": 0.9785, "mean_token_accuracy": 0.8596470355987549, "num_tokens": 547007025.0, "step": 14412 }, { "epoch": 1.8334817453250223, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.540977478027344, "learning_rate": 5e-05, "loss": 0.9764, "mean_token_accuracy": 0.8592348098754883, "num_tokens": 547051371.0, "step": 14413 }, { "epoch": 1.8336089556036128, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.571388244628906, "learning_rate": 5e-05, "loss": 0.9791, "mean_token_accuracy": 0.86640864610672, "num_tokens": 547087476.0, "step": 14414 }, { "epoch": 1.8337361658822031, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.4355292320251465, "learning_rate": 5e-05, "loss": 0.8, "mean_token_accuracy": 0.8767033815383911, "num_tokens": 547126922.0, "step": 14415 }, { "epoch": 1.8338633761607936, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.634839057922363, "learning_rate": 5e-05, "loss": 0.8819, "mean_token_accuracy": 0.8778576254844666, "num_tokens": 547164678.0, "step": 14416 }, { "epoch": 1.8339905864393842, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.190842628479004, "learning_rate": 5e-05, "loss": 0.7812, "mean_token_accuracy": 0.8703506588935852, "num_tokens": 547200039.0, "step": 14417 }, { "epoch": 1.8341177967179747, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.314197540283203, "learning_rate": 5e-05, "loss": 0.7836, "mean_token_accuracy": 0.8905142545700073, "num_tokens": 547233325.0, "step": 14418 }, { "epoch": 1.8342450069965652, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.441565036773682, "learning_rate": 5e-05, "loss": 0.7331, "mean_token_accuracy": 0.8849586248397827, "num_tokens": 547266658.0, "step": 14419 }, { "epoch": 1.8343722172751558, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.899649143218994, "learning_rate": 5e-05, "loss": 0.827, "mean_token_accuracy": 0.8680795431137085, "num_tokens": 547305876.0, "step": 14420 }, { "epoch": 1.8344994275537463, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.594014644622803, "learning_rate": 5e-05, "loss": 0.7533, "mean_token_accuracy": 0.8787015676498413, "num_tokens": 547346847.0, "step": 14421 }, { "epoch": 1.8346266378323368, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.6909871101379395, "learning_rate": 5e-05, "loss": 0.731, "mean_token_accuracy": 0.8824627995491028, "num_tokens": 547379896.0, "step": 14422 }, { "epoch": 1.8347538481109273, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.6534342765808105, "learning_rate": 5e-05, "loss": 0.7831, "mean_token_accuracy": 0.8681159615516663, "num_tokens": 547425666.0, "step": 14423 }, { "epoch": 1.8348810583895179, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.540303707122803, "learning_rate": 5e-05, "loss": 0.7199, "mean_token_accuracy": 0.8784252405166626, "num_tokens": 547467007.0, "step": 14424 }, { "epoch": 1.8350082686681084, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.721646785736084, "learning_rate": 5e-05, "loss": 0.8179, "mean_token_accuracy": 0.854094922542572, "num_tokens": 547507718.0, "step": 14425 }, { "epoch": 1.835135478946699, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.463254451751709, "learning_rate": 5e-05, "loss": 0.7146, "mean_token_accuracy": 0.878818929195404, "num_tokens": 547543445.0, "step": 14426 }, { "epoch": 1.8352626892252895, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.673064708709717, "learning_rate": 5e-05, "loss": 0.7354, "mean_token_accuracy": 0.8754883408546448, "num_tokens": 547576000.0, "step": 14427 }, { "epoch": 1.83538989950388, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.55069637298584, "learning_rate": 5e-05, "loss": 0.7173, "mean_token_accuracy": 0.8775076866149902, "num_tokens": 547609950.0, "step": 14428 }, { "epoch": 1.8355171097824705, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.499757766723633, "learning_rate": 5e-05, "loss": 0.7104, "mean_token_accuracy": 0.8778912425041199, "num_tokens": 547650133.0, "step": 14429 }, { "epoch": 1.835644320061061, "ewc_loss": 0.375, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.700686454772949, "learning_rate": 5e-05, "loss": 0.752, "mean_token_accuracy": 0.8680633306503296, "num_tokens": 547687754.0, "step": 14430 }, { "epoch": 1.8357715303396516, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.329733371734619, "learning_rate": 5e-05, "loss": 0.7167, "mean_token_accuracy": 0.8753193616867065, "num_tokens": 547732119.0, "step": 14431 }, { "epoch": 1.835898740618242, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.832193851470947, "learning_rate": 5e-05, "loss": 0.7336, "mean_token_accuracy": 0.8776400089263916, "num_tokens": 547770779.0, "step": 14432 }, { "epoch": 1.8360259508968326, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.154040336608887, "learning_rate": 5e-05, "loss": 0.6926, "mean_token_accuracy": 0.8787626028060913, "num_tokens": 547812524.0, "step": 14433 }, { "epoch": 1.836153161175423, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.104366302490234, "learning_rate": 5e-05, "loss": 0.8076, "mean_token_accuracy": 0.8656781911849976, "num_tokens": 547850955.0, "step": 14434 }, { "epoch": 1.8362803714540135, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.824948787689209, "learning_rate": 5e-05, "loss": 0.6872, "mean_token_accuracy": 0.8751136660575867, "num_tokens": 547891390.0, "step": 14435 }, { "epoch": 1.836407581732604, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.56110668182373, "learning_rate": 5e-05, "loss": 0.8341, "mean_token_accuracy": 0.8673446178436279, "num_tokens": 547930306.0, "step": 14436 }, { "epoch": 1.8365347920111945, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00023746490478515625, "grad_norm": 6.351085186004639, "learning_rate": 5e-05, "loss": 0.6839, "mean_token_accuracy": 0.862752377986908, "num_tokens": 547970940.0, "step": 14437 }, { "epoch": 1.836662002289785, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 9.271688461303711, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.8800725340843201, "num_tokens": 548007908.0, "step": 14438 }, { "epoch": 1.8367892125683756, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00023365020751953125, "grad_norm": 5.869412899017334, "learning_rate": 5e-05, "loss": 0.6685, "mean_token_accuracy": 0.8660088777542114, "num_tokens": 548045040.0, "step": 14439 }, { "epoch": 1.8369164228469659, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.24332332611084, "learning_rate": 5e-05, "loss": 1.0104, "mean_token_accuracy": 0.871862530708313, "num_tokens": 548088501.0, "step": 14440 }, { "epoch": 1.8370436331255564, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.155477046966553, "learning_rate": 5e-05, "loss": 0.7614, "mean_token_accuracy": 0.8608115315437317, "num_tokens": 548121888.0, "step": 14441 }, { "epoch": 1.837170843404147, "ewc_loss": 0.78125, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 10.79381275177002, "learning_rate": 5e-05, "loss": 1.1373, "mean_token_accuracy": 0.8759476542472839, "num_tokens": 548161117.0, "step": 14442 }, { "epoch": 1.8372980536827375, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.116493225097656, "learning_rate": 5e-05, "loss": 0.9945, "mean_token_accuracy": 0.881822943687439, "num_tokens": 548201709.0, "step": 14443 }, { "epoch": 1.837425263961328, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.434564590454102, "learning_rate": 5e-05, "loss": 0.9888, "mean_token_accuracy": 0.8711955547332764, "num_tokens": 548241843.0, "step": 14444 }, { "epoch": 1.8375524742399185, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 9.377241134643555, "learning_rate": 5e-05, "loss": 1.1086, "mean_token_accuracy": 0.878937304019928, "num_tokens": 548278785.0, "step": 14445 }, { "epoch": 1.837679684518509, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.254210472106934, "learning_rate": 5e-05, "loss": 0.7745, "mean_token_accuracy": 0.8759522438049316, "num_tokens": 548315766.0, "step": 14446 }, { "epoch": 1.8378068947970996, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 10.877729415893555, "learning_rate": 5e-05, "loss": 1.1762, "mean_token_accuracy": 0.8695756196975708, "num_tokens": 548353971.0, "step": 14447 }, { "epoch": 1.83793410507569, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.283857345581055, "learning_rate": 5e-05, "loss": 0.872, "mean_token_accuracy": 0.8722941875457764, "num_tokens": 548391261.0, "step": 14448 }, { "epoch": 1.8380613153542806, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0008087158203125, "grad_norm": 11.649025917053223, "learning_rate": 5e-05, "loss": 1.1966, "mean_token_accuracy": 0.8841647505760193, "num_tokens": 548424773.0, "step": 14449 }, { "epoch": 1.8381885256328712, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.245987892150879, "learning_rate": 5e-05, "loss": 1.154, "mean_token_accuracy": 0.887401819229126, "num_tokens": 548462404.0, "step": 14450 }, { "epoch": 1.8383157359114617, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.53702449798584, "learning_rate": 5e-05, "loss": 1.0326, "mean_token_accuracy": 0.8740444183349609, "num_tokens": 548504595.0, "step": 14451 }, { "epoch": 1.8384429461900522, "ewc_loss": 0.765625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 9.207762718200684, "learning_rate": 5e-05, "loss": 1.1066, "mean_token_accuracy": 0.8795996904373169, "num_tokens": 548547945.0, "step": 14452 }, { "epoch": 1.8385701564686427, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.678629875183105, "learning_rate": 5e-05, "loss": 1.0422, "mean_token_accuracy": 0.8632093667984009, "num_tokens": 548590302.0, "step": 14453 }, { "epoch": 1.8386973667472333, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.595711708068848, "learning_rate": 5e-05, "loss": 0.9264, "mean_token_accuracy": 0.8816401362419128, "num_tokens": 548629688.0, "step": 14454 }, { "epoch": 1.8388245770258238, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.178818702697754, "learning_rate": 5e-05, "loss": 1.0399, "mean_token_accuracy": 0.8518521189689636, "num_tokens": 548666294.0, "step": 14455 }, { "epoch": 1.8389517873044143, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.869439125061035, "learning_rate": 5e-05, "loss": 0.8819, "mean_token_accuracy": 0.8607772588729858, "num_tokens": 548707536.0, "step": 14456 }, { "epoch": 1.8390789975830049, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.228976249694824, "learning_rate": 5e-05, "loss": 0.961, "mean_token_accuracy": 0.8774074912071228, "num_tokens": 548751443.0, "step": 14457 }, { "epoch": 1.8392062078615952, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.313917636871338, "learning_rate": 5e-05, "loss": 0.7995, "mean_token_accuracy": 0.8705267906188965, "num_tokens": 548789259.0, "step": 14458 }, { "epoch": 1.8393334181401857, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.598467826843262, "learning_rate": 5e-05, "loss": 0.9631, "mean_token_accuracy": 0.8769873976707458, "num_tokens": 548822045.0, "step": 14459 }, { "epoch": 1.8394606284187762, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.243297100067139, "learning_rate": 5e-05, "loss": 0.8515, "mean_token_accuracy": 0.8604678511619568, "num_tokens": 548854722.0, "step": 14460 }, { "epoch": 1.8395878386973668, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.78928279876709, "learning_rate": 5e-05, "loss": 1.0141, "mean_token_accuracy": 0.8670480251312256, "num_tokens": 548888903.0, "step": 14461 }, { "epoch": 1.8397150489759573, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.6300048828125, "learning_rate": 5e-05, "loss": 0.8736, "mean_token_accuracy": 0.8738438487052917, "num_tokens": 548925611.0, "step": 14462 }, { "epoch": 1.8398422592545478, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.566644668579102, "learning_rate": 5e-05, "loss": 0.9786, "mean_token_accuracy": 0.8798803091049194, "num_tokens": 548967239.0, "step": 14463 }, { "epoch": 1.8399694695331381, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.998593807220459, "learning_rate": 5e-05, "loss": 0.9124, "mean_token_accuracy": 0.8767094016075134, "num_tokens": 549005453.0, "step": 14464 }, { "epoch": 1.8400966798117286, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.993144035339355, "learning_rate": 5e-05, "loss": 0.9696, "mean_token_accuracy": 0.8689883947372437, "num_tokens": 549041418.0, "step": 14465 }, { "epoch": 1.8402238900903192, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.103348731994629, "learning_rate": 5e-05, "loss": 0.8871, "mean_token_accuracy": 0.8795217275619507, "num_tokens": 549079223.0, "step": 14466 }, { "epoch": 1.8403511003689097, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.409831047058105, "learning_rate": 5e-05, "loss": 0.972, "mean_token_accuracy": 0.8490902781486511, "num_tokens": 549124542.0, "step": 14467 }, { "epoch": 1.8404783106475002, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.888791084289551, "learning_rate": 5e-05, "loss": 0.8723, "mean_token_accuracy": 0.8610215783119202, "num_tokens": 549165374.0, "step": 14468 }, { "epoch": 1.8406055209260908, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.328169822692871, "learning_rate": 5e-05, "loss": 0.9162, "mean_token_accuracy": 0.8607223033905029, "num_tokens": 549203887.0, "step": 14469 }, { "epoch": 1.8407327312046813, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.389284610748291, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.8713446855545044, "num_tokens": 549247583.0, "step": 14470 }, { "epoch": 1.8408599414832718, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.690455436706543, "learning_rate": 5e-05, "loss": 0.8489, "mean_token_accuracy": 0.8877986073493958, "num_tokens": 549282874.0, "step": 14471 }, { "epoch": 1.8409871517618623, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000568389892578125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.041224479675293, "learning_rate": 5e-05, "loss": 0.7091, "mean_token_accuracy": 0.88130784034729, "num_tokens": 549323133.0, "step": 14472 }, { "epoch": 1.8411143620404529, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0005645751953125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 14.901015281677246, "learning_rate": 5e-05, "loss": 1.1418, "mean_token_accuracy": 0.8557406663894653, "num_tokens": 549359794.0, "step": 14473 }, { "epoch": 1.8412415723190434, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.00057220458984375, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 10.132026672363281, "learning_rate": 5e-05, "loss": 1.233, "mean_token_accuracy": 0.854831874370575, "num_tokens": 549393281.0, "step": 14474 }, { "epoch": 1.841368782597634, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 9.361331939697266, "learning_rate": 5e-05, "loss": 0.9116, "mean_token_accuracy": 0.8670499324798584, "num_tokens": 549428181.0, "step": 14475 }, { "epoch": 1.8414959928762245, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.984322547912598, "learning_rate": 5e-05, "loss": 1.0142, "mean_token_accuracy": 0.8786987066268921, "num_tokens": 549465521.0, "step": 14476 }, { "epoch": 1.841623203154815, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.675289154052734, "learning_rate": 5e-05, "loss": 0.787, "mean_token_accuracy": 0.8698935508728027, "num_tokens": 549503800.0, "step": 14477 }, { "epoch": 1.8417504134334055, "ewc_loss": 0.75, "ewc_loss_diag": 0.000606536865234375, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.99587345123291, "learning_rate": 5e-05, "loss": 1.0873, "mean_token_accuracy": 0.8811818361282349, "num_tokens": 549536725.0, "step": 14478 }, { "epoch": 1.841877623711996, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.461609840393066, "learning_rate": 5e-05, "loss": 0.8278, "mean_token_accuracy": 0.8726080656051636, "num_tokens": 549575774.0, "step": 14479 }, { "epoch": 1.8420048339905866, "ewc_loss": 0.90625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000843048095703125, "grad_norm": 12.451776504516602, "learning_rate": 5e-05, "loss": 1.2774, "mean_token_accuracy": 0.8733227849006653, "num_tokens": 549616715.0, "step": 14480 }, { "epoch": 1.842132044269177, "ewc_loss": 0.796875, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 9.797118186950684, "learning_rate": 5e-05, "loss": 1.1723, "mean_token_accuracy": 0.871283769607544, "num_tokens": 549652005.0, "step": 14481 }, { "epoch": 1.8422592545477676, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 11.499286651611328, "learning_rate": 5e-05, "loss": 1.1219, "mean_token_accuracy": 0.870096743106842, "num_tokens": 549683708.0, "step": 14482 }, { "epoch": 1.842386464826358, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.00064849853515625, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 9.496003150939941, "learning_rate": 5e-05, "loss": 1.0849, "mean_token_accuracy": 0.8847774267196655, "num_tokens": 549722961.0, "step": 14483 }, { "epoch": 1.8425136751049485, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 11.2454833984375, "learning_rate": 5e-05, "loss": 1.1191, "mean_token_accuracy": 0.86292564868927, "num_tokens": 549758857.0, "step": 14484 }, { "epoch": 1.842640885383539, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.594115257263184, "learning_rate": 5e-05, "loss": 0.9847, "mean_token_accuracy": 0.8705403804779053, "num_tokens": 549793835.0, "step": 14485 }, { "epoch": 1.8427680956621295, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 11.143148422241211, "learning_rate": 5e-05, "loss": 1.0943, "mean_token_accuracy": 0.8735018968582153, "num_tokens": 549832691.0, "step": 14486 }, { "epoch": 1.84289530594072, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.198677062988281, "learning_rate": 5e-05, "loss": 1.0571, "mean_token_accuracy": 0.8338098526000977, "num_tokens": 549874981.0, "step": 14487 }, { "epoch": 1.8430225162193106, "ewc_loss": 0.78125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 11.671488761901855, "learning_rate": 5e-05, "loss": 1.1511, "mean_token_accuracy": 0.8709472417831421, "num_tokens": 549917904.0, "step": 14488 }, { "epoch": 1.8431497264979009, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.776063919067383, "learning_rate": 5e-05, "loss": 0.9912, "mean_token_accuracy": 0.8699787259101868, "num_tokens": 549955421.0, "step": 14489 }, { "epoch": 1.8432769367764914, "ewc_loss": 0.83203125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 11.99973201751709, "learning_rate": 5e-05, "loss": 1.1856, "mean_token_accuracy": 0.8759608268737793, "num_tokens": 549986615.0, "step": 14490 }, { "epoch": 1.843404147055082, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 9.760464668273926, "learning_rate": 5e-05, "loss": 1.1017, "mean_token_accuracy": 0.8650107383728027, "num_tokens": 550022747.0, "step": 14491 }, { "epoch": 1.8435313573336725, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 11.407711029052734, "learning_rate": 5e-05, "loss": 1.1303, "mean_token_accuracy": 0.8717036843299866, "num_tokens": 550060423.0, "step": 14492 }, { "epoch": 1.843658567612263, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 9.929715156555176, "learning_rate": 5e-05, "loss": 1.0885, "mean_token_accuracy": 0.8714809417724609, "num_tokens": 550102086.0, "step": 14493 }, { "epoch": 1.8437857778908535, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.539318084716797, "learning_rate": 5e-05, "loss": 1.0152, "mean_token_accuracy": 0.8856734037399292, "num_tokens": 550138281.0, "step": 14494 }, { "epoch": 1.843912988169444, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.524965286254883, "learning_rate": 5e-05, "loss": 1.0635, "mean_token_accuracy": 0.8558216094970703, "num_tokens": 550180219.0, "step": 14495 }, { "epoch": 1.8440401984480346, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.797747611999512, "learning_rate": 5e-05, "loss": 1.0, "mean_token_accuracy": 0.8718631267547607, "num_tokens": 550218095.0, "step": 14496 }, { "epoch": 1.844167408726625, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.845098495483398, "learning_rate": 5e-05, "loss": 0.9192, "mean_token_accuracy": 0.8658517599105835, "num_tokens": 550256080.0, "step": 14497 }, { "epoch": 1.8442946190052156, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.561497688293457, "learning_rate": 5e-05, "loss": 1.0181, "mean_token_accuracy": 0.8568300604820251, "num_tokens": 550297471.0, "step": 14498 }, { "epoch": 1.8444218292838062, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.041828155517578, "learning_rate": 5e-05, "loss": 0.7657, "mean_token_accuracy": 0.8895387649536133, "num_tokens": 550330495.0, "step": 14499 }, { "epoch": 1.8445490395623967, "ewc_loss": 0.625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 10.003238677978516, "learning_rate": 5e-05, "loss": 0.9958, "mean_token_accuracy": 0.8721380829811096, "num_tokens": 550362810.0, "step": 14500 }, { "epoch": 1.8446762498409872, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.740958213806152, "learning_rate": 5e-05, "loss": 0.8106, "mean_token_accuracy": 0.8702034950256348, "num_tokens": 550409408.0, "step": 14501 }, { "epoch": 1.8448034601195777, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.652482032775879, "learning_rate": 5e-05, "loss": 1.0844, "mean_token_accuracy": 0.8669115900993347, "num_tokens": 550441366.0, "step": 14502 }, { "epoch": 1.8449306703981683, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.51047134399414, "learning_rate": 5e-05, "loss": 0.8792, "mean_token_accuracy": 0.8777366876602173, "num_tokens": 550478247.0, "step": 14503 }, { "epoch": 1.8450578806767588, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.401079177856445, "learning_rate": 5e-05, "loss": 1.0273, "mean_token_accuracy": 0.8778955340385437, "num_tokens": 550511861.0, "step": 14504 }, { "epoch": 1.8451850909553493, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.237874984741211, "learning_rate": 5e-05, "loss": 0.9741, "mean_token_accuracy": 0.8656233549118042, "num_tokens": 550554959.0, "step": 14505 }, { "epoch": 1.8453123012339399, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.206034660339355, "learning_rate": 5e-05, "loss": 0.9086, "mean_token_accuracy": 0.8766583204269409, "num_tokens": 550589960.0, "step": 14506 }, { "epoch": 1.8454395115125302, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.981411933898926, "learning_rate": 5e-05, "loss": 0.9107, "mean_token_accuracy": 0.8718146085739136, "num_tokens": 550630481.0, "step": 14507 }, { "epoch": 1.8455667217911207, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.337220191955566, "learning_rate": 5e-05, "loss": 0.8687, "mean_token_accuracy": 0.8627832531929016, "num_tokens": 550671819.0, "step": 14508 }, { "epoch": 1.8456939320697112, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.423564910888672, "learning_rate": 5e-05, "loss": 0.7999, "mean_token_accuracy": 0.8807188272476196, "num_tokens": 550712230.0, "step": 14509 }, { "epoch": 1.8458211423483017, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.113673210144043, "learning_rate": 5e-05, "loss": 0.8184, "mean_token_accuracy": 0.8704342842102051, "num_tokens": 550751052.0, "step": 14510 }, { "epoch": 1.8459483526268923, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.863317489624023, "learning_rate": 5e-05, "loss": 0.7546, "mean_token_accuracy": 0.8794542551040649, "num_tokens": 550787374.0, "step": 14511 }, { "epoch": 1.8460755629054828, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.133124351501465, "learning_rate": 5e-05, "loss": 0.7841, "mean_token_accuracy": 0.8791931867599487, "num_tokens": 550829398.0, "step": 14512 }, { "epoch": 1.846202773184073, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.622817039489746, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.8631653785705566, "num_tokens": 550868818.0, "step": 14513 }, { "epoch": 1.8463299834626636, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.075285911560059, "learning_rate": 5e-05, "loss": 0.7791, "mean_token_accuracy": 0.8769925236701965, "num_tokens": 550904136.0, "step": 14514 }, { "epoch": 1.8464571937412542, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.504824161529541, "learning_rate": 5e-05, "loss": 0.8438, "mean_token_accuracy": 0.8396723866462708, "num_tokens": 550943505.0, "step": 14515 }, { "epoch": 1.8465844040198447, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00064849853515625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 8.108650207519531, "learning_rate": 5e-05, "loss": 0.7765, "mean_token_accuracy": 0.8713362216949463, "num_tokens": 550983633.0, "step": 14516 }, { "epoch": 1.8467116142984352, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00064849853515625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.3032917976379395, "learning_rate": 5e-05, "loss": 0.6876, "mean_token_accuracy": 0.881908655166626, "num_tokens": 551017683.0, "step": 14517 }, { "epoch": 1.8468388245770258, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00064849853515625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.349841117858887, "learning_rate": 5e-05, "loss": 0.7943, "mean_token_accuracy": 0.8709272146224976, "num_tokens": 551051731.0, "step": 14518 }, { "epoch": 1.8469660348556163, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00064849853515625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.9334635734558105, "learning_rate": 5e-05, "loss": 0.6869, "mean_token_accuracy": 0.8760872483253479, "num_tokens": 551090818.0, "step": 14519 }, { "epoch": 1.8470932451342068, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.918318748474121, "learning_rate": 5e-05, "loss": 0.8212, "mean_token_accuracy": 0.8792210221290588, "num_tokens": 551122032.0, "step": 14520 }, { "epoch": 1.8472204554127973, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.525718688964844, "learning_rate": 5e-05, "loss": 0.7093, "mean_token_accuracy": 0.86406409740448, "num_tokens": 551157708.0, "step": 14521 }, { "epoch": 1.8473476656913879, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.559885025024414, "learning_rate": 5e-05, "loss": 0.9265, "mean_token_accuracy": 0.8709431886672974, "num_tokens": 551194222.0, "step": 14522 }, { "epoch": 1.8474748759699784, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.644997596740723, "learning_rate": 5e-05, "loss": 0.7227, "mean_token_accuracy": 0.8713058233261108, "num_tokens": 551234387.0, "step": 14523 }, { "epoch": 1.847602086248569, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.758813858032227, "learning_rate": 5e-05, "loss": 0.9751, "mean_token_accuracy": 0.875371515750885, "num_tokens": 551273330.0, "step": 14524 }, { "epoch": 1.8477292965271594, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.728719711303711, "learning_rate": 5e-05, "loss": 0.7986, "mean_token_accuracy": 0.8717386722564697, "num_tokens": 551309180.0, "step": 14525 }, { "epoch": 1.84785650680575, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.721003532409668, "learning_rate": 5e-05, "loss": 0.8448, "mean_token_accuracy": 0.883135199546814, "num_tokens": 551344455.0, "step": 14526 }, { "epoch": 1.8479837170843405, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.664531707763672, "learning_rate": 5e-05, "loss": 0.9129, "mean_token_accuracy": 0.8624921441078186, "num_tokens": 551380795.0, "step": 14527 }, { "epoch": 1.848110927362931, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.199867248535156, "learning_rate": 5e-05, "loss": 0.7882, "mean_token_accuracy": 0.8597217798233032, "num_tokens": 551421314.0, "step": 14528 }, { "epoch": 1.8482381376415216, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 9.341472625732422, "learning_rate": 5e-05, "loss": 0.9063, "mean_token_accuracy": 0.8753544092178345, "num_tokens": 551459021.0, "step": 14529 }, { "epoch": 1.848365347920112, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.77958345413208, "learning_rate": 5e-05, "loss": 0.7209, "mean_token_accuracy": 0.8730282187461853, "num_tokens": 551498055.0, "step": 14530 }, { "epoch": 1.8484925581987026, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.615701675415039, "learning_rate": 5e-05, "loss": 0.9855, "mean_token_accuracy": 0.8681452870368958, "num_tokens": 551536746.0, "step": 14531 }, { "epoch": 1.848619768477293, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.481003284454346, "learning_rate": 5e-05, "loss": 0.8049, "mean_token_accuracy": 0.8614580631256104, "num_tokens": 551576230.0, "step": 14532 }, { "epoch": 1.8487469787558835, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.916015625, "learning_rate": 5e-05, "loss": 0.9013, "mean_token_accuracy": 0.8722397089004517, "num_tokens": 551613476.0, "step": 14533 }, { "epoch": 1.848874189034474, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.260083198547363, "learning_rate": 5e-05, "loss": 0.8354, "mean_token_accuracy": 0.8729111552238464, "num_tokens": 551655177.0, "step": 14534 }, { "epoch": 1.8490013993130645, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.8614983558654785, "learning_rate": 5e-05, "loss": 0.8202, "mean_token_accuracy": 0.8670916557312012, "num_tokens": 551691236.0, "step": 14535 }, { "epoch": 1.849128609591655, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.416996955871582, "learning_rate": 5e-05, "loss": 0.8578, "mean_token_accuracy": 0.8619588613510132, "num_tokens": 551732152.0, "step": 14536 }, { "epoch": 1.8492558198702456, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.640442848205566, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.8774296045303345, "num_tokens": 551765221.0, "step": 14537 }, { "epoch": 1.8493830301488359, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.959761619567871, "learning_rate": 5e-05, "loss": 0.7767, "mean_token_accuracy": 0.872588038444519, "num_tokens": 551805321.0, "step": 14538 }, { "epoch": 1.8495102404274264, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 8.014817237854004, "learning_rate": 5e-05, "loss": 0.7944, "mean_token_accuracy": 0.8646961450576782, "num_tokens": 551846752.0, "step": 14539 }, { "epoch": 1.849637450706017, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.251994609832764, "learning_rate": 5e-05, "loss": 0.7424, "mean_token_accuracy": 0.8686870336532593, "num_tokens": 551887877.0, "step": 14540 }, { "epoch": 1.8497646609846075, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000614166259765625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.5826997756958, "learning_rate": 5e-05, "loss": 0.8238, "mean_token_accuracy": 0.8660484552383423, "num_tokens": 551919669.0, "step": 14541 }, { "epoch": 1.849891871263198, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000614166259765625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.7096662521362305, "learning_rate": 5e-05, "loss": 0.6897, "mean_token_accuracy": 0.8739966154098511, "num_tokens": 551955459.0, "step": 14542 }, { "epoch": 1.8500190815417885, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000606536865234375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 9.166579246520996, "learning_rate": 5e-05, "loss": 0.8958, "mean_token_accuracy": 0.8700032234191895, "num_tokens": 551994889.0, "step": 14543 }, { "epoch": 1.850146291820379, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000606536865234375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.488593101501465, "learning_rate": 5e-05, "loss": 0.7017, "mean_token_accuracy": 0.8738068342208862, "num_tokens": 552030176.0, "step": 14544 }, { "epoch": 1.8502735020989696, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000606536865234375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.558968544006348, "learning_rate": 5e-05, "loss": 0.911, "mean_token_accuracy": 0.8851381540298462, "num_tokens": 552063317.0, "step": 14545 }, { "epoch": 1.85040071237756, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000606536865234375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.809326171875, "learning_rate": 5e-05, "loss": 0.6824, "mean_token_accuracy": 0.8886667490005493, "num_tokens": 552100959.0, "step": 14546 }, { "epoch": 1.8505279226561506, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.510624885559082, "learning_rate": 5e-05, "loss": 0.9476, "mean_token_accuracy": 0.8767650127410889, "num_tokens": 552139181.0, "step": 14547 }, { "epoch": 1.8506551329347412, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.414134979248047, "learning_rate": 5e-05, "loss": 0.7985, "mean_token_accuracy": 0.86695396900177, "num_tokens": 552172293.0, "step": 14548 }, { "epoch": 1.8507823432133317, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.178731918334961, "learning_rate": 5e-05, "loss": 0.964, "mean_token_accuracy": 0.8682613372802734, "num_tokens": 552214655.0, "step": 14549 }, { "epoch": 1.8509095534919222, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.636704444885254, "learning_rate": 5e-05, "loss": 0.7722, "mean_token_accuracy": 0.890404462814331, "num_tokens": 552254188.0, "step": 14550 }, { "epoch": 1.8510367637705127, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.025016784667969, "learning_rate": 5e-05, "loss": 1.0021, "mean_token_accuracy": 0.8520464897155762, "num_tokens": 552286712.0, "step": 14551 }, { "epoch": 1.8511639740491033, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.535113334655762, "learning_rate": 5e-05, "loss": 0.8249, "mean_token_accuracy": 0.8702579140663147, "num_tokens": 552326846.0, "step": 14552 }, { "epoch": 1.8512911843276938, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.042542457580566, "learning_rate": 5e-05, "loss": 0.9123, "mean_token_accuracy": 0.8839250802993774, "num_tokens": 552356753.0, "step": 14553 }, { "epoch": 1.8514183946062843, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.610525608062744, "learning_rate": 5e-05, "loss": 0.8602, "mean_token_accuracy": 0.8606922626495361, "num_tokens": 552391312.0, "step": 14554 }, { "epoch": 1.8515456048848749, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.742521286010742, "learning_rate": 5e-05, "loss": 0.9889, "mean_token_accuracy": 0.8488982915878296, "num_tokens": 552428782.0, "step": 14555 }, { "epoch": 1.8516728151634652, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.82801628112793, "learning_rate": 5e-05, "loss": 0.852, "mean_token_accuracy": 0.8716182708740234, "num_tokens": 552471131.0, "step": 14556 }, { "epoch": 1.8518000254420557, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.236425399780273, "learning_rate": 5e-05, "loss": 0.8528, "mean_token_accuracy": 0.8754241466522217, "num_tokens": 552512242.0, "step": 14557 }, { "epoch": 1.8519272357206462, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.856505393981934, "learning_rate": 5e-05, "loss": 0.8924, "mean_token_accuracy": 0.8525099754333496, "num_tokens": 552545655.0, "step": 14558 }, { "epoch": 1.8520544459992367, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.034483909606934, "learning_rate": 5e-05, "loss": 0.8404, "mean_token_accuracy": 0.8707950711250305, "num_tokens": 552586340.0, "step": 14559 }, { "epoch": 1.8521816562778273, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.638071060180664, "learning_rate": 5e-05, "loss": 0.787, "mean_token_accuracy": 0.8703964352607727, "num_tokens": 552624461.0, "step": 14560 }, { "epoch": 1.8523088665564178, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.13253116607666, "learning_rate": 5e-05, "loss": 0.7893, "mean_token_accuracy": 0.8825491666793823, "num_tokens": 552661566.0, "step": 14561 }, { "epoch": 1.852436076835008, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.400583744049072, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8839007616043091, "num_tokens": 552698184.0, "step": 14562 }, { "epoch": 1.8525632871135986, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.195079803466797, "learning_rate": 5e-05, "loss": 0.7628, "mean_token_accuracy": 0.8894661664962769, "num_tokens": 552741407.0, "step": 14563 }, { "epoch": 1.8526904973921892, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.3622612953186035, "learning_rate": 5e-05, "loss": 0.7728, "mean_token_accuracy": 0.865068793296814, "num_tokens": 552780038.0, "step": 14564 }, { "epoch": 1.8528177076707797, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.101215362548828, "learning_rate": 5e-05, "loss": 0.7858, "mean_token_accuracy": 0.878068745136261, "num_tokens": 552818479.0, "step": 14565 }, { "epoch": 1.8529449179493702, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.390092849731445, "learning_rate": 5e-05, "loss": 0.7264, "mean_token_accuracy": 0.8764867782592773, "num_tokens": 552853762.0, "step": 14566 }, { "epoch": 1.8530721282279607, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.069496154785156, "learning_rate": 5e-05, "loss": 0.866, "mean_token_accuracy": 0.8484443426132202, "num_tokens": 552892693.0, "step": 14567 }, { "epoch": 1.8531993385065513, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.324377059936523, "learning_rate": 5e-05, "loss": 0.7204, "mean_token_accuracy": 0.8791361451148987, "num_tokens": 552930127.0, "step": 14568 }, { "epoch": 1.8533265487851418, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.15771770477295, "learning_rate": 5e-05, "loss": 0.7962, "mean_token_accuracy": 0.8723583221435547, "num_tokens": 552969809.0, "step": 14569 }, { "epoch": 1.8534537590637323, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.209390163421631, "learning_rate": 5e-05, "loss": 0.7169, "mean_token_accuracy": 0.8786635398864746, "num_tokens": 553004747.0, "step": 14570 }, { "epoch": 1.8535809693423229, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.245138168334961, "learning_rate": 5e-05, "loss": 0.8541, "mean_token_accuracy": 0.8603286743164062, "num_tokens": 553049727.0, "step": 14571 }, { "epoch": 1.8537081796209134, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.170042991638184, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.8713082075119019, "num_tokens": 553093263.0, "step": 14572 }, { "epoch": 1.853835389899504, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.258383750915527, "learning_rate": 5e-05, "loss": 0.823, "mean_token_accuracy": 0.8705310225486755, "num_tokens": 553126198.0, "step": 14573 }, { "epoch": 1.8539626001780944, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.214283466339111, "learning_rate": 5e-05, "loss": 0.7127, "mean_token_accuracy": 0.8782089948654175, "num_tokens": 553156798.0, "step": 14574 }, { "epoch": 1.854089810456685, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.161128044128418, "learning_rate": 5e-05, "loss": 0.7897, "mean_token_accuracy": 0.8734647631645203, "num_tokens": 553193612.0, "step": 14575 }, { "epoch": 1.8542170207352755, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.3591837882995605, "learning_rate": 5e-05, "loss": 0.7108, "mean_token_accuracy": 0.8769858479499817, "num_tokens": 553227376.0, "step": 14576 }, { "epoch": 1.854344231013866, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.934566497802734, "learning_rate": 5e-05, "loss": 0.7854, "mean_token_accuracy": 0.8694726824760437, "num_tokens": 553263045.0, "step": 14577 }, { "epoch": 1.8544714412924566, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.554956436157227, "learning_rate": 5e-05, "loss": 0.7652, "mean_token_accuracy": 0.8656308650970459, "num_tokens": 553301834.0, "step": 14578 }, { "epoch": 1.854598651571047, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.667396545410156, "learning_rate": 5e-05, "loss": 0.7883, "mean_token_accuracy": 0.8591870665550232, "num_tokens": 553345980.0, "step": 14579 }, { "epoch": 1.8547258618496376, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.763894081115723, "learning_rate": 5e-05, "loss": 0.7711, "mean_token_accuracy": 0.8691012263298035, "num_tokens": 553380645.0, "step": 14580 }, { "epoch": 1.854853072128228, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.398398399353027, "learning_rate": 5e-05, "loss": 0.7496, "mean_token_accuracy": 0.8638964891433716, "num_tokens": 553417952.0, "step": 14581 }, { "epoch": 1.8549802824068184, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 8.005249977111816, "learning_rate": 5e-05, "loss": 0.7646, "mean_token_accuracy": 0.8754382729530334, "num_tokens": 553454575.0, "step": 14582 }, { "epoch": 1.855107492685409, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.12860107421875, "learning_rate": 5e-05, "loss": 0.641, "mean_token_accuracy": 0.8932836055755615, "num_tokens": 553489846.0, "step": 14583 }, { "epoch": 1.8552347029639995, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.279619216918945, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8790332078933716, "num_tokens": 553524444.0, "step": 14584 }, { "epoch": 1.85536191324259, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.800811290740967, "learning_rate": 5e-05, "loss": 0.6869, "mean_token_accuracy": 0.8769533634185791, "num_tokens": 553562295.0, "step": 14585 }, { "epoch": 1.8554891235211806, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.710433006286621, "learning_rate": 5e-05, "loss": 0.8857, "mean_token_accuracy": 0.8563179969787598, "num_tokens": 553600240.0, "step": 14586 }, { "epoch": 1.8556163337997709, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.4119415283203125, "learning_rate": 5e-05, "loss": 0.6213, "mean_token_accuracy": 0.8878787755966187, "num_tokens": 553641215.0, "step": 14587 }, { "epoch": 1.8557435440783614, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 9.274653434753418, "learning_rate": 5e-05, "loss": 0.8506, "mean_token_accuracy": 0.8863736391067505, "num_tokens": 553675599.0, "step": 14588 }, { "epoch": 1.855870754356952, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.1528143882751465, "learning_rate": 5e-05, "loss": 0.643, "mean_token_accuracy": 0.8841147422790527, "num_tokens": 553706681.0, "step": 14589 }, { "epoch": 1.8559979646355425, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.729454040527344, "learning_rate": 5e-05, "loss": 0.9691, "mean_token_accuracy": 0.8737447261810303, "num_tokens": 553740919.0, "step": 14590 }, { "epoch": 1.856125174914133, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.620728969573975, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.8764830231666565, "num_tokens": 553775509.0, "step": 14591 }, { "epoch": 1.8562523851927235, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.264395713806152, "learning_rate": 5e-05, "loss": 0.9987, "mean_token_accuracy": 0.8632010221481323, "num_tokens": 553809168.0, "step": 14592 }, { "epoch": 1.856379595471314, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.872645378112793, "learning_rate": 5e-05, "loss": 0.87, "mean_token_accuracy": 0.8830212354660034, "num_tokens": 553848186.0, "step": 14593 }, { "epoch": 1.8565068057499046, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.638331890106201, "learning_rate": 5e-05, "loss": 0.8184, "mean_token_accuracy": 0.8751037120819092, "num_tokens": 553888928.0, "step": 14594 }, { "epoch": 1.856634016028495, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.127840042114258, "learning_rate": 5e-05, "loss": 0.9941, "mean_token_accuracy": 0.8606969714164734, "num_tokens": 553929095.0, "step": 14595 }, { "epoch": 1.8567612263070856, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.3148579597473145, "learning_rate": 5e-05, "loss": 0.6934, "mean_token_accuracy": 0.8794116377830505, "num_tokens": 553969610.0, "step": 14596 }, { "epoch": 1.8568884365856761, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 10.129425048828125, "learning_rate": 5e-05, "loss": 1.0198, "mean_token_accuracy": 0.872698187828064, "num_tokens": 554006707.0, "step": 14597 }, { "epoch": 1.8570156468642667, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.558908462524414, "learning_rate": 5e-05, "loss": 0.7986, "mean_token_accuracy": 0.8621211051940918, "num_tokens": 554049006.0, "step": 14598 }, { "epoch": 1.8571428571428572, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.040298461914062, "learning_rate": 5e-05, "loss": 1.0249, "mean_token_accuracy": 0.8732317090034485, "num_tokens": 554086682.0, "step": 14599 }, { "epoch": 1.8572700674214477, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.063704490661621, "learning_rate": 5e-05, "loss": 0.985, "mean_token_accuracy": 0.8658374547958374, "num_tokens": 554122478.0, "step": 14600 }, { "epoch": 1.8573972777000383, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.525283813476562, "learning_rate": 5e-05, "loss": 0.9036, "mean_token_accuracy": 0.8769427537918091, "num_tokens": 554162640.0, "step": 14601 }, { "epoch": 1.8575244879786288, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.985541343688965, "learning_rate": 5e-05, "loss": 1.0048, "mean_token_accuracy": 0.8677121996879578, "num_tokens": 554202161.0, "step": 14602 }, { "epoch": 1.8576516982572193, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.347503662109375, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8766618371009827, "num_tokens": 554235401.0, "step": 14603 }, { "epoch": 1.8577789085358098, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.139841079711914, "learning_rate": 5e-05, "loss": 1.0351, "mean_token_accuracy": 0.8512908220291138, "num_tokens": 554276887.0, "step": 14604 }, { "epoch": 1.8579061188144002, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.582058906555176, "learning_rate": 5e-05, "loss": 0.8997, "mean_token_accuracy": 0.8551038503646851, "num_tokens": 554310513.0, "step": 14605 }, { "epoch": 1.8580333290929907, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.181532859802246, "learning_rate": 5e-05, "loss": 0.8421, "mean_token_accuracy": 0.8740912079811096, "num_tokens": 554354225.0, "step": 14606 }, { "epoch": 1.8581605393715812, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000598907470703125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.309109687805176, "learning_rate": 5e-05, "loss": 0.919, "mean_token_accuracy": 0.8646202087402344, "num_tokens": 554395198.0, "step": 14607 }, { "epoch": 1.8582877496501717, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.223130702972412, "learning_rate": 5e-05, "loss": 0.7775, "mean_token_accuracy": 0.8627328276634216, "num_tokens": 554430234.0, "step": 14608 }, { "epoch": 1.8584149599287623, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.836431503295898, "learning_rate": 5e-05, "loss": 0.9181, "mean_token_accuracy": 0.8720313310623169, "num_tokens": 554469381.0, "step": 14609 }, { "epoch": 1.8585421702073528, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.927953720092773, "learning_rate": 5e-05, "loss": 0.736, "mean_token_accuracy": 0.8681255578994751, "num_tokens": 554500409.0, "step": 14610 }, { "epoch": 1.858669380485943, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.998040199279785, "learning_rate": 5e-05, "loss": 0.8957, "mean_token_accuracy": 0.8803370594978333, "num_tokens": 554537785.0, "step": 14611 }, { "epoch": 1.8587965907645336, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.040561199188232, "learning_rate": 5e-05, "loss": 0.8064, "mean_token_accuracy": 0.8607284426689148, "num_tokens": 554578420.0, "step": 14612 }, { "epoch": 1.8589238010431242, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.912568092346191, "learning_rate": 5e-05, "loss": 0.9279, "mean_token_accuracy": 0.872501015663147, "num_tokens": 554615264.0, "step": 14613 }, { "epoch": 1.8590510113217147, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.119750022888184, "learning_rate": 5e-05, "loss": 0.798, "mean_token_accuracy": 0.8703710436820984, "num_tokens": 554647327.0, "step": 14614 }, { "epoch": 1.8591782216003052, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.90778923034668, "learning_rate": 5e-05, "loss": 0.9139, "mean_token_accuracy": 0.8741869330406189, "num_tokens": 554682552.0, "step": 14615 }, { "epoch": 1.8593054318788957, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.0488409996032715, "learning_rate": 5e-05, "loss": 0.8205, "mean_token_accuracy": 0.8632499575614929, "num_tokens": 554721348.0, "step": 14616 }, { "epoch": 1.8594326421574863, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.047402381896973, "learning_rate": 5e-05, "loss": 0.9192, "mean_token_accuracy": 0.8798251152038574, "num_tokens": 554760618.0, "step": 14617 }, { "epoch": 1.8595598524360768, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.923404216766357, "learning_rate": 5e-05, "loss": 0.8116, "mean_token_accuracy": 0.8659213781356812, "num_tokens": 554801831.0, "step": 14618 }, { "epoch": 1.8596870627146673, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.325026512145996, "learning_rate": 5e-05, "loss": 0.9607, "mean_token_accuracy": 0.877022385597229, "num_tokens": 554838307.0, "step": 14619 }, { "epoch": 1.8598142729932579, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.04494571685791, "learning_rate": 5e-05, "loss": 0.8358, "mean_token_accuracy": 0.868439793586731, "num_tokens": 554883833.0, "step": 14620 }, { "epoch": 1.8599414832718484, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.318397521972656, "learning_rate": 5e-05, "loss": 1.0034, "mean_token_accuracy": 0.8694499731063843, "num_tokens": 554920944.0, "step": 14621 }, { "epoch": 1.860068693550439, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.574819087982178, "learning_rate": 5e-05, "loss": 0.9067, "mean_token_accuracy": 0.8658100366592407, "num_tokens": 554958656.0, "step": 14622 }, { "epoch": 1.8601959038290294, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.720467567443848, "learning_rate": 5e-05, "loss": 0.9736, "mean_token_accuracy": 0.8635243773460388, "num_tokens": 555001864.0, "step": 14623 }, { "epoch": 1.86032311410762, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.035874366760254, "learning_rate": 5e-05, "loss": 0.9064, "mean_token_accuracy": 0.8759058117866516, "num_tokens": 555034268.0, "step": 14624 }, { "epoch": 1.8604503243862105, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.982141017913818, "learning_rate": 5e-05, "loss": 0.8834, "mean_token_accuracy": 0.8661459684371948, "num_tokens": 555075275.0, "step": 14625 }, { "epoch": 1.860577534664801, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.086095809936523, "learning_rate": 5e-05, "loss": 0.991, "mean_token_accuracy": 0.84148108959198, "num_tokens": 555110961.0, "step": 14626 }, { "epoch": 1.8607047449433916, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.728020191192627, "learning_rate": 5e-05, "loss": 0.8572, "mean_token_accuracy": 0.8590166568756104, "num_tokens": 555150369.0, "step": 14627 }, { "epoch": 1.860831955221982, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.8529863357543945, "learning_rate": 5e-05, "loss": 0.8724, "mean_token_accuracy": 0.8629556894302368, "num_tokens": 555182555.0, "step": 14628 }, { "epoch": 1.8609591655005726, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.796407222747803, "learning_rate": 5e-05, "loss": 0.7824, "mean_token_accuracy": 0.8861857652664185, "num_tokens": 555223267.0, "step": 14629 }, { "epoch": 1.861086375779163, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.638197422027588, "learning_rate": 5e-05, "loss": 0.8006, "mean_token_accuracy": 0.8689727783203125, "num_tokens": 555263461.0, "step": 14630 }, { "epoch": 1.8612135860577534, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.84501314163208, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.8703107833862305, "num_tokens": 555306201.0, "step": 14631 }, { "epoch": 1.861340796336344, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.54157829284668, "learning_rate": 5e-05, "loss": 0.7685, "mean_token_accuracy": 0.8725907802581787, "num_tokens": 555344418.0, "step": 14632 }, { "epoch": 1.8614680066149345, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.8039069175720215, "learning_rate": 5e-05, "loss": 0.8043, "mean_token_accuracy": 0.8689120411872864, "num_tokens": 555386096.0, "step": 14633 }, { "epoch": 1.861595216893525, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.544938564300537, "learning_rate": 5e-05, "loss": 0.7535, "mean_token_accuracy": 0.8756603002548218, "num_tokens": 555423368.0, "step": 14634 }, { "epoch": 1.8617224271721156, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.766927719116211, "learning_rate": 5e-05, "loss": 0.7689, "mean_token_accuracy": 0.8670473098754883, "num_tokens": 555458624.0, "step": 14635 }, { "epoch": 1.8618496374507059, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.490389347076416, "learning_rate": 5e-05, "loss": 0.7275, "mean_token_accuracy": 0.8782402276992798, "num_tokens": 555491433.0, "step": 14636 }, { "epoch": 1.8619768477292964, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.842281818389893, "learning_rate": 5e-05, "loss": 0.7211, "mean_token_accuracy": 0.8877735733985901, "num_tokens": 555527198.0, "step": 14637 }, { "epoch": 1.862104058007887, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.35967493057251, "learning_rate": 5e-05, "loss": 0.7072, "mean_token_accuracy": 0.8790795803070068, "num_tokens": 555563232.0, "step": 14638 }, { "epoch": 1.8622312682864774, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.979792594909668, "learning_rate": 5e-05, "loss": 0.8183, "mean_token_accuracy": 0.8603315353393555, "num_tokens": 555599805.0, "step": 14639 }, { "epoch": 1.862358478565068, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.180368423461914, "learning_rate": 5e-05, "loss": 0.7301, "mean_token_accuracy": 0.8680672645568848, "num_tokens": 555628835.0, "step": 14640 }, { "epoch": 1.8624856888436585, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.167487144470215, "learning_rate": 5e-05, "loss": 0.7671, "mean_token_accuracy": 0.8788251876831055, "num_tokens": 555668120.0, "step": 14641 }, { "epoch": 1.862612899122249, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.962831497192383, "learning_rate": 5e-05, "loss": 0.7574, "mean_token_accuracy": 0.8565244078636169, "num_tokens": 555704662.0, "step": 14642 }, { "epoch": 1.8627401094008396, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.497636795043945, "learning_rate": 5e-05, "loss": 0.79, "mean_token_accuracy": 0.8812955617904663, "num_tokens": 555738867.0, "step": 14643 }, { "epoch": 1.86286731967943, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.627100944519043, "learning_rate": 5e-05, "loss": 0.7144, "mean_token_accuracy": 0.8659696578979492, "num_tokens": 555773500.0, "step": 14644 }, { "epoch": 1.8629945299580206, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.976836204528809, "learning_rate": 5e-05, "loss": 0.8376, "mean_token_accuracy": 0.8852239847183228, "num_tokens": 555812139.0, "step": 14645 }, { "epoch": 1.8631217402366111, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.300455570220947, "learning_rate": 5e-05, "loss": 0.723, "mean_token_accuracy": 0.8608214855194092, "num_tokens": 555850823.0, "step": 14646 }, { "epoch": 1.8632489505152017, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.500018119812012, "learning_rate": 5e-05, "loss": 0.9031, "mean_token_accuracy": 0.8907183408737183, "num_tokens": 555888554.0, "step": 14647 }, { "epoch": 1.8633761607937922, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.438878059387207, "learning_rate": 5e-05, "loss": 0.7424, "mean_token_accuracy": 0.8701068758964539, "num_tokens": 555932629.0, "step": 14648 }, { "epoch": 1.8635033710723827, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.478208541870117, "learning_rate": 5e-05, "loss": 1.0142, "mean_token_accuracy": 0.862935483455658, "num_tokens": 555970722.0, "step": 14649 }, { "epoch": 1.8636305813509733, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.413348197937012, "learning_rate": 5e-05, "loss": 0.8414, "mean_token_accuracy": 0.8797160387039185, "num_tokens": 556005784.0, "step": 14650 }, { "epoch": 1.8637577916295638, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.314371109008789, "learning_rate": 5e-05, "loss": 0.8744, "mean_token_accuracy": 0.8791691064834595, "num_tokens": 556046053.0, "step": 14651 }, { "epoch": 1.8638850019081543, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.560615539550781, "learning_rate": 5e-05, "loss": 0.9603, "mean_token_accuracy": 0.8668214082717896, "num_tokens": 556084265.0, "step": 14652 }, { "epoch": 1.8640122121867448, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.735677242279053, "learning_rate": 5e-05, "loss": 0.713, "mean_token_accuracy": 0.8762643337249756, "num_tokens": 556125062.0, "step": 14653 }, { "epoch": 1.8641394224653351, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.69446849822998, "learning_rate": 5e-05, "loss": 1.0334, "mean_token_accuracy": 0.8571489453315735, "num_tokens": 556162590.0, "step": 14654 }, { "epoch": 1.8642666327439257, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.128228187561035, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8688348531723022, "num_tokens": 556200560.0, "step": 14655 }, { "epoch": 1.8643938430225162, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.412900924682617, "learning_rate": 5e-05, "loss": 1.0668, "mean_token_accuracy": 0.8785155415534973, "num_tokens": 556237199.0, "step": 14656 }, { "epoch": 1.8645210533011067, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.245455265045166, "learning_rate": 5e-05, "loss": 0.9117, "mean_token_accuracy": 0.8703389167785645, "num_tokens": 556278620.0, "step": 14657 }, { "epoch": 1.8646482635796973, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.82345962524414, "learning_rate": 5e-05, "loss": 1.0219, "mean_token_accuracy": 0.8755501508712769, "num_tokens": 556312844.0, "step": 14658 }, { "epoch": 1.8647754738582878, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 8.694796562194824, "learning_rate": 5e-05, "loss": 1.0412, "mean_token_accuracy": 0.8767398595809937, "num_tokens": 556351218.0, "step": 14659 }, { "epoch": 1.864902684136878, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.099078178405762, "learning_rate": 5e-05, "loss": 0.8519, "mean_token_accuracy": 0.882074236869812, "num_tokens": 556388176.0, "step": 14660 }, { "epoch": 1.8650298944154686, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.311841011047363, "learning_rate": 5e-05, "loss": 1.033, "mean_token_accuracy": 0.8677631616592407, "num_tokens": 556424996.0, "step": 14661 }, { "epoch": 1.8651571046940592, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.310143947601318, "learning_rate": 5e-05, "loss": 0.8954, "mean_token_accuracy": 0.8525031805038452, "num_tokens": 556465961.0, "step": 14662 }, { "epoch": 1.8652843149726497, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.116978645324707, "learning_rate": 5e-05, "loss": 0.9556, "mean_token_accuracy": 0.8726130723953247, "num_tokens": 556504777.0, "step": 14663 }, { "epoch": 1.8654115252512402, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.800059795379639, "learning_rate": 5e-05, "loss": 0.8756, "mean_token_accuracy": 0.8783807754516602, "num_tokens": 556539737.0, "step": 14664 }, { "epoch": 1.8655387355298307, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.971108913421631, "learning_rate": 5e-05, "loss": 0.814, "mean_token_accuracy": 0.8840617537498474, "num_tokens": 556578028.0, "step": 14665 }, { "epoch": 1.8656659458084213, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.470377922058105, "learning_rate": 5e-05, "loss": 0.9623, "mean_token_accuracy": 0.8614406585693359, "num_tokens": 556611362.0, "step": 14666 }, { "epoch": 1.8657931560870118, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.186554431915283, "learning_rate": 5e-05, "loss": 0.782, "mean_token_accuracy": 0.8690598011016846, "num_tokens": 556652329.0, "step": 14667 }, { "epoch": 1.8659203663656023, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.769139289855957, "learning_rate": 5e-05, "loss": 0.8594, "mean_token_accuracy": 0.8939673900604248, "num_tokens": 556689531.0, "step": 14668 }, { "epoch": 1.8660475766441929, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.046389579772949, "learning_rate": 5e-05, "loss": 0.7453, "mean_token_accuracy": 0.8807182908058167, "num_tokens": 556722309.0, "step": 14669 }, { "epoch": 1.8661747869227834, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.830078125, "learning_rate": 5e-05, "loss": 0.8856, "mean_token_accuracy": 0.8814167976379395, "num_tokens": 556755682.0, "step": 14670 }, { "epoch": 1.866301997201374, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.049667835235596, "learning_rate": 5e-05, "loss": 0.764, "mean_token_accuracy": 0.880831241607666, "num_tokens": 556792591.0, "step": 14671 }, { "epoch": 1.8664292074799644, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.8796968460083, "learning_rate": 5e-05, "loss": 0.9012, "mean_token_accuracy": 0.8771339058876038, "num_tokens": 556835168.0, "step": 14672 }, { "epoch": 1.866556417758555, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.082042217254639, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8735456466674805, "num_tokens": 556876106.0, "step": 14673 }, { "epoch": 1.8666836280371455, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.824545860290527, "learning_rate": 5e-05, "loss": 0.9257, "mean_token_accuracy": 0.8684300184249878, "num_tokens": 556917154.0, "step": 14674 }, { "epoch": 1.866810838315736, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.142207145690918, "learning_rate": 5e-05, "loss": 0.8324, "mean_token_accuracy": 0.8670275211334229, "num_tokens": 556953706.0, "step": 14675 }, { "epoch": 1.8669380485943265, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.770532608032227, "learning_rate": 5e-05, "loss": 0.9076, "mean_token_accuracy": 0.8738299608230591, "num_tokens": 556992035.0, "step": 14676 }, { "epoch": 1.867065258872917, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.1079020500183105, "learning_rate": 5e-05, "loss": 0.7966, "mean_token_accuracy": 0.8731811046600342, "num_tokens": 557034785.0, "step": 14677 }, { "epoch": 1.8671924691515076, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.874082565307617, "learning_rate": 5e-05, "loss": 0.9572, "mean_token_accuracy": 0.8637799620628357, "num_tokens": 557074910.0, "step": 14678 }, { "epoch": 1.867319679430098, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.0455780029296875, "learning_rate": 5e-05, "loss": 0.7824, "mean_token_accuracy": 0.8753689527511597, "num_tokens": 557105327.0, "step": 14679 }, { "epoch": 1.8674468897086884, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.937112808227539, "learning_rate": 5e-05, "loss": 0.9583, "mean_token_accuracy": 0.8698943853378296, "num_tokens": 557149261.0, "step": 14680 }, { "epoch": 1.867574099987279, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.210377216339111, "learning_rate": 5e-05, "loss": 0.8354, "mean_token_accuracy": 0.8694254159927368, "num_tokens": 557190550.0, "step": 14681 }, { "epoch": 1.8677013102658695, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.742292404174805, "learning_rate": 5e-05, "loss": 0.9679, "mean_token_accuracy": 0.8573986887931824, "num_tokens": 557222828.0, "step": 14682 }, { "epoch": 1.86782852054446, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.458216190338135, "learning_rate": 5e-05, "loss": 0.879, "mean_token_accuracy": 0.8661472797393799, "num_tokens": 557264228.0, "step": 14683 }, { "epoch": 1.8679557308230506, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.387755393981934, "learning_rate": 5e-05, "loss": 0.8588, "mean_token_accuracy": 0.889102041721344, "num_tokens": 557301310.0, "step": 14684 }, { "epoch": 1.8680829411016409, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.656191825866699, "learning_rate": 5e-05, "loss": 0.8586, "mean_token_accuracy": 0.8755618333816528, "num_tokens": 557338596.0, "step": 14685 }, { "epoch": 1.8682101513802314, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.035506248474121, "learning_rate": 5e-05, "loss": 0.8587, "mean_token_accuracy": 0.8697049617767334, "num_tokens": 557374939.0, "step": 14686 }, { "epoch": 1.868337361658822, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.724691390991211, "learning_rate": 5e-05, "loss": 0.8423, "mean_token_accuracy": 0.8725429773330688, "num_tokens": 557411711.0, "step": 14687 }, { "epoch": 1.8684645719374124, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.837678909301758, "learning_rate": 5e-05, "loss": 0.8178, "mean_token_accuracy": 0.874620795249939, "num_tokens": 557448892.0, "step": 14688 }, { "epoch": 1.868591782216003, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.668001174926758, "learning_rate": 5e-05, "loss": 0.8184, "mean_token_accuracy": 0.8644270896911621, "num_tokens": 557490911.0, "step": 14689 }, { "epoch": 1.8687189924945935, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.81781530380249, "learning_rate": 5e-05, "loss": 0.8001, "mean_token_accuracy": 0.8734539747238159, "num_tokens": 557529721.0, "step": 14690 }, { "epoch": 1.868846202773184, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.521660804748535, "learning_rate": 5e-05, "loss": 0.7551, "mean_token_accuracy": 0.8750890493392944, "num_tokens": 557567591.0, "step": 14691 }, { "epoch": 1.8689734130517746, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.891143321990967, "learning_rate": 5e-05, "loss": 0.8127, "mean_token_accuracy": 0.8691602945327759, "num_tokens": 557609105.0, "step": 14692 }, { "epoch": 1.869100623330365, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.418722629547119, "learning_rate": 5e-05, "loss": 0.8576, "mean_token_accuracy": 0.8412266969680786, "num_tokens": 557658798.0, "step": 14693 }, { "epoch": 1.8692278336089556, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.927618980407715, "learning_rate": 5e-05, "loss": 0.7473, "mean_token_accuracy": 0.8872625231742859, "num_tokens": 557695572.0, "step": 14694 }, { "epoch": 1.8693550438875461, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.34586238861084, "learning_rate": 5e-05, "loss": 0.748, "mean_token_accuracy": 0.875645637512207, "num_tokens": 557728796.0, "step": 14695 }, { "epoch": 1.8694822541661367, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.968968391418457, "learning_rate": 5e-05, "loss": 0.7478, "mean_token_accuracy": 0.8840973377227783, "num_tokens": 557761786.0, "step": 14696 }, { "epoch": 1.8696094644447272, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.258426189422607, "learning_rate": 5e-05, "loss": 0.7551, "mean_token_accuracy": 0.8657100796699524, "num_tokens": 557809454.0, "step": 14697 }, { "epoch": 1.8697366747233177, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.062276840209961, "learning_rate": 5e-05, "loss": 0.7238, "mean_token_accuracy": 0.891984224319458, "num_tokens": 557846959.0, "step": 14698 }, { "epoch": 1.8698638850019083, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.152133941650391, "learning_rate": 5e-05, "loss": 0.698, "mean_token_accuracy": 0.8824797868728638, "num_tokens": 557879873.0, "step": 14699 }, { "epoch": 1.8699910952804988, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.160493850708008, "learning_rate": 5e-05, "loss": 0.83, "mean_token_accuracy": 0.866629958152771, "num_tokens": 557918020.0, "step": 14700 }, { "epoch": 1.8701183055590893, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.051753520965576, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.867676317691803, "num_tokens": 557960633.0, "step": 14701 }, { "epoch": 1.8702455158376798, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.297462463378906, "learning_rate": 5e-05, "loss": 0.8125, "mean_token_accuracy": 0.8719112873077393, "num_tokens": 557998430.0, "step": 14702 }, { "epoch": 1.8703727261162701, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.936864376068115, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.8639395236968994, "num_tokens": 558038283.0, "step": 14703 }, { "epoch": 1.8704999363948607, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.444211959838867, "learning_rate": 5e-05, "loss": 0.8396, "mean_token_accuracy": 0.8707391023635864, "num_tokens": 558079636.0, "step": 14704 }, { "epoch": 1.8706271466734512, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.8735671043396, "learning_rate": 5e-05, "loss": 0.6922, "mean_token_accuracy": 0.8810259103775024, "num_tokens": 558117670.0, "step": 14705 }, { "epoch": 1.8707543569520417, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.506782531738281, "learning_rate": 5e-05, "loss": 0.8891, "mean_token_accuracy": 0.8571276664733887, "num_tokens": 558155732.0, "step": 14706 }, { "epoch": 1.8708815672306323, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.931663513183594, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.8683611154556274, "num_tokens": 558195389.0, "step": 14707 }, { "epoch": 1.8710087775092228, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.366243362426758, "learning_rate": 5e-05, "loss": 0.8115, "mean_token_accuracy": 0.8831883668899536, "num_tokens": 558239535.0, "step": 14708 }, { "epoch": 1.871135987787813, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.220053672790527, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8701236844062805, "num_tokens": 558277441.0, "step": 14709 }, { "epoch": 1.8712631980664036, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.947854995727539, "learning_rate": 5e-05, "loss": 0.8302, "mean_token_accuracy": 0.8663996458053589, "num_tokens": 558319331.0, "step": 14710 }, { "epoch": 1.8713904083449941, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.656822681427002, "learning_rate": 5e-05, "loss": 0.767, "mean_token_accuracy": 0.8738113641738892, "num_tokens": 558365494.0, "step": 14711 }, { "epoch": 1.8715176186235847, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.367422580718994, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.8651182055473328, "num_tokens": 558407608.0, "step": 14712 }, { "epoch": 1.8716448289021752, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.172991752624512, "learning_rate": 5e-05, "loss": 0.8708, "mean_token_accuracy": 0.8570054173469543, "num_tokens": 558444848.0, "step": 14713 }, { "epoch": 1.8717720391807657, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.765464782714844, "learning_rate": 5e-05, "loss": 0.7724, "mean_token_accuracy": 0.851493239402771, "num_tokens": 558486630.0, "step": 14714 }, { "epoch": 1.8718992494593563, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.79899787902832, "learning_rate": 5e-05, "loss": 0.8082, "mean_token_accuracy": 0.882111668586731, "num_tokens": 558523609.0, "step": 14715 }, { "epoch": 1.8720264597379468, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.222018241882324, "learning_rate": 5e-05, "loss": 0.665, "mean_token_accuracy": 0.8788248300552368, "num_tokens": 558561065.0, "step": 14716 }, { "epoch": 1.8721536700165373, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.573735237121582, "learning_rate": 5e-05, "loss": 0.9197, "mean_token_accuracy": 0.8782540559768677, "num_tokens": 558602007.0, "step": 14717 }, { "epoch": 1.8722808802951278, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 5.964156150817871, "learning_rate": 5e-05, "loss": 0.6951, "mean_token_accuracy": 0.8725805282592773, "num_tokens": 558637712.0, "step": 14718 }, { "epoch": 1.8724080905737184, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.175745964050293, "learning_rate": 5e-05, "loss": 1.0127, "mean_token_accuracy": 0.8761473894119263, "num_tokens": 558671939.0, "step": 14719 }, { "epoch": 1.872535300852309, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 6.887062072753906, "learning_rate": 5e-05, "loss": 0.8542, "mean_token_accuracy": 0.8677294254302979, "num_tokens": 558708413.0, "step": 14720 }, { "epoch": 1.8726625111308994, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.630758285522461, "learning_rate": 5e-05, "loss": 1.0993, "mean_token_accuracy": 0.8461227416992188, "num_tokens": 558748993.0, "step": 14721 }, { "epoch": 1.87278972140949, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.46627140045166, "learning_rate": 5e-05, "loss": 1.0346, "mean_token_accuracy": 0.864465594291687, "num_tokens": 558788058.0, "step": 14722 }, { "epoch": 1.8729169316880805, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.59208869934082, "learning_rate": 5e-05, "loss": 0.8973, "mean_token_accuracy": 0.8474088907241821, "num_tokens": 558828046.0, "step": 14723 }, { "epoch": 1.873044141966671, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.826289176940918, "learning_rate": 5e-05, "loss": 1.0242, "mean_token_accuracy": 0.8742715120315552, "num_tokens": 558862445.0, "step": 14724 }, { "epoch": 1.8731713522452615, "ewc_loss": 0.375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.359469413757324, "learning_rate": 5e-05, "loss": 0.7183, "mean_token_accuracy": 0.8812105655670166, "num_tokens": 558907645.0, "step": 14725 }, { "epoch": 1.873298562523852, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 10.935269355773926, "learning_rate": 5e-05, "loss": 1.1501, "mean_token_accuracy": 0.8724316358566284, "num_tokens": 558943253.0, "step": 14726 }, { "epoch": 1.8734257728024426, "ewc_loss": 0.625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 7.716526031494141, "learning_rate": 5e-05, "loss": 0.9812, "mean_token_accuracy": 0.8762310147285461, "num_tokens": 558980341.0, "step": 14727 }, { "epoch": 1.873552983081033, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.325552940368652, "learning_rate": 5e-05, "loss": 1.0846, "mean_token_accuracy": 0.860516369342804, "num_tokens": 559018659.0, "step": 14728 }, { "epoch": 1.8736801933596234, "ewc_loss": 0.78125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 8.952744483947754, "learning_rate": 5e-05, "loss": 1.1495, "mean_token_accuracy": 0.8735983371734619, "num_tokens": 559060254.0, "step": 14729 }, { "epoch": 1.873807403638214, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.769774436950684, "learning_rate": 5e-05, "loss": 0.9193, "mean_token_accuracy": 0.8740169405937195, "num_tokens": 559099166.0, "step": 14730 }, { "epoch": 1.8739346139168045, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.135481834411621, "learning_rate": 5e-05, "loss": 1.0012, "mean_token_accuracy": 0.8807729482650757, "num_tokens": 559135189.0, "step": 14731 }, { "epoch": 1.874061824195395, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.06571102142334, "learning_rate": 5e-05, "loss": 0.8851, "mean_token_accuracy": 0.8756687045097351, "num_tokens": 559178491.0, "step": 14732 }, { "epoch": 1.8741890344739855, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.67579460144043, "learning_rate": 5e-05, "loss": 0.9747, "mean_token_accuracy": 0.8641722202301025, "num_tokens": 559213977.0, "step": 14733 }, { "epoch": 1.8743162447525759, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.194369316101074, "learning_rate": 5e-05, "loss": 0.9373, "mean_token_accuracy": 0.8697630167007446, "num_tokens": 559254309.0, "step": 14734 }, { "epoch": 1.8744434550311664, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.729010581970215, "learning_rate": 5e-05, "loss": 0.7666, "mean_token_accuracy": 0.8824104070663452, "num_tokens": 559289756.0, "step": 14735 }, { "epoch": 1.874570665309757, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.657428741455078, "learning_rate": 5e-05, "loss": 0.9177, "mean_token_accuracy": 0.8725916743278503, "num_tokens": 559331143.0, "step": 14736 }, { "epoch": 1.8746978755883474, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.0661396980285645, "learning_rate": 5e-05, "loss": 0.7597, "mean_token_accuracy": 0.8688766956329346, "num_tokens": 559368256.0, "step": 14737 }, { "epoch": 1.874825085866938, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.102287292480469, "learning_rate": 5e-05, "loss": 1.0022, "mean_token_accuracy": 0.8526827096939087, "num_tokens": 559408853.0, "step": 14738 }, { "epoch": 1.8749522961455285, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.872351169586182, "learning_rate": 5e-05, "loss": 0.8089, "mean_token_accuracy": 0.8618254065513611, "num_tokens": 559445606.0, "step": 14739 }, { "epoch": 1.875079506424119, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.349928855895996, "learning_rate": 5e-05, "loss": 0.9456, "mean_token_accuracy": 0.8753616809844971, "num_tokens": 559482482.0, "step": 14740 }, { "epoch": 1.8752067167027096, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.031556129455566, "learning_rate": 5e-05, "loss": 0.8366, "mean_token_accuracy": 0.8704800009727478, "num_tokens": 559522757.0, "step": 14741 }, { "epoch": 1.8753339269813, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.32546615600586, "learning_rate": 5e-05, "loss": 0.9522, "mean_token_accuracy": 0.8792331218719482, "num_tokens": 559556684.0, "step": 14742 }, { "epoch": 1.8754611372598906, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.348047256469727, "learning_rate": 5e-05, "loss": 0.8931, "mean_token_accuracy": 0.8704061508178711, "num_tokens": 559597548.0, "step": 14743 }, { "epoch": 1.8755883475384811, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.979364395141602, "learning_rate": 5e-05, "loss": 0.9752, "mean_token_accuracy": 0.8663829565048218, "num_tokens": 559638440.0, "step": 14744 }, { "epoch": 1.8757155578170717, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.5469489097595215, "learning_rate": 5e-05, "loss": 0.9587, "mean_token_accuracy": 0.8493767380714417, "num_tokens": 559680819.0, "step": 14745 }, { "epoch": 1.8758427680956622, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.671427726745605, "learning_rate": 5e-05, "loss": 0.941, "mean_token_accuracy": 0.8694601058959961, "num_tokens": 559721861.0, "step": 14746 }, { "epoch": 1.8759699783742527, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.450999736785889, "learning_rate": 5e-05, "loss": 0.8713, "mean_token_accuracy": 0.8709498643875122, "num_tokens": 559757370.0, "step": 14747 }, { "epoch": 1.8760971886528433, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.68731689453125, "learning_rate": 5e-05, "loss": 0.9588, "mean_token_accuracy": 0.8647236824035645, "num_tokens": 559795570.0, "step": 14748 }, { "epoch": 1.8762243989314338, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.233219146728516, "learning_rate": 5e-05, "loss": 0.802, "mean_token_accuracy": 0.876643180847168, "num_tokens": 559835115.0, "step": 14749 }, { "epoch": 1.8763516092100243, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.842885971069336, "learning_rate": 5e-05, "loss": 0.9714, "mean_token_accuracy": 0.8668839335441589, "num_tokens": 559876119.0, "step": 14750 }, { "epoch": 1.8764788194886148, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.19992208480835, "learning_rate": 5e-05, "loss": 0.8134, "mean_token_accuracy": 0.87534499168396, "num_tokens": 559908349.0, "step": 14751 }, { "epoch": 1.8766060297672051, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.828104972839355, "learning_rate": 5e-05, "loss": 0.9763, "mean_token_accuracy": 0.8603661060333252, "num_tokens": 559947336.0, "step": 14752 }, { "epoch": 1.8767332400457957, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.396773338317871, "learning_rate": 5e-05, "loss": 0.8747, "mean_token_accuracy": 0.8630012273788452, "num_tokens": 559989571.0, "step": 14753 }, { "epoch": 1.8768604503243862, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.584576606750488, "learning_rate": 5e-05, "loss": 0.9857, "mean_token_accuracy": 0.8517311811447144, "num_tokens": 560022840.0, "step": 14754 }, { "epoch": 1.8769876606029767, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.624323844909668, "learning_rate": 5e-05, "loss": 0.8885, "mean_token_accuracy": 0.8630304336547852, "num_tokens": 560061939.0, "step": 14755 }, { "epoch": 1.8771148708815673, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.171489715576172, "learning_rate": 5e-05, "loss": 0.8971, "mean_token_accuracy": 0.8679234981536865, "num_tokens": 560097480.0, "step": 14756 }, { "epoch": 1.8772420811601578, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.7663679122924805, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.8687934875488281, "num_tokens": 560137957.0, "step": 14757 }, { "epoch": 1.877369291438748, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.862772464752197, "learning_rate": 5e-05, "loss": 0.8515, "mean_token_accuracy": 0.8663263916969299, "num_tokens": 560173717.0, "step": 14758 }, { "epoch": 1.8774965017173386, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.733474254608154, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.8697715401649475, "num_tokens": 560209833.0, "step": 14759 }, { "epoch": 1.8776237119959291, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.8147873878479, "learning_rate": 5e-05, "loss": 0.903, "mean_token_accuracy": 0.8447704315185547, "num_tokens": 560250406.0, "step": 14760 }, { "epoch": 1.8777509222745197, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.577368259429932, "learning_rate": 5e-05, "loss": 0.786, "mean_token_accuracy": 0.8695653080940247, "num_tokens": 560288828.0, "step": 14761 }, { "epoch": 1.8778781325531102, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.884112358093262, "learning_rate": 5e-05, "loss": 0.7921, "mean_token_accuracy": 0.8803510069847107, "num_tokens": 560327431.0, "step": 14762 }, { "epoch": 1.8780053428317007, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.455077648162842, "learning_rate": 5e-05, "loss": 0.7273, "mean_token_accuracy": 0.8870106339454651, "num_tokens": 560367767.0, "step": 14763 }, { "epoch": 1.8781325531102913, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.951058864593506, "learning_rate": 5e-05, "loss": 0.7765, "mean_token_accuracy": 0.8814337253570557, "num_tokens": 560401179.0, "step": 14764 }, { "epoch": 1.8782597633888818, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.314833641052246, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.8658878803253174, "num_tokens": 560431198.0, "step": 14765 }, { "epoch": 1.8783869736674723, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.057964324951172, "learning_rate": 5e-05, "loss": 0.8061, "mean_token_accuracy": 0.8693462610244751, "num_tokens": 560470244.0, "step": 14766 }, { "epoch": 1.8785141839460628, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.206394672393799, "learning_rate": 5e-05, "loss": 0.7212, "mean_token_accuracy": 0.8790134191513062, "num_tokens": 560511696.0, "step": 14767 }, { "epoch": 1.8786413942246534, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.152981758117676, "learning_rate": 5e-05, "loss": 0.746, "mean_token_accuracy": 0.8933159112930298, "num_tokens": 560547555.0, "step": 14768 }, { "epoch": 1.878768604503244, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.100512504577637, "learning_rate": 5e-05, "loss": 0.7412, "mean_token_accuracy": 0.8692834377288818, "num_tokens": 560584349.0, "step": 14769 }, { "epoch": 1.8788958147818344, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.243399620056152, "learning_rate": 5e-05, "loss": 0.7983, "mean_token_accuracy": 0.8780938386917114, "num_tokens": 560628403.0, "step": 14770 }, { "epoch": 1.879023025060425, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.032965660095215, "learning_rate": 5e-05, "loss": 0.7291, "mean_token_accuracy": 0.8738279342651367, "num_tokens": 560663454.0, "step": 14771 }, { "epoch": 1.8791502353390155, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.33625602722168, "learning_rate": 5e-05, "loss": 0.8301, "mean_token_accuracy": 0.8687044978141785, "num_tokens": 560700506.0, "step": 14772 }, { "epoch": 1.879277445617606, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.007597923278809, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8772194385528564, "num_tokens": 560739690.0, "step": 14773 }, { "epoch": 1.8794046558961965, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.349319458007812, "learning_rate": 5e-05, "loss": 0.8257, "mean_token_accuracy": 0.8726227283477783, "num_tokens": 560786364.0, "step": 14774 }, { "epoch": 1.879531866174787, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.090641498565674, "learning_rate": 5e-05, "loss": 0.7464, "mean_token_accuracy": 0.8659088611602783, "num_tokens": 560828850.0, "step": 14775 }, { "epoch": 1.8796590764533776, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.208208084106445, "learning_rate": 5e-05, "loss": 0.8043, "mean_token_accuracy": 0.8789079785346985, "num_tokens": 560863401.0, "step": 14776 }, { "epoch": 1.879786286731968, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.296694278717041, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.8731209635734558, "num_tokens": 560894347.0, "step": 14777 }, { "epoch": 1.8799134970105584, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.878856182098389, "learning_rate": 5e-05, "loss": 0.8471, "mean_token_accuracy": 0.8557677268981934, "num_tokens": 560935302.0, "step": 14778 }, { "epoch": 1.880040707289149, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.604973316192627, "learning_rate": 5e-05, "loss": 0.7303, "mean_token_accuracy": 0.8794040083885193, "num_tokens": 560971960.0, "step": 14779 }, { "epoch": 1.8801679175677395, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.4729108810424805, "learning_rate": 5e-05, "loss": 0.7667, "mean_token_accuracy": 0.863508403301239, "num_tokens": 561012267.0, "step": 14780 }, { "epoch": 1.88029512784633, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.953568935394287, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.8703083992004395, "num_tokens": 561048861.0, "step": 14781 }, { "epoch": 1.8804223381249205, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.043368339538574, "learning_rate": 5e-05, "loss": 0.7029, "mean_token_accuracy": 0.8742659687995911, "num_tokens": 561087832.0, "step": 14782 }, { "epoch": 1.8805495484035109, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.37821102142334, "learning_rate": 5e-05, "loss": 0.7901, "mean_token_accuracy": 0.8793745040893555, "num_tokens": 561120657.0, "step": 14783 }, { "epoch": 1.8806767586821014, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.584446907043457, "learning_rate": 5e-05, "loss": 0.6608, "mean_token_accuracy": 0.8807228207588196, "num_tokens": 561161762.0, "step": 14784 }, { "epoch": 1.880803968960692, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.930739402770996, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8727113008499146, "num_tokens": 561199459.0, "step": 14785 }, { "epoch": 1.8809311792392824, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.161421775817871, "learning_rate": 5e-05, "loss": 0.7364, "mean_token_accuracy": 0.8544921278953552, "num_tokens": 561237958.0, "step": 14786 }, { "epoch": 1.881058389517873, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.610709190368652, "learning_rate": 5e-05, "loss": 0.9651, "mean_token_accuracy": 0.866780698299408, "num_tokens": 561272086.0, "step": 14787 }, { "epoch": 1.8811855997964635, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.102154731750488, "learning_rate": 5e-05, "loss": 0.6866, "mean_token_accuracy": 0.879330039024353, "num_tokens": 561309821.0, "step": 14788 }, { "epoch": 1.881312810075054, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.929394721984863, "learning_rate": 5e-05, "loss": 1.017, "mean_token_accuracy": 0.8697871565818787, "num_tokens": 561351565.0, "step": 14789 }, { "epoch": 1.8814400203536445, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.1628098487854, "learning_rate": 5e-05, "loss": 0.9338, "mean_token_accuracy": 0.8501947522163391, "num_tokens": 561387399.0, "step": 14790 }, { "epoch": 1.881567230632235, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.954902648925781, "learning_rate": 5e-05, "loss": 0.9832, "mean_token_accuracy": 0.867133378982544, "num_tokens": 561427917.0, "step": 14791 }, { "epoch": 1.8816944409108256, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.553048133850098, "learning_rate": 5e-05, "loss": 1.008, "mean_token_accuracy": 0.8649535179138184, "num_tokens": 561465398.0, "step": 14792 }, { "epoch": 1.8818216511894161, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.98281192779541, "learning_rate": 5e-05, "loss": 0.7442, "mean_token_accuracy": 0.8731794357299805, "num_tokens": 561510264.0, "step": 14793 }, { "epoch": 1.8819488614680067, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.929770469665527, "learning_rate": 5e-05, "loss": 1.0493, "mean_token_accuracy": 0.8644540905952454, "num_tokens": 561550942.0, "step": 14794 }, { "epoch": 1.8820760717465972, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.243652820587158, "learning_rate": 5e-05, "loss": 0.7213, "mean_token_accuracy": 0.8778911828994751, "num_tokens": 561580571.0, "step": 14795 }, { "epoch": 1.8822032820251877, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.746169090270996, "learning_rate": 5e-05, "loss": 1.1316, "mean_token_accuracy": 0.8744158744812012, "num_tokens": 561618993.0, "step": 14796 }, { "epoch": 1.8823304923037782, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.742618083953857, "learning_rate": 5e-05, "loss": 1.0438, "mean_token_accuracy": 0.8535133600234985, "num_tokens": 561659389.0, "step": 14797 }, { "epoch": 1.8824577025823688, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.869644165039062, "learning_rate": 5e-05, "loss": 1.0286, "mean_token_accuracy": 0.8728996515274048, "num_tokens": 561707703.0, "step": 14798 }, { "epoch": 1.8825849128609593, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 8.91270637512207, "learning_rate": 5e-05, "loss": 1.1077, "mean_token_accuracy": 0.8686747550964355, "num_tokens": 561749367.0, "step": 14799 }, { "epoch": 1.8827121231395498, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.273255348205566, "learning_rate": 5e-05, "loss": 0.908, "mean_token_accuracy": 0.8636430501937866, "num_tokens": 561785234.0, "step": 14800 }, { "epoch": 1.8828393334181401, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.21761417388916, "learning_rate": 5e-05, "loss": 0.9984, "mean_token_accuracy": 0.8796191215515137, "num_tokens": 561825801.0, "step": 14801 }, { "epoch": 1.8829665436967307, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.777267932891846, "learning_rate": 5e-05, "loss": 0.9156, "mean_token_accuracy": 0.8617671132087708, "num_tokens": 561865030.0, "step": 14802 }, { "epoch": 1.8830937539753212, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00058746337890625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.68695068359375, "learning_rate": 5e-05, "loss": 0.914, "mean_token_accuracy": 0.8802545070648193, "num_tokens": 561899767.0, "step": 14803 }, { "epoch": 1.8832209642539117, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.154136657714844, "learning_rate": 5e-05, "loss": 0.9569, "mean_token_accuracy": 0.864470362663269, "num_tokens": 561936794.0, "step": 14804 }, { "epoch": 1.8833481745325023, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.606359958648682, "learning_rate": 5e-05, "loss": 0.794, "mean_token_accuracy": 0.8707313537597656, "num_tokens": 561971770.0, "step": 14805 }, { "epoch": 1.8834753848110928, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.719263076782227, "learning_rate": 5e-05, "loss": 0.9545, "mean_token_accuracy": 0.8652313351631165, "num_tokens": 562016400.0, "step": 14806 }, { "epoch": 1.883602595089683, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.994736671447754, "learning_rate": 5e-05, "loss": 0.7304, "mean_token_accuracy": 0.8783637285232544, "num_tokens": 562050732.0, "step": 14807 }, { "epoch": 1.8837298053682736, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.101167678833008, "learning_rate": 5e-05, "loss": 0.9544, "mean_token_accuracy": 0.8709574341773987, "num_tokens": 562093619.0, "step": 14808 }, { "epoch": 1.8838570156468641, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 6.941784858703613, "learning_rate": 5e-05, "loss": 0.7983, "mean_token_accuracy": 0.8708059787750244, "num_tokens": 562131434.0, "step": 14809 }, { "epoch": 1.8839842259254547, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.218459129333496, "learning_rate": 5e-05, "loss": 0.9833, "mean_token_accuracy": 0.8600431084632874, "num_tokens": 562169914.0, "step": 14810 }, { "epoch": 1.8841114362040452, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.116672039031982, "learning_rate": 5e-05, "loss": 0.8324, "mean_token_accuracy": 0.8745844960212708, "num_tokens": 562207190.0, "step": 14811 }, { "epoch": 1.8842386464826357, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.139910697937012, "learning_rate": 5e-05, "loss": 0.9796, "mean_token_accuracy": 0.8649864196777344, "num_tokens": 562246236.0, "step": 14812 }, { "epoch": 1.8843658567612263, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.266098499298096, "learning_rate": 5e-05, "loss": 0.8438, "mean_token_accuracy": 0.8796539902687073, "num_tokens": 562286875.0, "step": 14813 }, { "epoch": 1.8844930670398168, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.953129768371582, "learning_rate": 5e-05, "loss": 0.9337, "mean_token_accuracy": 0.8753325939178467, "num_tokens": 562325807.0, "step": 14814 }, { "epoch": 1.8846202773184073, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.312953472137451, "learning_rate": 5e-05, "loss": 0.8591, "mean_token_accuracy": 0.872266411781311, "num_tokens": 562363663.0, "step": 14815 }, { "epoch": 1.8847474875969978, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.879975318908691, "learning_rate": 5e-05, "loss": 0.9485, "mean_token_accuracy": 0.8706004023551941, "num_tokens": 562399700.0, "step": 14816 }, { "epoch": 1.8848746978755884, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.211249828338623, "learning_rate": 5e-05, "loss": 0.8767, "mean_token_accuracy": 0.8624172210693359, "num_tokens": 562434395.0, "step": 14817 }, { "epoch": 1.885001908154179, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.000286102294922, "learning_rate": 5e-05, "loss": 1.0103, "mean_token_accuracy": 0.8550487756729126, "num_tokens": 562472140.0, "step": 14818 }, { "epoch": 1.8851291184327694, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.201533794403076, "learning_rate": 5e-05, "loss": 0.8516, "mean_token_accuracy": 0.8688692450523376, "num_tokens": 562510556.0, "step": 14819 }, { "epoch": 1.88525632871136, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.951473236083984, "learning_rate": 5e-05, "loss": 0.9911, "mean_token_accuracy": 0.8663017749786377, "num_tokens": 562546679.0, "step": 14820 }, { "epoch": 1.8853835389899505, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.470655918121338, "learning_rate": 5e-05, "loss": 0.8869, "mean_token_accuracy": 0.8678885698318481, "num_tokens": 562580707.0, "step": 14821 }, { "epoch": 1.885510749268541, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.595932960510254, "learning_rate": 5e-05, "loss": 0.9375, "mean_token_accuracy": 0.8703263998031616, "num_tokens": 562621753.0, "step": 14822 }, { "epoch": 1.8856379595471315, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.702613830566406, "learning_rate": 5e-05, "loss": 0.8162, "mean_token_accuracy": 0.8902072310447693, "num_tokens": 562654064.0, "step": 14823 }, { "epoch": 1.885765169825722, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.17153263092041, "learning_rate": 5e-05, "loss": 0.9277, "mean_token_accuracy": 0.8606632351875305, "num_tokens": 562692575.0, "step": 14824 }, { "epoch": 1.8858923801043126, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.769830703735352, "learning_rate": 5e-05, "loss": 0.8343, "mean_token_accuracy": 0.8813397288322449, "num_tokens": 562732655.0, "step": 14825 }, { "epoch": 1.886019590382903, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.898606777191162, "learning_rate": 5e-05, "loss": 0.8732, "mean_token_accuracy": 0.867890477180481, "num_tokens": 562768618.0, "step": 14826 }, { "epoch": 1.8861468006614934, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.703069686889648, "learning_rate": 5e-05, "loss": 0.8788, "mean_token_accuracy": 0.8563067317008972, "num_tokens": 562805357.0, "step": 14827 }, { "epoch": 1.886274010940084, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.84236478805542, "learning_rate": 5e-05, "loss": 0.8584, "mean_token_accuracy": 0.8642402291297913, "num_tokens": 562848798.0, "step": 14828 }, { "epoch": 1.8864012212186745, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.557519912719727, "learning_rate": 5e-05, "loss": 0.7664, "mean_token_accuracy": 0.8791854381561279, "num_tokens": 562890108.0, "step": 14829 }, { "epoch": 1.886528431497265, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.89970588684082, "learning_rate": 5e-05, "loss": 0.7851, "mean_token_accuracy": 0.8771649599075317, "num_tokens": 562921089.0, "step": 14830 }, { "epoch": 1.8866556417758555, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.436682224273682, "learning_rate": 5e-05, "loss": 0.8682, "mean_token_accuracy": 0.8387981653213501, "num_tokens": 562961705.0, "step": 14831 }, { "epoch": 1.8867828520544458, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.9350199699401855, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8850834965705872, "num_tokens": 563000011.0, "step": 14832 }, { "epoch": 1.8869100623330364, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.358764171600342, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.8742858171463013, "num_tokens": 563042568.0, "step": 14833 }, { "epoch": 1.887037272611627, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.963599681854248, "learning_rate": 5e-05, "loss": 0.7854, "mean_token_accuracy": 0.8784183263778687, "num_tokens": 563075892.0, "step": 14834 }, { "epoch": 1.8871644828902174, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.263211727142334, "learning_rate": 5e-05, "loss": 0.7738, "mean_token_accuracy": 0.8686503767967224, "num_tokens": 563117276.0, "step": 14835 }, { "epoch": 1.887291693168808, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.056381225585938, "learning_rate": 5e-05, "loss": 0.8054, "mean_token_accuracy": 0.8724623918533325, "num_tokens": 563152557.0, "step": 14836 }, { "epoch": 1.8874189034473985, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.14118766784668, "learning_rate": 5e-05, "loss": 0.7248, "mean_token_accuracy": 0.8732292056083679, "num_tokens": 563188620.0, "step": 14837 }, { "epoch": 1.887546113725989, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.189230918884277, "learning_rate": 5e-05, "loss": 0.8351, "mean_token_accuracy": 0.8662755489349365, "num_tokens": 563227582.0, "step": 14838 }, { "epoch": 1.8876733240045795, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.028253555297852, "learning_rate": 5e-05, "loss": 0.6818, "mean_token_accuracy": 0.8846321105957031, "num_tokens": 563260057.0, "step": 14839 }, { "epoch": 1.88780053428317, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.322734832763672, "learning_rate": 5e-05, "loss": 0.8284, "mean_token_accuracy": 0.8690255880355835, "num_tokens": 563299612.0, "step": 14840 }, { "epoch": 1.8879277445617606, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.931840419769287, "learning_rate": 5e-05, "loss": 0.7126, "mean_token_accuracy": 0.8721387386322021, "num_tokens": 563331659.0, "step": 14841 }, { "epoch": 1.8880549548403511, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.433928489685059, "learning_rate": 5e-05, "loss": 0.86, "mean_token_accuracy": 0.8650515675544739, "num_tokens": 563371411.0, "step": 14842 }, { "epoch": 1.8881821651189417, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.949045181274414, "learning_rate": 5e-05, "loss": 0.7271, "mean_token_accuracy": 0.8699175715446472, "num_tokens": 563401498.0, "step": 14843 }, { "epoch": 1.8883093753975322, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.371885299682617, "learning_rate": 5e-05, "loss": 0.7909, "mean_token_accuracy": 0.8869335651397705, "num_tokens": 563441383.0, "step": 14844 }, { "epoch": 1.8884365856761227, "ewc_loss": 0.375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.136268615722656, "learning_rate": 5e-05, "loss": 0.7728, "mean_token_accuracy": 0.8601214289665222, "num_tokens": 563478547.0, "step": 14845 }, { "epoch": 1.8885637959547132, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.086106300354004, "learning_rate": 5e-05, "loss": 0.8172, "mean_token_accuracy": 0.8714785575866699, "num_tokens": 563521366.0, "step": 14846 }, { "epoch": 1.8886910062333038, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.4647135734558105, "learning_rate": 5e-05, "loss": 0.821, "mean_token_accuracy": 0.8534353375434875, "num_tokens": 563560062.0, "step": 14847 }, { "epoch": 1.8888182165118943, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.60347843170166, "learning_rate": 5e-05, "loss": 0.7935, "mean_token_accuracy": 0.8619703054428101, "num_tokens": 563598448.0, "step": 14848 }, { "epoch": 1.8889454267904848, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.879944801330566, "learning_rate": 5e-05, "loss": 0.7753, "mean_token_accuracy": 0.8777309060096741, "num_tokens": 563639274.0, "step": 14849 }, { "epoch": 1.8890726370690751, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.0977983474731445, "learning_rate": 5e-05, "loss": 0.7472, "mean_token_accuracy": 0.8687736392021179, "num_tokens": 563679309.0, "step": 14850 }, { "epoch": 1.8891998473476657, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.362712860107422, "learning_rate": 5e-05, "loss": 0.8178, "mean_token_accuracy": 0.8698770999908447, "num_tokens": 563717594.0, "step": 14851 }, { "epoch": 1.8893270576262562, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.597548007965088, "learning_rate": 5e-05, "loss": 0.6874, "mean_token_accuracy": 0.8798631429672241, "num_tokens": 563749442.0, "step": 14852 }, { "epoch": 1.8894542679048467, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.924203872680664, "learning_rate": 5e-05, "loss": 0.8801, "mean_token_accuracy": 0.8650195598602295, "num_tokens": 563788519.0, "step": 14853 }, { "epoch": 1.8895814781834372, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.154452323913574, "learning_rate": 5e-05, "loss": 0.7125, "mean_token_accuracy": 0.8636127710342407, "num_tokens": 563827233.0, "step": 14854 }, { "epoch": 1.8897086884620278, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.600776672363281, "learning_rate": 5e-05, "loss": 0.9636, "mean_token_accuracy": 0.8679401874542236, "num_tokens": 563863534.0, "step": 14855 }, { "epoch": 1.889835898740618, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.134707927703857, "learning_rate": 5e-05, "loss": 0.766, "mean_token_accuracy": 0.8543547987937927, "num_tokens": 563903237.0, "step": 14856 }, { "epoch": 1.8899631090192086, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.872171401977539, "learning_rate": 5e-05, "loss": 0.9462, "mean_token_accuracy": 0.8929124474525452, "num_tokens": 563944334.0, "step": 14857 }, { "epoch": 1.8900903192977991, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.204395294189453, "learning_rate": 5e-05, "loss": 0.8905, "mean_token_accuracy": 0.8650832176208496, "num_tokens": 563979801.0, "step": 14858 }, { "epoch": 1.8902175295763897, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.863886833190918, "learning_rate": 5e-05, "loss": 0.9816, "mean_token_accuracy": 0.8642062544822693, "num_tokens": 564022348.0, "step": 14859 }, { "epoch": 1.8903447398549802, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.52076530456543, "learning_rate": 5e-05, "loss": 0.9426, "mean_token_accuracy": 0.8828871846199036, "num_tokens": 564063285.0, "step": 14860 }, { "epoch": 1.8904719501335707, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.034619331359863, "learning_rate": 5e-05, "loss": 0.7783, "mean_token_accuracy": 0.8704838752746582, "num_tokens": 564105323.0, "step": 14861 }, { "epoch": 1.8905991604121613, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.710691452026367, "learning_rate": 5e-05, "loss": 1.0369, "mean_token_accuracy": 0.8612889051437378, "num_tokens": 564140238.0, "step": 14862 }, { "epoch": 1.8907263706907518, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.367650508880615, "learning_rate": 5e-05, "loss": 0.8252, "mean_token_accuracy": 0.8434512615203857, "num_tokens": 564178441.0, "step": 14863 }, { "epoch": 1.8908535809693423, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.212465286254883, "learning_rate": 5e-05, "loss": 1.0666, "mean_token_accuracy": 0.8712535500526428, "num_tokens": 564217586.0, "step": 14864 }, { "epoch": 1.8909807912479328, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.482630729675293, "learning_rate": 5e-05, "loss": 0.9728, "mean_token_accuracy": 0.8605157136917114, "num_tokens": 564259198.0, "step": 14865 }, { "epoch": 1.8911080015265234, "ewc_loss": 0.625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.258920669555664, "learning_rate": 5e-05, "loss": 0.9571, "mean_token_accuracy": 0.8843116164207458, "num_tokens": 564298622.0, "step": 14866 }, { "epoch": 1.891235211805114, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.559085845947266, "learning_rate": 5e-05, "loss": 0.9824, "mean_token_accuracy": 0.8784435987472534, "num_tokens": 564331573.0, "step": 14867 }, { "epoch": 1.8913624220837044, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.848328590393066, "learning_rate": 5e-05, "loss": 0.8655, "mean_token_accuracy": 0.8649147748947144, "num_tokens": 564371461.0, "step": 14868 }, { "epoch": 1.891489632362295, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.882890701293945, "learning_rate": 5e-05, "loss": 0.923, "mean_token_accuracy": 0.8879626989364624, "num_tokens": 564410160.0, "step": 14869 }, { "epoch": 1.8916168426408855, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.525614261627197, "learning_rate": 5e-05, "loss": 0.8664, "mean_token_accuracy": 0.8586242198944092, "num_tokens": 564455954.0, "step": 14870 }, { "epoch": 1.891744052919476, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.371756553649902, "learning_rate": 5e-05, "loss": 0.8982, "mean_token_accuracy": 0.8700288534164429, "num_tokens": 564490196.0, "step": 14871 }, { "epoch": 1.8918712631980665, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.912853717803955, "learning_rate": 5e-05, "loss": 0.8647, "mean_token_accuracy": 0.8753343820571899, "num_tokens": 564530362.0, "step": 14872 }, { "epoch": 1.891998473476657, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.551026344299316, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.868690013885498, "num_tokens": 564563528.0, "step": 14873 }, { "epoch": 1.8921256837552476, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.331363677978516, "learning_rate": 5e-05, "loss": 0.8842, "mean_token_accuracy": 0.872718095779419, "num_tokens": 564596865.0, "step": 14874 }, { "epoch": 1.892252894033838, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.101704120635986, "learning_rate": 5e-05, "loss": 0.7712, "mean_token_accuracy": 0.8650974631309509, "num_tokens": 564635740.0, "step": 14875 }, { "epoch": 1.8923801043124284, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.5698823928833, "learning_rate": 5e-05, "loss": 0.9044, "mean_token_accuracy": 0.8597952127456665, "num_tokens": 564679007.0, "step": 14876 }, { "epoch": 1.892507314591019, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.889247417449951, "learning_rate": 5e-05, "loss": 0.7309, "mean_token_accuracy": 0.8739131689071655, "num_tokens": 564719211.0, "step": 14877 }, { "epoch": 1.8926345248696095, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.850519180297852, "learning_rate": 5e-05, "loss": 0.8542, "mean_token_accuracy": 0.8851639032363892, "num_tokens": 564756482.0, "step": 14878 }, { "epoch": 1.8927617351482, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.743544578552246, "learning_rate": 5e-05, "loss": 0.7582, "mean_token_accuracy": 0.8719514608383179, "num_tokens": 564796304.0, "step": 14879 }, { "epoch": 1.8928889454267905, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.086882591247559, "learning_rate": 5e-05, "loss": 0.95, "mean_token_accuracy": 0.866987943649292, "num_tokens": 564835444.0, "step": 14880 }, { "epoch": 1.8930161557053808, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.778630256652832, "learning_rate": 5e-05, "loss": 0.8193, "mean_token_accuracy": 0.8556175231933594, "num_tokens": 564873637.0, "step": 14881 }, { "epoch": 1.8931433659839714, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.153377532958984, "learning_rate": 5e-05, "loss": 0.9435, "mean_token_accuracy": 0.8700988292694092, "num_tokens": 564914742.0, "step": 14882 }, { "epoch": 1.893270576262562, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 6.994657039642334, "learning_rate": 5e-05, "loss": 0.8163, "mean_token_accuracy": 0.8715512752532959, "num_tokens": 564953519.0, "step": 14883 }, { "epoch": 1.8933977865411524, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.064045906066895, "learning_rate": 5e-05, "loss": 0.974, "mean_token_accuracy": 0.8631464242935181, "num_tokens": 564984994.0, "step": 14884 }, { "epoch": 1.893524996819743, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.207827568054199, "learning_rate": 5e-05, "loss": 0.8841, "mean_token_accuracy": 0.8610645532608032, "num_tokens": 565028584.0, "step": 14885 }, { "epoch": 1.8936522070983335, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.933271408081055, "learning_rate": 5e-05, "loss": 0.9738, "mean_token_accuracy": 0.869009256362915, "num_tokens": 565068726.0, "step": 14886 }, { "epoch": 1.893779417376924, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.398290157318115, "learning_rate": 5e-05, "loss": 0.9058, "mean_token_accuracy": 0.861419141292572, "num_tokens": 565112156.0, "step": 14887 }, { "epoch": 1.8939066276555145, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.653453826904297, "learning_rate": 5e-05, "loss": 0.9308, "mean_token_accuracy": 0.8747470378875732, "num_tokens": 565146158.0, "step": 14888 }, { "epoch": 1.894033837934105, "ewc_loss": 0.5, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.562557697296143, "learning_rate": 5e-05, "loss": 0.8326, "mean_token_accuracy": 0.8819077014923096, "num_tokens": 565176571.0, "step": 14889 }, { "epoch": 1.8941610482126956, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.332413673400879, "learning_rate": 5e-05, "loss": 0.9114, "mean_token_accuracy": 0.8690360188484192, "num_tokens": 565216162.0, "step": 14890 }, { "epoch": 1.8942882584912861, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.591341972351074, "learning_rate": 5e-05, "loss": 0.8094, "mean_token_accuracy": 0.88480544090271, "num_tokens": 565252010.0, "step": 14891 }, { "epoch": 1.8944154687698767, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.133569717407227, "learning_rate": 5e-05, "loss": 0.9159, "mean_token_accuracy": 0.8555705547332764, "num_tokens": 565297159.0, "step": 14892 }, { "epoch": 1.8945426790484672, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.5160980224609375, "learning_rate": 5e-05, "loss": 0.815, "mean_token_accuracy": 0.8749960660934448, "num_tokens": 565337074.0, "step": 14893 }, { "epoch": 1.8946698893270577, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.046758651733398, "learning_rate": 5e-05, "loss": 0.8827, "mean_token_accuracy": 0.8644348382949829, "num_tokens": 565379551.0, "step": 14894 }, { "epoch": 1.8947970996056482, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.491623878479004, "learning_rate": 5e-05, "loss": 0.7965, "mean_token_accuracy": 0.8744429349899292, "num_tokens": 565418611.0, "step": 14895 }, { "epoch": 1.8949243098842388, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.947091579437256, "learning_rate": 5e-05, "loss": 0.8261, "mean_token_accuracy": 0.8710935115814209, "num_tokens": 565452804.0, "step": 14896 }, { "epoch": 1.8950515201628293, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.4696574211120605, "learning_rate": 5e-05, "loss": 0.8008, "mean_token_accuracy": 0.8662474155426025, "num_tokens": 565490628.0, "step": 14897 }, { "epoch": 1.8951787304414198, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.888349533081055, "learning_rate": 5e-05, "loss": 0.8274, "mean_token_accuracy": 0.8641860485076904, "num_tokens": 565523990.0, "step": 14898 }, { "epoch": 1.8953059407200101, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.439529895782471, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.8772254586219788, "num_tokens": 565562684.0, "step": 14899 }, { "epoch": 1.8954331509986007, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.845485210418701, "learning_rate": 5e-05, "loss": 0.7286, "mean_token_accuracy": 0.890743613243103, "num_tokens": 565596479.0, "step": 14900 }, { "epoch": 1.8955603612771912, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.430796146392822, "learning_rate": 5e-05, "loss": 0.7595, "mean_token_accuracy": 0.8682231903076172, "num_tokens": 565639112.0, "step": 14901 }, { "epoch": 1.8956875715557817, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.808070182800293, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8798171281814575, "num_tokens": 565671379.0, "step": 14902 }, { "epoch": 1.8958147818343722, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.393723964691162, "learning_rate": 5e-05, "loss": 0.7349, "mean_token_accuracy": 0.8781725764274597, "num_tokens": 565708115.0, "step": 14903 }, { "epoch": 1.8959419921129628, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.8300957679748535, "learning_rate": 5e-05, "loss": 0.7965, "mean_token_accuracy": 0.8640682697296143, "num_tokens": 565744767.0, "step": 14904 }, { "epoch": 1.896069202391553, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.315052032470703, "learning_rate": 5e-05, "loss": 0.7399, "mean_token_accuracy": 0.869945764541626, "num_tokens": 565783709.0, "step": 14905 }, { "epoch": 1.8961964126701436, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.93088960647583, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.8753436803817749, "num_tokens": 565818503.0, "step": 14906 }, { "epoch": 1.8963236229487341, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.171441555023193, "learning_rate": 5e-05, "loss": 0.6652, "mean_token_accuracy": 0.8895039558410645, "num_tokens": 565850188.0, "step": 14907 }, { "epoch": 1.8964508332273247, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.090226173400879, "learning_rate": 5e-05, "loss": 0.809, "mean_token_accuracy": 0.8641963005065918, "num_tokens": 565889177.0, "step": 14908 }, { "epoch": 1.8965780435059152, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.997267723083496, "learning_rate": 5e-05, "loss": 0.7888, "mean_token_accuracy": 0.8491177558898926, "num_tokens": 565929316.0, "step": 14909 }, { "epoch": 1.8967052537845057, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.281708717346191, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.887326717376709, "num_tokens": 565961874.0, "step": 14910 }, { "epoch": 1.8968324640630962, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.831711769104004, "learning_rate": 5e-05, "loss": 0.7568, "mean_token_accuracy": 0.8564156293869019, "num_tokens": 565995143.0, "step": 14911 }, { "epoch": 1.8969596743416868, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.51830768585205, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.8713765144348145, "num_tokens": 566032297.0, "step": 14912 }, { "epoch": 1.8970868846202773, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.694398880004883, "learning_rate": 5e-05, "loss": 0.6973, "mean_token_accuracy": 0.8747063875198364, "num_tokens": 566073138.0, "step": 14913 }, { "epoch": 1.8972140948988678, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.693466186523438, "learning_rate": 5e-05, "loss": 0.8627, "mean_token_accuracy": 0.873639702796936, "num_tokens": 566115471.0, "step": 14914 }, { "epoch": 1.8973413051774584, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.715446949005127, "learning_rate": 5e-05, "loss": 0.7286, "mean_token_accuracy": 0.8673112988471985, "num_tokens": 566152644.0, "step": 14915 }, { "epoch": 1.8974685154560489, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.646410942077637, "learning_rate": 5e-05, "loss": 0.8785, "mean_token_accuracy": 0.8686716556549072, "num_tokens": 566197411.0, "step": 14916 }, { "epoch": 1.8975957257346394, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.0485758781433105, "learning_rate": 5e-05, "loss": 0.6825, "mean_token_accuracy": 0.89459228515625, "num_tokens": 566234738.0, "step": 14917 }, { "epoch": 1.89772293601323, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.116982460021973, "learning_rate": 5e-05, "loss": 0.8904, "mean_token_accuracy": 0.852954626083374, "num_tokens": 566274445.0, "step": 14918 }, { "epoch": 1.8978501462918205, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.65972375869751, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.8904772996902466, "num_tokens": 566312710.0, "step": 14919 }, { "epoch": 1.897977356570411, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.286213397979736, "learning_rate": 5e-05, "loss": 0.7712, "mean_token_accuracy": 0.8669090867042542, "num_tokens": 566349229.0, "step": 14920 }, { "epoch": 1.8981045668490015, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.398364067077637, "learning_rate": 5e-05, "loss": 0.8769, "mean_token_accuracy": 0.863010048866272, "num_tokens": 566387208.0, "step": 14921 }, { "epoch": 1.898231777127592, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.479826927185059, "learning_rate": 5e-05, "loss": 0.696, "mean_token_accuracy": 0.8709733486175537, "num_tokens": 566423616.0, "step": 14922 }, { "epoch": 1.8983589874061826, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 9.239349365234375, "learning_rate": 5e-05, "loss": 0.8953, "mean_token_accuracy": 0.8739245533943176, "num_tokens": 566457984.0, "step": 14923 }, { "epoch": 1.898486197684773, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.917732238769531, "learning_rate": 5e-05, "loss": 0.7007, "mean_token_accuracy": 0.8660496473312378, "num_tokens": 566495664.0, "step": 14924 }, { "epoch": 1.8986134079633634, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 10.098135948181152, "learning_rate": 5e-05, "loss": 1.0338, "mean_token_accuracy": 0.8635464906692505, "num_tokens": 566533750.0, "step": 14925 }, { "epoch": 1.898740618241954, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.329323768615723, "learning_rate": 5e-05, "loss": 0.7861, "mean_token_accuracy": 0.862956166267395, "num_tokens": 566576247.0, "step": 14926 }, { "epoch": 1.8988678285205445, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.212873458862305, "learning_rate": 5e-05, "loss": 1.1416, "mean_token_accuracy": 0.853489875793457, "num_tokens": 566618185.0, "step": 14927 }, { "epoch": 1.898995038799135, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.020600318908691, "learning_rate": 5e-05, "loss": 1.04, "mean_token_accuracy": 0.8573969602584839, "num_tokens": 566659219.0, "step": 14928 }, { "epoch": 1.8991222490777255, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.62381649017334, "learning_rate": 5e-05, "loss": 0.9263, "mean_token_accuracy": 0.8763853907585144, "num_tokens": 566695101.0, "step": 14929 }, { "epoch": 1.8992494593563158, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.269399642944336, "learning_rate": 5e-05, "loss": 1.0685, "mean_token_accuracy": 0.8691426515579224, "num_tokens": 566735286.0, "step": 14930 }, { "epoch": 1.8993766696349064, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.778877258300781, "learning_rate": 5e-05, "loss": 0.7712, "mean_token_accuracy": 0.8712167739868164, "num_tokens": 566773534.0, "step": 14931 }, { "epoch": 1.899503879913497, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.474266052246094, "learning_rate": 5e-05, "loss": 1.0939, "mean_token_accuracy": 0.8768557906150818, "num_tokens": 566819141.0, "step": 14932 }, { "epoch": 1.8996310901920874, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.1332550048828125, "learning_rate": 5e-05, "loss": 0.8498, "mean_token_accuracy": 0.8858370780944824, "num_tokens": 566855095.0, "step": 14933 }, { "epoch": 1.899758300470678, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.393885612487793, "learning_rate": 5e-05, "loss": 1.0977, "mean_token_accuracy": 0.8638309240341187, "num_tokens": 566888989.0, "step": 14934 }, { "epoch": 1.8998855107492685, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 8.525556564331055, "learning_rate": 5e-05, "loss": 1.0544, "mean_token_accuracy": 0.8827559947967529, "num_tokens": 566923049.0, "step": 14935 }, { "epoch": 1.900012721027859, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.9679536819458, "learning_rate": 5e-05, "loss": 1.0208, "mean_token_accuracy": 0.851470947265625, "num_tokens": 566962049.0, "step": 14936 }, { "epoch": 1.9001399313064495, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 8.900402069091797, "learning_rate": 5e-05, "loss": 0.9902, "mean_token_accuracy": 0.8846192359924316, "num_tokens": 567003538.0, "step": 14937 }, { "epoch": 1.90026714158504, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.023824691772461, "learning_rate": 5e-05, "loss": 0.8542, "mean_token_accuracy": 0.8812218904495239, "num_tokens": 567035821.0, "step": 14938 }, { "epoch": 1.9003943518636306, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.659964561462402, "learning_rate": 5e-05, "loss": 0.9308, "mean_token_accuracy": 0.8799660205841064, "num_tokens": 567068095.0, "step": 14939 }, { "epoch": 1.9005215621422211, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.001407623291016, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.875043511390686, "num_tokens": 567104741.0, "step": 14940 }, { "epoch": 1.9006487724208116, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.874689102172852, "learning_rate": 5e-05, "loss": 0.8342, "mean_token_accuracy": 0.8674019575119019, "num_tokens": 567142500.0, "step": 14941 }, { "epoch": 1.9007759826994022, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.358259201049805, "learning_rate": 5e-05, "loss": 0.9194, "mean_token_accuracy": 0.8696824312210083, "num_tokens": 567182001.0, "step": 14942 }, { "epoch": 1.9009031929779927, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.160966396331787, "learning_rate": 5e-05, "loss": 0.7728, "mean_token_accuracy": 0.8685904741287231, "num_tokens": 567219809.0, "step": 14943 }, { "epoch": 1.9010304032565832, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.823417663574219, "learning_rate": 5e-05, "loss": 1.0003, "mean_token_accuracy": 0.8436019420623779, "num_tokens": 567260309.0, "step": 14944 }, { "epoch": 1.9011576135351738, "ewc_loss": 0.375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.802567958831787, "learning_rate": 5e-05, "loss": 0.7819, "mean_token_accuracy": 0.8584800958633423, "num_tokens": 567295840.0, "step": 14945 }, { "epoch": 1.9012848238137643, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.17013931274414, "learning_rate": 5e-05, "loss": 1.0226, "mean_token_accuracy": 0.8493605852127075, "num_tokens": 567335546.0, "step": 14946 }, { "epoch": 1.9014120340923548, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.804783821105957, "learning_rate": 5e-05, "loss": 0.7385, "mean_token_accuracy": 0.8873183727264404, "num_tokens": 567372996.0, "step": 14947 }, { "epoch": 1.9015392443709451, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.293303489685059, "learning_rate": 5e-05, "loss": 0.9401, "mean_token_accuracy": 0.8791079521179199, "num_tokens": 567411758.0, "step": 14948 }, { "epoch": 1.9016664546495357, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.068098068237305, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8676633834838867, "num_tokens": 567446859.0, "step": 14949 }, { "epoch": 1.9017936649281262, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.089463233947754, "learning_rate": 5e-05, "loss": 0.9331, "mean_token_accuracy": 0.8807634115219116, "num_tokens": 567480890.0, "step": 14950 }, { "epoch": 1.9019208752067167, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.350273132324219, "learning_rate": 5e-05, "loss": 0.8616, "mean_token_accuracy": 0.8761789202690125, "num_tokens": 567523381.0, "step": 14951 }, { "epoch": 1.9020480854853072, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.765689849853516, "learning_rate": 5e-05, "loss": 0.9569, "mean_token_accuracy": 0.8707402348518372, "num_tokens": 567564926.0, "step": 14952 }, { "epoch": 1.9021752957638978, "ewc_loss": 0.5, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.402515411376953, "learning_rate": 5e-05, "loss": 0.8778, "mean_token_accuracy": 0.8689587116241455, "num_tokens": 567603351.0, "step": 14953 }, { "epoch": 1.902302506042488, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.634809494018555, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.8839772343635559, "num_tokens": 567636710.0, "step": 14954 }, { "epoch": 1.9024297163210786, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.2993011474609375, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.8846098184585571, "num_tokens": 567676325.0, "step": 14955 }, { "epoch": 1.9025569265996691, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.620589256286621, "learning_rate": 5e-05, "loss": 0.8658, "mean_token_accuracy": 0.8908374905586243, "num_tokens": 567720663.0, "step": 14956 }, { "epoch": 1.9026841368782597, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.2613091468811035, "learning_rate": 5e-05, "loss": 0.8644, "mean_token_accuracy": 0.8625274300575256, "num_tokens": 567761589.0, "step": 14957 }, { "epoch": 1.9028113471568502, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.542655944824219, "learning_rate": 5e-05, "loss": 0.8781, "mean_token_accuracy": 0.8822412490844727, "num_tokens": 567801649.0, "step": 14958 }, { "epoch": 1.9029385574354407, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.29951810836792, "learning_rate": 5e-05, "loss": 0.7977, "mean_token_accuracy": 0.8801181316375732, "num_tokens": 567839953.0, "step": 14959 }, { "epoch": 1.9030657677140312, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.433086395263672, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8793195486068726, "num_tokens": 567877634.0, "step": 14960 }, { "epoch": 1.9031929779926218, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.3830389976501465, "learning_rate": 5e-05, "loss": 0.8758, "mean_token_accuracy": 0.8562024235725403, "num_tokens": 567911504.0, "step": 14961 }, { "epoch": 1.9033201882712123, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.211613655090332, "learning_rate": 5e-05, "loss": 0.8734, "mean_token_accuracy": 0.8763089179992676, "num_tokens": 567954904.0, "step": 14962 }, { "epoch": 1.9034473985498028, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.518863677978516, "learning_rate": 5e-05, "loss": 0.858, "mean_token_accuracy": 0.8645785450935364, "num_tokens": 567992384.0, "step": 14963 }, { "epoch": 1.9035746088283934, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.957007884979248, "learning_rate": 5e-05, "loss": 0.8794, "mean_token_accuracy": 0.8661783933639526, "num_tokens": 568027692.0, "step": 14964 }, { "epoch": 1.9037018191069839, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.555881500244141, "learning_rate": 5e-05, "loss": 0.8236, "mean_token_accuracy": 0.8640075325965881, "num_tokens": 568063385.0, "step": 14965 }, { "epoch": 1.9038290293855744, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.857297897338867, "learning_rate": 5e-05, "loss": 0.7909, "mean_token_accuracy": 0.8799546957015991, "num_tokens": 568098315.0, "step": 14966 }, { "epoch": 1.903956239664165, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.509440898895264, "learning_rate": 5e-05, "loss": 0.8272, "mean_token_accuracy": 0.860476016998291, "num_tokens": 568144032.0, "step": 14967 }, { "epoch": 1.9040834499427555, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.826234817504883, "learning_rate": 5e-05, "loss": 0.8125, "mean_token_accuracy": 0.8675824403762817, "num_tokens": 568174988.0, "step": 14968 }, { "epoch": 1.904210660221346, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.469837665557861, "learning_rate": 5e-05, "loss": 0.8151, "mean_token_accuracy": 0.8585017323493958, "num_tokens": 568219794.0, "step": 14969 }, { "epoch": 1.9043378704999365, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.796144485473633, "learning_rate": 5e-05, "loss": 0.7721, "mean_token_accuracy": 0.8778234124183655, "num_tokens": 568264995.0, "step": 14970 }, { "epoch": 1.904465080778527, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.44032096862793, "learning_rate": 5e-05, "loss": 0.7412, "mean_token_accuracy": 0.8744657635688782, "num_tokens": 568300876.0, "step": 14971 }, { "epoch": 1.9045922910571176, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.806279182434082, "learning_rate": 5e-05, "loss": 0.7992, "mean_token_accuracy": 0.869541585445404, "num_tokens": 568337341.0, "step": 14972 }, { "epoch": 1.9047195013357079, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.369115829467773, "learning_rate": 5e-05, "loss": 0.7547, "mean_token_accuracy": 0.8666660189628601, "num_tokens": 568373524.0, "step": 14973 }, { "epoch": 1.9048467116142984, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.860837936401367, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.8852423429489136, "num_tokens": 568412575.0, "step": 14974 }, { "epoch": 1.904973921892889, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.279487133026123, "learning_rate": 5e-05, "loss": 0.7268, "mean_token_accuracy": 0.8763285875320435, "num_tokens": 568446794.0, "step": 14975 }, { "epoch": 1.9051011321714795, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.914172649383545, "learning_rate": 5e-05, "loss": 0.7981, "mean_token_accuracy": 0.86832594871521, "num_tokens": 568490645.0, "step": 14976 }, { "epoch": 1.90522834245007, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.1840972900390625, "learning_rate": 5e-05, "loss": 0.733, "mean_token_accuracy": 0.8723006844520569, "num_tokens": 568529281.0, "step": 14977 }, { "epoch": 1.9053555527286605, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.046905517578125, "learning_rate": 5e-05, "loss": 0.7673, "mean_token_accuracy": 0.8795990943908691, "num_tokens": 568568722.0, "step": 14978 }, { "epoch": 1.9054827630072508, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.034651279449463, "learning_rate": 5e-05, "loss": 0.704, "mean_token_accuracy": 0.876482367515564, "num_tokens": 568604736.0, "step": 14979 }, { "epoch": 1.9056099732858414, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.213617324829102, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.8852479457855225, "num_tokens": 568643720.0, "step": 14980 }, { "epoch": 1.905737183564432, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.874603271484375, "learning_rate": 5e-05, "loss": 0.6805, "mean_token_accuracy": 0.8819013833999634, "num_tokens": 568681805.0, "step": 14981 }, { "epoch": 1.9058643938430224, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.404279708862305, "learning_rate": 5e-05, "loss": 0.8339, "mean_token_accuracy": 0.8664537072181702, "num_tokens": 568723107.0, "step": 14982 }, { "epoch": 1.905991604121613, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.756181240081787, "learning_rate": 5e-05, "loss": 0.7764, "mean_token_accuracy": 0.8522522449493408, "num_tokens": 568760229.0, "step": 14983 }, { "epoch": 1.9061188144002035, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.562429428100586, "learning_rate": 5e-05, "loss": 0.8896, "mean_token_accuracy": 0.860640287399292, "num_tokens": 568794539.0, "step": 14984 }, { "epoch": 1.906246024678794, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.735481262207031, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.86319500207901, "num_tokens": 568834808.0, "step": 14985 }, { "epoch": 1.9063732349573845, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.548494338989258, "learning_rate": 5e-05, "loss": 0.8608, "mean_token_accuracy": 0.8713719844818115, "num_tokens": 568875004.0, "step": 14986 }, { "epoch": 1.906500445235975, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.9788360595703125, "learning_rate": 5e-05, "loss": 0.7334, "mean_token_accuracy": 0.8742733597755432, "num_tokens": 568917372.0, "step": 14987 }, { "epoch": 1.9066276555145656, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.172297477722168, "learning_rate": 5e-05, "loss": 0.8442, "mean_token_accuracy": 0.8651452660560608, "num_tokens": 568953567.0, "step": 14988 }, { "epoch": 1.9067548657931561, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.4768266677856445, "learning_rate": 5e-05, "loss": 0.8126, "mean_token_accuracy": 0.8604632616043091, "num_tokens": 568988634.0, "step": 14989 }, { "epoch": 1.9068820760717466, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.445295810699463, "learning_rate": 5e-05, "loss": 0.7372, "mean_token_accuracy": 0.876273512840271, "num_tokens": 569023929.0, "step": 14990 }, { "epoch": 1.9070092863503372, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.175466537475586, "learning_rate": 5e-05, "loss": 0.7939, "mean_token_accuracy": 0.8786011934280396, "num_tokens": 569054483.0, "step": 14991 }, { "epoch": 1.9071364966289277, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.634302139282227, "learning_rate": 5e-05, "loss": 0.7327, "mean_token_accuracy": 0.8609737157821655, "num_tokens": 569096919.0, "step": 14992 }, { "epoch": 1.9072637069075182, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 9.024667739868164, "learning_rate": 5e-05, "loss": 0.9034, "mean_token_accuracy": 0.8668129444122314, "num_tokens": 569134161.0, "step": 14993 }, { "epoch": 1.9073909171861088, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00023174285888671875, "grad_norm": 5.916439056396484, "learning_rate": 5e-05, "loss": 0.6529, "mean_token_accuracy": 0.8710131645202637, "num_tokens": 569173391.0, "step": 14994 }, { "epoch": 1.9075181274646993, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 10.102974891662598, "learning_rate": 5e-05, "loss": 1.0727, "mean_token_accuracy": 0.8506807088851929, "num_tokens": 569206173.0, "step": 14995 }, { "epoch": 1.9076453377432898, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 5.903565883636475, "learning_rate": 5e-05, "loss": 0.7282, "mean_token_accuracy": 0.8633748888969421, "num_tokens": 569241692.0, "step": 14996 }, { "epoch": 1.9077725480218801, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.810445785522461, "learning_rate": 5e-05, "loss": 1.1331, "mean_token_accuracy": 0.8783142566680908, "num_tokens": 569276361.0, "step": 14997 }, { "epoch": 1.9078997583004706, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 7.789438724517822, "learning_rate": 5e-05, "loss": 1.0845, "mean_token_accuracy": 0.8456225395202637, "num_tokens": 569314478.0, "step": 14998 }, { "epoch": 1.9080269685790612, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.696194648742676, "learning_rate": 5e-05, "loss": 1.009, "mean_token_accuracy": 0.8730458617210388, "num_tokens": 569354867.0, "step": 14999 }, { "epoch": 1.9081541788576517, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 9.254327774047852, "learning_rate": 5e-05, "loss": 1.151, "mean_token_accuracy": 0.8788127899169922, "num_tokens": 569394031.0, "step": 15000 }, { "epoch": 1.9082813891362422, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.414649486541748, "learning_rate": 5e-05, "loss": 0.7323, "mean_token_accuracy": 0.8965216875076294, "num_tokens": 569430899.0, "step": 15001 }, { "epoch": 1.9084085994148328, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 10.828495979309082, "learning_rate": 5e-05, "loss": 1.2388, "mean_token_accuracy": 0.8510645627975464, "num_tokens": 569469871.0, "step": 15002 }, { "epoch": 1.908535809693423, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.236964225769043, "learning_rate": 5e-05, "loss": 0.9005, "mean_token_accuracy": 0.8688515424728394, "num_tokens": 569508385.0, "step": 15003 }, { "epoch": 1.9086630199720136, "ewc_loss": 0.859375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 11.784520149230957, "learning_rate": 5e-05, "loss": 1.233, "mean_token_accuracy": 0.8728054165840149, "num_tokens": 569549705.0, "step": 15004 }, { "epoch": 1.9087902302506041, "ewc_loss": 0.85546875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 9.221881866455078, "learning_rate": 5e-05, "loss": 1.264, "mean_token_accuracy": 0.859822690486908, "num_tokens": 569584092.0, "step": 15005 }, { "epoch": 1.9089174405291947, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.861554145812988, "learning_rate": 5e-05, "loss": 1.0588, "mean_token_accuracy": 0.8689087629318237, "num_tokens": 569619683.0, "step": 15006 }, { "epoch": 1.9090446508077852, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 9.054560661315918, "learning_rate": 5e-05, "loss": 1.1229, "mean_token_accuracy": 0.8775718808174133, "num_tokens": 569656347.0, "step": 15007 }, { "epoch": 1.9091718610863757, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.29904842376709, "learning_rate": 5e-05, "loss": 1.0524, "mean_token_accuracy": 0.8777520656585693, "num_tokens": 569689633.0, "step": 15008 }, { "epoch": 1.9092990713649662, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.343648910522461, "learning_rate": 5e-05, "loss": 1.0126, "mean_token_accuracy": 0.8586684465408325, "num_tokens": 569730685.0, "step": 15009 }, { "epoch": 1.9094262816435568, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.067669868469238, "learning_rate": 5e-05, "loss": 1.113, "mean_token_accuracy": 0.8591769933700562, "num_tokens": 569773343.0, "step": 15010 }, { "epoch": 1.9095534919221473, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.8585028648376465, "learning_rate": 5e-05, "loss": 0.9144, "mean_token_accuracy": 0.8689368367195129, "num_tokens": 569814417.0, "step": 15011 }, { "epoch": 1.9096807022007378, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.361328125, "learning_rate": 5e-05, "loss": 1.0753, "mean_token_accuracy": 0.8783395886421204, "num_tokens": 569854436.0, "step": 15012 }, { "epoch": 1.9098079124793284, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.000014305114746, "learning_rate": 5e-05, "loss": 1.0034, "mean_token_accuracy": 0.853308916091919, "num_tokens": 569897530.0, "step": 15013 }, { "epoch": 1.9099351227579189, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.54182243347168, "learning_rate": 5e-05, "loss": 1.0888, "mean_token_accuracy": 0.8759856820106506, "num_tokens": 569931927.0, "step": 15014 }, { "epoch": 1.9100623330365094, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.515387535095215, "learning_rate": 5e-05, "loss": 0.9915, "mean_token_accuracy": 0.8723388910293579, "num_tokens": 569968283.0, "step": 15015 }, { "epoch": 1.9101895433151, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.158788681030273, "learning_rate": 5e-05, "loss": 1.072, "mean_token_accuracy": 0.8671870231628418, "num_tokens": 570009073.0, "step": 15016 }, { "epoch": 1.9103167535936905, "ewc_loss": 0.625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.6969633102417, "learning_rate": 5e-05, "loss": 0.9985, "mean_token_accuracy": 0.8699787855148315, "num_tokens": 570046421.0, "step": 15017 }, { "epoch": 1.910443963872281, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.587242126464844, "learning_rate": 5e-05, "loss": 1.0327, "mean_token_accuracy": 0.8662529587745667, "num_tokens": 570093679.0, "step": 15018 }, { "epoch": 1.9105711741508715, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.500711441040039, "learning_rate": 5e-05, "loss": 0.9206, "mean_token_accuracy": 0.8780971169471741, "num_tokens": 570132575.0, "step": 15019 }, { "epoch": 1.910698384429462, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.15322208404541, "learning_rate": 5e-05, "loss": 0.9836, "mean_token_accuracy": 0.8675615787506104, "num_tokens": 570169782.0, "step": 15020 }, { "epoch": 1.9108255947080524, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.088689804077148, "learning_rate": 5e-05, "loss": 0.8693, "mean_token_accuracy": 0.876410961151123, "num_tokens": 570205034.0, "step": 15021 }, { "epoch": 1.9109528049866429, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.182695388793945, "learning_rate": 5e-05, "loss": 0.9657, "mean_token_accuracy": 0.8732255697250366, "num_tokens": 570236804.0, "step": 15022 }, { "epoch": 1.9110800152652334, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.646889686584473, "learning_rate": 5e-05, "loss": 0.8175, "mean_token_accuracy": 0.8715178966522217, "num_tokens": 570266893.0, "step": 15023 }, { "epoch": 1.911207225543824, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.558784484863281, "learning_rate": 5e-05, "loss": 0.973, "mean_token_accuracy": 0.8828039169311523, "num_tokens": 570297779.0, "step": 15024 }, { "epoch": 1.9113344358224145, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.685389995574951, "learning_rate": 5e-05, "loss": 0.8469, "mean_token_accuracy": 0.8765048980712891, "num_tokens": 570334774.0, "step": 15025 }, { "epoch": 1.911461646101005, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.62192153930664, "learning_rate": 5e-05, "loss": 1.0373, "mean_token_accuracy": 0.8664265275001526, "num_tokens": 570373788.0, "step": 15026 }, { "epoch": 1.9115888563795955, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.152480125427246, "learning_rate": 5e-05, "loss": 0.8929, "mean_token_accuracy": 0.8763507604598999, "num_tokens": 570412156.0, "step": 15027 }, { "epoch": 1.9117160666581858, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.076318740844727, "learning_rate": 5e-05, "loss": 0.9724, "mean_token_accuracy": 0.8766400218009949, "num_tokens": 570442702.0, "step": 15028 }, { "epoch": 1.9118432769367764, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.303976058959961, "learning_rate": 5e-05, "loss": 0.9502, "mean_token_accuracy": 0.8595643043518066, "num_tokens": 570485302.0, "step": 15029 }, { "epoch": 1.9119704872153669, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.467105865478516, "learning_rate": 5e-05, "loss": 0.9068, "mean_token_accuracy": 0.8737791776657104, "num_tokens": 570524820.0, "step": 15030 }, { "epoch": 1.9120976974939574, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.9950079917907715, "learning_rate": 5e-05, "loss": 0.8904, "mean_token_accuracy": 0.8595151901245117, "num_tokens": 570567973.0, "step": 15031 }, { "epoch": 1.912224907772548, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.354302406311035, "learning_rate": 5e-05, "loss": 0.8756, "mean_token_accuracy": 0.8755061626434326, "num_tokens": 570610369.0, "step": 15032 }, { "epoch": 1.9123521180511385, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.531142234802246, "learning_rate": 5e-05, "loss": 0.8151, "mean_token_accuracy": 0.866899311542511, "num_tokens": 570653226.0, "step": 15033 }, { "epoch": 1.912479328329729, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.639954566955566, "learning_rate": 5e-05, "loss": 0.9297, "mean_token_accuracy": 0.8681542873382568, "num_tokens": 570691505.0, "step": 15034 }, { "epoch": 1.9126065386083195, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.194329261779785, "learning_rate": 5e-05, "loss": 0.7884, "mean_token_accuracy": 0.8636298775672913, "num_tokens": 570733781.0, "step": 15035 }, { "epoch": 1.91273374888691, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.923039436340332, "learning_rate": 5e-05, "loss": 0.9296, "mean_token_accuracy": 0.8764740228652954, "num_tokens": 570772675.0, "step": 15036 }, { "epoch": 1.9128609591655006, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.255863189697266, "learning_rate": 5e-05, "loss": 0.7347, "mean_token_accuracy": 0.8851076364517212, "num_tokens": 570813967.0, "step": 15037 }, { "epoch": 1.9129881694440911, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.854096412658691, "learning_rate": 5e-05, "loss": 0.9144, "mean_token_accuracy": 0.8814237117767334, "num_tokens": 570850315.0, "step": 15038 }, { "epoch": 1.9131153797226816, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.568861961364746, "learning_rate": 5e-05, "loss": 0.8323, "mean_token_accuracy": 0.8649216890335083, "num_tokens": 570890191.0, "step": 15039 }, { "epoch": 1.9132425900012722, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.49725341796875, "learning_rate": 5e-05, "loss": 0.8947, "mean_token_accuracy": 0.8765454292297363, "num_tokens": 570926564.0, "step": 15040 }, { "epoch": 1.9133698002798627, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.772361755371094, "learning_rate": 5e-05, "loss": 0.8103, "mean_token_accuracy": 0.8746150732040405, "num_tokens": 570966334.0, "step": 15041 }, { "epoch": 1.9134970105584532, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.056628227233887, "learning_rate": 5e-05, "loss": 0.8761, "mean_token_accuracy": 0.8653085231781006, "num_tokens": 571008008.0, "step": 15042 }, { "epoch": 1.9136242208370438, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.827882766723633, "learning_rate": 5e-05, "loss": 0.8448, "mean_token_accuracy": 0.8628288507461548, "num_tokens": 571046915.0, "step": 15043 }, { "epoch": 1.9137514311156343, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.783766269683838, "learning_rate": 5e-05, "loss": 0.8108, "mean_token_accuracy": 0.8702983856201172, "num_tokens": 571085945.0, "step": 15044 }, { "epoch": 1.9138786413942248, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.726008892059326, "learning_rate": 5e-05, "loss": 0.7439, "mean_token_accuracy": 0.8838364481925964, "num_tokens": 571119839.0, "step": 15045 }, { "epoch": 1.9140058516728151, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.791974067687988, "learning_rate": 5e-05, "loss": 0.7654, "mean_token_accuracy": 0.8807719945907593, "num_tokens": 571156888.0, "step": 15046 }, { "epoch": 1.9141330619514056, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.5436201095581055, "learning_rate": 5e-05, "loss": 0.6949, "mean_token_accuracy": 0.8901326656341553, "num_tokens": 571197022.0, "step": 15047 }, { "epoch": 1.9142602722299962, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.871461391448975, "learning_rate": 5e-05, "loss": 0.8051, "mean_token_accuracy": 0.8641954660415649, "num_tokens": 571234780.0, "step": 15048 }, { "epoch": 1.9143874825085867, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.430021286010742, "learning_rate": 5e-05, "loss": 0.7333, "mean_token_accuracy": 0.8744797706604004, "num_tokens": 571272804.0, "step": 15049 }, { "epoch": 1.9145146927871772, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.887807846069336, "learning_rate": 5e-05, "loss": 0.7874, "mean_token_accuracy": 0.8716872334480286, "num_tokens": 571315455.0, "step": 15050 }, { "epoch": 1.9146419030657678, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.368043899536133, "learning_rate": 5e-05, "loss": 0.7149, "mean_token_accuracy": 0.876408576965332, "num_tokens": 571350522.0, "step": 15051 }, { "epoch": 1.914769113344358, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.968852519989014, "learning_rate": 5e-05, "loss": 0.7558, "mean_token_accuracy": 0.8763525485992432, "num_tokens": 571384452.0, "step": 15052 }, { "epoch": 1.9148963236229486, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.2012128829956055, "learning_rate": 5e-05, "loss": 0.6703, "mean_token_accuracy": 0.8876345157623291, "num_tokens": 571423041.0, "step": 15053 }, { "epoch": 1.9150235339015391, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.139119148254395, "learning_rate": 5e-05, "loss": 0.753, "mean_token_accuracy": 0.8835505247116089, "num_tokens": 571454744.0, "step": 15054 }, { "epoch": 1.9151507441801296, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.050604820251465, "learning_rate": 5e-05, "loss": 0.6812, "mean_token_accuracy": 0.8814948797225952, "num_tokens": 571492937.0, "step": 15055 }, { "epoch": 1.9152779544587202, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.262592315673828, "learning_rate": 5e-05, "loss": 0.7704, "mean_token_accuracy": 0.8843953013420105, "num_tokens": 571531756.0, "step": 15056 }, { "epoch": 1.9154051647373107, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.000009059906006, "learning_rate": 5e-05, "loss": 0.7311, "mean_token_accuracy": 0.862348198890686, "num_tokens": 571566095.0, "step": 15057 }, { "epoch": 1.9155323750159012, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.310090065002441, "learning_rate": 5e-05, "loss": 0.8031, "mean_token_accuracy": 0.8743008375167847, "num_tokens": 571599169.0, "step": 15058 }, { "epoch": 1.9156595852944918, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.019527912139893, "learning_rate": 5e-05, "loss": 0.6899, "mean_token_accuracy": 0.8794676065444946, "num_tokens": 571632063.0, "step": 15059 }, { "epoch": 1.9157867955730823, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.245454788208008, "learning_rate": 5e-05, "loss": 0.8129, "mean_token_accuracy": 0.8745500445365906, "num_tokens": 571670867.0, "step": 15060 }, { "epoch": 1.9159140058516728, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.168863296508789, "learning_rate": 5e-05, "loss": 0.7262, "mean_token_accuracy": 0.8722001314163208, "num_tokens": 571710445.0, "step": 15061 }, { "epoch": 1.9160412161302633, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.968462944030762, "learning_rate": 5e-05, "loss": 0.7291, "mean_token_accuracy": 0.8888627290725708, "num_tokens": 571748410.0, "step": 15062 }, { "epoch": 1.9161684264088539, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.5083489418029785, "learning_rate": 5e-05, "loss": 0.7835, "mean_token_accuracy": 0.8588491678237915, "num_tokens": 571786386.0, "step": 15063 }, { "epoch": 1.9162956366874444, "ewc_loss": 0.375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.493891716003418, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.8663639426231384, "num_tokens": 571824414.0, "step": 15064 }, { "epoch": 1.916422846966035, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.948143005371094, "learning_rate": 5e-05, "loss": 0.8354, "mean_token_accuracy": 0.8500638008117676, "num_tokens": 571863085.0, "step": 15065 }, { "epoch": 1.9165500572446255, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.9732441902160645, "learning_rate": 5e-05, "loss": 0.713, "mean_token_accuracy": 0.8696116209030151, "num_tokens": 571894749.0, "step": 15066 }, { "epoch": 1.916677267523216, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.515336990356445, "learning_rate": 5e-05, "loss": 0.8285, "mean_token_accuracy": 0.861229419708252, "num_tokens": 571925962.0, "step": 15067 }, { "epoch": 1.9168044778018065, "ewc_loss": 0.298828125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.384172439575195, "learning_rate": 5e-05, "loss": 0.6951, "mean_token_accuracy": 0.8608648180961609, "num_tokens": 571966021.0, "step": 15068 }, { "epoch": 1.916931688080397, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 9.277680397033691, "learning_rate": 5e-05, "loss": 0.8626, "mean_token_accuracy": 0.8801226019859314, "num_tokens": 572002684.0, "step": 15069 }, { "epoch": 1.9170588983589874, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00023365020751953125, "grad_norm": 5.891689300537109, "learning_rate": 5e-05, "loss": 0.6609, "mean_token_accuracy": 0.8730279803276062, "num_tokens": 572038054.0, "step": 15070 }, { "epoch": 1.9171861086375779, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 10.122739791870117, "learning_rate": 5e-05, "loss": 1.008, "mean_token_accuracy": 0.8692027926445007, "num_tokens": 572076851.0, "step": 15071 }, { "epoch": 1.9173133189161684, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.32499885559082, "learning_rate": 5e-05, "loss": 0.7213, "mean_token_accuracy": 0.8761889934539795, "num_tokens": 572115642.0, "step": 15072 }, { "epoch": 1.917440529194759, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.12318229675293, "learning_rate": 5e-05, "loss": 1.0249, "mean_token_accuracy": 0.8782137036323547, "num_tokens": 572154803.0, "step": 15073 }, { "epoch": 1.9175677394733495, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.059305191040039, "learning_rate": 5e-05, "loss": 0.9935, "mean_token_accuracy": 0.865625262260437, "num_tokens": 572189241.0, "step": 15074 }, { "epoch": 1.91769494975194, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.316774368286133, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.8848972320556641, "num_tokens": 572230464.0, "step": 15075 }, { "epoch": 1.9178221600305305, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.528970718383789, "learning_rate": 5e-05, "loss": 1.0096, "mean_token_accuracy": 0.8821977376937866, "num_tokens": 572268863.0, "step": 15076 }, { "epoch": 1.9179493703091208, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.383595943450928, "learning_rate": 5e-05, "loss": 0.7373, "mean_token_accuracy": 0.8672765493392944, "num_tokens": 572307484.0, "step": 15077 }, { "epoch": 1.9180765805877114, "ewc_loss": 0.78125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 11.288026809692383, "learning_rate": 5e-05, "loss": 1.1514, "mean_token_accuracy": 0.8697339296340942, "num_tokens": 572347653.0, "step": 15078 }, { "epoch": 1.9182037908663019, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 7.51258659362793, "learning_rate": 5e-05, "loss": 0.9724, "mean_token_accuracy": 0.8595197200775146, "num_tokens": 572386850.0, "step": 15079 }, { "epoch": 1.9183310011448924, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 11.507033348083496, "learning_rate": 5e-05, "loss": 1.2263, "mean_token_accuracy": 0.859446108341217, "num_tokens": 572422416.0, "step": 15080 }, { "epoch": 1.918458211423483, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0007781982421875, "grad_norm": 9.425521850585938, "learning_rate": 5e-05, "loss": 1.2362, "mean_token_accuracy": 0.8590999841690063, "num_tokens": 572458845.0, "step": 15081 }, { "epoch": 1.9185854217020735, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 10.13869571685791, "learning_rate": 5e-05, "loss": 1.0032, "mean_token_accuracy": 0.8744393587112427, "num_tokens": 572500553.0, "step": 15082 }, { "epoch": 1.918712631980664, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.13681411743164, "learning_rate": 5e-05, "loss": 1.0753, "mean_token_accuracy": 0.8635097742080688, "num_tokens": 572541315.0, "step": 15083 }, { "epoch": 1.9188398422592545, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.279547691345215, "learning_rate": 5e-05, "loss": 0.9692, "mean_token_accuracy": 0.879753589630127, "num_tokens": 572574461.0, "step": 15084 }, { "epoch": 1.918967052537845, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.425411224365234, "learning_rate": 5e-05, "loss": 0.9574, "mean_token_accuracy": 0.8635627031326294, "num_tokens": 572616860.0, "step": 15085 }, { "epoch": 1.9190942628164356, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.171538352966309, "learning_rate": 5e-05, "loss": 0.9589, "mean_token_accuracy": 0.8757469654083252, "num_tokens": 572654228.0, "step": 15086 }, { "epoch": 1.919221473095026, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.504175662994385, "learning_rate": 5e-05, "loss": 0.8168, "mean_token_accuracy": 0.8705461621284485, "num_tokens": 572693603.0, "step": 15087 }, { "epoch": 1.9193486833736166, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.78696346282959, "learning_rate": 5e-05, "loss": 1.0399, "mean_token_accuracy": 0.8633599281311035, "num_tokens": 572730292.0, "step": 15088 }, { "epoch": 1.9194758936522072, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.329610347747803, "learning_rate": 5e-05, "loss": 0.8485, "mean_token_accuracy": 0.8660438656806946, "num_tokens": 572764348.0, "step": 15089 }, { "epoch": 1.9196031039307977, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.16278076171875, "learning_rate": 5e-05, "loss": 1.0281, "mean_token_accuracy": 0.8775555491447449, "num_tokens": 572798410.0, "step": 15090 }, { "epoch": 1.9197303142093882, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.971273422241211, "learning_rate": 5e-05, "loss": 0.9917, "mean_token_accuracy": 0.8578640222549438, "num_tokens": 572832067.0, "step": 15091 }, { "epoch": 1.9198575244879788, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.728013038635254, "learning_rate": 5e-05, "loss": 1.0047, "mean_token_accuracy": 0.879393458366394, "num_tokens": 572866676.0, "step": 15092 }, { "epoch": 1.9199847347665693, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.299144744873047, "learning_rate": 5e-05, "loss": 0.9894, "mean_token_accuracy": 0.8600947856903076, "num_tokens": 572913631.0, "step": 15093 }, { "epoch": 1.9201119450451598, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.09560489654541, "learning_rate": 5e-05, "loss": 0.96, "mean_token_accuracy": 0.8765575289726257, "num_tokens": 572953154.0, "step": 15094 }, { "epoch": 1.9202391553237501, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.17636489868164, "learning_rate": 5e-05, "loss": 0.9115, "mean_token_accuracy": 0.8659025430679321, "num_tokens": 572992259.0, "step": 15095 }, { "epoch": 1.9203663656023406, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.678689002990723, "learning_rate": 5e-05, "loss": 0.9146, "mean_token_accuracy": 0.8727948665618896, "num_tokens": 573024464.0, "step": 15096 }, { "epoch": 1.9204935758809312, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.8182549476623535, "learning_rate": 5e-05, "loss": 0.8172, "mean_token_accuracy": 0.8769619464874268, "num_tokens": 573060620.0, "step": 15097 }, { "epoch": 1.9206207861595217, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.789344787597656, "learning_rate": 5e-05, "loss": 0.9519, "mean_token_accuracy": 0.8659571409225464, "num_tokens": 573098284.0, "step": 15098 }, { "epoch": 1.9207479964381122, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.301090717315674, "learning_rate": 5e-05, "loss": 0.7775, "mean_token_accuracy": 0.8726184964179993, "num_tokens": 573136253.0, "step": 15099 }, { "epoch": 1.9208752067167028, "ewc_loss": 0.625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.357298851013184, "learning_rate": 5e-05, "loss": 1.0151, "mean_token_accuracy": 0.8700485825538635, "num_tokens": 573174843.0, "step": 15100 }, { "epoch": 1.921002416995293, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.279353141784668, "learning_rate": 5e-05, "loss": 0.8391, "mean_token_accuracy": 0.8598093390464783, "num_tokens": 573215759.0, "step": 15101 }, { "epoch": 1.9211296272738836, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.511731147766113, "learning_rate": 5e-05, "loss": 1.0131, "mean_token_accuracy": 0.87457275390625, "num_tokens": 573255297.0, "step": 15102 }, { "epoch": 1.9212568375524741, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.86775541305542, "learning_rate": 5e-05, "loss": 0.7963, "mean_token_accuracy": 0.8977445363998413, "num_tokens": 573293152.0, "step": 15103 }, { "epoch": 1.9213840478310646, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.875146865844727, "learning_rate": 5e-05, "loss": 0.965, "mean_token_accuracy": 0.8701528906822205, "num_tokens": 573330456.0, "step": 15104 }, { "epoch": 1.9215112581096552, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.268366813659668, "learning_rate": 5e-05, "loss": 0.9313, "mean_token_accuracy": 0.8648982048034668, "num_tokens": 573371595.0, "step": 15105 }, { "epoch": 1.9216384683882457, "ewc_loss": 0.5, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.990011692047119, "learning_rate": 5e-05, "loss": 0.8916, "mean_token_accuracy": 0.8695169687271118, "num_tokens": 573417414.0, "step": 15106 }, { "epoch": 1.9217656786668362, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.253422737121582, "learning_rate": 5e-05, "loss": 0.9001, "mean_token_accuracy": 0.8652382493019104, "num_tokens": 573452146.0, "step": 15107 }, { "epoch": 1.9218928889454268, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.647839069366455, "learning_rate": 5e-05, "loss": 0.8382, "mean_token_accuracy": 0.8642199039459229, "num_tokens": 573485469.0, "step": 15108 }, { "epoch": 1.9220200992240173, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.899425029754639, "learning_rate": 5e-05, "loss": 0.7943, "mean_token_accuracy": 0.8761730194091797, "num_tokens": 573524763.0, "step": 15109 }, { "epoch": 1.9221473095026078, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.902851104736328, "learning_rate": 5e-05, "loss": 0.8224, "mean_token_accuracy": 0.868382453918457, "num_tokens": 573559326.0, "step": 15110 }, { "epoch": 1.9222745197811983, "ewc_loss": 0.375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.3483171463012695, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.866043210029602, "num_tokens": 573604700.0, "step": 15111 }, { "epoch": 1.9224017300597889, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.373224258422852, "learning_rate": 5e-05, "loss": 0.8037, "mean_token_accuracy": 0.8851429224014282, "num_tokens": 573639570.0, "step": 15112 }, { "epoch": 1.9225289403383794, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.976468563079834, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8677653670310974, "num_tokens": 573669849.0, "step": 15113 }, { "epoch": 1.92265615061697, "ewc_loss": 0.5, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.648417472839355, "learning_rate": 5e-05, "loss": 0.8759, "mean_token_accuracy": 0.8679588437080383, "num_tokens": 573702402.0, "step": 15114 }, { "epoch": 1.9227833608955605, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.894221305847168, "learning_rate": 5e-05, "loss": 0.7344, "mean_token_accuracy": 0.8689841032028198, "num_tokens": 573737428.0, "step": 15115 }, { "epoch": 1.922910571174151, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.75731086730957, "learning_rate": 5e-05, "loss": 0.8751, "mean_token_accuracy": 0.8775162696838379, "num_tokens": 573775111.0, "step": 15116 }, { "epoch": 1.9230377814527415, "ewc_loss": 0.375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.996500015258789, "learning_rate": 5e-05, "loss": 0.7715, "mean_token_accuracy": 0.8610759377479553, "num_tokens": 573814550.0, "step": 15117 }, { "epoch": 1.923164991731332, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.701001167297363, "learning_rate": 5e-05, "loss": 0.9291, "mean_token_accuracy": 0.8604874610900879, "num_tokens": 573852932.0, "step": 15118 }, { "epoch": 1.9232922020099223, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.287081241607666, "learning_rate": 5e-05, "loss": 0.7493, "mean_token_accuracy": 0.8768360614776611, "num_tokens": 573887769.0, "step": 15119 }, { "epoch": 1.9234194122885129, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.304508209228516, "learning_rate": 5e-05, "loss": 0.8797, "mean_token_accuracy": 0.866682767868042, "num_tokens": 573925258.0, "step": 15120 }, { "epoch": 1.9235466225671034, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.6712236404418945, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.8812853097915649, "num_tokens": 573956690.0, "step": 15121 }, { "epoch": 1.923673832845694, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.767419815063477, "learning_rate": 5e-05, "loss": 0.8402, "mean_token_accuracy": 0.8582309484481812, "num_tokens": 573993822.0, "step": 15122 }, { "epoch": 1.9238010431242845, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.917365550994873, "learning_rate": 5e-05, "loss": 0.8083, "mean_token_accuracy": 0.8693066835403442, "num_tokens": 574030087.0, "step": 15123 }, { "epoch": 1.923928253402875, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.427642822265625, "learning_rate": 5e-05, "loss": 0.7406, "mean_token_accuracy": 0.8761857151985168, "num_tokens": 574071964.0, "step": 15124 }, { "epoch": 1.9240554636814655, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.021303176879883, "learning_rate": 5e-05, "loss": 0.8219, "mean_token_accuracy": 0.8657317161560059, "num_tokens": 574105706.0, "step": 15125 }, { "epoch": 1.9241826739600558, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.265198230743408, "learning_rate": 5e-05, "loss": 0.7883, "mean_token_accuracy": 0.8534311056137085, "num_tokens": 574145798.0, "step": 15126 }, { "epoch": 1.9243098842386464, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.065725326538086, "learning_rate": 5e-05, "loss": 0.7534, "mean_token_accuracy": 0.8861610889434814, "num_tokens": 574192631.0, "step": 15127 }, { "epoch": 1.9244370945172369, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.188915252685547, "learning_rate": 5e-05, "loss": 0.7757, "mean_token_accuracy": 0.8528074026107788, "num_tokens": 574228601.0, "step": 15128 }, { "epoch": 1.9245643047958274, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.144536018371582, "learning_rate": 5e-05, "loss": 0.7844, "mean_token_accuracy": 0.8770533800125122, "num_tokens": 574266416.0, "step": 15129 }, { "epoch": 1.924691515074418, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.065573215484619, "learning_rate": 5e-05, "loss": 0.7567, "mean_token_accuracy": 0.8608512878417969, "num_tokens": 574297414.0, "step": 15130 }, { "epoch": 1.9248187253530085, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.294269561767578, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8879173994064331, "num_tokens": 574333749.0, "step": 15131 }, { "epoch": 1.924945935631599, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.900125026702881, "learning_rate": 5e-05, "loss": 0.7118, "mean_token_accuracy": 0.8703463077545166, "num_tokens": 574377186.0, "step": 15132 }, { "epoch": 1.9250731459101895, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.516088485717773, "learning_rate": 5e-05, "loss": 0.8172, "mean_token_accuracy": 0.875281572341919, "num_tokens": 574414814.0, "step": 15133 }, { "epoch": 1.92520035618878, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.75013542175293, "learning_rate": 5e-05, "loss": 0.731, "mean_token_accuracy": 0.8650879859924316, "num_tokens": 574453070.0, "step": 15134 }, { "epoch": 1.9253275664673706, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.674455642700195, "learning_rate": 5e-05, "loss": 0.8219, "mean_token_accuracy": 0.877683162689209, "num_tokens": 574487564.0, "step": 15135 }, { "epoch": 1.925454776745961, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.7612433433532715, "learning_rate": 5e-05, "loss": 0.7404, "mean_token_accuracy": 0.8680342435836792, "num_tokens": 574528456.0, "step": 15136 }, { "epoch": 1.9255819870245516, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.665502548217773, "learning_rate": 5e-05, "loss": 0.864, "mean_token_accuracy": 0.8734995126724243, "num_tokens": 574567724.0, "step": 15137 }, { "epoch": 1.9257091973031422, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.989929676055908, "learning_rate": 5e-05, "loss": 0.7232, "mean_token_accuracy": 0.871492862701416, "num_tokens": 574599702.0, "step": 15138 }, { "epoch": 1.9258364075817327, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.361255645751953, "learning_rate": 5e-05, "loss": 0.834, "mean_token_accuracy": 0.8771963119506836, "num_tokens": 574633709.0, "step": 15139 }, { "epoch": 1.9259636178603232, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.370187282562256, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8677408695220947, "num_tokens": 574670490.0, "step": 15140 }, { "epoch": 1.9260908281389137, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.863731861114502, "learning_rate": 5e-05, "loss": 0.8481, "mean_token_accuracy": 0.8579674959182739, "num_tokens": 574706563.0, "step": 15141 }, { "epoch": 1.9262180384175043, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.7476606369018555, "learning_rate": 5e-05, "loss": 0.7935, "mean_token_accuracy": 0.8736615180969238, "num_tokens": 574744743.0, "step": 15142 }, { "epoch": 1.9263452486960948, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.461818218231201, "learning_rate": 5e-05, "loss": 0.7615, "mean_token_accuracy": 0.8703410625457764, "num_tokens": 574782236.0, "step": 15143 }, { "epoch": 1.926472458974685, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.908023357391357, "learning_rate": 5e-05, "loss": 0.7563, "mean_token_accuracy": 0.8829129934310913, "num_tokens": 574826457.0, "step": 15144 }, { "epoch": 1.9265996692532756, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.3624138832092285, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8639794588088989, "num_tokens": 574870362.0, "step": 15145 }, { "epoch": 1.9267268795318662, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.802328109741211, "learning_rate": 5e-05, "loss": 0.753, "mean_token_accuracy": 0.8778666257858276, "num_tokens": 574907984.0, "step": 15146 }, { "epoch": 1.9268540898104567, "ewc_loss": 0.375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.533172607421875, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8634770512580872, "num_tokens": 574950813.0, "step": 15147 }, { "epoch": 1.9269813000890472, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.509352207183838, "learning_rate": 5e-05, "loss": 0.7375, "mean_token_accuracy": 0.8742920160293579, "num_tokens": 574985210.0, "step": 15148 }, { "epoch": 1.9271085103676378, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.853762149810791, "learning_rate": 5e-05, "loss": 0.7251, "mean_token_accuracy": 0.8779205083847046, "num_tokens": 575020263.0, "step": 15149 }, { "epoch": 1.927235720646228, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.099324703216553, "learning_rate": 5e-05, "loss": 0.7334, "mean_token_accuracy": 0.8679417371749878, "num_tokens": 575060632.0, "step": 15150 }, { "epoch": 1.9273629309248186, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.322128295898438, "learning_rate": 5e-05, "loss": 0.8115, "mean_token_accuracy": 0.8685277700424194, "num_tokens": 575102725.0, "step": 15151 }, { "epoch": 1.9274901412034091, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.580306053161621, "learning_rate": 5e-05, "loss": 0.6306, "mean_token_accuracy": 0.8854775428771973, "num_tokens": 575143069.0, "step": 15152 }, { "epoch": 1.9276173514819996, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 9.022344589233398, "learning_rate": 5e-05, "loss": 0.8494, "mean_token_accuracy": 0.8743473291397095, "num_tokens": 575182618.0, "step": 15153 }, { "epoch": 1.9277445617605902, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 6.00043249130249, "learning_rate": 5e-05, "loss": 0.6681, "mean_token_accuracy": 0.8695228099822998, "num_tokens": 575224576.0, "step": 15154 }, { "epoch": 1.9278717720391807, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.954222679138184, "learning_rate": 5e-05, "loss": 0.9526, "mean_token_accuracy": 0.8815951943397522, "num_tokens": 575260152.0, "step": 15155 }, { "epoch": 1.9279989823177712, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 5.990638732910156, "learning_rate": 5e-05, "loss": 0.7104, "mean_token_accuracy": 0.8678281307220459, "num_tokens": 575292308.0, "step": 15156 }, { "epoch": 1.9281261925963618, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.417421340942383, "learning_rate": 5e-05, "loss": 1.0835, "mean_token_accuracy": 0.8655728101730347, "num_tokens": 575326305.0, "step": 15157 }, { "epoch": 1.9282534028749523, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.5577216148376465, "learning_rate": 5e-05, "loss": 0.96, "mean_token_accuracy": 0.866104006767273, "num_tokens": 575361775.0, "step": 15158 }, { "epoch": 1.9283806131535428, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.154253959655762, "learning_rate": 5e-05, "loss": 0.9364, "mean_token_accuracy": 0.8797439932823181, "num_tokens": 575397508.0, "step": 15159 }, { "epoch": 1.9285078234321333, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.122365951538086, "learning_rate": 5e-05, "loss": 1.0532, "mean_token_accuracy": 0.8682276010513306, "num_tokens": 575437770.0, "step": 15160 }, { "epoch": 1.9286350337107239, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.787072658538818, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8746057748794556, "num_tokens": 575478322.0, "step": 15161 }, { "epoch": 1.9287622439893144, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000743865966796875, "grad_norm": 11.167637825012207, "learning_rate": 5e-05, "loss": 1.2204, "mean_token_accuracy": 0.8590512871742249, "num_tokens": 575518816.0, "step": 15162 }, { "epoch": 1.928889454267905, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 6.928296089172363, "learning_rate": 5e-05, "loss": 0.8723, "mean_token_accuracy": 0.8606364727020264, "num_tokens": 575554276.0, "step": 15163 }, { "epoch": 1.9290166645464955, "ewc_loss": 0.96875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000911712646484375, "grad_norm": 12.575765609741211, "learning_rate": 5e-05, "loss": 1.3672, "mean_token_accuracy": 0.8646708726882935, "num_tokens": 575591331.0, "step": 15164 }, { "epoch": 1.929143874825086, "ewc_loss": 0.9765625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000919342041015625, "grad_norm": 10.003239631652832, "learning_rate": 5e-05, "loss": 1.3671, "mean_token_accuracy": 0.8667973279953003, "num_tokens": 575629449.0, "step": 15165 }, { "epoch": 1.9292710851036765, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 11.264461517333984, "learning_rate": 5e-05, "loss": 1.1435, "mean_token_accuracy": 0.8569040298461914, "num_tokens": 575668169.0, "step": 15166 }, { "epoch": 1.929398295382267, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 9.54971694946289, "learning_rate": 5e-05, "loss": 1.2104, "mean_token_accuracy": 0.8614755272865295, "num_tokens": 575708481.0, "step": 15167 }, { "epoch": 1.9295255056608573, "ewc_loss": 0.75, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.86202621459961, "learning_rate": 5e-05, "loss": 1.187, "mean_token_accuracy": 0.8495672941207886, "num_tokens": 575744879.0, "step": 15168 }, { "epoch": 1.9296527159394479, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.658082962036133, "learning_rate": 5e-05, "loss": 0.9209, "mean_token_accuracy": 0.8917726278305054, "num_tokens": 575781587.0, "step": 15169 }, { "epoch": 1.9297799262180384, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.60872745513916, "learning_rate": 5e-05, "loss": 1.1191, "mean_token_accuracy": 0.8700529932975769, "num_tokens": 575816662.0, "step": 15170 }, { "epoch": 1.929907136496629, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.360054016113281, "learning_rate": 5e-05, "loss": 0.9689, "mean_token_accuracy": 0.8634263277053833, "num_tokens": 575855169.0, "step": 15171 }, { "epoch": 1.9300343467752195, "ewc_loss": 0.75, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.770750999450684, "learning_rate": 5e-05, "loss": 1.0949, "mean_token_accuracy": 0.8818502426147461, "num_tokens": 575896939.0, "step": 15172 }, { "epoch": 1.93016155705381, "ewc_loss": 0.625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.664612770080566, "learning_rate": 5e-05, "loss": 0.9916, "mean_token_accuracy": 0.875857412815094, "num_tokens": 575930587.0, "step": 15173 }, { "epoch": 1.9302887673324005, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.632343292236328, "learning_rate": 5e-05, "loss": 1.1418, "mean_token_accuracy": 0.8572037816047668, "num_tokens": 575964432.0, "step": 15174 }, { "epoch": 1.9304159776109908, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.905230522155762, "learning_rate": 5e-05, "loss": 1.0016, "mean_token_accuracy": 0.8750227689743042, "num_tokens": 576004916.0, "step": 15175 }, { "epoch": 1.9305431878895813, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.151900291442871, "learning_rate": 5e-05, "loss": 1.128, "mean_token_accuracy": 0.8521540760993958, "num_tokens": 576047857.0, "step": 15176 }, { "epoch": 1.9306703981681719, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.772262573242188, "learning_rate": 5e-05, "loss": 1.0122, "mean_token_accuracy": 0.8608808517456055, "num_tokens": 576089087.0, "step": 15177 }, { "epoch": 1.9307976084467624, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.778706550598145, "learning_rate": 5e-05, "loss": 1.0319, "mean_token_accuracy": 0.8738385438919067, "num_tokens": 576134106.0, "step": 15178 }, { "epoch": 1.930924818725353, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.433977127075195, "learning_rate": 5e-05, "loss": 0.9304, "mean_token_accuracy": 0.8679736256599426, "num_tokens": 576172566.0, "step": 15179 }, { "epoch": 1.9310520290039435, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.631512641906738, "learning_rate": 5e-05, "loss": 1.0897, "mean_token_accuracy": 0.8475297689437866, "num_tokens": 576214469.0, "step": 15180 }, { "epoch": 1.931179239282534, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.16044807434082, "learning_rate": 5e-05, "loss": 0.8175, "mean_token_accuracy": 0.892807126045227, "num_tokens": 576249316.0, "step": 15181 }, { "epoch": 1.9313064495611245, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.679234504699707, "learning_rate": 5e-05, "loss": 1.0633, "mean_token_accuracy": 0.8510129451751709, "num_tokens": 576290748.0, "step": 15182 }, { "epoch": 1.931433659839715, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.054046630859375, "learning_rate": 5e-05, "loss": 0.8723, "mean_token_accuracy": 0.8729898929595947, "num_tokens": 576329137.0, "step": 15183 }, { "epoch": 1.9315608701183056, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.737639427185059, "learning_rate": 5e-05, "loss": 0.9839, "mean_token_accuracy": 0.8834967613220215, "num_tokens": 576367147.0, "step": 15184 }, { "epoch": 1.931688080396896, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.207634925842285, "learning_rate": 5e-05, "loss": 0.9447, "mean_token_accuracy": 0.8635557293891907, "num_tokens": 576406068.0, "step": 15185 }, { "epoch": 1.9318152906754866, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.508246421813965, "learning_rate": 5e-05, "loss": 1.0273, "mean_token_accuracy": 0.8646166324615479, "num_tokens": 576438150.0, "step": 15186 }, { "epoch": 1.9319425009540772, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.334068298339844, "learning_rate": 5e-05, "loss": 0.8968, "mean_token_accuracy": 0.8775776028633118, "num_tokens": 576470630.0, "step": 15187 }, { "epoch": 1.9320697112326677, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.05223560333252, "learning_rate": 5e-05, "loss": 0.938, "mean_token_accuracy": 0.879040002822876, "num_tokens": 576505896.0, "step": 15188 }, { "epoch": 1.9321969215112582, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.153243064880371, "learning_rate": 5e-05, "loss": 0.8354, "mean_token_accuracy": 0.8818781971931458, "num_tokens": 576545204.0, "step": 15189 }, { "epoch": 1.9323241317898487, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.872172355651855, "learning_rate": 5e-05, "loss": 1.0087, "mean_token_accuracy": 0.8607935309410095, "num_tokens": 576590746.0, "step": 15190 }, { "epoch": 1.9324513420684393, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.793434143066406, "learning_rate": 5e-05, "loss": 0.8314, "mean_token_accuracy": 0.8688710331916809, "num_tokens": 576631185.0, "step": 15191 }, { "epoch": 1.9325785523470298, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.967637062072754, "learning_rate": 5e-05, "loss": 0.9669, "mean_token_accuracy": 0.8694256544113159, "num_tokens": 576668945.0, "step": 15192 }, { "epoch": 1.93270576262562, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.6185383796691895, "learning_rate": 5e-05, "loss": 0.809, "mean_token_accuracy": 0.8721233010292053, "num_tokens": 576706052.0, "step": 15193 }, { "epoch": 1.9328329729042106, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.027037620544434, "learning_rate": 5e-05, "loss": 0.959, "mean_token_accuracy": 0.8712714910507202, "num_tokens": 576740668.0, "step": 15194 }, { "epoch": 1.9329601831828012, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.661848068237305, "learning_rate": 5e-05, "loss": 0.8472, "mean_token_accuracy": 0.860603928565979, "num_tokens": 576780312.0, "step": 15195 }, { "epoch": 1.9330873934613917, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.91266918182373, "learning_rate": 5e-05, "loss": 0.9139, "mean_token_accuracy": 0.882482647895813, "num_tokens": 576817130.0, "step": 15196 }, { "epoch": 1.9332146037399822, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.764966011047363, "learning_rate": 5e-05, "loss": 0.7711, "mean_token_accuracy": 0.8871050477027893, "num_tokens": 576849823.0, "step": 15197 }, { "epoch": 1.9333418140185727, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.644869804382324, "learning_rate": 5e-05, "loss": 0.9352, "mean_token_accuracy": 0.8634381294250488, "num_tokens": 576890890.0, "step": 15198 }, { "epoch": 1.933469024297163, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.814845561981201, "learning_rate": 5e-05, "loss": 0.8602, "mean_token_accuracy": 0.8581699132919312, "num_tokens": 576927926.0, "step": 15199 }, { "epoch": 1.9335962345757536, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.339054107666016, "learning_rate": 5e-05, "loss": 0.9162, "mean_token_accuracy": 0.8624439239501953, "num_tokens": 576966194.0, "step": 15200 }, { "epoch": 1.933723444854344, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.734220027923584, "learning_rate": 5e-05, "loss": 0.8393, "mean_token_accuracy": 0.8598675727844238, "num_tokens": 577007218.0, "step": 15201 }, { "epoch": 1.9338506551329346, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.192359924316406, "learning_rate": 5e-05, "loss": 0.7915, "mean_token_accuracy": 0.8894427418708801, "num_tokens": 577040551.0, "step": 15202 }, { "epoch": 1.9339778654115252, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.568634986877441, "learning_rate": 5e-05, "loss": 0.7832, "mean_token_accuracy": 0.8714567422866821, "num_tokens": 577082676.0, "step": 15203 }, { "epoch": 1.9341050756901157, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.184568405151367, "learning_rate": 5e-05, "loss": 0.8244, "mean_token_accuracy": 0.875648021697998, "num_tokens": 577124572.0, "step": 15204 }, { "epoch": 1.9342322859687062, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.451369762420654, "learning_rate": 5e-05, "loss": 0.7182, "mean_token_accuracy": 0.8836789727210999, "num_tokens": 577167571.0, "step": 15205 }, { "epoch": 1.9343594962472968, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.200270652770996, "learning_rate": 5e-05, "loss": 0.8828, "mean_token_accuracy": 0.8529456257820129, "num_tokens": 577211976.0, "step": 15206 }, { "epoch": 1.9344867065258873, "ewc_loss": 0.375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.408539295196533, "learning_rate": 5e-05, "loss": 0.7261, "mean_token_accuracy": 0.8781828880310059, "num_tokens": 577252572.0, "step": 15207 }, { "epoch": 1.9346139168044778, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.14233684539795, "learning_rate": 5e-05, "loss": 0.8484, "mean_token_accuracy": 0.8582308888435364, "num_tokens": 577286923.0, "step": 15208 }, { "epoch": 1.9347411270830683, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.387372970581055, "learning_rate": 5e-05, "loss": 0.7458, "mean_token_accuracy": 0.8690043687820435, "num_tokens": 577324520.0, "step": 15209 }, { "epoch": 1.9348683373616589, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.083518981933594, "learning_rate": 5e-05, "loss": 0.8059, "mean_token_accuracy": 0.8688097596168518, "num_tokens": 577361100.0, "step": 15210 }, { "epoch": 1.9349955476402494, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.408647537231445, "learning_rate": 5e-05, "loss": 0.7392, "mean_token_accuracy": 0.869621217250824, "num_tokens": 577402714.0, "step": 15211 }, { "epoch": 1.93512275791884, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.001147270202637, "learning_rate": 5e-05, "loss": 0.8476, "mean_token_accuracy": 0.8605074882507324, "num_tokens": 577437226.0, "step": 15212 }, { "epoch": 1.9352499681974304, "ewc_loss": 0.375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.4490556716918945, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8743575215339661, "num_tokens": 577475564.0, "step": 15213 }, { "epoch": 1.935377178476021, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.909422397613525, "learning_rate": 5e-05, "loss": 0.8515, "mean_token_accuracy": 0.854210376739502, "num_tokens": 577512623.0, "step": 15214 }, { "epoch": 1.9355043887546115, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.513689994812012, "learning_rate": 5e-05, "loss": 0.7477, "mean_token_accuracy": 0.8699698448181152, "num_tokens": 577545515.0, "step": 15215 }, { "epoch": 1.935631599033202, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.776479721069336, "learning_rate": 5e-05, "loss": 0.7711, "mean_token_accuracy": 0.8702264428138733, "num_tokens": 577577043.0, "step": 15216 }, { "epoch": 1.9357588093117923, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.592867374420166, "learning_rate": 5e-05, "loss": 0.723, "mean_token_accuracy": 0.8795692324638367, "num_tokens": 577616311.0, "step": 15217 }, { "epoch": 1.9358860195903829, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.624824523925781, "learning_rate": 5e-05, "loss": 0.7256, "mean_token_accuracy": 0.8743343949317932, "num_tokens": 577654963.0, "step": 15218 }, { "epoch": 1.9360132298689734, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.719141960144043, "learning_rate": 5e-05, "loss": 0.7132, "mean_token_accuracy": 0.8809398412704468, "num_tokens": 577688283.0, "step": 15219 }, { "epoch": 1.936140440147564, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.454660415649414, "learning_rate": 5e-05, "loss": 0.7521, "mean_token_accuracy": 0.8640789985656738, "num_tokens": 577726855.0, "step": 15220 }, { "epoch": 1.9362676504261545, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.854281902313232, "learning_rate": 5e-05, "loss": 0.8038, "mean_token_accuracy": 0.8605490326881409, "num_tokens": 577765699.0, "step": 15221 }, { "epoch": 1.936394860704745, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.316656112670898, "learning_rate": 5e-05, "loss": 0.6528, "mean_token_accuracy": 0.8946251273155212, "num_tokens": 577797393.0, "step": 15222 }, { "epoch": 1.9365220709833355, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.956379413604736, "learning_rate": 5e-05, "loss": 0.7467, "mean_token_accuracy": 0.8796572685241699, "num_tokens": 577838768.0, "step": 15223 }, { "epoch": 1.9366492812619258, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.1985602378845215, "learning_rate": 5e-05, "loss": 0.7068, "mean_token_accuracy": 0.8747802376747131, "num_tokens": 577882455.0, "step": 15224 }, { "epoch": 1.9367764915405163, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.076054573059082, "learning_rate": 5e-05, "loss": 0.7356, "mean_token_accuracy": 0.8844828009605408, "num_tokens": 577922649.0, "step": 15225 }, { "epoch": 1.9369037018191069, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 7.0712714195251465, "learning_rate": 5e-05, "loss": 0.6995, "mean_token_accuracy": 0.8739575743675232, "num_tokens": 577961679.0, "step": 15226 }, { "epoch": 1.9370309120976974, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.220726013183594, "learning_rate": 5e-05, "loss": 0.7642, "mean_token_accuracy": 0.876166582107544, "num_tokens": 578001848.0, "step": 15227 }, { "epoch": 1.937158122376288, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.914435863494873, "learning_rate": 5e-05, "loss": 0.678, "mean_token_accuracy": 0.8735637068748474, "num_tokens": 578039278.0, "step": 15228 }, { "epoch": 1.9372853326548785, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.448654174804688, "learning_rate": 5e-05, "loss": 0.8024, "mean_token_accuracy": 0.8705028295516968, "num_tokens": 578078152.0, "step": 15229 }, { "epoch": 1.937412542933469, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.716195583343506, "learning_rate": 5e-05, "loss": 0.7087, "mean_token_accuracy": 0.8642486333847046, "num_tokens": 578115927.0, "step": 15230 }, { "epoch": 1.9375397532120595, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.69381332397461, "learning_rate": 5e-05, "loss": 0.8155, "mean_token_accuracy": 0.8732858896255493, "num_tokens": 578157425.0, "step": 15231 }, { "epoch": 1.93766696349065, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.571167469024658, "learning_rate": 5e-05, "loss": 0.6635, "mean_token_accuracy": 0.8776121139526367, "num_tokens": 578201985.0, "step": 15232 }, { "epoch": 1.9377941737692406, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.893331527709961, "learning_rate": 5e-05, "loss": 0.8619, "mean_token_accuracy": 0.8700329661369324, "num_tokens": 578239471.0, "step": 15233 }, { "epoch": 1.937921384047831, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.630074977874756, "learning_rate": 5e-05, "loss": 0.6836, "mean_token_accuracy": 0.8797557353973389, "num_tokens": 578272302.0, "step": 15234 }, { "epoch": 1.9380485943264216, "ewc_loss": 0.5, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.799988746643066, "learning_rate": 5e-05, "loss": 0.8716, "mean_token_accuracy": 0.8697068095207214, "num_tokens": 578314585.0, "step": 15235 }, { "epoch": 1.9381758046050122, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.053867816925049, "learning_rate": 5e-05, "loss": 0.7984, "mean_token_accuracy": 0.8521250486373901, "num_tokens": 578355561.0, "step": 15236 }, { "epoch": 1.9383030148836027, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.154021263122559, "learning_rate": 5e-05, "loss": 0.8347, "mean_token_accuracy": 0.867526650428772, "num_tokens": 578398903.0, "step": 15237 }, { "epoch": 1.9384302251621932, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.796090126037598, "learning_rate": 5e-05, "loss": 0.8769, "mean_token_accuracy": 0.845925509929657, "num_tokens": 578439293.0, "step": 15238 }, { "epoch": 1.9385574354407837, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.196043968200684, "learning_rate": 5e-05, "loss": 0.7783, "mean_token_accuracy": 0.8552849292755127, "num_tokens": 578477965.0, "step": 15239 }, { "epoch": 1.9386846457193743, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.589717864990234, "learning_rate": 5e-05, "loss": 0.8273, "mean_token_accuracy": 0.8761477470397949, "num_tokens": 578515577.0, "step": 15240 }, { "epoch": 1.9388118559979648, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.434828281402588, "learning_rate": 5e-05, "loss": 0.6954, "mean_token_accuracy": 0.873076319694519, "num_tokens": 578557415.0, "step": 15241 }, { "epoch": 1.938939066276555, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 9.319766998291016, "learning_rate": 5e-05, "loss": 0.8942, "mean_token_accuracy": 0.8758590221405029, "num_tokens": 578594556.0, "step": 15242 }, { "epoch": 1.9390662765551456, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.133890151977539, "learning_rate": 5e-05, "loss": 0.6646, "mean_token_accuracy": 0.8811037540435791, "num_tokens": 578627219.0, "step": 15243 }, { "epoch": 1.9391934868337362, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.751855850219727, "learning_rate": 5e-05, "loss": 1.0255, "mean_token_accuracy": 0.8565894365310669, "num_tokens": 578669135.0, "step": 15244 }, { "epoch": 1.9393206971123267, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.691572666168213, "learning_rate": 5e-05, "loss": 0.7464, "mean_token_accuracy": 0.8779168128967285, "num_tokens": 578706174.0, "step": 15245 }, { "epoch": 1.9394479073909172, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.21158218383789, "learning_rate": 5e-05, "loss": 0.9673, "mean_token_accuracy": 0.8700985908508301, "num_tokens": 578738942.0, "step": 15246 }, { "epoch": 1.9395751176695077, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.946036338806152, "learning_rate": 5e-05, "loss": 0.8562, "mean_token_accuracy": 0.892072856426239, "num_tokens": 578771553.0, "step": 15247 }, { "epoch": 1.939702327948098, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.634044647216797, "learning_rate": 5e-05, "loss": 0.849, "mean_token_accuracy": 0.866631031036377, "num_tokens": 578811532.0, "step": 15248 }, { "epoch": 1.9398295382266886, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.064346313476562, "learning_rate": 5e-05, "loss": 0.9678, "mean_token_accuracy": 0.8678994178771973, "num_tokens": 578852858.0, "step": 15249 }, { "epoch": 1.939956748505279, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.570744514465332, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8656883239746094, "num_tokens": 578890340.0, "step": 15250 }, { "epoch": 1.9400839587838696, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.573431015014648, "learning_rate": 5e-05, "loss": 0.9541, "mean_token_accuracy": 0.8793632984161377, "num_tokens": 578928953.0, "step": 15251 }, { "epoch": 1.9402111690624602, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.011843681335449, "learning_rate": 5e-05, "loss": 0.7769, "mean_token_accuracy": 0.8807894587516785, "num_tokens": 578967978.0, "step": 15252 }, { "epoch": 1.9403383793410507, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.780890464782715, "learning_rate": 5e-05, "loss": 0.9237, "mean_token_accuracy": 0.8772331476211548, "num_tokens": 579005484.0, "step": 15253 }, { "epoch": 1.9404655896196412, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.064311027526855, "learning_rate": 5e-05, "loss": 0.9699, "mean_token_accuracy": 0.8540874123573303, "num_tokens": 579045172.0, "step": 15254 }, { "epoch": 1.9405927998982317, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.447797775268555, "learning_rate": 5e-05, "loss": 0.7522, "mean_token_accuracy": 0.8816248774528503, "num_tokens": 579075001.0, "step": 15255 }, { "epoch": 1.9407200101768223, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.78123950958252, "learning_rate": 5e-05, "loss": 0.9325, "mean_token_accuracy": 0.8696210384368896, "num_tokens": 579116029.0, "step": 15256 }, { "epoch": 1.9408472204554128, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.996274948120117, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.869686484336853, "num_tokens": 579157741.0, "step": 15257 }, { "epoch": 1.9409744307340033, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.76185417175293, "learning_rate": 5e-05, "loss": 0.9017, "mean_token_accuracy": 0.8711378574371338, "num_tokens": 579196033.0, "step": 15258 }, { "epoch": 1.9411016410125939, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.366715908050537, "learning_rate": 5e-05, "loss": 0.7698, "mean_token_accuracy": 0.880611777305603, "num_tokens": 579235060.0, "step": 15259 }, { "epoch": 1.9412288512911844, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.104815483093262, "learning_rate": 5e-05, "loss": 0.8421, "mean_token_accuracy": 0.8725744485855103, "num_tokens": 579272184.0, "step": 15260 }, { "epoch": 1.941356061569775, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.798187255859375, "learning_rate": 5e-05, "loss": 0.8286, "mean_token_accuracy": 0.8734959363937378, "num_tokens": 579309461.0, "step": 15261 }, { "epoch": 1.9414832718483654, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.638819694519043, "learning_rate": 5e-05, "loss": 0.8237, "mean_token_accuracy": 0.8545703887939453, "num_tokens": 579351871.0, "step": 15262 }, { "epoch": 1.941610482126956, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.874514102935791, "learning_rate": 5e-05, "loss": 0.8567, "mean_token_accuracy": 0.855894923210144, "num_tokens": 579390123.0, "step": 15263 }, { "epoch": 1.9417376924055465, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.608139514923096, "learning_rate": 5e-05, "loss": 0.8059, "mean_token_accuracy": 0.8598251342773438, "num_tokens": 579426015.0, "step": 15264 }, { "epoch": 1.941864902684137, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.664764404296875, "learning_rate": 5e-05, "loss": 0.8067, "mean_token_accuracy": 0.8586676120758057, "num_tokens": 579465853.0, "step": 15265 }, { "epoch": 1.9419921129627273, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.756053447723389, "learning_rate": 5e-05, "loss": 0.7984, "mean_token_accuracy": 0.8671082854270935, "num_tokens": 579499740.0, "step": 15266 }, { "epoch": 1.9421193232413179, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.4087395668029785, "learning_rate": 5e-05, "loss": 0.7003, "mean_token_accuracy": 0.8859568238258362, "num_tokens": 579538908.0, "step": 15267 }, { "epoch": 1.9422465335199084, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.96996545791626, "learning_rate": 5e-05, "loss": 0.7715, "mean_token_accuracy": 0.8738768100738525, "num_tokens": 579576026.0, "step": 15268 }, { "epoch": 1.942373743798499, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.158168315887451, "learning_rate": 5e-05, "loss": 0.7094, "mean_token_accuracy": 0.8766249418258667, "num_tokens": 579618980.0, "step": 15269 }, { "epoch": 1.9425009540770894, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.255749702453613, "learning_rate": 5e-05, "loss": 0.8409, "mean_token_accuracy": 0.8576281070709229, "num_tokens": 579659205.0, "step": 15270 }, { "epoch": 1.94262816435568, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.847855567932129, "learning_rate": 5e-05, "loss": 0.668, "mean_token_accuracy": 0.8824938535690308, "num_tokens": 579695598.0, "step": 15271 }, { "epoch": 1.9427553746342705, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.62463092803955, "learning_rate": 5e-05, "loss": 0.8315, "mean_token_accuracy": 0.8731582164764404, "num_tokens": 579730871.0, "step": 15272 }, { "epoch": 1.9428825849128608, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.497138500213623, "learning_rate": 5e-05, "loss": 0.7088, "mean_token_accuracy": 0.8671096563339233, "num_tokens": 579772879.0, "step": 15273 }, { "epoch": 1.9430097951914513, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 9.136735916137695, "learning_rate": 5e-05, "loss": 0.8994, "mean_token_accuracy": 0.8682079315185547, "num_tokens": 579812854.0, "step": 15274 }, { "epoch": 1.9431370054700419, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.172461986541748, "learning_rate": 5e-05, "loss": 0.7181, "mean_token_accuracy": 0.8651705980300903, "num_tokens": 579857324.0, "step": 15275 }, { "epoch": 1.9432642157486324, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.778783798217773, "learning_rate": 5e-05, "loss": 0.9731, "mean_token_accuracy": 0.8740668296813965, "num_tokens": 579897103.0, "step": 15276 }, { "epoch": 1.943391426027223, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.222357273101807, "learning_rate": 5e-05, "loss": 0.7474, "mean_token_accuracy": 0.86488938331604, "num_tokens": 579937382.0, "step": 15277 }, { "epoch": 1.9435186363058135, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.203466415405273, "learning_rate": 5e-05, "loss": 1.0647, "mean_token_accuracy": 0.8763318061828613, "num_tokens": 579978591.0, "step": 15278 }, { "epoch": 1.943645846584404, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.321771144866943, "learning_rate": 5e-05, "loss": 0.9211, "mean_token_accuracy": 0.8741458058357239, "num_tokens": 580014354.0, "step": 15279 }, { "epoch": 1.9437730568629945, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.555967330932617, "learning_rate": 5e-05, "loss": 1.0121, "mean_token_accuracy": 0.8774694800376892, "num_tokens": 580055575.0, "step": 15280 }, { "epoch": 1.943900267141585, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.53151798248291, "learning_rate": 5e-05, "loss": 1.0503, "mean_token_accuracy": 0.8640491962432861, "num_tokens": 580105007.0, "step": 15281 }, { "epoch": 1.9440274774201756, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.883307933807373, "learning_rate": 5e-05, "loss": 0.8926, "mean_token_accuracy": 0.8647738099098206, "num_tokens": 580150820.0, "step": 15282 }, { "epoch": 1.944154687698766, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.210092544555664, "learning_rate": 5e-05, "loss": 0.9758, "mean_token_accuracy": 0.8796163201332092, "num_tokens": 580188303.0, "step": 15283 }, { "epoch": 1.9442818979773566, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.986082553863525, "learning_rate": 5e-05, "loss": 0.7938, "mean_token_accuracy": 0.8668628334999084, "num_tokens": 580221824.0, "step": 15284 }, { "epoch": 1.9444091082559471, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.241545677185059, "learning_rate": 5e-05, "loss": 0.9872, "mean_token_accuracy": 0.868899941444397, "num_tokens": 580262615.0, "step": 15285 }, { "epoch": 1.9445363185345377, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.303039073944092, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.8775761723518372, "num_tokens": 580293535.0, "step": 15286 }, { "epoch": 1.9446635288131282, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.556668281555176, "learning_rate": 5e-05, "loss": 0.9418, "mean_token_accuracy": 0.8685721158981323, "num_tokens": 580328645.0, "step": 15287 }, { "epoch": 1.9447907390917187, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.815496444702148, "learning_rate": 5e-05, "loss": 0.8832, "mean_token_accuracy": 0.8716940879821777, "num_tokens": 580365157.0, "step": 15288 }, { "epoch": 1.9449179493703093, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.895027160644531, "learning_rate": 5e-05, "loss": 0.8065, "mean_token_accuracy": 0.8768907785415649, "num_tokens": 580398195.0, "step": 15289 }, { "epoch": 1.9450451596488998, "ewc_loss": 0.5, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.920226573944092, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8670724034309387, "num_tokens": 580440238.0, "step": 15290 }, { "epoch": 1.94517236992749, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.657519817352295, "learning_rate": 5e-05, "loss": 0.8218, "mean_token_accuracy": 0.864494800567627, "num_tokens": 580477997.0, "step": 15291 }, { "epoch": 1.9452995802060806, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.842828750610352, "learning_rate": 5e-05, "loss": 0.8235, "mean_token_accuracy": 0.8736488819122314, "num_tokens": 580516335.0, "step": 15292 }, { "epoch": 1.9454267904846712, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.587142467498779, "learning_rate": 5e-05, "loss": 0.7665, "mean_token_accuracy": 0.8819010257720947, "num_tokens": 580546492.0, "step": 15293 }, { "epoch": 1.9455540007632617, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.831394672393799, "learning_rate": 5e-05, "loss": 0.7645, "mean_token_accuracy": 0.884100079536438, "num_tokens": 580583804.0, "step": 15294 }, { "epoch": 1.9456812110418522, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.463796615600586, "learning_rate": 5e-05, "loss": 0.7944, "mean_token_accuracy": 0.8604443073272705, "num_tokens": 580618111.0, "step": 15295 }, { "epoch": 1.9458084213204427, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.889114856719971, "learning_rate": 5e-05, "loss": 0.8019, "mean_token_accuracy": 0.8698543310165405, "num_tokens": 580657058.0, "step": 15296 }, { "epoch": 1.945935631599033, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.3112993240356445, "learning_rate": 5e-05, "loss": 0.7673, "mean_token_accuracy": 0.8623266220092773, "num_tokens": 580701628.0, "step": 15297 }, { "epoch": 1.9460628418776236, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.046031951904297, "learning_rate": 5e-05, "loss": 0.7826, "mean_token_accuracy": 0.8762785196304321, "num_tokens": 580738766.0, "step": 15298 }, { "epoch": 1.946190052156214, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.081049919128418, "learning_rate": 5e-05, "loss": 0.7256, "mean_token_accuracy": 0.8720565438270569, "num_tokens": 580780110.0, "step": 15299 }, { "epoch": 1.9463172624348046, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.285451889038086, "learning_rate": 5e-05, "loss": 0.7952, "mean_token_accuracy": 0.8771530389785767, "num_tokens": 580825208.0, "step": 15300 }, { "epoch": 1.9464444727133952, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.8273444175720215, "learning_rate": 5e-05, "loss": 0.7743, "mean_token_accuracy": 0.8529841899871826, "num_tokens": 580870339.0, "step": 15301 }, { "epoch": 1.9465716829919857, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.63608455657959, "learning_rate": 5e-05, "loss": 0.8175, "mean_token_accuracy": 0.8803280591964722, "num_tokens": 580899509.0, "step": 15302 }, { "epoch": 1.9466988932705762, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.534316062927246, "learning_rate": 5e-05, "loss": 0.7029, "mean_token_accuracy": 0.8727394342422485, "num_tokens": 580941345.0, "step": 15303 }, { "epoch": 1.9468261035491667, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 9.0939359664917, "learning_rate": 5e-05, "loss": 0.8877, "mean_token_accuracy": 0.8778461813926697, "num_tokens": 580976819.0, "step": 15304 }, { "epoch": 1.9469533138277573, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.3364949226379395, "learning_rate": 5e-05, "loss": 0.6973, "mean_token_accuracy": 0.8784202337265015, "num_tokens": 581011971.0, "step": 15305 }, { "epoch": 1.9470805241063478, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.489752769470215, "learning_rate": 5e-05, "loss": 1.019, "mean_token_accuracy": 0.8567078709602356, "num_tokens": 581045871.0, "step": 15306 }, { "epoch": 1.9472077343849383, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 6.659934043884277, "learning_rate": 5e-05, "loss": 0.7665, "mean_token_accuracy": 0.881179690361023, "num_tokens": 581077986.0, "step": 15307 }, { "epoch": 1.9473349446635289, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.295238494873047, "learning_rate": 5e-05, "loss": 0.9749, "mean_token_accuracy": 0.8735504150390625, "num_tokens": 581112988.0, "step": 15308 }, { "epoch": 1.9474621549421194, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000579833984375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 22.210729598999023, "learning_rate": 5e-05, "loss": 1.0292, "mean_token_accuracy": 0.8459713459014893, "num_tokens": 581155314.0, "step": 15309 }, { "epoch": 1.94758936522071, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000583648681640625, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.5908203125, "learning_rate": 5e-05, "loss": 1.0617, "mean_token_accuracy": 0.8767244815826416, "num_tokens": 581203622.0, "step": 15310 }, { "epoch": 1.9477165754993004, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000591278076171875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 9.370926856994629, "learning_rate": 5e-05, "loss": 0.8938, "mean_token_accuracy": 0.8739632368087769, "num_tokens": 581234306.0, "step": 15311 }, { "epoch": 1.947843785777891, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00060272216796875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 6.8947434425354, "learning_rate": 5e-05, "loss": 0.8749, "mean_token_accuracy": 0.8537158966064453, "num_tokens": 581275884.0, "step": 15312 }, { "epoch": 1.9479709960564815, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.0006103515625, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 11.69050407409668, "learning_rate": 5e-05, "loss": 1.083, "mean_token_accuracy": 0.8840998411178589, "num_tokens": 581312674.0, "step": 15313 }, { "epoch": 1.948098206335072, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.448153972625732, "learning_rate": 5e-05, "loss": 0.9645, "mean_token_accuracy": 0.8631046414375305, "num_tokens": 581352400.0, "step": 15314 }, { "epoch": 1.9482254166136623, "ewc_loss": 0.95703125, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.00089263916015625, "grad_norm": 13.015718460083008, "learning_rate": 5e-05, "loss": 1.3521, "mean_token_accuracy": 0.8653559684753418, "num_tokens": 581390136.0, "step": 15315 }, { "epoch": 1.9483526268922529, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 9.493672370910645, "learning_rate": 5e-05, "loss": 1.1875, "mean_token_accuracy": 0.8802935481071472, "num_tokens": 581424864.0, "step": 15316 }, { "epoch": 1.9484798371708434, "ewc_loss": 0.96875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000904083251953125, "grad_norm": 13.147350311279297, "learning_rate": 5e-05, "loss": 1.4046, "mean_token_accuracy": 0.8540271520614624, "num_tokens": 581467791.0, "step": 15317 }, { "epoch": 1.948607047449434, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 10.572670936584473, "learning_rate": 5e-05, "loss": 1.2733, "mean_token_accuracy": 0.871727466583252, "num_tokens": 581505121.0, "step": 15318 }, { "epoch": 1.9487342577280244, "ewc_loss": 0.98046875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000911712646484375, "grad_norm": 12.844009399414062, "learning_rate": 5e-05, "loss": 1.3384, "mean_token_accuracy": 0.8740515112876892, "num_tokens": 581543306.0, "step": 15319 }, { "epoch": 1.948861468006615, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 10.764408111572266, "learning_rate": 5e-05, "loss": 1.19, "mean_token_accuracy": 0.8673893213272095, "num_tokens": 581573476.0, "step": 15320 }, { "epoch": 1.9489886782852053, "ewc_loss": 0.91796875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00084686279296875, "grad_norm": 12.27354621887207, "learning_rate": 5e-05, "loss": 1.2592, "mean_token_accuracy": 0.881949245929718, "num_tokens": 581606693.0, "step": 15321 }, { "epoch": 1.9491158885637958, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.65372371673584, "learning_rate": 5e-05, "loss": 1.1469, "mean_token_accuracy": 0.8704133629798889, "num_tokens": 581637826.0, "step": 15322 }, { "epoch": 1.9492430988423863, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 11.582549095153809, "learning_rate": 5e-05, "loss": 1.1377, "mean_token_accuracy": 0.8833507895469666, "num_tokens": 581670590.0, "step": 15323 }, { "epoch": 1.9493703091209769, "ewc_loss": 0.703125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.181654930114746, "learning_rate": 5e-05, "loss": 1.0811, "mean_token_accuracy": 0.8711201548576355, "num_tokens": 581704091.0, "step": 15324 }, { "epoch": 1.9494975193995674, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 11.068656921386719, "learning_rate": 5e-05, "loss": 1.0949, "mean_token_accuracy": 0.8766031265258789, "num_tokens": 581747362.0, "step": 15325 }, { "epoch": 1.949624729678158, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.470380783081055, "learning_rate": 5e-05, "loss": 1.0128, "mean_token_accuracy": 0.8733921051025391, "num_tokens": 581781105.0, "step": 15326 }, { "epoch": 1.9497519399567484, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000736236572265625, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 11.029215812683105, "learning_rate": 5e-05, "loss": 1.0862, "mean_token_accuracy": 0.8750399351119995, "num_tokens": 581824800.0, "step": 15327 }, { "epoch": 1.949879150235339, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000736236572265625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.975428581237793, "learning_rate": 5e-05, "loss": 0.9878, "mean_token_accuracy": 0.8572635054588318, "num_tokens": 581865984.0, "step": 15328 }, { "epoch": 1.9500063605139295, "ewc_loss": 0.765625, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 11.370125770568848, "learning_rate": 5e-05, "loss": 1.1425, "mean_token_accuracy": 0.8725584149360657, "num_tokens": 581903098.0, "step": 15329 }, { "epoch": 1.95013357079252, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.059382438659668, "learning_rate": 5e-05, "loss": 0.9182, "mean_token_accuracy": 0.8850953578948975, "num_tokens": 581940704.0, "step": 15330 }, { "epoch": 1.9502607810711106, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.000732421875, "grad_norm": 11.703280448913574, "learning_rate": 5e-05, "loss": 1.2182, "mean_token_accuracy": 0.8592941164970398, "num_tokens": 581983598.0, "step": 15331 }, { "epoch": 1.950387991349701, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.77751636505127, "learning_rate": 5e-05, "loss": 1.02, "mean_token_accuracy": 0.8733290433883667, "num_tokens": 582020073.0, "step": 15332 }, { "epoch": 1.9505152016282916, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000736236572265625, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 11.178893089294434, "learning_rate": 5e-05, "loss": 1.059, "mean_token_accuracy": 0.884417712688446, "num_tokens": 582056597.0, "step": 15333 }, { "epoch": 1.9506424119068821, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000736236572265625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.903423309326172, "learning_rate": 5e-05, "loss": 0.9897, "mean_token_accuracy": 0.8854592442512512, "num_tokens": 582090164.0, "step": 15334 }, { "epoch": 1.9507696221854727, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000736236572265625, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 10.390527725219727, "learning_rate": 5e-05, "loss": 1.0142, "mean_token_accuracy": 0.8790525197982788, "num_tokens": 582130553.0, "step": 15335 }, { "epoch": 1.9508968324640632, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.388954162597656, "learning_rate": 5e-05, "loss": 0.9419, "mean_token_accuracy": 0.8785492181777954, "num_tokens": 582161175.0, "step": 15336 }, { "epoch": 1.9510240427426537, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.91866397857666, "learning_rate": 5e-05, "loss": 0.9949, "mean_token_accuracy": 0.8731759190559387, "num_tokens": 582194424.0, "step": 15337 }, { "epoch": 1.9511512530212443, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.681353569030762, "learning_rate": 5e-05, "loss": 0.8938, "mean_token_accuracy": 0.8681647777557373, "num_tokens": 582231446.0, "step": 15338 }, { "epoch": 1.9512784632998348, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.968562126159668, "learning_rate": 5e-05, "loss": 1.017, "mean_token_accuracy": 0.8678017854690552, "num_tokens": 582270486.0, "step": 15339 }, { "epoch": 1.951405673578425, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.1673002243042, "learning_rate": 5e-05, "loss": 0.8163, "mean_token_accuracy": 0.8736926317214966, "num_tokens": 582303583.0, "step": 15340 }, { "epoch": 1.9515328838570156, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.442319869995117, "learning_rate": 5e-05, "loss": 1.0472, "mean_token_accuracy": 0.8707559704780579, "num_tokens": 582344982.0, "step": 15341 }, { "epoch": 1.9516600941356061, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.35456657409668, "learning_rate": 5e-05, "loss": 0.8241, "mean_token_accuracy": 0.8802592158317566, "num_tokens": 582383373.0, "step": 15342 }, { "epoch": 1.9517873044141967, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.676750183105469, "learning_rate": 5e-05, "loss": 1.0683, "mean_token_accuracy": 0.8726667165756226, "num_tokens": 582426671.0, "step": 15343 }, { "epoch": 1.9519145146927872, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.138238906860352, "learning_rate": 5e-05, "loss": 0.9627, "mean_token_accuracy": 0.8644712567329407, "num_tokens": 582464801.0, "step": 15344 }, { "epoch": 1.9520417249713777, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.808835983276367, "learning_rate": 5e-05, "loss": 0.9865, "mean_token_accuracy": 0.8697535991668701, "num_tokens": 582507627.0, "step": 15345 }, { "epoch": 1.952168935249968, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 9.11458969116211, "learning_rate": 5e-05, "loss": 0.8915, "mean_token_accuracy": 0.8789448738098145, "num_tokens": 582544072.0, "step": 15346 }, { "epoch": 1.9522961455285586, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.886309623718262, "learning_rate": 5e-05, "loss": 0.9063, "mean_token_accuracy": 0.8668608665466309, "num_tokens": 582587928.0, "step": 15347 }, { "epoch": 1.952423355807149, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.565046310424805, "learning_rate": 5e-05, "loss": 0.8481, "mean_token_accuracy": 0.8718048930168152, "num_tokens": 582622182.0, "step": 15348 }, { "epoch": 1.9525505660857396, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.486778259277344, "learning_rate": 5e-05, "loss": 0.794, "mean_token_accuracy": 0.8889398574829102, "num_tokens": 582654890.0, "step": 15349 }, { "epoch": 1.9526777763643302, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.940128326416016, "learning_rate": 5e-05, "loss": 0.7666, "mean_token_accuracy": 0.8761081695556641, "num_tokens": 582690874.0, "step": 15350 }, { "epoch": 1.9528049866429207, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.443028450012207, "learning_rate": 5e-05, "loss": 0.8075, "mean_token_accuracy": 0.8760875463485718, "num_tokens": 582726543.0, "step": 15351 }, { "epoch": 1.9529321969215112, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.564105033874512, "learning_rate": 5e-05, "loss": 0.7804, "mean_token_accuracy": 0.8610265254974365, "num_tokens": 582764552.0, "step": 15352 }, { "epoch": 1.9530594072001017, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.456575393676758, "learning_rate": 5e-05, "loss": 0.8568, "mean_token_accuracy": 0.8595396280288696, "num_tokens": 582804613.0, "step": 15353 }, { "epoch": 1.9531866174786923, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.346042633056641, "learning_rate": 5e-05, "loss": 0.7155, "mean_token_accuracy": 0.8762193918228149, "num_tokens": 582841510.0, "step": 15354 }, { "epoch": 1.9533138277572828, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.632765769958496, "learning_rate": 5e-05, "loss": 0.9199, "mean_token_accuracy": 0.8452333211898804, "num_tokens": 582879371.0, "step": 15355 }, { "epoch": 1.9534410380358733, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.083817481994629, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.8643496036529541, "num_tokens": 582915354.0, "step": 15356 }, { "epoch": 1.9535682483144639, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 9.032768249511719, "learning_rate": 5e-05, "loss": 0.8584, "mean_token_accuracy": 0.8782105445861816, "num_tokens": 582949897.0, "step": 15357 }, { "epoch": 1.9536954585930544, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.00596809387207, "learning_rate": 5e-05, "loss": 0.7589, "mean_token_accuracy": 0.8606720566749573, "num_tokens": 582984383.0, "step": 15358 }, { "epoch": 1.953822668871645, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 9.308162689208984, "learning_rate": 5e-05, "loss": 0.9485, "mean_token_accuracy": 0.8681260347366333, "num_tokens": 583023146.0, "step": 15359 }, { "epoch": 1.9539498791502354, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.48180627822876, "learning_rate": 5e-05, "loss": 0.7186, "mean_token_accuracy": 0.8822201490402222, "num_tokens": 583061485.0, "step": 15360 }, { "epoch": 1.954077089428826, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.805668830871582, "learning_rate": 5e-05, "loss": 0.8882, "mean_token_accuracy": 0.8678677082061768, "num_tokens": 583097278.0, "step": 15361 }, { "epoch": 1.9542042997074165, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.1477632522583, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.877646803855896, "num_tokens": 583140458.0, "step": 15362 }, { "epoch": 1.954331509986007, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.687228202819824, "learning_rate": 5e-05, "loss": 0.7499, "mean_token_accuracy": 0.8784221410751343, "num_tokens": 583184798.0, "step": 15363 }, { "epoch": 1.9544587202645973, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.583625793457031, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.8804240822792053, "num_tokens": 583220517.0, "step": 15364 }, { "epoch": 1.9545859305431879, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.869144916534424, "learning_rate": 5e-05, "loss": 0.7514, "mean_token_accuracy": 0.8613582849502563, "num_tokens": 583256615.0, "step": 15365 }, { "epoch": 1.9547131408217784, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 9.250622749328613, "learning_rate": 5e-05, "loss": 0.8882, "mean_token_accuracy": 0.8668674826622009, "num_tokens": 583294533.0, "step": 15366 }, { "epoch": 1.954840351100369, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.496777057647705, "learning_rate": 5e-05, "loss": 0.658, "mean_token_accuracy": 0.8821784257888794, "num_tokens": 583333610.0, "step": 15367 }, { "epoch": 1.9549675613789594, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.953725814819336, "learning_rate": 5e-05, "loss": 0.9729, "mean_token_accuracy": 0.8780117034912109, "num_tokens": 583366424.0, "step": 15368 }, { "epoch": 1.95509477165755, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.326032638549805, "learning_rate": 5e-05, "loss": 0.7756, "mean_token_accuracy": 0.8698540925979614, "num_tokens": 583403939.0, "step": 15369 }, { "epoch": 1.9552219819361403, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.400604248046875, "learning_rate": 5e-05, "loss": 0.946, "mean_token_accuracy": 0.8717364072799683, "num_tokens": 583441053.0, "step": 15370 }, { "epoch": 1.9553491922147308, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.577068328857422, "learning_rate": 5e-05, "loss": 0.8055, "mean_token_accuracy": 0.8929716348648071, "num_tokens": 583471168.0, "step": 15371 }, { "epoch": 1.9554764024933213, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.932572364807129, "learning_rate": 5e-05, "loss": 0.8228, "mean_token_accuracy": 0.8653653860092163, "num_tokens": 583510604.0, "step": 15372 }, { "epoch": 1.9556036127719119, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.818137168884277, "learning_rate": 5e-05, "loss": 0.8792, "mean_token_accuracy": 0.8730688095092773, "num_tokens": 583546186.0, "step": 15373 }, { "epoch": 1.9557308230505024, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.583524703979492, "learning_rate": 5e-05, "loss": 0.8073, "mean_token_accuracy": 0.8579316139221191, "num_tokens": 583581301.0, "step": 15374 }, { "epoch": 1.955858033329093, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.176020622253418, "learning_rate": 5e-05, "loss": 0.8665, "mean_token_accuracy": 0.8556344509124756, "num_tokens": 583624005.0, "step": 15375 }, { "epoch": 1.9559852436076834, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.06554889678955, "learning_rate": 5e-05, "loss": 0.7785, "mean_token_accuracy": 0.8709689378738403, "num_tokens": 583662738.0, "step": 15376 }, { "epoch": 1.956112453886274, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.351109981536865, "learning_rate": 5e-05, "loss": 0.7147, "mean_token_accuracy": 0.8801038265228271, "num_tokens": 583700742.0, "step": 15377 }, { "epoch": 1.9562396641648645, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.474299430847168, "learning_rate": 5e-05, "loss": 0.7865, "mean_token_accuracy": 0.8790299892425537, "num_tokens": 583740529.0, "step": 15378 }, { "epoch": 1.956366874443455, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.073254585266113, "learning_rate": 5e-05, "loss": 0.6759, "mean_token_accuracy": 0.8839995265007019, "num_tokens": 583776866.0, "step": 15379 }, { "epoch": 1.9564940847220456, "ewc_loss": 0.453125, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.507072448730469, "learning_rate": 5e-05, "loss": 0.8092, "mean_token_accuracy": 0.8753213882446289, "num_tokens": 583815561.0, "step": 15380 }, { "epoch": 1.956621295000636, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.158975601196289, "learning_rate": 5e-05, "loss": 0.6921, "mean_token_accuracy": 0.8790181875228882, "num_tokens": 583846701.0, "step": 15381 }, { "epoch": 1.9567485052792266, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.405291557312012, "learning_rate": 5e-05, "loss": 0.7774, "mean_token_accuracy": 0.8842209577560425, "num_tokens": 583885561.0, "step": 15382 }, { "epoch": 1.9568757155578171, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.193212032318115, "learning_rate": 5e-05, "loss": 0.7349, "mean_token_accuracy": 0.8662525415420532, "num_tokens": 583928662.0, "step": 15383 }, { "epoch": 1.9570029258364077, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.43009090423584, "learning_rate": 5e-05, "loss": 0.7944, "mean_token_accuracy": 0.8785277009010315, "num_tokens": 583969097.0, "step": 15384 }, { "epoch": 1.9571301361149982, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.098509311676025, "learning_rate": 5e-05, "loss": 0.7319, "mean_token_accuracy": 0.8668581247329712, "num_tokens": 584008547.0, "step": 15385 }, { "epoch": 1.9572573463935887, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.551717758178711, "learning_rate": 5e-05, "loss": 0.904, "mean_token_accuracy": 0.8529719710350037, "num_tokens": 584046982.0, "step": 15386 }, { "epoch": 1.9573845566721793, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.966317176818848, "learning_rate": 5e-05, "loss": 0.6746, "mean_token_accuracy": 0.8824495673179626, "num_tokens": 584081011.0, "step": 15387 }, { "epoch": 1.9575117669507698, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.73983383178711, "learning_rate": 5e-05, "loss": 0.8899, "mean_token_accuracy": 0.8603867292404175, "num_tokens": 584115550.0, "step": 15388 }, { "epoch": 1.95763897722936, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.841585159301758, "learning_rate": 5e-05, "loss": 0.6938, "mean_token_accuracy": 0.88093501329422, "num_tokens": 584153733.0, "step": 15389 }, { "epoch": 1.9577661875079506, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.927518844604492, "learning_rate": 5e-05, "loss": 0.8876, "mean_token_accuracy": 0.8683795928955078, "num_tokens": 584182097.0, "step": 15390 }, { "epoch": 1.9578933977865411, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.952159881591797, "learning_rate": 5e-05, "loss": 0.6618, "mean_token_accuracy": 0.892505407333374, "num_tokens": 584215089.0, "step": 15391 }, { "epoch": 1.9580206080651317, "ewc_loss": 0.5, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.71951961517334, "learning_rate": 5e-05, "loss": 0.8244, "mean_token_accuracy": 0.887181282043457, "num_tokens": 584244635.0, "step": 15392 }, { "epoch": 1.9581478183437222, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.39317512512207, "learning_rate": 5e-05, "loss": 0.7404, "mean_token_accuracy": 0.879177987575531, "num_tokens": 584280267.0, "step": 15393 }, { "epoch": 1.9582750286223127, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.103860855102539, "learning_rate": 5e-05, "loss": 0.7662, "mean_token_accuracy": 0.8853887319564819, "num_tokens": 584311998.0, "step": 15394 }, { "epoch": 1.958402238900903, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.892037868499756, "learning_rate": 5e-05, "loss": 0.8059, "mean_token_accuracy": 0.8707866072654724, "num_tokens": 584342975.0, "step": 15395 }, { "epoch": 1.9585294491794936, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.491275310516357, "learning_rate": 5e-05, "loss": 0.7444, "mean_token_accuracy": 0.8715407848358154, "num_tokens": 584379087.0, "step": 15396 }, { "epoch": 1.958656659458084, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.22903823852539, "learning_rate": 5e-05, "loss": 0.8242, "mean_token_accuracy": 0.8656443953514099, "num_tokens": 584420126.0, "step": 15397 }, { "epoch": 1.9587838697366746, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.072664260864258, "learning_rate": 5e-05, "loss": 0.7184, "mean_token_accuracy": 0.8746967315673828, "num_tokens": 584457742.0, "step": 15398 }, { "epoch": 1.9589110800152651, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.555170059204102, "learning_rate": 5e-05, "loss": 0.8103, "mean_token_accuracy": 0.8761389255523682, "num_tokens": 584497439.0, "step": 15399 }, { "epoch": 1.9590382902938557, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.733632564544678, "learning_rate": 5e-05, "loss": 0.678, "mean_token_accuracy": 0.8815109133720398, "num_tokens": 584532922.0, "step": 15400 }, { "epoch": 1.9591655005724462, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 9.00928783416748, "learning_rate": 5e-05, "loss": 0.8498, "mean_token_accuracy": 0.8829883337020874, "num_tokens": 584574834.0, "step": 15401 }, { "epoch": 1.9592927108510367, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.508650302886963, "learning_rate": 5e-05, "loss": 0.7093, "mean_token_accuracy": 0.8731743097305298, "num_tokens": 584615494.0, "step": 15402 }, { "epoch": 1.9594199211296273, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.333471298217773, "learning_rate": 5e-05, "loss": 0.9537, "mean_token_accuracy": 0.862864077091217, "num_tokens": 584651494.0, "step": 15403 }, { "epoch": 1.9595471314082178, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.772556781768799, "learning_rate": 5e-05, "loss": 0.7376, "mean_token_accuracy": 0.8641700744628906, "num_tokens": 584691556.0, "step": 15404 }, { "epoch": 1.9596743416868083, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.076858520507812, "learning_rate": 5e-05, "loss": 0.998, "mean_token_accuracy": 0.8518184423446655, "num_tokens": 584735960.0, "step": 15405 }, { "epoch": 1.9598015519653988, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.518787860870361, "learning_rate": 5e-05, "loss": 0.8359, "mean_token_accuracy": 0.8648912906646729, "num_tokens": 584770808.0, "step": 15406 }, { "epoch": 1.9599287622439894, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.220000267028809, "learning_rate": 5e-05, "loss": 0.8471, "mean_token_accuracy": 0.8746123909950256, "num_tokens": 584811451.0, "step": 15407 }, { "epoch": 1.96005597252258, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.08913803100586, "learning_rate": 5e-05, "loss": 0.8509, "mean_token_accuracy": 0.8687149286270142, "num_tokens": 584845050.0, "step": 15408 }, { "epoch": 1.9601831828011704, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.580521583557129, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8736746311187744, "num_tokens": 584883857.0, "step": 15409 }, { "epoch": 1.960310393079761, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.077184677124023, "learning_rate": 5e-05, "loss": 0.837, "mean_token_accuracy": 0.8677504062652588, "num_tokens": 584918118.0, "step": 15410 }, { "epoch": 1.9604376033583515, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.709356784820557, "learning_rate": 5e-05, "loss": 0.7949, "mean_token_accuracy": 0.8704108595848083, "num_tokens": 584960789.0, "step": 15411 }, { "epoch": 1.960564813636942, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.512214660644531, "learning_rate": 5e-05, "loss": 0.7124, "mean_token_accuracy": 0.8864270448684692, "num_tokens": 585000631.0, "step": 15412 }, { "epoch": 1.9606920239155323, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.262338638305664, "learning_rate": 5e-05, "loss": 0.8145, "mean_token_accuracy": 0.876045286655426, "num_tokens": 585036615.0, "step": 15413 }, { "epoch": 1.9608192341941229, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.897984504699707, "learning_rate": 5e-05, "loss": 0.6828, "mean_token_accuracy": 0.8833441734313965, "num_tokens": 585075076.0, "step": 15414 }, { "epoch": 1.9609464444727134, "ewc_loss": 0.5, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.820631980895996, "learning_rate": 5e-05, "loss": 0.8466, "mean_token_accuracy": 0.8805925846099854, "num_tokens": 585112215.0, "step": 15415 }, { "epoch": 1.961073654751304, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.5888261795043945, "learning_rate": 5e-05, "loss": 0.6887, "mean_token_accuracy": 0.8764914870262146, "num_tokens": 585148920.0, "step": 15416 }, { "epoch": 1.9612008650298944, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 9.117619514465332, "learning_rate": 5e-05, "loss": 0.9392, "mean_token_accuracy": 0.8643639087677002, "num_tokens": 585191128.0, "step": 15417 }, { "epoch": 1.961328075308485, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.722119331359863, "learning_rate": 5e-05, "loss": 0.728, "mean_token_accuracy": 0.8682446479797363, "num_tokens": 585234130.0, "step": 15418 }, { "epoch": 1.9614552855870753, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.986055374145508, "learning_rate": 5e-05, "loss": 0.9212, "mean_token_accuracy": 0.8696243762969971, "num_tokens": 585267089.0, "step": 15419 }, { "epoch": 1.9615824958656658, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.252167701721191, "learning_rate": 5e-05, "loss": 0.7751, "mean_token_accuracy": 0.8771549463272095, "num_tokens": 585310508.0, "step": 15420 }, { "epoch": 1.9617097061442563, "ewc_loss": 0.5, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.355072975158691, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8780776858329773, "num_tokens": 585346579.0, "step": 15421 }, { "epoch": 1.9618369164228469, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.868082046508789, "learning_rate": 5e-05, "loss": 0.8765, "mean_token_accuracy": 0.86647629737854, "num_tokens": 585383914.0, "step": 15422 }, { "epoch": 1.9619641267014374, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.5826029777526855, "learning_rate": 5e-05, "loss": 0.8097, "mean_token_accuracy": 0.8639899492263794, "num_tokens": 585419322.0, "step": 15423 }, { "epoch": 1.962091336980028, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.231854438781738, "learning_rate": 5e-05, "loss": 0.821, "mean_token_accuracy": 0.8780532479286194, "num_tokens": 585459081.0, "step": 15424 }, { "epoch": 1.9622185472586184, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.187716960906982, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8676085472106934, "num_tokens": 585497116.0, "step": 15425 }, { "epoch": 1.962345757537209, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.360726356506348, "learning_rate": 5e-05, "loss": 0.8529, "mean_token_accuracy": 0.861642599105835, "num_tokens": 585534704.0, "step": 15426 }, { "epoch": 1.9624729678157995, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.049314022064209, "learning_rate": 5e-05, "loss": 0.7073, "mean_token_accuracy": 0.8788566589355469, "num_tokens": 585575806.0, "step": 15427 }, { "epoch": 1.96260017809439, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.506237983703613, "learning_rate": 5e-05, "loss": 0.8422, "mean_token_accuracy": 0.8785350322723389, "num_tokens": 585612585.0, "step": 15428 }, { "epoch": 1.9627273883729806, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.858847618103027, "learning_rate": 5e-05, "loss": 0.6953, "mean_token_accuracy": 0.8778886795043945, "num_tokens": 585647414.0, "step": 15429 }, { "epoch": 1.962854598651571, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.802470207214355, "learning_rate": 5e-05, "loss": 0.8939, "mean_token_accuracy": 0.8691292405128479, "num_tokens": 585693802.0, "step": 15430 }, { "epoch": 1.9629818089301616, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.669334888458252, "learning_rate": 5e-05, "loss": 0.7034, "mean_token_accuracy": 0.8740781545639038, "num_tokens": 585727955.0, "step": 15431 }, { "epoch": 1.9631090192087521, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.080846786499023, "learning_rate": 5e-05, "loss": 0.9571, "mean_token_accuracy": 0.8612238168716431, "num_tokens": 585769784.0, "step": 15432 }, { "epoch": 1.9632362294873427, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.776855945587158, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.878389298915863, "num_tokens": 585809989.0, "step": 15433 }, { "epoch": 1.9633634397659332, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.977571487426758, "learning_rate": 5e-05, "loss": 0.9369, "mean_token_accuracy": 0.8719481825828552, "num_tokens": 585851054.0, "step": 15434 }, { "epoch": 1.9634906500445237, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.259060382843018, "learning_rate": 5e-05, "loss": 0.847, "mean_token_accuracy": 0.8594951629638672, "num_tokens": 585886638.0, "step": 15435 }, { "epoch": 1.9636178603231143, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.414299011230469, "learning_rate": 5e-05, "loss": 0.9211, "mean_token_accuracy": 0.8643980026245117, "num_tokens": 585925572.0, "step": 15436 }, { "epoch": 1.9637450706017048, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.717772960662842, "learning_rate": 5e-05, "loss": 0.8308, "mean_token_accuracy": 0.8725277185440063, "num_tokens": 585961939.0, "step": 15437 }, { "epoch": 1.963872280880295, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.900362014770508, "learning_rate": 5e-05, "loss": 0.8455, "mean_token_accuracy": 0.8678317070007324, "num_tokens": 585996804.0, "step": 15438 }, { "epoch": 1.9639994911588856, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000637054443359375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.788112640380859, "learning_rate": 5e-05, "loss": 0.79, "mean_token_accuracy": 0.880930483341217, "num_tokens": 586040347.0, "step": 15439 }, { "epoch": 1.9641267014374761, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.884138107299805, "learning_rate": 5e-05, "loss": 0.8012, "mean_token_accuracy": 0.8791283369064331, "num_tokens": 586078495.0, "step": 15440 }, { "epoch": 1.9642539117160667, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.471092700958252, "learning_rate": 5e-05, "loss": 0.7473, "mean_token_accuracy": 0.8794666528701782, "num_tokens": 586114743.0, "step": 15441 }, { "epoch": 1.9643811219946572, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.20448112487793, "learning_rate": 5e-05, "loss": 0.8182, "mean_token_accuracy": 0.8779312372207642, "num_tokens": 586150462.0, "step": 15442 }, { "epoch": 1.9645083322732477, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.029646396636963, "learning_rate": 5e-05, "loss": 0.7335, "mean_token_accuracy": 0.8712565302848816, "num_tokens": 586182860.0, "step": 15443 }, { "epoch": 1.964635542551838, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.581155776977539, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.8686648607254028, "num_tokens": 586226438.0, "step": 15444 }, { "epoch": 1.9647627528304286, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.822286128997803, "learning_rate": 5e-05, "loss": 0.7092, "mean_token_accuracy": 0.8739132881164551, "num_tokens": 586267022.0, "step": 15445 }, { "epoch": 1.964889963109019, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.75637435913086, "learning_rate": 5e-05, "loss": 0.8582, "mean_token_accuracy": 0.8799954056739807, "num_tokens": 586299415.0, "step": 15446 }, { "epoch": 1.9650171733876096, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.886702060699463, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8732659816741943, "num_tokens": 586337351.0, "step": 15447 }, { "epoch": 1.9651443836662001, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.67374038696289, "learning_rate": 5e-05, "loss": 0.9245, "mean_token_accuracy": 0.8642812967300415, "num_tokens": 586378311.0, "step": 15448 }, { "epoch": 1.9652715939447907, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.205453872680664, "learning_rate": 5e-05, "loss": 0.7782, "mean_token_accuracy": 0.8722748160362244, "num_tokens": 586414137.0, "step": 15449 }, { "epoch": 1.9653988042233812, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.246289253234863, "learning_rate": 5e-05, "loss": 0.886, "mean_token_accuracy": 0.8675031065940857, "num_tokens": 586451536.0, "step": 15450 }, { "epoch": 1.9655260145019717, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.637789726257324, "learning_rate": 5e-05, "loss": 0.8174, "mean_token_accuracy": 0.8688801527023315, "num_tokens": 586484643.0, "step": 15451 }, { "epoch": 1.9656532247805623, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.68686580657959, "learning_rate": 5e-05, "loss": 0.8249, "mean_token_accuracy": 0.8631317615509033, "num_tokens": 586523926.0, "step": 15452 }, { "epoch": 1.9657804350591528, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.9733123779296875, "learning_rate": 5e-05, "loss": 0.8104, "mean_token_accuracy": 0.8743700981140137, "num_tokens": 586564999.0, "step": 15453 }, { "epoch": 1.9659076453377433, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.299091339111328, "learning_rate": 5e-05, "loss": 0.7863, "mean_token_accuracy": 0.8646240830421448, "num_tokens": 586603455.0, "step": 15454 }, { "epoch": 1.9660348556163338, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.152592658996582, "learning_rate": 5e-05, "loss": 0.8223, "mean_token_accuracy": 0.8687307238578796, "num_tokens": 586646345.0, "step": 15455 }, { "epoch": 1.9661620658949244, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.11159610748291, "learning_rate": 5e-05, "loss": 0.7502, "mean_token_accuracy": 0.8671606779098511, "num_tokens": 586688065.0, "step": 15456 }, { "epoch": 1.966289276173515, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.265369415283203, "learning_rate": 5e-05, "loss": 0.7906, "mean_token_accuracy": 0.8773797750473022, "num_tokens": 586722631.0, "step": 15457 }, { "epoch": 1.9664164864521054, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.96056604385376, "learning_rate": 5e-05, "loss": 0.6648, "mean_token_accuracy": 0.8898054957389832, "num_tokens": 586762630.0, "step": 15458 }, { "epoch": 1.966543696730696, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.47741413116455, "learning_rate": 5e-05, "loss": 0.8399, "mean_token_accuracy": 0.8725787401199341, "num_tokens": 586799514.0, "step": 15459 }, { "epoch": 1.9666709070092865, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.754678726196289, "learning_rate": 5e-05, "loss": 0.6843, "mean_token_accuracy": 0.8842472434043884, "num_tokens": 586839507.0, "step": 15460 }, { "epoch": 1.966798117287877, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.790647506713867, "learning_rate": 5e-05, "loss": 0.9157, "mean_token_accuracy": 0.8597602844238281, "num_tokens": 586879263.0, "step": 15461 }, { "epoch": 1.9669253275664673, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.565545082092285, "learning_rate": 5e-05, "loss": 0.701, "mean_token_accuracy": 0.8735051155090332, "num_tokens": 586920297.0, "step": 15462 }, { "epoch": 1.9670525378450578, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.042877197265625, "learning_rate": 5e-05, "loss": 0.9467, "mean_token_accuracy": 0.8623304963111877, "num_tokens": 586961171.0, "step": 15463 }, { "epoch": 1.9671797481236484, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.694129467010498, "learning_rate": 5e-05, "loss": 0.7229, "mean_token_accuracy": 0.873863160610199, "num_tokens": 587003808.0, "step": 15464 }, { "epoch": 1.967306958402239, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.914726257324219, "learning_rate": 5e-05, "loss": 0.9671, "mean_token_accuracy": 0.8622697591781616, "num_tokens": 587042056.0, "step": 15465 }, { "epoch": 1.9674341686808294, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.243093967437744, "learning_rate": 5e-05, "loss": 0.8685, "mean_token_accuracy": 0.858065128326416, "num_tokens": 587083009.0, "step": 15466 }, { "epoch": 1.96756137895942, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.204907417297363, "learning_rate": 5e-05, "loss": 0.9355, "mean_token_accuracy": 0.8518086671829224, "num_tokens": 587125129.0, "step": 15467 }, { "epoch": 1.9676885892380103, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.9012908935546875, "learning_rate": 5e-05, "loss": 0.9111, "mean_token_accuracy": 0.8576554656028748, "num_tokens": 587163314.0, "step": 15468 }, { "epoch": 1.9678157995166008, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.382760524749756, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.8701278567314148, "num_tokens": 587208275.0, "step": 15469 }, { "epoch": 1.9679430097951913, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.312637329101562, "learning_rate": 5e-05, "loss": 0.8538, "mean_token_accuracy": 0.8708199858665466, "num_tokens": 587242885.0, "step": 15470 }, { "epoch": 1.9680702200737819, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.053808689117432, "learning_rate": 5e-05, "loss": 0.7687, "mean_token_accuracy": 0.8638880848884583, "num_tokens": 587287787.0, "step": 15471 }, { "epoch": 1.9681974303523724, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.311192512512207, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.8873330354690552, "num_tokens": 587325445.0, "step": 15472 }, { "epoch": 1.968324640630963, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.190158843994141, "learning_rate": 5e-05, "loss": 0.7512, "mean_token_accuracy": 0.8707116842269897, "num_tokens": 587366691.0, "step": 15473 }, { "epoch": 1.9684518509095534, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.027057647705078, "learning_rate": 5e-05, "loss": 0.8365, "mean_token_accuracy": 0.8639127016067505, "num_tokens": 587406310.0, "step": 15474 }, { "epoch": 1.968579061188144, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.42820930480957, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.8689866662025452, "num_tokens": 587446654.0, "step": 15475 }, { "epoch": 1.9687062714667345, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.763727188110352, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.867363691329956, "num_tokens": 587481089.0, "step": 15476 }, { "epoch": 1.968833481745325, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.563905239105225, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.871141791343689, "num_tokens": 587521293.0, "step": 15477 }, { "epoch": 1.9689606920239155, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.621077537536621, "learning_rate": 5e-05, "loss": 0.7879, "mean_token_accuracy": 0.862686276435852, "num_tokens": 587564193.0, "step": 15478 }, { "epoch": 1.969087902302506, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.635684967041016, "learning_rate": 5e-05, "loss": 0.7743, "mean_token_accuracy": 0.8639277219772339, "num_tokens": 587603487.0, "step": 15479 }, { "epoch": 1.9692151125810966, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.533354759216309, "learning_rate": 5e-05, "loss": 0.8005, "mean_token_accuracy": 0.8604038953781128, "num_tokens": 587641131.0, "step": 15480 }, { "epoch": 1.9693423228596871, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.67706298828125, "learning_rate": 5e-05, "loss": 0.813, "mean_token_accuracy": 0.8548921942710876, "num_tokens": 587680012.0, "step": 15481 }, { "epoch": 1.9694695331382777, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.468883514404297, "learning_rate": 5e-05, "loss": 0.7281, "mean_token_accuracy": 0.8730456829071045, "num_tokens": 587721904.0, "step": 15482 }, { "epoch": 1.9695967434168682, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.721332550048828, "learning_rate": 5e-05, "loss": 0.7414, "mean_token_accuracy": 0.874636709690094, "num_tokens": 587752080.0, "step": 15483 }, { "epoch": 1.9697239536954587, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.368924617767334, "learning_rate": 5e-05, "loss": 0.6911, "mean_token_accuracy": 0.8821486830711365, "num_tokens": 587787838.0, "step": 15484 }, { "epoch": 1.9698511639740492, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.8274431228637695, "learning_rate": 5e-05, "loss": 0.7585, "mean_token_accuracy": 0.8728184700012207, "num_tokens": 587826370.0, "step": 15485 }, { "epoch": 1.9699783742526398, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.216993808746338, "learning_rate": 5e-05, "loss": 0.6701, "mean_token_accuracy": 0.8875436186790466, "num_tokens": 587866716.0, "step": 15486 }, { "epoch": 1.97010558453123, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 8.037196159362793, "learning_rate": 5e-05, "loss": 0.755, "mean_token_accuracy": 0.8757176995277405, "num_tokens": 587902766.0, "step": 15487 }, { "epoch": 1.9702327948098206, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.932941913604736, "learning_rate": 5e-05, "loss": 0.7202, "mean_token_accuracy": 0.8607261180877686, "num_tokens": 587944584.0, "step": 15488 }, { "epoch": 1.9703600050884111, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.432575225830078, "learning_rate": 5e-05, "loss": 0.8456, "mean_token_accuracy": 0.8591885566711426, "num_tokens": 587983482.0, "step": 15489 }, { "epoch": 1.9704872153670017, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.456707000732422, "learning_rate": 5e-05, "loss": 0.6551, "mean_token_accuracy": 0.8792752027511597, "num_tokens": 588019621.0, "step": 15490 }, { "epoch": 1.9706144256455922, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 9.186280250549316, "learning_rate": 5e-05, "loss": 0.8605, "mean_token_accuracy": 0.8765408992767334, "num_tokens": 588056845.0, "step": 15491 }, { "epoch": 1.9707416359241827, "ewc_loss": 0.287109375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00022411346435546875, "grad_norm": 5.777312278747559, "learning_rate": 5e-05, "loss": 0.642, "mean_token_accuracy": 0.878315806388855, "num_tokens": 588093731.0, "step": 15492 }, { "epoch": 1.970868846202773, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 10.425603866577148, "learning_rate": 5e-05, "loss": 0.9698, "mean_token_accuracy": 0.8877250552177429, "num_tokens": 588128908.0, "step": 15493 }, { "epoch": 1.9709960564813636, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 5.7826457023620605, "learning_rate": 5e-05, "loss": 0.7526, "mean_token_accuracy": 0.858439028263092, "num_tokens": 588167729.0, "step": 15494 }, { "epoch": 1.971123266759954, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00084686279296875, "grad_norm": 11.523977279663086, "learning_rate": 5e-05, "loss": 1.3306, "mean_token_accuracy": 0.8548035621643066, "num_tokens": 588200337.0, "step": 15495 }, { "epoch": 1.9712504770385446, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 8.244725227355957, "learning_rate": 5e-05, "loss": 1.1235, "mean_token_accuracy": 0.8696032166481018, "num_tokens": 588240736.0, "step": 15496 }, { "epoch": 1.9713776873171351, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.426374435424805, "learning_rate": 5e-05, "loss": 1.0921, "mean_token_accuracy": 0.8620883822441101, "num_tokens": 588275187.0, "step": 15497 }, { "epoch": 1.9715048975957257, "ewc_loss": 0.9296875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000865936279296875, "grad_norm": 9.675518989562988, "learning_rate": 5e-05, "loss": 1.3194, "mean_token_accuracy": 0.8663051128387451, "num_tokens": 588308800.0, "step": 15498 }, { "epoch": 1.9716321078743162, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.129549026489258, "learning_rate": 5e-05, "loss": 0.8384, "mean_token_accuracy": 0.8845171332359314, "num_tokens": 588347688.0, "step": 15499 }, { "epoch": 1.9717593181529067, "ewc_loss": 0.94921875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000885009765625, "grad_norm": 11.269537925720215, "learning_rate": 5e-05, "loss": 1.3122, "mean_token_accuracy": 0.8724830746650696, "num_tokens": 588380791.0, "step": 15500 }, { "epoch": 1.9718865284314973, "ewc_loss": 0.625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 7.946932315826416, "learning_rate": 5e-05, "loss": 1.0287, "mean_token_accuracy": 0.8594604730606079, "num_tokens": 588413667.0, "step": 15501 }, { "epoch": 1.9720137387100878, "ewc_loss": 0.9765625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 12.47607707977295, "learning_rate": 5e-05, "loss": 1.3363, "mean_token_accuracy": 0.8714501857757568, "num_tokens": 588453884.0, "step": 15502 }, { "epoch": 1.9721409489886783, "ewc_loss": 1.0234375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0009613037109375, "grad_norm": 10.080687522888184, "learning_rate": 5e-05, "loss": 1.3978, "mean_token_accuracy": 0.8735468983650208, "num_tokens": 588488895.0, "step": 15503 }, { "epoch": 1.9722681592672688, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 11.56613540649414, "learning_rate": 5e-05, "loss": 1.1194, "mean_token_accuracy": 0.8771654367446899, "num_tokens": 588526807.0, "step": 15504 }, { "epoch": 1.9723953695458594, "ewc_loss": 0.828125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000766754150390625, "grad_norm": 9.445941925048828, "learning_rate": 5e-05, "loss": 1.1891, "mean_token_accuracy": 0.8763634562492371, "num_tokens": 588562272.0, "step": 15505 }, { "epoch": 1.97252257982445, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 11.539467811584473, "learning_rate": 5e-05, "loss": 1.196, "mean_token_accuracy": 0.8665267825126648, "num_tokens": 588598298.0, "step": 15506 }, { "epoch": 1.9726497901030404, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.543994903564453, "learning_rate": 5e-05, "loss": 1.0014, "mean_token_accuracy": 0.8720567226409912, "num_tokens": 588637833.0, "step": 15507 }, { "epoch": 1.972777000381631, "ewc_loss": 0.87109375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0008087158203125, "grad_norm": 11.891082763671875, "learning_rate": 5e-05, "loss": 1.2379, "mean_token_accuracy": 0.8673500418663025, "num_tokens": 588674933.0, "step": 15508 }, { "epoch": 1.9729042106602215, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.79437255859375, "learning_rate": 5e-05, "loss": 1.0269, "mean_token_accuracy": 0.8735159039497375, "num_tokens": 588711899.0, "step": 15509 }, { "epoch": 1.973031420938812, "ewc_loss": 0.91796875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 12.411133766174316, "learning_rate": 5e-05, "loss": 1.2914, "mean_token_accuracy": 0.8697450160980225, "num_tokens": 588749702.0, "step": 15510 }, { "epoch": 1.9731586312174023, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.07259464263916, "learning_rate": 5e-05, "loss": 1.139, "mean_token_accuracy": 0.8780999183654785, "num_tokens": 588787066.0, "step": 15511 }, { "epoch": 1.9732858414959928, "ewc_loss": 0.8125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 11.794733047485352, "learning_rate": 5e-05, "loss": 1.1815, "mean_token_accuracy": 0.8742815256118774, "num_tokens": 588824062.0, "step": 15512 }, { "epoch": 1.9734130517745834, "ewc_loss": 0.765625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.220749855041504, "learning_rate": 5e-05, "loss": 1.1466, "mean_token_accuracy": 0.869437038898468, "num_tokens": 588864858.0, "step": 15513 }, { "epoch": 1.973540262053174, "ewc_loss": 0.734375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 11.051928520202637, "learning_rate": 5e-05, "loss": 1.1151, "mean_token_accuracy": 0.8700530529022217, "num_tokens": 588900219.0, "step": 15514 }, { "epoch": 1.9736674723317644, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.755206108093262, "learning_rate": 5e-05, "loss": 1.0517, "mean_token_accuracy": 0.8670531511306763, "num_tokens": 588937256.0, "step": 15515 }, { "epoch": 1.973794682610355, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.443543434143066, "learning_rate": 5e-05, "loss": 1.0275, "mean_token_accuracy": 0.878244936466217, "num_tokens": 588971263.0, "step": 15516 }, { "epoch": 1.9739218928889453, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.170503616333008, "learning_rate": 5e-05, "loss": 0.9814, "mean_token_accuracy": 0.8680585622787476, "num_tokens": 589010710.0, "step": 15517 }, { "epoch": 1.9740491031675358, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.133995056152344, "learning_rate": 5e-05, "loss": 1.009, "mean_token_accuracy": 0.8753046989440918, "num_tokens": 589043940.0, "step": 15518 }, { "epoch": 1.9741763134461263, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.551251411437988, "learning_rate": 5e-05, "loss": 0.9015, "mean_token_accuracy": 0.8679118752479553, "num_tokens": 589082187.0, "step": 15519 }, { "epoch": 1.9743035237247168, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.2887544631958, "learning_rate": 5e-05, "loss": 1.0057, "mean_token_accuracy": 0.8794530630111694, "num_tokens": 589126870.0, "step": 15520 }, { "epoch": 1.9744307340033074, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.24209976196289, "learning_rate": 5e-05, "loss": 0.8604, "mean_token_accuracy": 0.8685862421989441, "num_tokens": 589168378.0, "step": 15521 }, { "epoch": 1.974557944281898, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.696452140808105, "learning_rate": 5e-05, "loss": 1.121, "mean_token_accuracy": 0.8578389286994934, "num_tokens": 589213070.0, "step": 15522 }, { "epoch": 1.9746851545604884, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.63750171661377, "learning_rate": 5e-05, "loss": 0.9258, "mean_token_accuracy": 0.8701696395874023, "num_tokens": 589253008.0, "step": 15523 }, { "epoch": 1.974812364839079, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.639323234558105, "learning_rate": 5e-05, "loss": 1.0742, "mean_token_accuracy": 0.8665553331375122, "num_tokens": 589291786.0, "step": 15524 }, { "epoch": 1.9749395751176695, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.184248924255371, "learning_rate": 5e-05, "loss": 0.9781, "mean_token_accuracy": 0.8733708262443542, "num_tokens": 589332045.0, "step": 15525 }, { "epoch": 1.97506678539626, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.849364280700684, "learning_rate": 5e-05, "loss": 1.0649, "mean_token_accuracy": 0.8514567613601685, "num_tokens": 589374983.0, "step": 15526 }, { "epoch": 1.9751939956748505, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.015384674072266, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.8948471546173096, "num_tokens": 589419786.0, "step": 15527 }, { "epoch": 1.975321205953441, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.130796432495117, "learning_rate": 5e-05, "loss": 0.8783, "mean_token_accuracy": 0.8913663625717163, "num_tokens": 589454365.0, "step": 15528 }, { "epoch": 1.9754484162320316, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.441486358642578, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.8768771886825562, "num_tokens": 589486829.0, "step": 15529 }, { "epoch": 1.9755756265106221, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.924641609191895, "learning_rate": 5e-05, "loss": 0.8958, "mean_token_accuracy": 0.8734699487686157, "num_tokens": 589527500.0, "step": 15530 }, { "epoch": 1.9757028367892127, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.770084857940674, "learning_rate": 5e-05, "loss": 0.7398, "mean_token_accuracy": 0.8849842548370361, "num_tokens": 589570064.0, "step": 15531 }, { "epoch": 1.9758300470678032, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.177145957946777, "learning_rate": 5e-05, "loss": 0.9703, "mean_token_accuracy": 0.858001172542572, "num_tokens": 589605327.0, "step": 15532 }, { "epoch": 1.9759572573463937, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.407415390014648, "learning_rate": 5e-05, "loss": 0.7131, "mean_token_accuracy": 0.8842117786407471, "num_tokens": 589640886.0, "step": 15533 }, { "epoch": 1.9760844676249842, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.539144515991211, "learning_rate": 5e-05, "loss": 1.0122, "mean_token_accuracy": 0.8579995036125183, "num_tokens": 589681178.0, "step": 15534 }, { "epoch": 1.9762116779035748, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.561966896057129, "learning_rate": 5e-05, "loss": 0.7284, "mean_token_accuracy": 0.8841006755828857, "num_tokens": 589719073.0, "step": 15535 }, { "epoch": 1.976338888182165, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.612421035766602, "learning_rate": 5e-05, "loss": 1.0037, "mean_token_accuracy": 0.862978458404541, "num_tokens": 589760528.0, "step": 15536 }, { "epoch": 1.9764660984607556, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.121496200561523, "learning_rate": 5e-05, "loss": 0.8215, "mean_token_accuracy": 0.8713994026184082, "num_tokens": 589795946.0, "step": 15537 }, { "epoch": 1.9765933087393461, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.985539436340332, "learning_rate": 5e-05, "loss": 0.9217, "mean_token_accuracy": 0.8710812926292419, "num_tokens": 589829139.0, "step": 15538 }, { "epoch": 1.9767205190179367, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.403975486755371, "learning_rate": 5e-05, "loss": 0.7988, "mean_token_accuracy": 0.8825546503067017, "num_tokens": 589862916.0, "step": 15539 }, { "epoch": 1.9768477292965272, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.064536094665527, "learning_rate": 5e-05, "loss": 0.8142, "mean_token_accuracy": 0.878406822681427, "num_tokens": 589900230.0, "step": 15540 }, { "epoch": 1.9769749395751177, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.316588401794434, "learning_rate": 5e-05, "loss": 0.8051, "mean_token_accuracy": 0.8728251457214355, "num_tokens": 589938737.0, "step": 15541 }, { "epoch": 1.977102149853708, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.607126235961914, "learning_rate": 5e-05, "loss": 0.7537, "mean_token_accuracy": 0.8779424428939819, "num_tokens": 589977989.0, "step": 15542 }, { "epoch": 1.9772293601322986, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.199507713317871, "learning_rate": 5e-05, "loss": 0.7304, "mean_token_accuracy": 0.8886455297470093, "num_tokens": 590015547.0, "step": 15543 }, { "epoch": 1.977356570410889, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.422952651977539, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.8748581409454346, "num_tokens": 590055257.0, "step": 15544 }, { "epoch": 1.9774837806894796, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.323188781738281, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8755009174346924, "num_tokens": 590087911.0, "step": 15545 }, { "epoch": 1.9776109909680701, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.161066055297852, "learning_rate": 5e-05, "loss": 0.7315, "mean_token_accuracy": 0.8684639930725098, "num_tokens": 590128889.0, "step": 15546 }, { "epoch": 1.9777382012466607, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.58963680267334, "learning_rate": 5e-05, "loss": 0.8616, "mean_token_accuracy": 0.8620409965515137, "num_tokens": 590168078.0, "step": 15547 }, { "epoch": 1.9778654115252512, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.081107139587402, "learning_rate": 5e-05, "loss": 0.7782, "mean_token_accuracy": 0.8541361093521118, "num_tokens": 590211361.0, "step": 15548 }, { "epoch": 1.9779926218038417, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.560676574707031, "learning_rate": 5e-05, "loss": 0.8602, "mean_token_accuracy": 0.86253821849823, "num_tokens": 590249922.0, "step": 15549 }, { "epoch": 1.9781198320824323, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.28653621673584, "learning_rate": 5e-05, "loss": 0.7304, "mean_token_accuracy": 0.8736602067947388, "num_tokens": 590288284.0, "step": 15550 }, { "epoch": 1.9782470423610228, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.291884422302246, "learning_rate": 5e-05, "loss": 0.8416, "mean_token_accuracy": 0.8604499101638794, "num_tokens": 590324373.0, "step": 15551 }, { "epoch": 1.9783742526396133, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.495321750640869, "learning_rate": 5e-05, "loss": 0.7345, "mean_token_accuracy": 0.871880292892456, "num_tokens": 590365374.0, "step": 15552 }, { "epoch": 1.9785014629182038, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.050349235534668, "learning_rate": 5e-05, "loss": 0.7924, "mean_token_accuracy": 0.8695568442344666, "num_tokens": 590402712.0, "step": 15553 }, { "epoch": 1.9786286731967944, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.560770034790039, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.854227602481842, "num_tokens": 590438434.0, "step": 15554 }, { "epoch": 1.978755883475385, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.893312454223633, "learning_rate": 5e-05, "loss": 0.8091, "mean_token_accuracy": 0.8629388809204102, "num_tokens": 590482619.0, "step": 15555 }, { "epoch": 1.9788830937539754, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.570886611938477, "learning_rate": 5e-05, "loss": 0.7885, "mean_token_accuracy": 0.856442928314209, "num_tokens": 590527828.0, "step": 15556 }, { "epoch": 1.979010304032566, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.8946638107299805, "learning_rate": 5e-05, "loss": 0.7515, "mean_token_accuracy": 0.8753928542137146, "num_tokens": 590561758.0, "step": 15557 }, { "epoch": 1.9791375143111565, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.367992401123047, "learning_rate": 5e-05, "loss": 0.6799, "mean_token_accuracy": 0.8834089040756226, "num_tokens": 590598323.0, "step": 15558 }, { "epoch": 1.979264724589747, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.154668807983398, "learning_rate": 5e-05, "loss": 0.8037, "mean_token_accuracy": 0.866481363773346, "num_tokens": 590642469.0, "step": 15559 }, { "epoch": 1.9793919348683373, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.98747444152832, "learning_rate": 5e-05, "loss": 0.7231, "mean_token_accuracy": 0.8663114309310913, "num_tokens": 590683603.0, "step": 15560 }, { "epoch": 1.9795191451469278, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.62790584564209, "learning_rate": 5e-05, "loss": 0.7867, "mean_token_accuracy": 0.8796736001968384, "num_tokens": 590710136.0, "step": 15561 }, { "epoch": 1.9796463554255184, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.607990264892578, "learning_rate": 5e-05, "loss": 0.6933, "mean_token_accuracy": 0.8676498532295227, "num_tokens": 590743104.0, "step": 15562 }, { "epoch": 1.979773565704109, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 9.089468955993652, "learning_rate": 5e-05, "loss": 0.8878, "mean_token_accuracy": 0.8668511509895325, "num_tokens": 590781851.0, "step": 15563 }, { "epoch": 1.9799007759826994, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.515479564666748, "learning_rate": 5e-05, "loss": 0.7334, "mean_token_accuracy": 0.8619825839996338, "num_tokens": 590819865.0, "step": 15564 }, { "epoch": 1.98002798626129, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 9.197190284729004, "learning_rate": 5e-05, "loss": 0.9116, "mean_token_accuracy": 0.8787443041801453, "num_tokens": 590856478.0, "step": 15565 }, { "epoch": 1.9801551965398803, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.071448802947998, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8610765337944031, "num_tokens": 590895765.0, "step": 15566 }, { "epoch": 1.9802824068184708, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.489726066589355, "learning_rate": 5e-05, "loss": 0.8656, "mean_token_accuracy": 0.8754125833511353, "num_tokens": 590931238.0, "step": 15567 }, { "epoch": 1.9804096170970613, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.943582057952881, "learning_rate": 5e-05, "loss": 0.7909, "mean_token_accuracy": 0.8751023411750793, "num_tokens": 590969875.0, "step": 15568 }, { "epoch": 1.9805368273756518, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.349704742431641, "learning_rate": 5e-05, "loss": 0.7799, "mean_token_accuracy": 0.8588411211967468, "num_tokens": 591007964.0, "step": 15569 }, { "epoch": 1.9806640376542424, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.675185203552246, "learning_rate": 5e-05, "loss": 0.8442, "mean_token_accuracy": 0.8755921721458435, "num_tokens": 591049339.0, "step": 15570 }, { "epoch": 1.980791247932833, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.635779857635498, "learning_rate": 5e-05, "loss": 0.6779, "mean_token_accuracy": 0.8795145750045776, "num_tokens": 591086429.0, "step": 15571 }, { "epoch": 1.9809184582114234, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 9.208890914916992, "learning_rate": 5e-05, "loss": 0.8967, "mean_token_accuracy": 0.8740659356117249, "num_tokens": 591128252.0, "step": 15572 }, { "epoch": 1.981045668490014, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.497591018676758, "learning_rate": 5e-05, "loss": 0.6613, "mean_token_accuracy": 0.8836424350738525, "num_tokens": 591165049.0, "step": 15573 }, { "epoch": 1.9811728787686045, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.37016773223877, "learning_rate": 5e-05, "loss": 0.9319, "mean_token_accuracy": 0.8820241689682007, "num_tokens": 591196805.0, "step": 15574 }, { "epoch": 1.981300089047195, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.008543968200684, "learning_rate": 5e-05, "loss": 0.7443, "mean_token_accuracy": 0.8747427463531494, "num_tokens": 591242897.0, "step": 15575 }, { "epoch": 1.9814272993257855, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.869823455810547, "learning_rate": 5e-05, "loss": 0.9297, "mean_token_accuracy": 0.8697748184204102, "num_tokens": 591280684.0, "step": 15576 }, { "epoch": 1.981554509604376, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.7442545890808105, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.8749804496765137, "num_tokens": 591315264.0, "step": 15577 }, { "epoch": 1.9816817198829666, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.02444839477539, "learning_rate": 5e-05, "loss": 0.8577, "mean_token_accuracy": 0.8694343566894531, "num_tokens": 591358028.0, "step": 15578 }, { "epoch": 1.9818089301615571, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.054205894470215, "learning_rate": 5e-05, "loss": 0.8995, "mean_token_accuracy": 0.8506643772125244, "num_tokens": 591404272.0, "step": 15579 }, { "epoch": 1.9819361404401477, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.748080253601074, "learning_rate": 5e-05, "loss": 0.757, "mean_token_accuracy": 0.8855882287025452, "num_tokens": 591439896.0, "step": 15580 }, { "epoch": 1.9820633507187382, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.702173709869385, "learning_rate": 5e-05, "loss": 0.7656, "mean_token_accuracy": 0.8776870965957642, "num_tokens": 591477318.0, "step": 15581 }, { "epoch": 1.9821905609973287, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.140083312988281, "learning_rate": 5e-05, "loss": 0.8468, "mean_token_accuracy": 0.866443932056427, "num_tokens": 591518600.0, "step": 15582 }, { "epoch": 1.9823177712759192, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.056859970092773, "learning_rate": 5e-05, "loss": 0.725, "mean_token_accuracy": 0.8672958016395569, "num_tokens": 591547414.0, "step": 15583 }, { "epoch": 1.9824449815545098, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.685434341430664, "learning_rate": 5e-05, "loss": 0.9027, "mean_token_accuracy": 0.8619463443756104, "num_tokens": 591594270.0, "step": 15584 }, { "epoch": 1.9825721918331, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.632157325744629, "learning_rate": 5e-05, "loss": 0.6911, "mean_token_accuracy": 0.8752831220626831, "num_tokens": 591638252.0, "step": 15585 }, { "epoch": 1.9826994021116906, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 9.131159782409668, "learning_rate": 5e-05, "loss": 0.9846, "mean_token_accuracy": 0.8506392240524292, "num_tokens": 591680268.0, "step": 15586 }, { "epoch": 1.9828266123902811, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.595029354095459, "learning_rate": 5e-05, "loss": 0.7112, "mean_token_accuracy": 0.8745993375778198, "num_tokens": 591717724.0, "step": 15587 }, { "epoch": 1.9829538226688717, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.253799438476562, "learning_rate": 5e-05, "loss": 0.917, "mean_token_accuracy": 0.8863757848739624, "num_tokens": 591754562.0, "step": 15588 }, { "epoch": 1.9830810329474622, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.057374000549316, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8756318092346191, "num_tokens": 591794838.0, "step": 15589 }, { "epoch": 1.9832082432260527, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.819872856140137, "learning_rate": 5e-05, "loss": 0.9987, "mean_token_accuracy": 0.8527567386627197, "num_tokens": 591834771.0, "step": 15590 }, { "epoch": 1.983335453504643, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.723713397979736, "learning_rate": 5e-05, "loss": 0.8804, "mean_token_accuracy": 0.8600587844848633, "num_tokens": 591872764.0, "step": 15591 }, { "epoch": 1.9834626637832335, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.982161521911621, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.8657304048538208, "num_tokens": 591907035.0, "step": 15592 }, { "epoch": 1.983589874061824, "ewc_loss": 0.5, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.153047561645508, "learning_rate": 5e-05, "loss": 0.8491, "mean_token_accuracy": 0.8778206706047058, "num_tokens": 591940797.0, "step": 15593 }, { "epoch": 1.9837170843404146, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.4247870445251465, "learning_rate": 5e-05, "loss": 0.7269, "mean_token_accuracy": 0.886843204498291, "num_tokens": 591976872.0, "step": 15594 }, { "epoch": 1.9838442946190051, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.210597038269043, "learning_rate": 5e-05, "loss": 0.8637, "mean_token_accuracy": 0.8582999110221863, "num_tokens": 592015933.0, "step": 15595 }, { "epoch": 1.9839715048975957, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.325646877288818, "learning_rate": 5e-05, "loss": 0.7267, "mean_token_accuracy": 0.8784317374229431, "num_tokens": 592046796.0, "step": 15596 }, { "epoch": 1.9840987151761862, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.175063133239746, "learning_rate": 5e-05, "loss": 0.8261, "mean_token_accuracy": 0.8718271851539612, "num_tokens": 592092841.0, "step": 15597 }, { "epoch": 1.9842259254547767, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.1947407722473145, "learning_rate": 5e-05, "loss": 0.795, "mean_token_accuracy": 0.8559475541114807, "num_tokens": 592134118.0, "step": 15598 }, { "epoch": 1.9843531357333672, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.373169898986816, "learning_rate": 5e-05, "loss": 0.8511, "mean_token_accuracy": 0.8646881580352783, "num_tokens": 592167292.0, "step": 15599 }, { "epoch": 1.9844803460119578, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.900068759918213, "learning_rate": 5e-05, "loss": 0.69, "mean_token_accuracy": 0.882820188999176, "num_tokens": 592205616.0, "step": 15600 }, { "epoch": 1.9846075562905483, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.742570877075195, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8793717622756958, "num_tokens": 592246650.0, "step": 15601 }, { "epoch": 1.9847347665691388, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.674320697784424, "learning_rate": 5e-05, "loss": 0.6991, "mean_token_accuracy": 0.8755071759223938, "num_tokens": 592284729.0, "step": 15602 }, { "epoch": 1.9848619768477294, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 9.0061616897583, "learning_rate": 5e-05, "loss": 0.8979, "mean_token_accuracy": 0.8738406300544739, "num_tokens": 592318423.0, "step": 15603 }, { "epoch": 1.9849891871263199, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.783230304718018, "learning_rate": 5e-05, "loss": 0.7265, "mean_token_accuracy": 0.8692854642868042, "num_tokens": 592351304.0, "step": 15604 }, { "epoch": 1.9851163974049104, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.857789993286133, "learning_rate": 5e-05, "loss": 0.9159, "mean_token_accuracy": 0.8747905492782593, "num_tokens": 592388353.0, "step": 15605 }, { "epoch": 1.985243607683501, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.252523422241211, "learning_rate": 5e-05, "loss": 0.7817, "mean_token_accuracy": 0.8773655295372009, "num_tokens": 592427651.0, "step": 15606 }, { "epoch": 1.9853708179620915, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.317265510559082, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.870725154876709, "num_tokens": 592463638.0, "step": 15607 }, { "epoch": 1.985498028240682, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.712143421173096, "learning_rate": 5e-05, "loss": 0.8626, "mean_token_accuracy": 0.861751914024353, "num_tokens": 592501508.0, "step": 15608 }, { "epoch": 1.9856252385192723, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.770566463470459, "learning_rate": 5e-05, "loss": 0.7904, "mean_token_accuracy": 0.8793848156929016, "num_tokens": 592538418.0, "step": 15609 }, { "epoch": 1.9857524487978628, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.882205486297607, "learning_rate": 5e-05, "loss": 0.7714, "mean_token_accuracy": 0.8862342834472656, "num_tokens": 592573552.0, "step": 15610 }, { "epoch": 1.9858796590764534, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.634435653686523, "learning_rate": 5e-05, "loss": 0.7717, "mean_token_accuracy": 0.8726296424865723, "num_tokens": 592607212.0, "step": 15611 }, { "epoch": 1.986006869355044, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.707998275756836, "learning_rate": 5e-05, "loss": 0.8039, "mean_token_accuracy": 0.8648834824562073, "num_tokens": 592649910.0, "step": 15612 }, { "epoch": 1.9861340796336344, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.81437349319458, "learning_rate": 5e-05, "loss": 0.7868, "mean_token_accuracy": 0.8732429146766663, "num_tokens": 592689555.0, "step": 15613 }, { "epoch": 1.986261289912225, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.378520965576172, "learning_rate": 5e-05, "loss": 0.7312, "mean_token_accuracy": 0.8774394989013672, "num_tokens": 592731921.0, "step": 15614 }, { "epoch": 1.9863885001908153, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.10147476196289, "learning_rate": 5e-05, "loss": 0.8077, "mean_token_accuracy": 0.8759356141090393, "num_tokens": 592763631.0, "step": 15615 }, { "epoch": 1.9865157104694058, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.059769153594971, "learning_rate": 5e-05, "loss": 0.6727, "mean_token_accuracy": 0.8911091685295105, "num_tokens": 592796961.0, "step": 15616 }, { "epoch": 1.9866429207479963, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.40275764465332, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.8792733550071716, "num_tokens": 592831380.0, "step": 15617 }, { "epoch": 1.9867701310265868, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.787732124328613, "learning_rate": 5e-05, "loss": 0.7367, "mean_token_accuracy": 0.8671882748603821, "num_tokens": 592873606.0, "step": 15618 }, { "epoch": 1.9868973413051774, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.768915176391602, "learning_rate": 5e-05, "loss": 0.899, "mean_token_accuracy": 0.8654727935791016, "num_tokens": 592919311.0, "step": 15619 }, { "epoch": 1.987024551583768, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.558562278747559, "learning_rate": 5e-05, "loss": 0.7755, "mean_token_accuracy": 0.8495069742202759, "num_tokens": 592961818.0, "step": 15620 }, { "epoch": 1.9871517618623584, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 9.085017204284668, "learning_rate": 5e-05, "loss": 0.9152, "mean_token_accuracy": 0.8710085153579712, "num_tokens": 593003320.0, "step": 15621 }, { "epoch": 1.987278972140949, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.610652923583984, "learning_rate": 5e-05, "loss": 0.7588, "mean_token_accuracy": 0.861405611038208, "num_tokens": 593050887.0, "step": 15622 }, { "epoch": 1.9874061824195395, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.065241813659668, "learning_rate": 5e-05, "loss": 0.9855, "mean_token_accuracy": 0.8615288734436035, "num_tokens": 593086296.0, "step": 15623 }, { "epoch": 1.98753339269813, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.166633605957031, "learning_rate": 5e-05, "loss": 0.8621, "mean_token_accuracy": 0.8547859191894531, "num_tokens": 593122135.0, "step": 15624 }, { "epoch": 1.9876606029767205, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.381129264831543, "learning_rate": 5e-05, "loss": 0.9143, "mean_token_accuracy": 0.8626704216003418, "num_tokens": 593158845.0, "step": 15625 }, { "epoch": 1.987787813255311, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.93751335144043, "learning_rate": 5e-05, "loss": 0.8364, "mean_token_accuracy": 0.8778008222579956, "num_tokens": 593199386.0, "step": 15626 }, { "epoch": 1.9879150235339016, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.325196743011475, "learning_rate": 5e-05, "loss": 0.7439, "mean_token_accuracy": 0.8807958364486694, "num_tokens": 593237061.0, "step": 15627 }, { "epoch": 1.9880422338124921, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.635924339294434, "learning_rate": 5e-05, "loss": 0.8907, "mean_token_accuracy": 0.8714637756347656, "num_tokens": 593275220.0, "step": 15628 }, { "epoch": 1.9881694440910826, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.659711837768555, "learning_rate": 5e-05, "loss": 0.7801, "mean_token_accuracy": 0.850990891456604, "num_tokens": 593314576.0, "step": 15629 }, { "epoch": 1.9882966543696732, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.099346160888672, "learning_rate": 5e-05, "loss": 0.9022, "mean_token_accuracy": 0.8789260387420654, "num_tokens": 593362905.0, "step": 15630 }, { "epoch": 1.9884238646482637, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.506033897399902, "learning_rate": 5e-05, "loss": 0.6922, "mean_token_accuracy": 0.8789410591125488, "num_tokens": 593402486.0, "step": 15631 }, { "epoch": 1.9885510749268542, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.345401763916016, "learning_rate": 5e-05, "loss": 0.9368, "mean_token_accuracy": 0.8823969960212708, "num_tokens": 593440555.0, "step": 15632 }, { "epoch": 1.9886782852054448, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.810988426208496, "learning_rate": 5e-05, "loss": 0.7783, "mean_token_accuracy": 0.8712958097457886, "num_tokens": 593480488.0, "step": 15633 }, { "epoch": 1.988805495484035, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.156046867370605, "learning_rate": 5e-05, "loss": 0.9459, "mean_token_accuracy": 0.8775044083595276, "num_tokens": 593518615.0, "step": 15634 }, { "epoch": 1.9889327057626256, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.51382303237915, "learning_rate": 5e-05, "loss": 0.8625, "mean_token_accuracy": 0.8777418732643127, "num_tokens": 593556990.0, "step": 15635 }, { "epoch": 1.9890599160412161, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.418927192687988, "learning_rate": 5e-05, "loss": 0.9649, "mean_token_accuracy": 0.8638031482696533, "num_tokens": 593599438.0, "step": 15636 }, { "epoch": 1.9891871263198067, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.985798358917236, "learning_rate": 5e-05, "loss": 0.9448, "mean_token_accuracy": 0.8585498332977295, "num_tokens": 593635991.0, "step": 15637 }, { "epoch": 1.9893143365983972, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.823105812072754, "learning_rate": 5e-05, "loss": 0.8638, "mean_token_accuracy": 0.870730459690094, "num_tokens": 593676600.0, "step": 15638 }, { "epoch": 1.9894415468769877, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.8954572677612305, "learning_rate": 5e-05, "loss": 0.8298, "mean_token_accuracy": 0.8817464113235474, "num_tokens": 593709069.0, "step": 15639 }, { "epoch": 1.989568757155578, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.970674514770508, "learning_rate": 5e-05, "loss": 0.7848, "mean_token_accuracy": 0.8957807421684265, "num_tokens": 593740361.0, "step": 15640 }, { "epoch": 1.9896959674341685, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.270660400390625, "learning_rate": 5e-05, "loss": 0.719, "mean_token_accuracy": 0.8866816759109497, "num_tokens": 593775424.0, "step": 15641 }, { "epoch": 1.989823177712759, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.581791877746582, "learning_rate": 5e-05, "loss": 0.9159, "mean_token_accuracy": 0.86250901222229, "num_tokens": 593816562.0, "step": 15642 }, { "epoch": 1.9899503879913496, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.714881420135498, "learning_rate": 5e-05, "loss": 0.7379, "mean_token_accuracy": 0.8732258677482605, "num_tokens": 593853272.0, "step": 15643 }, { "epoch": 1.9900775982699401, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.100683212280273, "learning_rate": 5e-05, "loss": 0.9079, "mean_token_accuracy": 0.8787364959716797, "num_tokens": 593885308.0, "step": 15644 }, { "epoch": 1.9902048085485307, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.6222968101501465, "learning_rate": 5e-05, "loss": 0.7023, "mean_token_accuracy": 0.8836988210678101, "num_tokens": 593918150.0, "step": 15645 }, { "epoch": 1.9903320188271212, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.237885475158691, "learning_rate": 5e-05, "loss": 0.9975, "mean_token_accuracy": 0.8656783103942871, "num_tokens": 593961947.0, "step": 15646 }, { "epoch": 1.9904592291057117, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.111697673797607, "learning_rate": 5e-05, "loss": 0.8753, "mean_token_accuracy": 0.8581106662750244, "num_tokens": 594000659.0, "step": 15647 }, { "epoch": 1.9905864393843022, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.766490936279297, "learning_rate": 5e-05, "loss": 0.9508, "mean_token_accuracy": 0.8726361989974976, "num_tokens": 594040033.0, "step": 15648 }, { "epoch": 1.9907136496628928, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.796830654144287, "learning_rate": 5e-05, "loss": 0.8952, "mean_token_accuracy": 0.876435399055481, "num_tokens": 594080534.0, "step": 15649 }, { "epoch": 1.9908408599414833, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.888330936431885, "learning_rate": 5e-05, "loss": 0.882, "mean_token_accuracy": 0.8672614097595215, "num_tokens": 594118015.0, "step": 15650 }, { "epoch": 1.9909680702200738, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.221456527709961, "learning_rate": 5e-05, "loss": 0.8959, "mean_token_accuracy": 0.8747109770774841, "num_tokens": 594153839.0, "step": 15651 }, { "epoch": 1.9910952804986644, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.280551910400391, "learning_rate": 5e-05, "loss": 0.759, "mean_token_accuracy": 0.8822198510169983, "num_tokens": 594195672.0, "step": 15652 }, { "epoch": 1.9912224907772549, "ewc_loss": 0.5, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.339642524719238, "learning_rate": 5e-05, "loss": 0.8862, "mean_token_accuracy": 0.8653895854949951, "num_tokens": 594239178.0, "step": 15653 }, { "epoch": 1.9913497010558454, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.132538795471191, "learning_rate": 5e-05, "loss": 0.7442, "mean_token_accuracy": 0.8752548694610596, "num_tokens": 594274938.0, "step": 15654 }, { "epoch": 1.991476911334436, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.3388671875, "learning_rate": 5e-05, "loss": 0.8279, "mean_token_accuracy": 0.8836839199066162, "num_tokens": 594310632.0, "step": 15655 }, { "epoch": 1.9916041216130265, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.054874420166016, "learning_rate": 5e-05, "loss": 0.7634, "mean_token_accuracy": 0.8668773174285889, "num_tokens": 594344308.0, "step": 15656 }, { "epoch": 1.991731331891617, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.4901123046875, "learning_rate": 5e-05, "loss": 0.8745, "mean_token_accuracy": 0.8701752424240112, "num_tokens": 594385287.0, "step": 15657 }, { "epoch": 1.9918585421702073, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.851768970489502, "learning_rate": 5e-05, "loss": 0.7071, "mean_token_accuracy": 0.8824706673622131, "num_tokens": 594422521.0, "step": 15658 }, { "epoch": 1.9919857524487978, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.794238090515137, "learning_rate": 5e-05, "loss": 0.9573, "mean_token_accuracy": 0.8580824136734009, "num_tokens": 594466837.0, "step": 15659 }, { "epoch": 1.9921129627273884, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.745723724365234, "learning_rate": 5e-05, "loss": 0.7572, "mean_token_accuracy": 0.8708943128585815, "num_tokens": 594501283.0, "step": 15660 }, { "epoch": 1.9922401730059789, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.916047096252441, "learning_rate": 5e-05, "loss": 0.9703, "mean_token_accuracy": 0.859296441078186, "num_tokens": 594544627.0, "step": 15661 }, { "epoch": 1.9923673832845694, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.003441333770752, "learning_rate": 5e-05, "loss": 0.8521, "mean_token_accuracy": 0.8516561985015869, "num_tokens": 594586363.0, "step": 15662 }, { "epoch": 1.99249459356316, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.605298042297363, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.8704319000244141, "num_tokens": 594624950.0, "step": 15663 }, { "epoch": 1.9926218038417502, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.462069988250732, "learning_rate": 5e-05, "loss": 0.8965, "mean_token_accuracy": 0.8573262691497803, "num_tokens": 594665103.0, "step": 15664 }, { "epoch": 1.9927490141203408, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.052845001220703, "learning_rate": 5e-05, "loss": 0.824, "mean_token_accuracy": 0.87986820936203, "num_tokens": 594699506.0, "step": 15665 }, { "epoch": 1.9928762243989313, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.782784461975098, "learning_rate": 5e-05, "loss": 0.8759, "mean_token_accuracy": 0.8668148517608643, "num_tokens": 594741771.0, "step": 15666 }, { "epoch": 1.9930034346775218, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.676610469818115, "learning_rate": 5e-05, "loss": 0.7887, "mean_token_accuracy": 0.8793400526046753, "num_tokens": 594784705.0, "step": 15667 }, { "epoch": 1.9931306449561124, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.7892279624938965, "learning_rate": 5e-05, "loss": 0.7903, "mean_token_accuracy": 0.8784433603286743, "num_tokens": 594827800.0, "step": 15668 }, { "epoch": 1.993257855234703, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.68850564956665, "learning_rate": 5e-05, "loss": 0.7976, "mean_token_accuracy": 0.8700852394104004, "num_tokens": 594865150.0, "step": 15669 }, { "epoch": 1.9933850655132934, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.548172473907471, "learning_rate": 5e-05, "loss": 0.8253, "mean_token_accuracy": 0.8579955101013184, "num_tokens": 594903440.0, "step": 15670 }, { "epoch": 1.993512275791884, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.908899784088135, "learning_rate": 5e-05, "loss": 0.7857, "mean_token_accuracy": 0.8740450143814087, "num_tokens": 594936761.0, "step": 15671 }, { "epoch": 1.9936394860704745, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.2340569496154785, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8588598966598511, "num_tokens": 594972298.0, "step": 15672 }, { "epoch": 1.993766696349065, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.179387092590332, "learning_rate": 5e-05, "loss": 0.8229, "mean_token_accuracy": 0.8721197247505188, "num_tokens": 595009530.0, "step": 15673 }, { "epoch": 1.9938939066276555, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.964349746704102, "learning_rate": 5e-05, "loss": 0.7252, "mean_token_accuracy": 0.8742085695266724, "num_tokens": 595048147.0, "step": 15674 }, { "epoch": 1.994021116906246, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.476297378540039, "learning_rate": 5e-05, "loss": 0.8112, "mean_token_accuracy": 0.8836214542388916, "num_tokens": 595083580.0, "step": 15675 }, { "epoch": 1.9941483271848366, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.7020583152771, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.8513159155845642, "num_tokens": 595122989.0, "step": 15676 }, { "epoch": 1.9942755374634271, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.845304489135742, "learning_rate": 5e-05, "loss": 0.8938, "mean_token_accuracy": 0.8687736988067627, "num_tokens": 595165378.0, "step": 15677 }, { "epoch": 1.9944027477420176, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.534520149230957, "learning_rate": 5e-05, "loss": 0.7953, "mean_token_accuracy": 0.8497968912124634, "num_tokens": 595203802.0, "step": 15678 }, { "epoch": 1.9945299580206082, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.109583854675293, "learning_rate": 5e-05, "loss": 0.9409, "mean_token_accuracy": 0.8676673769950867, "num_tokens": 595242056.0, "step": 15679 }, { "epoch": 1.9946571682991987, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.740899562835693, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.8567298650741577, "num_tokens": 595276381.0, "step": 15680 }, { "epoch": 1.9947843785777892, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.88032054901123, "learning_rate": 5e-05, "loss": 0.9497, "mean_token_accuracy": 0.8727383017539978, "num_tokens": 595318551.0, "step": 15681 }, { "epoch": 1.9949115888563798, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.41023063659668, "learning_rate": 5e-05, "loss": 0.8731, "mean_token_accuracy": 0.8630678653717041, "num_tokens": 595361802.0, "step": 15682 }, { "epoch": 1.99503879913497, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.976115703582764, "learning_rate": 5e-05, "loss": 0.8612, "mean_token_accuracy": 0.8711141347885132, "num_tokens": 595402853.0, "step": 15683 }, { "epoch": 1.9951660094135606, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.168015480041504, "learning_rate": 5e-05, "loss": 0.8709, "mean_token_accuracy": 0.8776047825813293, "num_tokens": 595443774.0, "step": 15684 }, { "epoch": 1.9952932196921511, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.023212432861328, "learning_rate": 5e-05, "loss": 0.7666, "mean_token_accuracy": 0.8643306493759155, "num_tokens": 595485395.0, "step": 15685 }, { "epoch": 1.9954204299707416, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.776103973388672, "learning_rate": 5e-05, "loss": 0.92, "mean_token_accuracy": 0.8676624894142151, "num_tokens": 595527724.0, "step": 15686 }, { "epoch": 1.9955476402493322, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.500303745269775, "learning_rate": 5e-05, "loss": 0.7564, "mean_token_accuracy": 0.860075831413269, "num_tokens": 595569081.0, "step": 15687 }, { "epoch": 1.9956748505279227, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.235481262207031, "learning_rate": 5e-05, "loss": 0.9863, "mean_token_accuracy": 0.8630144596099854, "num_tokens": 595611400.0, "step": 15688 }, { "epoch": 1.995802060806513, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.456703186035156, "learning_rate": 5e-05, "loss": 0.7584, "mean_token_accuracy": 0.8677834272384644, "num_tokens": 595648063.0, "step": 15689 }, { "epoch": 1.9959292710851035, "ewc_loss": 0.625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.432502746582031, "learning_rate": 5e-05, "loss": 1.0427, "mean_token_accuracy": 0.8579928874969482, "num_tokens": 595688679.0, "step": 15690 }, { "epoch": 1.996056481363694, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 6.993010520935059, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.8612858653068542, "num_tokens": 595726120.0, "step": 15691 }, { "epoch": 1.9961836916422846, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.029240608215332, "learning_rate": 5e-05, "loss": 0.97, "mean_token_accuracy": 0.8726370930671692, "num_tokens": 595762020.0, "step": 15692 }, { "epoch": 1.9963109019208751, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.820401191711426, "learning_rate": 5e-05, "loss": 0.995, "mean_token_accuracy": 0.8496299982070923, "num_tokens": 595797909.0, "step": 15693 }, { "epoch": 1.9964381121994657, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.042484283447266, "learning_rate": 5e-05, "loss": 0.9036, "mean_token_accuracy": 0.8686807155609131, "num_tokens": 595837542.0, "step": 15694 }, { "epoch": 1.9965653224780562, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.251954078674316, "learning_rate": 5e-05, "loss": 0.9186, "mean_token_accuracy": 0.8733282685279846, "num_tokens": 595874006.0, "step": 15695 }, { "epoch": 1.9966925327566467, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.519934177398682, "learning_rate": 5e-05, "loss": 0.8506, "mean_token_accuracy": 0.8630754351615906, "num_tokens": 595904098.0, "step": 15696 }, { "epoch": 1.9968197430352372, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.973934650421143, "learning_rate": 5e-05, "loss": 0.8448, "mean_token_accuracy": 0.8752832412719727, "num_tokens": 595945725.0, "step": 15697 }, { "epoch": 1.9969469533138278, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.873480319976807, "learning_rate": 5e-05, "loss": 0.8468, "mean_token_accuracy": 0.8701317310333252, "num_tokens": 595977826.0, "step": 15698 }, { "epoch": 1.9970741635924183, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.258379936218262, "learning_rate": 5e-05, "loss": 0.8011, "mean_token_accuracy": 0.8621126413345337, "num_tokens": 596011350.0, "step": 15699 }, { "epoch": 1.9972013738710088, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.447464942932129, "learning_rate": 5e-05, "loss": 0.8298, "mean_token_accuracy": 0.8857678771018982, "num_tokens": 596050688.0, "step": 15700 }, { "epoch": 1.9973285841495994, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.749807357788086, "learning_rate": 5e-05, "loss": 0.7003, "mean_token_accuracy": 0.8810338377952576, "num_tokens": 596090711.0, "step": 15701 }, { "epoch": 1.9974557944281899, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.90311336517334, "learning_rate": 5e-05, "loss": 0.9275, "mean_token_accuracy": 0.8705332279205322, "num_tokens": 596130346.0, "step": 15702 }, { "epoch": 1.9975830047067804, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.564517021179199, "learning_rate": 5e-05, "loss": 0.7772, "mean_token_accuracy": 0.861680269241333, "num_tokens": 596173035.0, "step": 15703 }, { "epoch": 1.997710214985371, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.17011547088623, "learning_rate": 5e-05, "loss": 1.0092, "mean_token_accuracy": 0.853763222694397, "num_tokens": 596208105.0, "step": 15704 }, { "epoch": 1.9978374252639615, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.7709808349609375, "learning_rate": 5e-05, "loss": 0.7418, "mean_token_accuracy": 0.8841009140014648, "num_tokens": 596247367.0, "step": 15705 }, { "epoch": 1.997964635542552, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.12639045715332, "learning_rate": 5e-05, "loss": 0.9909, "mean_token_accuracy": 0.8688086867332458, "num_tokens": 596288226.0, "step": 15706 }, { "epoch": 1.9980918458211423, "ewc_loss": 0.5, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.34825325012207, "learning_rate": 5e-05, "loss": 0.9054, "mean_token_accuracy": 0.8595072031021118, "num_tokens": 596329422.0, "step": 15707 }, { "epoch": 1.9982190560997328, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.506293296813965, "learning_rate": 5e-05, "loss": 0.9301, "mean_token_accuracy": 0.8748074769973755, "num_tokens": 596367447.0, "step": 15708 }, { "epoch": 1.9983462663783234, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.924389362335205, "learning_rate": 5e-05, "loss": 0.9252, "mean_token_accuracy": 0.8679186701774597, "num_tokens": 596401564.0, "step": 15709 }, { "epoch": 1.9984734766569139, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.680511474609375, "learning_rate": 5e-05, "loss": 0.8688, "mean_token_accuracy": 0.8629841208457947, "num_tokens": 596441076.0, "step": 15710 }, { "epoch": 1.9986006869355044, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.2158842086792, "learning_rate": 5e-05, "loss": 0.864, "mean_token_accuracy": 0.8791837096214294, "num_tokens": 596480360.0, "step": 15711 }, { "epoch": 1.998727897214095, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.3009514808654785, "learning_rate": 5e-05, "loss": 0.8143, "mean_token_accuracy": 0.8602712154388428, "num_tokens": 596512836.0, "step": 15712 }, { "epoch": 1.9988551074926852, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.186141014099121, "learning_rate": 5e-05, "loss": 0.8278, "mean_token_accuracy": 0.8839365243911743, "num_tokens": 596549154.0, "step": 15713 }, { "epoch": 1.9989823177712758, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.208189010620117, "learning_rate": 5e-05, "loss": 0.7516, "mean_token_accuracy": 0.8781219720840454, "num_tokens": 596590528.0, "step": 15714 }, { "epoch": 1.9991095280498663, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.26099681854248, "learning_rate": 5e-05, "loss": 0.8569, "mean_token_accuracy": 0.8716387748718262, "num_tokens": 596623926.0, "step": 15715 }, { "epoch": 1.9992367383284568, "ewc_loss": 0.375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.986865520477295, "learning_rate": 5e-05, "loss": 0.7411, "mean_token_accuracy": 0.8740440607070923, "num_tokens": 596661102.0, "step": 15716 }, { "epoch": 1.9993639486070474, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.605931282043457, "learning_rate": 5e-05, "loss": 0.8452, "mean_token_accuracy": 0.8837645053863525, "num_tokens": 596701113.0, "step": 15717 }, { "epoch": 1.9994911588856379, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.650826930999756, "learning_rate": 5e-05, "loss": 0.704, "mean_token_accuracy": 0.881494402885437, "num_tokens": 596735130.0, "step": 15718 }, { "epoch": 1.9996183691642284, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.08423137664795, "learning_rate": 5e-05, "loss": 0.9859, "mean_token_accuracy": 0.8579916954040527, "num_tokens": 596774612.0, "step": 15719 }, { "epoch": 1.999745579442819, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.569876670837402, "learning_rate": 5e-05, "loss": 0.7447, "mean_token_accuracy": 0.871192216873169, "num_tokens": 596811033.0, "step": 15720 }, { "epoch": 1.9998727897214095, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.249977111816406, "learning_rate": 5e-05, "loss": 1.0122, "mean_token_accuracy": 0.8599762320518494, "num_tokens": 596849169.0, "step": 15721 }, { "epoch": 2.0, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.069736480712891, "learning_rate": 5e-05, "loss": 0.8167, "mean_token_accuracy": 0.8768000602722168, "num_tokens": 596886661.0, "step": 15722 }, { "epoch": 2.0001272102785905, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.769548416137695, "learning_rate": 5e-05, "loss": 0.9026, "mean_token_accuracy": 0.8895301818847656, "num_tokens": 596927769.0, "step": 15723 }, { "epoch": 2.000254420557181, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.752819061279297, "learning_rate": 5e-05, "loss": 0.8653, "mean_token_accuracy": 0.8868852853775024, "num_tokens": 596966558.0, "step": 15724 }, { "epoch": 2.0003816308357716, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.9110941886901855, "learning_rate": 5e-05, "loss": 0.8075, "mean_token_accuracy": 0.889438807964325, "num_tokens": 597007424.0, "step": 15725 }, { "epoch": 2.000508841114362, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.139425277709961, "learning_rate": 5e-05, "loss": 0.8456, "mean_token_accuracy": 0.886710524559021, "num_tokens": 597043544.0, "step": 15726 }, { "epoch": 2.0006360513929526, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.431862831115723, "learning_rate": 5e-05, "loss": 0.7542, "mean_token_accuracy": 0.8848599195480347, "num_tokens": 597081079.0, "step": 15727 }, { "epoch": 2.000763261671543, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.014878273010254, "learning_rate": 5e-05, "loss": 0.7723, "mean_token_accuracy": 0.8940496444702148, "num_tokens": 597122188.0, "step": 15728 }, { "epoch": 2.0008904719501337, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.606545448303223, "learning_rate": 5e-05, "loss": 0.7723, "mean_token_accuracy": 0.880932629108429, "num_tokens": 597164835.0, "step": 15729 }, { "epoch": 2.0010176822287242, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.52860689163208, "learning_rate": 5e-05, "loss": 0.7177, "mean_token_accuracy": 0.8920499086380005, "num_tokens": 597201266.0, "step": 15730 }, { "epoch": 2.0011448925073148, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.011795043945312, "learning_rate": 5e-05, "loss": 0.7612, "mean_token_accuracy": 0.8882174491882324, "num_tokens": 597244095.0, "step": 15731 }, { "epoch": 2.0012721027859053, "ewc_loss": 0.375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.069194316864014, "learning_rate": 5e-05, "loss": 0.6964, "mean_token_accuracy": 0.8846101760864258, "num_tokens": 597285462.0, "step": 15732 }, { "epoch": 2.001399313064496, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.396668434143066, "learning_rate": 5e-05, "loss": 0.7879, "mean_token_accuracy": 0.8899121880531311, "num_tokens": 597323436.0, "step": 15733 }, { "epoch": 2.0015265233430863, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.769076347351074, "learning_rate": 5e-05, "loss": 0.6807, "mean_token_accuracy": 0.8823645114898682, "num_tokens": 597361775.0, "step": 15734 }, { "epoch": 2.0016537336216764, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.716445922851562, "learning_rate": 5e-05, "loss": 0.8604, "mean_token_accuracy": 0.8747969269752502, "num_tokens": 597402772.0, "step": 15735 }, { "epoch": 2.001780943900267, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.563842296600342, "learning_rate": 5e-05, "loss": 0.6549, "mean_token_accuracy": 0.8909385204315186, "num_tokens": 597444245.0, "step": 15736 }, { "epoch": 2.0019081541788575, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.084190368652344, "learning_rate": 5e-05, "loss": 0.8905, "mean_token_accuracy": 0.885326623916626, "num_tokens": 597481742.0, "step": 15737 }, { "epoch": 2.002035364457448, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.59971809387207, "learning_rate": 5e-05, "loss": 0.7286, "mean_token_accuracy": 0.8781371712684631, "num_tokens": 597520339.0, "step": 15738 }, { "epoch": 2.0021625747360385, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.142467498779297, "learning_rate": 5e-05, "loss": 0.8999, "mean_token_accuracy": 0.8929120898246765, "num_tokens": 597552756.0, "step": 15739 }, { "epoch": 2.002289785014629, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.144933700561523, "learning_rate": 5e-05, "loss": 0.7557, "mean_token_accuracy": 0.8941055536270142, "num_tokens": 597582224.0, "step": 15740 }, { "epoch": 2.0024169952932196, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.506497383117676, "learning_rate": 5e-05, "loss": 0.8652, "mean_token_accuracy": 0.8866233825683594, "num_tokens": 597619121.0, "step": 15741 }, { "epoch": 2.00254420557181, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.885256767272949, "learning_rate": 5e-05, "loss": 0.8856, "mean_token_accuracy": 0.875038206577301, "num_tokens": 597654074.0, "step": 15742 }, { "epoch": 2.0026714158504006, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.523910045623779, "learning_rate": 5e-05, "loss": 0.7553, "mean_token_accuracy": 0.8887625932693481, "num_tokens": 597695873.0, "step": 15743 }, { "epoch": 2.002798626128991, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.392763137817383, "learning_rate": 5e-05, "loss": 0.836, "mean_token_accuracy": 0.8891811370849609, "num_tokens": 597733072.0, "step": 15744 }, { "epoch": 2.0029258364075817, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.952550888061523, "learning_rate": 5e-05, "loss": 0.645, "mean_token_accuracy": 0.9044084548950195, "num_tokens": 597772777.0, "step": 15745 }, { "epoch": 2.0030530466861722, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.59090518951416, "learning_rate": 5e-05, "loss": 0.8669, "mean_token_accuracy": 0.8764569759368896, "num_tokens": 597809276.0, "step": 15746 }, { "epoch": 2.0031802569647628, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.845421314239502, "learning_rate": 5e-05, "loss": 0.7034, "mean_token_accuracy": 0.8816418647766113, "num_tokens": 597848956.0, "step": 15747 }, { "epoch": 2.0033074672433533, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.680686950683594, "learning_rate": 5e-05, "loss": 0.8537, "mean_token_accuracy": 0.8907907009124756, "num_tokens": 597890664.0, "step": 15748 }, { "epoch": 2.003434677521944, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.837337493896484, "learning_rate": 5e-05, "loss": 0.7108, "mean_token_accuracy": 0.8854833841323853, "num_tokens": 597931017.0, "step": 15749 }, { "epoch": 2.0035618878005343, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.790315628051758, "learning_rate": 5e-05, "loss": 0.8538, "mean_token_accuracy": 0.8980271816253662, "num_tokens": 597971552.0, "step": 15750 }, { "epoch": 2.003689098079125, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.898085594177246, "learning_rate": 5e-05, "loss": 0.7561, "mean_token_accuracy": 0.8764520883560181, "num_tokens": 598011554.0, "step": 15751 }, { "epoch": 2.0038163083577154, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.807869911193848, "learning_rate": 5e-05, "loss": 0.8964, "mean_token_accuracy": 0.8841755390167236, "num_tokens": 598050633.0, "step": 15752 }, { "epoch": 2.003943518636306, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.105819225311279, "learning_rate": 5e-05, "loss": 0.7468, "mean_token_accuracy": 0.8943933248519897, "num_tokens": 598088022.0, "step": 15753 }, { "epoch": 2.0040707289148965, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.628678321838379, "learning_rate": 5e-05, "loss": 0.8862, "mean_token_accuracy": 0.8847384452819824, "num_tokens": 598126333.0, "step": 15754 }, { "epoch": 2.004197939193487, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.363404750823975, "learning_rate": 5e-05, "loss": 0.8044, "mean_token_accuracy": 0.883698582649231, "num_tokens": 598161725.0, "step": 15755 }, { "epoch": 2.0043251494720775, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.303934097290039, "learning_rate": 5e-05, "loss": 0.8535, "mean_token_accuracy": 0.8872870206832886, "num_tokens": 598203158.0, "step": 15756 }, { "epoch": 2.004452359750668, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.4945807456970215, "learning_rate": 5e-05, "loss": 0.7892, "mean_token_accuracy": 0.887214720249176, "num_tokens": 598241093.0, "step": 15757 }, { "epoch": 2.0045795700292586, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.149993896484375, "learning_rate": 5e-05, "loss": 0.8437, "mean_token_accuracy": 0.8854426145553589, "num_tokens": 598276389.0, "step": 15758 }, { "epoch": 2.0047067803078487, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.370537757873535, "learning_rate": 5e-05, "loss": 0.7523, "mean_token_accuracy": 0.8889491558074951, "num_tokens": 598316645.0, "step": 15759 }, { "epoch": 2.004833990586439, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.25825023651123, "learning_rate": 5e-05, "loss": 0.8423, "mean_token_accuracy": 0.884692370891571, "num_tokens": 598361443.0, "step": 15760 }, { "epoch": 2.0049612008650297, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.121674537658691, "learning_rate": 5e-05, "loss": 0.7311, "mean_token_accuracy": 0.8869388699531555, "num_tokens": 598402934.0, "step": 15761 }, { "epoch": 2.0050884111436202, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.49932861328125, "learning_rate": 5e-05, "loss": 0.8713, "mean_token_accuracy": 0.881146252155304, "num_tokens": 598436315.0, "step": 15762 }, { "epoch": 2.0052156214222108, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.979647159576416, "learning_rate": 5e-05, "loss": 0.7128, "mean_token_accuracy": 0.8932406306266785, "num_tokens": 598476562.0, "step": 15763 }, { "epoch": 2.0053428317008013, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.586695671081543, "learning_rate": 5e-05, "loss": 0.8279, "mean_token_accuracy": 0.8952913284301758, "num_tokens": 598515180.0, "step": 15764 }, { "epoch": 2.005470041979392, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.092401027679443, "learning_rate": 5e-05, "loss": 0.7029, "mean_token_accuracy": 0.8978153467178345, "num_tokens": 598554177.0, "step": 15765 }, { "epoch": 2.0055972522579824, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.385201454162598, "learning_rate": 5e-05, "loss": 0.8595, "mean_token_accuracy": 0.8855434656143188, "num_tokens": 598589091.0, "step": 15766 }, { "epoch": 2.005724462536573, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.380975246429443, "learning_rate": 5e-05, "loss": 0.7524, "mean_token_accuracy": 0.8917195796966553, "num_tokens": 598624570.0, "step": 15767 }, { "epoch": 2.0058516728151634, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.980844497680664, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.904547929763794, "num_tokens": 598660350.0, "step": 15768 }, { "epoch": 2.005978883093754, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.666525840759277, "learning_rate": 5e-05, "loss": 0.8164, "mean_token_accuracy": 0.8725749254226685, "num_tokens": 598701823.0, "step": 15769 }, { "epoch": 2.0061060933723445, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.602801322937012, "learning_rate": 5e-05, "loss": 0.7284, "mean_token_accuracy": 0.8919416069984436, "num_tokens": 598738637.0, "step": 15770 }, { "epoch": 2.006233303650935, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.82694149017334, "learning_rate": 5e-05, "loss": 0.7608, "mean_token_accuracy": 0.8866481781005859, "num_tokens": 598773459.0, "step": 15771 }, { "epoch": 2.0063605139295255, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.408973217010498, "learning_rate": 5e-05, "loss": 0.7021, "mean_token_accuracy": 0.892713725566864, "num_tokens": 598810098.0, "step": 15772 }, { "epoch": 2.006487724208116, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.839086532592773, "learning_rate": 5e-05, "loss": 0.7635, "mean_token_accuracy": 0.8800103068351746, "num_tokens": 598850701.0, "step": 15773 }, { "epoch": 2.0066149344867066, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.361392974853516, "learning_rate": 5e-05, "loss": 0.6656, "mean_token_accuracy": 0.8990901708602905, "num_tokens": 598887745.0, "step": 15774 }, { "epoch": 2.006742144765297, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.812105655670166, "learning_rate": 5e-05, "loss": 0.7343, "mean_token_accuracy": 0.8884628415107727, "num_tokens": 598928360.0, "step": 15775 }, { "epoch": 2.0068693550438876, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.358500957489014, "learning_rate": 5e-05, "loss": 0.6775, "mean_token_accuracy": 0.89234459400177, "num_tokens": 598965628.0, "step": 15776 }, { "epoch": 2.006996565322478, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.796032905578613, "learning_rate": 5e-05, "loss": 0.7327, "mean_token_accuracy": 0.8870779275894165, "num_tokens": 599008445.0, "step": 15777 }, { "epoch": 2.0071237756010687, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.3405680656433105, "learning_rate": 5e-05, "loss": 0.6474, "mean_token_accuracy": 0.9036860466003418, "num_tokens": 599041730.0, "step": 15778 }, { "epoch": 2.007250985879659, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.810500144958496, "learning_rate": 5e-05, "loss": 0.7061, "mean_token_accuracy": 0.8902068138122559, "num_tokens": 599082481.0, "step": 15779 }, { "epoch": 2.0073781961582498, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.295473098754883, "learning_rate": 5e-05, "loss": 0.6826, "mean_token_accuracy": 0.8833888173103333, "num_tokens": 599122503.0, "step": 15780 }, { "epoch": 2.0075054064368403, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.861451148986816, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.8749879002571106, "num_tokens": 599158558.0, "step": 15781 }, { "epoch": 2.007632616715431, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.172756671905518, "learning_rate": 5e-05, "loss": 0.6697, "mean_token_accuracy": 0.8901017904281616, "num_tokens": 599196769.0, "step": 15782 }, { "epoch": 2.0077598269940213, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.014765739440918, "learning_rate": 5e-05, "loss": 0.7144, "mean_token_accuracy": 0.8890117406845093, "num_tokens": 599231050.0, "step": 15783 }, { "epoch": 2.0078870372726114, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.985184669494629, "learning_rate": 5e-05, "loss": 0.6411, "mean_token_accuracy": 0.8909143805503845, "num_tokens": 599271826.0, "step": 15784 }, { "epoch": 2.008014247551202, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.281614303588867, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.884230375289917, "num_tokens": 599311134.0, "step": 15785 }, { "epoch": 2.0081414578297925, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.689045429229736, "learning_rate": 5e-05, "loss": 0.6549, "mean_token_accuracy": 0.8898521065711975, "num_tokens": 599349240.0, "step": 15786 }, { "epoch": 2.008268668108383, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.691104888916016, "learning_rate": 5e-05, "loss": 0.7892, "mean_token_accuracy": 0.8884203433990479, "num_tokens": 599381624.0, "step": 15787 }, { "epoch": 2.0083958783869735, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.297518253326416, "learning_rate": 5e-05, "loss": 0.6253, "mean_token_accuracy": 0.8859527707099915, "num_tokens": 599421524.0, "step": 15788 }, { "epoch": 2.008523088665564, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.340354919433594, "learning_rate": 5e-05, "loss": 0.8934, "mean_token_accuracy": 0.8804745078086853, "num_tokens": 599459285.0, "step": 15789 }, { "epoch": 2.0086502989441546, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.042977333068848, "learning_rate": 5e-05, "loss": 0.6053, "mean_token_accuracy": 0.9013029336929321, "num_tokens": 599501682.0, "step": 15790 }, { "epoch": 2.008777509222745, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.865358352661133, "learning_rate": 5e-05, "loss": 0.93, "mean_token_accuracy": 0.8953020572662354, "num_tokens": 599538518.0, "step": 15791 }, { "epoch": 2.0089047195013356, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 6.729170322418213, "learning_rate": 5e-05, "loss": 0.7619, "mean_token_accuracy": 0.8929122686386108, "num_tokens": 599572802.0, "step": 15792 }, { "epoch": 2.009031929779926, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.316120147705078, "learning_rate": 5e-05, "loss": 0.9381, "mean_token_accuracy": 0.8909671306610107, "num_tokens": 599614366.0, "step": 15793 }, { "epoch": 2.0091591400585167, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.169242858886719, "learning_rate": 5e-05, "loss": 0.9449, "mean_token_accuracy": 0.8852423429489136, "num_tokens": 599654814.0, "step": 15794 }, { "epoch": 2.0092863503371072, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.3562164306640625, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.8887310028076172, "num_tokens": 599689678.0, "step": 15795 }, { "epoch": 2.0094135606156978, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.76739501953125, "learning_rate": 5e-05, "loss": 0.984, "mean_token_accuracy": 0.8841139078140259, "num_tokens": 599725932.0, "step": 15796 }, { "epoch": 2.0095407708942883, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 5.868081092834473, "learning_rate": 5e-05, "loss": 0.6208, "mean_token_accuracy": 0.9002598524093628, "num_tokens": 599762041.0, "step": 15797 }, { "epoch": 2.009667981172879, "ewc_loss": 0.92578125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00086212158203125, "grad_norm": 11.724296569824219, "learning_rate": 5e-05, "loss": 1.2323, "mean_token_accuracy": 0.8888173699378967, "num_tokens": 599800706.0, "step": 15798 }, { "epoch": 2.0097951914514693, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 7.766026973724365, "learning_rate": 5e-05, "loss": 1.0014, "mean_token_accuracy": 0.8724191188812256, "num_tokens": 599833483.0, "step": 15799 }, { "epoch": 2.00992240173006, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 11.935771942138672, "learning_rate": 5e-05, "loss": 1.2101, "mean_token_accuracy": 0.8879982233047485, "num_tokens": 599872172.0, "step": 15800 }, { "epoch": 2.0100496120086504, "ewc_loss": 1.015625, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.000957489013671875, "grad_norm": 9.925063133239746, "learning_rate": 5e-05, "loss": 1.3438, "mean_token_accuracy": 0.8850374221801758, "num_tokens": 599914149.0, "step": 15801 }, { "epoch": 2.010176822287241, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.547369956970215, "learning_rate": 5e-05, "loss": 1.0119, "mean_token_accuracy": 0.8874123096466064, "num_tokens": 599950603.0, "step": 15802 }, { "epoch": 2.0103040325658315, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.00061798095703125, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 9.460793495178223, "learning_rate": 5e-05, "loss": 1.1357, "mean_token_accuracy": 0.882780134677887, "num_tokens": 599991844.0, "step": 15803 }, { "epoch": 2.010431242844422, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 9.569031715393066, "learning_rate": 5e-05, "loss": 1.0097, "mean_token_accuracy": 0.8914132118225098, "num_tokens": 600033303.0, "step": 15804 }, { "epoch": 2.0105584531230125, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.928491592407227, "learning_rate": 5e-05, "loss": 0.9146, "mean_token_accuracy": 0.8993380069732666, "num_tokens": 600065633.0, "step": 15805 }, { "epoch": 2.010685663401603, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.92926025390625, "learning_rate": 5e-05, "loss": 0.9643, "mean_token_accuracy": 0.8845801949501038, "num_tokens": 600104333.0, "step": 15806 }, { "epoch": 2.0108128736801936, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.223976135253906, "learning_rate": 5e-05, "loss": 0.8801, "mean_token_accuracy": 0.8809693455696106, "num_tokens": 600141329.0, "step": 15807 }, { "epoch": 2.0109400839587837, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.869882583618164, "learning_rate": 5e-05, "loss": 0.952, "mean_token_accuracy": 0.8750687837600708, "num_tokens": 600184171.0, "step": 15808 }, { "epoch": 2.011067294237374, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.65703821182251, "learning_rate": 5e-05, "loss": 0.8523, "mean_token_accuracy": 0.8729530572891235, "num_tokens": 600220249.0, "step": 15809 }, { "epoch": 2.0111945045159647, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.96920394897461, "learning_rate": 5e-05, "loss": 0.9263, "mean_token_accuracy": 0.8823447227478027, "num_tokens": 600256015.0, "step": 15810 }, { "epoch": 2.0113217147945552, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.346317768096924, "learning_rate": 5e-05, "loss": 0.7753, "mean_token_accuracy": 0.8874424695968628, "num_tokens": 600282893.0, "step": 15811 }, { "epoch": 2.0114489250731458, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.125566482543945, "learning_rate": 5e-05, "loss": 0.975, "mean_token_accuracy": 0.8709341883659363, "num_tokens": 600320940.0, "step": 15812 }, { "epoch": 2.0115761353517363, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.198540687561035, "learning_rate": 5e-05, "loss": 0.8406, "mean_token_accuracy": 0.8696228265762329, "num_tokens": 600362054.0, "step": 15813 }, { "epoch": 2.011703345630327, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.372042655944824, "learning_rate": 5e-05, "loss": 0.9904, "mean_token_accuracy": 0.8750065565109253, "num_tokens": 600395080.0, "step": 15814 }, { "epoch": 2.0118305559089174, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.227345943450928, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8957333564758301, "num_tokens": 600428060.0, "step": 15815 }, { "epoch": 2.011957766187508, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.587058067321777, "learning_rate": 5e-05, "loss": 0.972, "mean_token_accuracy": 0.8933106064796448, "num_tokens": 600469833.0, "step": 15816 }, { "epoch": 2.0120849764660984, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 7.568548202514648, "learning_rate": 5e-05, "loss": 0.8459, "mean_token_accuracy": 0.8892914056777954, "num_tokens": 600498006.0, "step": 15817 }, { "epoch": 2.012212186744689, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.39797306060791, "learning_rate": 5e-05, "loss": 0.9871, "mean_token_accuracy": 0.8856251239776611, "num_tokens": 600534744.0, "step": 15818 }, { "epoch": 2.0123393970232795, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.983377456665039, "learning_rate": 5e-05, "loss": 0.9189, "mean_token_accuracy": 0.8842126727104187, "num_tokens": 600570282.0, "step": 15819 }, { "epoch": 2.01246660730187, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.796009063720703, "learning_rate": 5e-05, "loss": 0.9035, "mean_token_accuracy": 0.8959675431251526, "num_tokens": 600603858.0, "step": 15820 }, { "epoch": 2.0125938175804605, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.097450256347656, "learning_rate": 5e-05, "loss": 0.9272, "mean_token_accuracy": 0.8741604089736938, "num_tokens": 600642078.0, "step": 15821 }, { "epoch": 2.012721027859051, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.269317626953125, "learning_rate": 5e-05, "loss": 0.8916, "mean_token_accuracy": 0.8815101981163025, "num_tokens": 600679424.0, "step": 15822 }, { "epoch": 2.0128482381376416, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.849304676055908, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8944909572601318, "num_tokens": 600716310.0, "step": 15823 }, { "epoch": 2.012975448416232, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.249933242797852, "learning_rate": 5e-05, "loss": 0.8997, "mean_token_accuracy": 0.8767368793487549, "num_tokens": 600757031.0, "step": 15824 }, { "epoch": 2.0131026586948226, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.378594398498535, "learning_rate": 5e-05, "loss": 0.7498, "mean_token_accuracy": 0.8844107389450073, "num_tokens": 600795414.0, "step": 15825 }, { "epoch": 2.013229868973413, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.672082901000977, "learning_rate": 5e-05, "loss": 0.9054, "mean_token_accuracy": 0.8791056871414185, "num_tokens": 600829239.0, "step": 15826 }, { "epoch": 2.0133570792520037, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.99350643157959, "learning_rate": 5e-05, "loss": 0.6935, "mean_token_accuracy": 0.894874632358551, "num_tokens": 600868155.0, "step": 15827 }, { "epoch": 2.013484289530594, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.08188533782959, "learning_rate": 5e-05, "loss": 0.9338, "mean_token_accuracy": 0.8794828653335571, "num_tokens": 600906797.0, "step": 15828 }, { "epoch": 2.0136114998091847, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.145682334899902, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8783947229385376, "num_tokens": 600948399.0, "step": 15829 }, { "epoch": 2.0137387100877753, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.953326225280762, "learning_rate": 5e-05, "loss": 0.9387, "mean_token_accuracy": 0.8843590021133423, "num_tokens": 600989682.0, "step": 15830 }, { "epoch": 2.013865920366366, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.6664581298828125, "learning_rate": 5e-05, "loss": 0.8314, "mean_token_accuracy": 0.883049726486206, "num_tokens": 601025387.0, "step": 15831 }, { "epoch": 2.0139931306449563, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.280964851379395, "learning_rate": 5e-05, "loss": 0.8375, "mean_token_accuracy": 0.8983477354049683, "num_tokens": 601061124.0, "step": 15832 }, { "epoch": 2.0141203409235464, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.031471252441406, "learning_rate": 5e-05, "loss": 0.8641, "mean_token_accuracy": 0.8841291666030884, "num_tokens": 601105789.0, "step": 15833 }, { "epoch": 2.014247551202137, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.594549179077148, "learning_rate": 5e-05, "loss": 0.8147, "mean_token_accuracy": 0.8718689680099487, "num_tokens": 601138697.0, "step": 15834 }, { "epoch": 2.0143747614807275, "ewc_loss": 0.5, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.146181106567383, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.8970656394958496, "num_tokens": 601172875.0, "step": 15835 }, { "epoch": 2.014501971759318, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.320470333099365, "learning_rate": 5e-05, "loss": 0.6961, "mean_token_accuracy": 0.8977354764938354, "num_tokens": 601206303.0, "step": 15836 }, { "epoch": 2.0146291820379085, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.065080642700195, "learning_rate": 5e-05, "loss": 0.7367, "mean_token_accuracy": 0.9006213545799255, "num_tokens": 601243879.0, "step": 15837 }, { "epoch": 2.014756392316499, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.367557525634766, "learning_rate": 5e-05, "loss": 0.7079, "mean_token_accuracy": 0.8898096084594727, "num_tokens": 601282639.0, "step": 15838 }, { "epoch": 2.0148836025950896, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.861420631408691, "learning_rate": 5e-05, "loss": 0.7614, "mean_token_accuracy": 0.8837878704071045, "num_tokens": 601326204.0, "step": 15839 }, { "epoch": 2.01501081287368, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.486178398132324, "learning_rate": 5e-05, "loss": 0.6935, "mean_token_accuracy": 0.890672504901886, "num_tokens": 601358936.0, "step": 15840 }, { "epoch": 2.0151380231522706, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.705763816833496, "learning_rate": 5e-05, "loss": 0.6875, "mean_token_accuracy": 0.8986142873764038, "num_tokens": 601398253.0, "step": 15841 }, { "epoch": 2.015265233430861, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.547107696533203, "learning_rate": 5e-05, "loss": 0.6268, "mean_token_accuracy": 0.9128807187080383, "num_tokens": 601434806.0, "step": 15842 }, { "epoch": 2.0153924437094517, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.619999408721924, "learning_rate": 5e-05, "loss": 0.6996, "mean_token_accuracy": 0.8896785974502563, "num_tokens": 601471834.0, "step": 15843 }, { "epoch": 2.0155196539880422, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.585836410522461, "learning_rate": 5e-05, "loss": 0.6816, "mean_token_accuracy": 0.8959698677062988, "num_tokens": 601511256.0, "step": 15844 }, { "epoch": 2.0156468642666328, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.534457206726074, "learning_rate": 5e-05, "loss": 0.672, "mean_token_accuracy": 0.8941360116004944, "num_tokens": 601547434.0, "step": 15845 }, { "epoch": 2.0157740745452233, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.634998798370361, "learning_rate": 5e-05, "loss": 0.6782, "mean_token_accuracy": 0.8955861330032349, "num_tokens": 601581375.0, "step": 15846 }, { "epoch": 2.015901284823814, "ewc_loss": 0.375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.466219425201416, "learning_rate": 5e-05, "loss": 0.7081, "mean_token_accuracy": 0.8813506364822388, "num_tokens": 601615067.0, "step": 15847 }, { "epoch": 2.0160284951024043, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.679328441619873, "learning_rate": 5e-05, "loss": 0.6364, "mean_token_accuracy": 0.9052014350891113, "num_tokens": 601654168.0, "step": 15848 }, { "epoch": 2.016155705380995, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.400872230529785, "learning_rate": 5e-05, "loss": 0.6516, "mean_token_accuracy": 0.8969888687133789, "num_tokens": 601687113.0, "step": 15849 }, { "epoch": 2.0162829156595854, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.707812786102295, "learning_rate": 5e-05, "loss": 0.7168, "mean_token_accuracy": 0.8802903890609741, "num_tokens": 601723993.0, "step": 15850 }, { "epoch": 2.016410125938176, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.340231418609619, "learning_rate": 5e-05, "loss": 0.7142, "mean_token_accuracy": 0.8730741739273071, "num_tokens": 601758522.0, "step": 15851 }, { "epoch": 2.0165373362167665, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.777232646942139, "learning_rate": 5e-05, "loss": 0.6431, "mean_token_accuracy": 0.9070339202880859, "num_tokens": 601794221.0, "step": 15852 }, { "epoch": 2.016664546495357, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.2350592613220215, "learning_rate": 5e-05, "loss": 0.6332, "mean_token_accuracy": 0.8970908522605896, "num_tokens": 601830430.0, "step": 15853 }, { "epoch": 2.0167917567739475, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.921504974365234, "learning_rate": 5e-05, "loss": 0.7016, "mean_token_accuracy": 0.889337420463562, "num_tokens": 601871834.0, "step": 15854 }, { "epoch": 2.016918967052538, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 7.030170440673828, "learning_rate": 5e-05, "loss": 0.6627, "mean_token_accuracy": 0.883120596408844, "num_tokens": 601911432.0, "step": 15855 }, { "epoch": 2.0170461773311286, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.228141784667969, "learning_rate": 5e-05, "loss": 0.7019, "mean_token_accuracy": 0.8969581127166748, "num_tokens": 601955592.0, "step": 15856 }, { "epoch": 2.0171733876097186, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.64216423034668, "learning_rate": 5e-05, "loss": 0.6475, "mean_token_accuracy": 0.8800078630447388, "num_tokens": 601996843.0, "step": 15857 }, { "epoch": 2.017300597888309, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.843247413635254, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8933406472206116, "num_tokens": 602034963.0, "step": 15858 }, { "epoch": 2.0174278081668997, "ewc_loss": 0.28515625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00022220611572265625, "grad_norm": 6.00527286529541, "learning_rate": 5e-05, "loss": 0.6448, "mean_token_accuracy": 0.8853187561035156, "num_tokens": 602073460.0, "step": 15859 }, { "epoch": 2.0175550184454902, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.874733924865723, "learning_rate": 5e-05, "loss": 0.9662, "mean_token_accuracy": 0.8709014654159546, "num_tokens": 602117442.0, "step": 15860 }, { "epoch": 2.0176822287240808, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 5.557884693145752, "learning_rate": 5e-05, "loss": 0.6016, "mean_token_accuracy": 0.8932752013206482, "num_tokens": 602161253.0, "step": 15861 }, { "epoch": 2.0178094390026713, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 11.156394958496094, "learning_rate": 5e-05, "loss": 1.1421, "mean_token_accuracy": 0.8783000707626343, "num_tokens": 602204324.0, "step": 15862 }, { "epoch": 2.017936649281262, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.09519624710083, "learning_rate": 5e-05, "loss": 0.8656, "mean_token_accuracy": 0.8921455144882202, "num_tokens": 602236335.0, "step": 15863 }, { "epoch": 2.0180638595598523, "ewc_loss": 0.828125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000766754150390625, "grad_norm": 11.13270378112793, "learning_rate": 5e-05, "loss": 1.1703, "mean_token_accuracy": 0.8784291744232178, "num_tokens": 602271302.0, "step": 15864 }, { "epoch": 2.018191069838443, "ewc_loss": 0.9296875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000865936279296875, "grad_norm": 9.364083290100098, "learning_rate": 5e-05, "loss": 1.2251, "mean_token_accuracy": 0.8927785158157349, "num_tokens": 602304103.0, "step": 15865 }, { "epoch": 2.0183182801170334, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.948896408081055, "learning_rate": 5e-05, "loss": 0.8551, "mean_token_accuracy": 0.8877331018447876, "num_tokens": 602343293.0, "step": 15866 }, { "epoch": 2.018445490395624, "ewc_loss": 0.91796875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 10.418174743652344, "learning_rate": 5e-05, "loss": 1.2545, "mean_token_accuracy": 0.8804587125778198, "num_tokens": 602383650.0, "step": 15867 }, { "epoch": 2.0185727006742145, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.530839443206787, "learning_rate": 5e-05, "loss": 0.8073, "mean_token_accuracy": 0.8959356546401978, "num_tokens": 602423436.0, "step": 15868 }, { "epoch": 2.018699910952805, "ewc_loss": 0.96875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00090789794921875, "grad_norm": 12.0841646194458, "learning_rate": 5e-05, "loss": 1.2862, "mean_token_accuracy": 0.8878466486930847, "num_tokens": 602460909.0, "step": 15869 }, { "epoch": 2.0188271212313955, "ewc_loss": 0.8828125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 9.467467308044434, "learning_rate": 5e-05, "loss": 1.1964, "mean_token_accuracy": 0.8876023292541504, "num_tokens": 602499610.0, "step": 15870 }, { "epoch": 2.018954331509986, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 11.364165306091309, "learning_rate": 5e-05, "loss": 1.0577, "mean_token_accuracy": 0.8841363191604614, "num_tokens": 602540783.0, "step": 15871 }, { "epoch": 2.0190815417885766, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 9.532455444335938, "learning_rate": 5e-05, "loss": 1.2254, "mean_token_accuracy": 0.8805310726165771, "num_tokens": 602579205.0, "step": 15872 }, { "epoch": 2.019208752067167, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.872392654418945, "learning_rate": 5e-05, "loss": 1.0392, "mean_token_accuracy": 0.8878943920135498, "num_tokens": 602613720.0, "step": 15873 }, { "epoch": 2.0193359623457576, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.618947982788086, "learning_rate": 5e-05, "loss": 0.9578, "mean_token_accuracy": 0.8897093534469604, "num_tokens": 602651954.0, "step": 15874 }, { "epoch": 2.019463172624348, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.608023643493652, "learning_rate": 5e-05, "loss": 1.0051, "mean_token_accuracy": 0.8972399830818176, "num_tokens": 602685553.0, "step": 15875 }, { "epoch": 2.0195903829029387, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.942466735839844, "learning_rate": 5e-05, "loss": 0.8846, "mean_token_accuracy": 0.8809054493904114, "num_tokens": 602727854.0, "step": 15876 }, { "epoch": 2.019717593181529, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.966525077819824, "learning_rate": 5e-05, "loss": 1.075, "mean_token_accuracy": 0.8881654739379883, "num_tokens": 602765993.0, "step": 15877 }, { "epoch": 2.0198448034601197, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.058541297912598, "learning_rate": 5e-05, "loss": 0.8658, "mean_token_accuracy": 0.9001016616821289, "num_tokens": 602807450.0, "step": 15878 }, { "epoch": 2.0199720137387103, "ewc_loss": 0.796875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000732421875, "grad_norm": 11.48832893371582, "learning_rate": 5e-05, "loss": 1.1493, "mean_token_accuracy": 0.874160885810852, "num_tokens": 602848689.0, "step": 15879 }, { "epoch": 2.020099224017301, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.02749252319336, "learning_rate": 5e-05, "loss": 0.9844, "mean_token_accuracy": 0.8995373249053955, "num_tokens": 602888010.0, "step": 15880 }, { "epoch": 2.0202264342958913, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 11.124085426330566, "learning_rate": 5e-05, "loss": 1.1314, "mean_token_accuracy": 0.8759901523590088, "num_tokens": 602923345.0, "step": 15881 }, { "epoch": 2.0203536445744814, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.428177833557129, "learning_rate": 5e-05, "loss": 1.0043, "mean_token_accuracy": 0.8897916078567505, "num_tokens": 602967320.0, "step": 15882 }, { "epoch": 2.020480854853072, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.368647575378418, "learning_rate": 5e-05, "loss": 0.9667, "mean_token_accuracy": 0.8964922428131104, "num_tokens": 603005177.0, "step": 15883 }, { "epoch": 2.0206080651316625, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.193934440612793, "learning_rate": 5e-05, "loss": 0.9197, "mean_token_accuracy": 0.8957597613334656, "num_tokens": 603039562.0, "step": 15884 }, { "epoch": 2.020735275410253, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.583585739135742, "learning_rate": 5e-05, "loss": 0.9462, "mean_token_accuracy": 0.8938586711883545, "num_tokens": 603079075.0, "step": 15885 }, { "epoch": 2.0208624856888435, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.670677185058594, "learning_rate": 5e-05, "loss": 0.8741, "mean_token_accuracy": 0.8859506845474243, "num_tokens": 603118077.0, "step": 15886 }, { "epoch": 2.020989695967434, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.22839069366455, "learning_rate": 5e-05, "loss": 0.9274, "mean_token_accuracy": 0.8805440068244934, "num_tokens": 603156921.0, "step": 15887 }, { "epoch": 2.0211169062460246, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.954514503479004, "learning_rate": 5e-05, "loss": 0.7839, "mean_token_accuracy": 0.886681854724884, "num_tokens": 603202062.0, "step": 15888 }, { "epoch": 2.021244116524615, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.479850769042969, "learning_rate": 5e-05, "loss": 0.9104, "mean_token_accuracy": 0.8945074677467346, "num_tokens": 603236240.0, "step": 15889 }, { "epoch": 2.0213713268032056, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.505341053009033, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8857988119125366, "num_tokens": 603277376.0, "step": 15890 }, { "epoch": 2.021498537081796, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.913044929504395, "learning_rate": 5e-05, "loss": 0.9672, "mean_token_accuracy": 0.886546790599823, "num_tokens": 603314740.0, "step": 15891 }, { "epoch": 2.0216257473603867, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.750032901763916, "learning_rate": 5e-05, "loss": 0.7756, "mean_token_accuracy": 0.8880044221878052, "num_tokens": 603354992.0, "step": 15892 }, { "epoch": 2.021752957638977, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.032382011413574, "learning_rate": 5e-05, "loss": 1.0053, "mean_token_accuracy": 0.8853259086608887, "num_tokens": 603387271.0, "step": 15893 }, { "epoch": 2.0218801679175677, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.395770072937012, "learning_rate": 5e-05, "loss": 0.8398, "mean_token_accuracy": 0.8960628509521484, "num_tokens": 603427092.0, "step": 15894 }, { "epoch": 2.0220073781961583, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.375244140625, "learning_rate": 5e-05, "loss": 0.9659, "mean_token_accuracy": 0.8759059906005859, "num_tokens": 603465188.0, "step": 15895 }, { "epoch": 2.022134588474749, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.636185646057129, "learning_rate": 5e-05, "loss": 0.8576, "mean_token_accuracy": 0.8872662782669067, "num_tokens": 603499290.0, "step": 15896 }, { "epoch": 2.0222617987533393, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.459120750427246, "learning_rate": 5e-05, "loss": 0.835, "mean_token_accuracy": 0.8888837099075317, "num_tokens": 603530594.0, "step": 15897 }, { "epoch": 2.02238900903193, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.348296165466309, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8892825841903687, "num_tokens": 603569337.0, "step": 15898 }, { "epoch": 2.0225162193105204, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.093537330627441, "learning_rate": 5e-05, "loss": 0.7867, "mean_token_accuracy": 0.8894449472427368, "num_tokens": 603609704.0, "step": 15899 }, { "epoch": 2.022643429589111, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.797413349151611, "learning_rate": 5e-05, "loss": 0.7282, "mean_token_accuracy": 0.8897971510887146, "num_tokens": 603647623.0, "step": 15900 }, { "epoch": 2.0227706398677014, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.227716445922852, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8896933794021606, "num_tokens": 603682237.0, "step": 15901 }, { "epoch": 2.022897850146292, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.354057788848877, "learning_rate": 5e-05, "loss": 0.6992, "mean_token_accuracy": 0.8826246857643127, "num_tokens": 603720348.0, "step": 15902 }, { "epoch": 2.0230250604248825, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.442000389099121, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.88718181848526, "num_tokens": 603759857.0, "step": 15903 }, { "epoch": 2.023152270703473, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.0592732429504395, "learning_rate": 5e-05, "loss": 0.6783, "mean_token_accuracy": 0.8860304951667786, "num_tokens": 603796970.0, "step": 15904 }, { "epoch": 2.0232794809820636, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.752872467041016, "learning_rate": 5e-05, "loss": 0.8393, "mean_token_accuracy": 0.8834248781204224, "num_tokens": 603838385.0, "step": 15905 }, { "epoch": 2.0234066912606536, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.841117858886719, "learning_rate": 5e-05, "loss": 0.6702, "mean_token_accuracy": 0.8862860202789307, "num_tokens": 603884262.0, "step": 15906 }, { "epoch": 2.023533901539244, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.129473686218262, "learning_rate": 5e-05, "loss": 0.8956, "mean_token_accuracy": 0.8812144994735718, "num_tokens": 603925013.0, "step": 15907 }, { "epoch": 2.0236611118178347, "ewc_loss": 0.375, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.027838706970215, "learning_rate": 5e-05, "loss": 0.7066, "mean_token_accuracy": 0.8812413811683655, "num_tokens": 603971173.0, "step": 15908 }, { "epoch": 2.0237883220964252, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000621795654296875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.003833770751953, "learning_rate": 5e-05, "loss": 0.8952, "mean_token_accuracy": 0.8853498697280884, "num_tokens": 604009091.0, "step": 15909 }, { "epoch": 2.0239155323750158, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.6685566902160645, "learning_rate": 5e-05, "loss": 0.7418, "mean_token_accuracy": 0.8906699419021606, "num_tokens": 604045163.0, "step": 15910 }, { "epoch": 2.0240427426536063, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.163307189941406, "learning_rate": 5e-05, "loss": 0.7825, "mean_token_accuracy": 0.8905161023139954, "num_tokens": 604091379.0, "step": 15911 }, { "epoch": 2.024169952932197, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.14178466796875, "learning_rate": 5e-05, "loss": 0.7841, "mean_token_accuracy": 0.8851764798164368, "num_tokens": 604132159.0, "step": 15912 }, { "epoch": 2.0242971632107873, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.396103382110596, "learning_rate": 5e-05, "loss": 0.7444, "mean_token_accuracy": 0.8773659467697144, "num_tokens": 604169689.0, "step": 15913 }, { "epoch": 2.024424373489378, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.396215438842773, "learning_rate": 5e-05, "loss": 0.7141, "mean_token_accuracy": 0.9054064750671387, "num_tokens": 604201913.0, "step": 15914 }, { "epoch": 2.0245515837679684, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.049032211303711, "learning_rate": 5e-05, "loss": 0.6995, "mean_token_accuracy": 0.8836711049079895, "num_tokens": 604238275.0, "step": 15915 }, { "epoch": 2.024678794046559, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.595351219177246, "learning_rate": 5e-05, "loss": 0.7789, "mean_token_accuracy": 0.8932852149009705, "num_tokens": 604279862.0, "step": 15916 }, { "epoch": 2.0248060043251495, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.948339939117432, "learning_rate": 5e-05, "loss": 0.6807, "mean_token_accuracy": 0.884178102016449, "num_tokens": 604319091.0, "step": 15917 }, { "epoch": 2.02493321460374, "ewc_loss": 0.5, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.643749237060547, "learning_rate": 5e-05, "loss": 0.8338, "mean_token_accuracy": 0.8802639245986938, "num_tokens": 604363614.0, "step": 15918 }, { "epoch": 2.0250604248823305, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.146336555480957, "learning_rate": 5e-05, "loss": 0.6878, "mean_token_accuracy": 0.8835930824279785, "num_tokens": 604394422.0, "step": 15919 }, { "epoch": 2.025187635160921, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.280712127685547, "learning_rate": 5e-05, "loss": 0.8038, "mean_token_accuracy": 0.8824570775032043, "num_tokens": 604432844.0, "step": 15920 }, { "epoch": 2.0253148454395116, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.533712863922119, "learning_rate": 5e-05, "loss": 0.7005, "mean_token_accuracy": 0.8909660577774048, "num_tokens": 604472074.0, "step": 15921 }, { "epoch": 2.025442055718102, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.77426290512085, "learning_rate": 5e-05, "loss": 0.6856, "mean_token_accuracy": 0.9019935131072998, "num_tokens": 604507827.0, "step": 15922 }, { "epoch": 2.0255692659966926, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.8638410568237305, "learning_rate": 5e-05, "loss": 0.7352, "mean_token_accuracy": 0.8816340565681458, "num_tokens": 604548600.0, "step": 15923 }, { "epoch": 2.025696476275283, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.391618251800537, "learning_rate": 5e-05, "loss": 0.7343, "mean_token_accuracy": 0.8761192560195923, "num_tokens": 604582701.0, "step": 15924 }, { "epoch": 2.0258236865538737, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.050188064575195, "learning_rate": 5e-05, "loss": 0.752, "mean_token_accuracy": 0.8782801628112793, "num_tokens": 604624089.0, "step": 15925 }, { "epoch": 2.025950896832464, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.266136169433594, "learning_rate": 5e-05, "loss": 0.6524, "mean_token_accuracy": 0.8968508839607239, "num_tokens": 604655764.0, "step": 15926 }, { "epoch": 2.0260781071110547, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.022037506103516, "learning_rate": 5e-05, "loss": 0.7101, "mean_token_accuracy": 0.893914520740509, "num_tokens": 604688000.0, "step": 15927 }, { "epoch": 2.0262053173896453, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.357584476470947, "learning_rate": 5e-05, "loss": 0.7114, "mean_token_accuracy": 0.8786345720291138, "num_tokens": 604724602.0, "step": 15928 }, { "epoch": 2.026332527668236, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.812272071838379, "learning_rate": 5e-05, "loss": 0.7062, "mean_token_accuracy": 0.8890827894210815, "num_tokens": 604761794.0, "step": 15929 }, { "epoch": 2.0264597379468263, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.5442681312561035, "learning_rate": 5e-05, "loss": 0.6419, "mean_token_accuracy": 0.9012753963470459, "num_tokens": 604800239.0, "step": 15930 }, { "epoch": 2.0265869482254164, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.574394702911377, "learning_rate": 5e-05, "loss": 0.6988, "mean_token_accuracy": 0.8894588947296143, "num_tokens": 604837143.0, "step": 15931 }, { "epoch": 2.026714158504007, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.736872673034668, "learning_rate": 5e-05, "loss": 0.7025, "mean_token_accuracy": 0.8868601322174072, "num_tokens": 604874074.0, "step": 15932 }, { "epoch": 2.0268413687825975, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.386153697967529, "learning_rate": 5e-05, "loss": 0.6927, "mean_token_accuracy": 0.8823367953300476, "num_tokens": 604911639.0, "step": 15933 }, { "epoch": 2.026968579061188, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.860753536224365, "learning_rate": 5e-05, "loss": 0.665, "mean_token_accuracy": 0.8986409902572632, "num_tokens": 604944963.0, "step": 15934 }, { "epoch": 2.0270957893397785, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0006256103515625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.260420799255371, "learning_rate": 5e-05, "loss": 0.6686, "mean_token_accuracy": 0.885883092880249, "num_tokens": 604986048.0, "step": 15935 }, { "epoch": 2.027222999618369, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.93646764755249, "learning_rate": 5e-05, "loss": 0.6907, "mean_token_accuracy": 0.8924235701560974, "num_tokens": 605023995.0, "step": 15936 }, { "epoch": 2.0273502098969596, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.182614803314209, "learning_rate": 5e-05, "loss": 0.6396, "mean_token_accuracy": 0.8942129611968994, "num_tokens": 605064612.0, "step": 15937 }, { "epoch": 2.02747742017555, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.030104637145996, "learning_rate": 5e-05, "loss": 0.7553, "mean_token_accuracy": 0.8807602524757385, "num_tokens": 605099332.0, "step": 15938 }, { "epoch": 2.0276046304541406, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 7.040288925170898, "learning_rate": 5e-05, "loss": 0.6459, "mean_token_accuracy": 0.8915555477142334, "num_tokens": 605137616.0, "step": 15939 }, { "epoch": 2.027731840732731, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.249651908874512, "learning_rate": 5e-05, "loss": 0.7466, "mean_token_accuracy": 0.8845257759094238, "num_tokens": 605177974.0, "step": 15940 }, { "epoch": 2.0278590510113217, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.743473052978516, "learning_rate": 5e-05, "loss": 0.6518, "mean_token_accuracy": 0.882866382598877, "num_tokens": 605219156.0, "step": 15941 }, { "epoch": 2.027986261289912, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.70256233215332, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8894681334495544, "num_tokens": 605261131.0, "step": 15942 }, { "epoch": 2.0281134715685027, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 6.259068965911865, "learning_rate": 5e-05, "loss": 0.621, "mean_token_accuracy": 0.8807369470596313, "num_tokens": 605303640.0, "step": 15943 }, { "epoch": 2.0282406818470933, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 9.50555419921875, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.8950295448303223, "num_tokens": 605341823.0, "step": 15944 }, { "epoch": 2.028367892125684, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 5.753530025482178, "learning_rate": 5e-05, "loss": 0.5836, "mean_token_accuracy": 0.897007405757904, "num_tokens": 605380282.0, "step": 15945 }, { "epoch": 2.0284951024042743, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.515583038330078, "learning_rate": 5e-05, "loss": 0.976, "mean_token_accuracy": 0.8945369720458984, "num_tokens": 605415077.0, "step": 15946 }, { "epoch": 2.028622312682865, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.41792106628418, "learning_rate": 5e-05, "loss": 0.7271, "mean_token_accuracy": 0.8847841024398804, "num_tokens": 605453805.0, "step": 15947 }, { "epoch": 2.0287495229614554, "ewc_loss": 0.78125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 10.70858383178711, "learning_rate": 5e-05, "loss": 1.1088, "mean_token_accuracy": 0.8851656913757324, "num_tokens": 605488740.0, "step": 15948 }, { "epoch": 2.028876733240046, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 8.501882553100586, "learning_rate": 5e-05, "loss": 1.0528, "mean_token_accuracy": 0.8782342672348022, "num_tokens": 605529155.0, "step": 15949 }, { "epoch": 2.0290039435186364, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.821578025817871, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8885424733161926, "num_tokens": 605567610.0, "step": 15950 }, { "epoch": 2.029131153797227, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 9.824710845947266, "learning_rate": 5e-05, "loss": 1.1062, "mean_token_accuracy": 0.8841580748558044, "num_tokens": 605609485.0, "step": 15951 }, { "epoch": 2.0292583640758175, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.750496864318848, "learning_rate": 5e-05, "loss": 0.6721, "mean_token_accuracy": 0.8971292972564697, "num_tokens": 605644718.0, "step": 15952 }, { "epoch": 2.029385574354408, "ewc_loss": 0.91796875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 11.813401222229004, "learning_rate": 5e-05, "loss": 1.2351, "mean_token_accuracy": 0.8877391219139099, "num_tokens": 605686893.0, "step": 15953 }, { "epoch": 2.0295127846329986, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.159200668334961, "learning_rate": 5e-05, "loss": 0.9767, "mean_token_accuracy": 0.8857808113098145, "num_tokens": 605724007.0, "step": 15954 }, { "epoch": 2.0296399949115886, "ewc_loss": 0.8828125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 12.160649299621582, "learning_rate": 5e-05, "loss": 1.2277, "mean_token_accuracy": 0.8789026737213135, "num_tokens": 605761213.0, "step": 15955 }, { "epoch": 2.029767205190179, "ewc_loss": 0.95703125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00089263916015625, "grad_norm": 9.907724380493164, "learning_rate": 5e-05, "loss": 1.2802, "mean_token_accuracy": 0.8869098424911499, "num_tokens": 605803274.0, "step": 15956 }, { "epoch": 2.0298944154687697, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 11.241925239562988, "learning_rate": 5e-05, "loss": 1.0876, "mean_token_accuracy": 0.8837518692016602, "num_tokens": 605838326.0, "step": 15957 }, { "epoch": 2.0300216257473602, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 9.2512845993042, "learning_rate": 5e-05, "loss": 1.0538, "mean_token_accuracy": 0.891721785068512, "num_tokens": 605878882.0, "step": 15958 }, { "epoch": 2.0301488360259508, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 10.882040977478027, "learning_rate": 5e-05, "loss": 1.0613, "mean_token_accuracy": 0.8899214267730713, "num_tokens": 605913574.0, "step": 15959 }, { "epoch": 2.0302760463045413, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.589920997619629, "learning_rate": 5e-05, "loss": 0.8847, "mean_token_accuracy": 0.8946901559829712, "num_tokens": 605944845.0, "step": 15960 }, { "epoch": 2.030403256583132, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 10.819129943847656, "learning_rate": 5e-05, "loss": 1.1554, "mean_token_accuracy": 0.8670016527175903, "num_tokens": 605982604.0, "step": 15961 }, { "epoch": 2.0305304668617223, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.454838752746582, "learning_rate": 5e-05, "loss": 0.9353, "mean_token_accuracy": 0.8802415728569031, "num_tokens": 606023506.0, "step": 15962 }, { "epoch": 2.030657677140313, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.987732887268066, "learning_rate": 5e-05, "loss": 1.134, "mean_token_accuracy": 0.877636194229126, "num_tokens": 606063319.0, "step": 15963 }, { "epoch": 2.0307848874189034, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.862563133239746, "learning_rate": 5e-05, "loss": 0.9231, "mean_token_accuracy": 0.9012777805328369, "num_tokens": 606103393.0, "step": 15964 }, { "epoch": 2.030912097697494, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.738496780395508, "learning_rate": 5e-05, "loss": 1.033, "mean_token_accuracy": 0.8956372737884521, "num_tokens": 606144191.0, "step": 15965 }, { "epoch": 2.0310393079760845, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.073431968688965, "learning_rate": 5e-05, "loss": 0.9648, "mean_token_accuracy": 0.8918018341064453, "num_tokens": 606183219.0, "step": 15966 }, { "epoch": 2.031166518254675, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.19865608215332, "learning_rate": 5e-05, "loss": 1.018, "mean_token_accuracy": 0.8843601942062378, "num_tokens": 606221429.0, "step": 15967 }, { "epoch": 2.0312937285332655, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.880846977233887, "learning_rate": 5e-05, "loss": 0.9108, "mean_token_accuracy": 0.8925881385803223, "num_tokens": 606260810.0, "step": 15968 }, { "epoch": 2.031420938811856, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.771547317504883, "learning_rate": 5e-05, "loss": 0.9568, "mean_token_accuracy": 0.8936454653739929, "num_tokens": 606299602.0, "step": 15969 }, { "epoch": 2.0315481490904466, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.493144989013672, "learning_rate": 5e-05, "loss": 0.8566, "mean_token_accuracy": 0.8917133808135986, "num_tokens": 606341130.0, "step": 15970 }, { "epoch": 2.031675359369037, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.620966911315918, "learning_rate": 5e-05, "loss": 0.9423, "mean_token_accuracy": 0.8905883431434631, "num_tokens": 606378341.0, "step": 15971 }, { "epoch": 2.0318025696476276, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.128868103027344, "learning_rate": 5e-05, "loss": 0.8685, "mean_token_accuracy": 0.8689507246017456, "num_tokens": 606413799.0, "step": 15972 }, { "epoch": 2.031929779926218, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.766683578491211, "learning_rate": 5e-05, "loss": 0.9339, "mean_token_accuracy": 0.8957526683807373, "num_tokens": 606447428.0, "step": 15973 }, { "epoch": 2.0320569902048087, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.016798973083496, "learning_rate": 5e-05, "loss": 0.8285, "mean_token_accuracy": 0.8845299482345581, "num_tokens": 606481652.0, "step": 15974 }, { "epoch": 2.032184200483399, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.904540061950684, "learning_rate": 5e-05, "loss": 0.9407, "mean_token_accuracy": 0.9032655358314514, "num_tokens": 606519279.0, "step": 15975 }, { "epoch": 2.0323114107619897, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.285390853881836, "learning_rate": 5e-05, "loss": 0.8552, "mean_token_accuracy": 0.8868634700775146, "num_tokens": 606556399.0, "step": 15976 }, { "epoch": 2.0324386210405803, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.606624603271484, "learning_rate": 5e-05, "loss": 1.0137, "mean_token_accuracy": 0.8706634044647217, "num_tokens": 606596807.0, "step": 15977 }, { "epoch": 2.032565831319171, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.467658042907715, "learning_rate": 5e-05, "loss": 0.8362, "mean_token_accuracy": 0.8995327353477478, "num_tokens": 606629231.0, "step": 15978 }, { "epoch": 2.032693041597761, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.073079109191895, "learning_rate": 5e-05, "loss": 0.9178, "mean_token_accuracy": 0.8823333978652954, "num_tokens": 606669042.0, "step": 15979 }, { "epoch": 2.0328202518763514, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.249107360839844, "learning_rate": 5e-05, "loss": 0.8626, "mean_token_accuracy": 0.8776001334190369, "num_tokens": 606705550.0, "step": 15980 }, { "epoch": 2.032947462154942, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.80900764465332, "learning_rate": 5e-05, "loss": 0.8932, "mean_token_accuracy": 0.8817886114120483, "num_tokens": 606739866.0, "step": 15981 }, { "epoch": 2.0330746724335325, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.834747791290283, "learning_rate": 5e-05, "loss": 0.7911, "mean_token_accuracy": 0.8843764066696167, "num_tokens": 606782923.0, "step": 15982 }, { "epoch": 2.033201882712123, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.894546508789062, "learning_rate": 5e-05, "loss": 0.8793, "mean_token_accuracy": 0.8910646438598633, "num_tokens": 606822831.0, "step": 15983 }, { "epoch": 2.0333290929907135, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.532479286193848, "learning_rate": 5e-05, "loss": 0.7336, "mean_token_accuracy": 0.8903293609619141, "num_tokens": 606861129.0, "step": 15984 }, { "epoch": 2.033456303269304, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.097248077392578, "learning_rate": 5e-05, "loss": 0.9613, "mean_token_accuracy": 0.8717030882835388, "num_tokens": 606902744.0, "step": 15985 }, { "epoch": 2.0335835135478946, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.54467248916626, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8812869787216187, "num_tokens": 606935137.0, "step": 15986 }, { "epoch": 2.033710723826485, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.059850692749023, "learning_rate": 5e-05, "loss": 0.8947, "mean_token_accuracy": 0.8897745013237, "num_tokens": 606972215.0, "step": 15987 }, { "epoch": 2.0338379341050756, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.765908241271973, "learning_rate": 5e-05, "loss": 0.7816, "mean_token_accuracy": 0.8895515203475952, "num_tokens": 607004489.0, "step": 15988 }, { "epoch": 2.033965144383666, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.717586517333984, "learning_rate": 5e-05, "loss": 0.8973, "mean_token_accuracy": 0.88169264793396, "num_tokens": 607039457.0, "step": 15989 }, { "epoch": 2.0340923546622567, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.878335952758789, "learning_rate": 5e-05, "loss": 0.7851, "mean_token_accuracy": 0.8847870826721191, "num_tokens": 607084863.0, "step": 15990 }, { "epoch": 2.034219564940847, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.34675407409668, "learning_rate": 5e-05, "loss": 0.8708, "mean_token_accuracy": 0.8759177923202515, "num_tokens": 607125013.0, "step": 15991 }, { "epoch": 2.0343467752194377, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.805583953857422, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8897586464881897, "num_tokens": 607160501.0, "step": 15992 }, { "epoch": 2.0344739854980283, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.126006126403809, "learning_rate": 5e-05, "loss": 0.8161, "mean_token_accuracy": 0.8799687623977661, "num_tokens": 607204621.0, "step": 15993 }, { "epoch": 2.034601195776619, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.652339935302734, "learning_rate": 5e-05, "loss": 0.7073, "mean_token_accuracy": 0.8929768800735474, "num_tokens": 607245904.0, "step": 15994 }, { "epoch": 2.0347284060552093, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.104543685913086, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8887519836425781, "num_tokens": 607286907.0, "step": 15995 }, { "epoch": 2.0348556163338, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.503970146179199, "learning_rate": 5e-05, "loss": 0.7226, "mean_token_accuracy": 0.8780044317245483, "num_tokens": 607325990.0, "step": 15996 }, { "epoch": 2.0349828266123904, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.099799156188965, "learning_rate": 5e-05, "loss": 0.7846, "mean_token_accuracy": 0.8821237087249756, "num_tokens": 607363000.0, "step": 15997 }, { "epoch": 2.035110036890981, "ewc_loss": 0.375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.435468673706055, "learning_rate": 5e-05, "loss": 0.6936, "mean_token_accuracy": 0.8862243890762329, "num_tokens": 607405913.0, "step": 15998 }, { "epoch": 2.0352372471695714, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.068557739257812, "learning_rate": 5e-05, "loss": 0.7496, "mean_token_accuracy": 0.8864365816116333, "num_tokens": 607440242.0, "step": 15999 }, { "epoch": 2.035364457448162, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.4009928703308105, "learning_rate": 5e-05, "loss": 0.6843, "mean_token_accuracy": 0.8867722749710083, "num_tokens": 607485705.0, "step": 16000 }, { "epoch": 2.0354916677267525, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.053505897521973, "learning_rate": 5e-05, "loss": 0.7721, "mean_token_accuracy": 0.8775143623352051, "num_tokens": 607524026.0, "step": 16001 }, { "epoch": 2.035618878005343, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.355044364929199, "learning_rate": 5e-05, "loss": 0.6988, "mean_token_accuracy": 0.8826431035995483, "num_tokens": 607562925.0, "step": 16002 }, { "epoch": 2.0357460882839336, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.060812950134277, "learning_rate": 5e-05, "loss": 0.7647, "mean_token_accuracy": 0.8805959224700928, "num_tokens": 607598126.0, "step": 16003 }, { "epoch": 2.0358732985625236, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.336319446563721, "learning_rate": 5e-05, "loss": 0.655, "mean_token_accuracy": 0.899827241897583, "num_tokens": 607631911.0, "step": 16004 }, { "epoch": 2.036000508841114, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.983257293701172, "learning_rate": 5e-05, "loss": 0.759, "mean_token_accuracy": 0.8766213059425354, "num_tokens": 607669817.0, "step": 16005 }, { "epoch": 2.0361277191197047, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.431695938110352, "learning_rate": 5e-05, "loss": 0.683, "mean_token_accuracy": 0.8875412344932556, "num_tokens": 607710713.0, "step": 16006 }, { "epoch": 2.036254929398295, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.832454204559326, "learning_rate": 5e-05, "loss": 0.6758, "mean_token_accuracy": 0.9021781086921692, "num_tokens": 607751540.0, "step": 16007 }, { "epoch": 2.0363821396768857, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.544947147369385, "learning_rate": 5e-05, "loss": 0.6869, "mean_token_accuracy": 0.8862364888191223, "num_tokens": 607792831.0, "step": 16008 }, { "epoch": 2.0365093499554763, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.677194118499756, "learning_rate": 5e-05, "loss": 0.7067, "mean_token_accuracy": 0.8884089589118958, "num_tokens": 607824175.0, "step": 16009 }, { "epoch": 2.036636560234067, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.65067195892334, "learning_rate": 5e-05, "loss": 0.7245, "mean_token_accuracy": 0.8775671720504761, "num_tokens": 607864302.0, "step": 16010 }, { "epoch": 2.0367637705126573, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.520918846130371, "learning_rate": 5e-05, "loss": 0.7081, "mean_token_accuracy": 0.8826314210891724, "num_tokens": 607903231.0, "step": 16011 }, { "epoch": 2.036890980791248, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.782201290130615, "learning_rate": 5e-05, "loss": 0.686, "mean_token_accuracy": 0.8919671773910522, "num_tokens": 607942795.0, "step": 16012 }, { "epoch": 2.0370181910698384, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.337681293487549, "learning_rate": 5e-05, "loss": 0.6727, "mean_token_accuracy": 0.8849611282348633, "num_tokens": 607978034.0, "step": 16013 }, { "epoch": 2.037145401348429, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.980779647827148, "learning_rate": 5e-05, "loss": 0.6901, "mean_token_accuracy": 0.89701247215271, "num_tokens": 608014260.0, "step": 16014 }, { "epoch": 2.0372726116270194, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.108279228210449, "learning_rate": 5e-05, "loss": 0.6712, "mean_token_accuracy": 0.8788561820983887, "num_tokens": 608052637.0, "step": 16015 }, { "epoch": 2.03739982190561, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.249960899353027, "learning_rate": 5e-05, "loss": 0.7392, "mean_token_accuracy": 0.8860416412353516, "num_tokens": 608085877.0, "step": 16016 }, { "epoch": 2.0375270321842005, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.842367649078369, "learning_rate": 5e-05, "loss": 0.6614, "mean_token_accuracy": 0.8850318789482117, "num_tokens": 608125059.0, "step": 16017 }, { "epoch": 2.037654242462791, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.583111763000488, "learning_rate": 5e-05, "loss": 0.7497, "mean_token_accuracy": 0.8898054361343384, "num_tokens": 608161357.0, "step": 16018 }, { "epoch": 2.0377814527413816, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.5499749183654785, "learning_rate": 5e-05, "loss": 0.6383, "mean_token_accuracy": 0.8837005496025085, "num_tokens": 608196534.0, "step": 16019 }, { "epoch": 2.037908663019972, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.96471881866455, "learning_rate": 5e-05, "loss": 0.815, "mean_token_accuracy": 0.8841639161109924, "num_tokens": 608237130.0, "step": 16020 }, { "epoch": 2.0380358732985626, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.372094631195068, "learning_rate": 5e-05, "loss": 0.5995, "mean_token_accuracy": 0.9020709991455078, "num_tokens": 608279064.0, "step": 16021 }, { "epoch": 2.038163083577153, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 9.229959487915039, "learning_rate": 5e-05, "loss": 0.8277, "mean_token_accuracy": 0.8958045244216919, "num_tokens": 608316019.0, "step": 16022 }, { "epoch": 2.0382902938557437, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.588864326477051, "learning_rate": 5e-05, "loss": 0.6345, "mean_token_accuracy": 0.8970924019813538, "num_tokens": 608357199.0, "step": 16023 }, { "epoch": 2.038417504134334, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.914304733276367, "learning_rate": 5e-05, "loss": 0.8567, "mean_token_accuracy": 0.8869885802268982, "num_tokens": 608392149.0, "step": 16024 }, { "epoch": 2.0385447144129247, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000629425048828125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.39312219619751, "learning_rate": 5e-05, "loss": 0.7247, "mean_token_accuracy": 0.8949590921401978, "num_tokens": 608431200.0, "step": 16025 }, { "epoch": 2.0386719246915153, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.810359001159668, "learning_rate": 5e-05, "loss": 0.7512, "mean_token_accuracy": 0.8914752006530762, "num_tokens": 608471832.0, "step": 16026 }, { "epoch": 2.038799134970106, "ewc_loss": 0.5, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.403058052062988, "learning_rate": 5e-05, "loss": 0.8462, "mean_token_accuracy": 0.8756637573242188, "num_tokens": 608508362.0, "step": 16027 }, { "epoch": 2.0389263452486963, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.651803016662598, "learning_rate": 5e-05, "loss": 0.682, "mean_token_accuracy": 0.8801065683364868, "num_tokens": 608547988.0, "step": 16028 }, { "epoch": 2.0390535555272864, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.32500171661377, "learning_rate": 5e-05, "loss": 0.8498, "mean_token_accuracy": 0.8968355655670166, "num_tokens": 608580448.0, "step": 16029 }, { "epoch": 2.039180765805877, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.178692817687988, "learning_rate": 5e-05, "loss": 0.6864, "mean_token_accuracy": 0.8762389421463013, "num_tokens": 608618573.0, "step": 16030 }, { "epoch": 2.0393079760844675, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.817296981811523, "learning_rate": 5e-05, "loss": 0.9472, "mean_token_accuracy": 0.8816523551940918, "num_tokens": 608658161.0, "step": 16031 }, { "epoch": 2.039435186363058, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 6.767846584320068, "learning_rate": 5e-05, "loss": 0.7585, "mean_token_accuracy": 0.8785757422447205, "num_tokens": 608696599.0, "step": 16032 }, { "epoch": 2.0395623966416485, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.323736190795898, "learning_rate": 5e-05, "loss": 0.9103, "mean_token_accuracy": 0.8940486907958984, "num_tokens": 608738347.0, "step": 16033 }, { "epoch": 2.039689606920239, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.004376411437988, "learning_rate": 5e-05, "loss": 0.8778, "mean_token_accuracy": 0.8891799449920654, "num_tokens": 608779483.0, "step": 16034 }, { "epoch": 2.0398168171988296, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.85395622253418, "learning_rate": 5e-05, "loss": 0.819, "mean_token_accuracy": 0.8830422163009644, "num_tokens": 608821369.0, "step": 16035 }, { "epoch": 2.03994402747742, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.855268478393555, "learning_rate": 5e-05, "loss": 0.9361, "mean_token_accuracy": 0.8830195069313049, "num_tokens": 608854749.0, "step": 16036 }, { "epoch": 2.0400712377560106, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.012306213378906, "learning_rate": 5e-05, "loss": 0.6956, "mean_token_accuracy": 0.89606112241745, "num_tokens": 608893692.0, "step": 16037 }, { "epoch": 2.040198448034601, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.914517402648926, "learning_rate": 5e-05, "loss": 0.8812, "mean_token_accuracy": 0.8844491839408875, "num_tokens": 608930721.0, "step": 16038 }, { "epoch": 2.0403256583131917, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.399233341217041, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.8779178261756897, "num_tokens": 608967022.0, "step": 16039 }, { "epoch": 2.040452868591782, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.992040634155273, "learning_rate": 5e-05, "loss": 0.8297, "mean_token_accuracy": 0.8689173460006714, "num_tokens": 609009995.0, "step": 16040 }, { "epoch": 2.0405800788703727, "ewc_loss": 0.5, "ewc_loss_diag": 0.00063323974609375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.16553020477295, "learning_rate": 5e-05, "loss": 0.8285, "mean_token_accuracy": 0.8837421536445618, "num_tokens": 609044636.0, "step": 16041 }, { "epoch": 2.0407072891489633, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.077183723449707, "learning_rate": 5e-05, "loss": 0.6703, "mean_token_accuracy": 0.8917990326881409, "num_tokens": 609076328.0, "step": 16042 }, { "epoch": 2.040834499427554, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.757955551147461, "learning_rate": 5e-05, "loss": 0.9017, "mean_token_accuracy": 0.8669700622558594, "num_tokens": 609118955.0, "step": 16043 }, { "epoch": 2.0409617097061443, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.757966995239258, "learning_rate": 5e-05, "loss": 0.7116, "mean_token_accuracy": 0.873508870601654, "num_tokens": 609157958.0, "step": 16044 }, { "epoch": 2.041088919984735, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.006671905517578, "learning_rate": 5e-05, "loss": 0.8652, "mean_token_accuracy": 0.8891135454177856, "num_tokens": 609197636.0, "step": 16045 }, { "epoch": 2.0412161302633254, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.8826727867126465, "learning_rate": 5e-05, "loss": 0.77, "mean_token_accuracy": 0.8672916889190674, "num_tokens": 609239648.0, "step": 16046 }, { "epoch": 2.041343340541916, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.908506393432617, "learning_rate": 5e-05, "loss": 0.8534, "mean_token_accuracy": 0.896557092666626, "num_tokens": 609281012.0, "step": 16047 }, { "epoch": 2.0414705508205064, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.143713474273682, "learning_rate": 5e-05, "loss": 0.7605, "mean_token_accuracy": 0.8844575881958008, "num_tokens": 609323775.0, "step": 16048 }, { "epoch": 2.041597761099097, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.719752311706543, "learning_rate": 5e-05, "loss": 0.8422, "mean_token_accuracy": 0.8992623090744019, "num_tokens": 609358960.0, "step": 16049 }, { "epoch": 2.0417249713776875, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.186938762664795, "learning_rate": 5e-05, "loss": 0.785, "mean_token_accuracy": 0.8743813037872314, "num_tokens": 609395447.0, "step": 16050 }, { "epoch": 2.041852181656278, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.786142349243164, "learning_rate": 5e-05, "loss": 0.8596, "mean_token_accuracy": 0.8956968784332275, "num_tokens": 609429811.0, "step": 16051 }, { "epoch": 2.0419793919348685, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.923496723175049, "learning_rate": 5e-05, "loss": 0.7134, "mean_token_accuracy": 0.8898012638092041, "num_tokens": 609466104.0, "step": 16052 }, { "epoch": 2.0421066022134586, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.254505157470703, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.8939588665962219, "num_tokens": 609503513.0, "step": 16053 }, { "epoch": 2.042233812492049, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.756995677947998, "learning_rate": 5e-05, "loss": 0.6971, "mean_token_accuracy": 0.8959169387817383, "num_tokens": 609546899.0, "step": 16054 }, { "epoch": 2.0423610227706397, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.628376960754395, "learning_rate": 5e-05, "loss": 1.0195, "mean_token_accuracy": 0.8783841133117676, "num_tokens": 609589297.0, "step": 16055 }, { "epoch": 2.04248823304923, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.356931209564209, "learning_rate": 5e-05, "loss": 0.8174, "mean_token_accuracy": 0.8931000828742981, "num_tokens": 609624839.0, "step": 16056 }, { "epoch": 2.0426154433278207, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.106829643249512, "learning_rate": 5e-05, "loss": 0.9436, "mean_token_accuracy": 0.8908061385154724, "num_tokens": 609661708.0, "step": 16057 }, { "epoch": 2.0427426536064113, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.206270217895508, "learning_rate": 5e-05, "loss": 0.9584, "mean_token_accuracy": 0.8762398362159729, "num_tokens": 609704133.0, "step": 16058 }, { "epoch": 2.042869863885002, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.903536319732666, "learning_rate": 5e-05, "loss": 0.8261, "mean_token_accuracy": 0.887809693813324, "num_tokens": 609740743.0, "step": 16059 }, { "epoch": 2.0429970741635923, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.566425323486328, "learning_rate": 5e-05, "loss": 0.9253, "mean_token_accuracy": 0.8776417374610901, "num_tokens": 609777520.0, "step": 16060 }, { "epoch": 2.043124284442183, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.276331901550293, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8865776658058167, "num_tokens": 609812974.0, "step": 16061 }, { "epoch": 2.0432514947207734, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.367115020751953, "learning_rate": 5e-05, "loss": 0.8345, "mean_token_accuracy": 0.8919483423233032, "num_tokens": 609853287.0, "step": 16062 }, { "epoch": 2.043378704999364, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.441885471343994, "learning_rate": 5e-05, "loss": 0.7205, "mean_token_accuracy": 0.8939424753189087, "num_tokens": 609888768.0, "step": 16063 }, { "epoch": 2.0435059152779544, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.927988052368164, "learning_rate": 5e-05, "loss": 0.7754, "mean_token_accuracy": 0.8940032720565796, "num_tokens": 609927976.0, "step": 16064 }, { "epoch": 2.043633125556545, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.655190467834473, "learning_rate": 5e-05, "loss": 0.723, "mean_token_accuracy": 0.893634021282196, "num_tokens": 609964404.0, "step": 16065 }, { "epoch": 2.0437603358351355, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.736089706420898, "learning_rate": 5e-05, "loss": 0.752, "mean_token_accuracy": 0.8869118690490723, "num_tokens": 610005188.0, "step": 16066 }, { "epoch": 2.043887546113726, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.554373741149902, "learning_rate": 5e-05, "loss": 0.7161, "mean_token_accuracy": 0.8942285776138306, "num_tokens": 610039118.0, "step": 16067 }, { "epoch": 2.0440147563923166, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.8790411949157715, "learning_rate": 5e-05, "loss": 0.7285, "mean_token_accuracy": 0.8949903249740601, "num_tokens": 610074867.0, "step": 16068 }, { "epoch": 2.044141966670907, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.330169677734375, "learning_rate": 5e-05, "loss": 0.6911, "mean_token_accuracy": 0.8902106285095215, "num_tokens": 610111382.0, "step": 16069 }, { "epoch": 2.0442691769494976, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.058266639709473, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8840731382369995, "num_tokens": 610148026.0, "step": 16070 }, { "epoch": 2.044396387228088, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.141348838806152, "learning_rate": 5e-05, "loss": 0.659, "mean_token_accuracy": 0.8942565321922302, "num_tokens": 610179841.0, "step": 16071 }, { "epoch": 2.0445235975066787, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.248014450073242, "learning_rate": 5e-05, "loss": 0.7767, "mean_token_accuracy": 0.8825611472129822, "num_tokens": 610214224.0, "step": 16072 }, { "epoch": 2.044650807785269, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.917089462280273, "learning_rate": 5e-05, "loss": 0.668, "mean_token_accuracy": 0.8891983032226562, "num_tokens": 610257640.0, "step": 16073 }, { "epoch": 2.0447780180638597, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.548312187194824, "learning_rate": 5e-05, "loss": 0.7716, "mean_token_accuracy": 0.8977550268173218, "num_tokens": 610295438.0, "step": 16074 }, { "epoch": 2.0449052283424503, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.666383743286133, "learning_rate": 5e-05, "loss": 0.6447, "mean_token_accuracy": 0.8932617902755737, "num_tokens": 610336390.0, "step": 16075 }, { "epoch": 2.045032438621041, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000640869140625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.910103797912598, "learning_rate": 5e-05, "loss": 0.8228, "mean_token_accuracy": 0.8911111354827881, "num_tokens": 610369860.0, "step": 16076 }, { "epoch": 2.045159648899631, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.54484224319458, "learning_rate": 5e-05, "loss": 0.6623, "mean_token_accuracy": 0.8885698318481445, "num_tokens": 610408116.0, "step": 16077 }, { "epoch": 2.0452868591782214, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.153342247009277, "learning_rate": 5e-05, "loss": 0.9348, "mean_token_accuracy": 0.8762772679328918, "num_tokens": 610444317.0, "step": 16078 }, { "epoch": 2.045414069456812, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.766911029815674, "learning_rate": 5e-05, "loss": 0.7354, "mean_token_accuracy": 0.8773341774940491, "num_tokens": 610485675.0, "step": 16079 }, { "epoch": 2.0455412797354025, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.910836219787598, "learning_rate": 5e-05, "loss": 0.8787, "mean_token_accuracy": 0.8917152285575867, "num_tokens": 610524697.0, "step": 16080 }, { "epoch": 2.045668490013993, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.404938697814941, "learning_rate": 5e-05, "loss": 0.8109, "mean_token_accuracy": 0.8885316252708435, "num_tokens": 610567470.0, "step": 16081 }, { "epoch": 2.0457957002925835, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.109201431274414, "learning_rate": 5e-05, "loss": 0.817, "mean_token_accuracy": 0.8888047933578491, "num_tokens": 610603168.0, "step": 16082 }, { "epoch": 2.045922910571174, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.074530601501465, "learning_rate": 5e-05, "loss": 0.8329, "mean_token_accuracy": 0.8873463869094849, "num_tokens": 610643416.0, "step": 16083 }, { "epoch": 2.0460501208497646, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.238272190093994, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8858985304832458, "num_tokens": 610676513.0, "step": 16084 }, { "epoch": 2.046177331128355, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.5182523727417, "learning_rate": 5e-05, "loss": 0.8237, "mean_token_accuracy": 0.8905876278877258, "num_tokens": 610710842.0, "step": 16085 }, { "epoch": 2.0463045414069456, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.881337642669678, "learning_rate": 5e-05, "loss": 0.7028, "mean_token_accuracy": 0.8771259188652039, "num_tokens": 610747770.0, "step": 16086 }, { "epoch": 2.046431751685536, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.619612693786621, "learning_rate": 5e-05, "loss": 0.8145, "mean_token_accuracy": 0.8915668725967407, "num_tokens": 610788780.0, "step": 16087 }, { "epoch": 2.0465589619641267, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.889669418334961, "learning_rate": 5e-05, "loss": 0.6735, "mean_token_accuracy": 0.8909983038902283, "num_tokens": 610832930.0, "step": 16088 }, { "epoch": 2.046686172242717, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.568065643310547, "learning_rate": 5e-05, "loss": 0.8175, "mean_token_accuracy": 0.8902268409729004, "num_tokens": 610875828.0, "step": 16089 }, { "epoch": 2.0468133825213077, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.0355939865112305, "learning_rate": 5e-05, "loss": 0.6826, "mean_token_accuracy": 0.896891713142395, "num_tokens": 610913920.0, "step": 16090 }, { "epoch": 2.0469405927998983, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.430622100830078, "learning_rate": 5e-05, "loss": 0.8244, "mean_token_accuracy": 0.8857810497283936, "num_tokens": 610956887.0, "step": 16091 }, { "epoch": 2.047067803078489, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.203074932098389, "learning_rate": 5e-05, "loss": 0.764, "mean_token_accuracy": 0.8768205642700195, "num_tokens": 611000889.0, "step": 16092 }, { "epoch": 2.0471950133570793, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.215251922607422, "learning_rate": 5e-05, "loss": 0.8085, "mean_token_accuracy": 0.886319637298584, "num_tokens": 611040573.0, "step": 16093 }, { "epoch": 2.04732222363567, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.372857570648193, "learning_rate": 5e-05, "loss": 0.7164, "mean_token_accuracy": 0.8956610560417175, "num_tokens": 611075925.0, "step": 16094 }, { "epoch": 2.0474494339142604, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.002013206481934, "learning_rate": 5e-05, "loss": 0.7989, "mean_token_accuracy": 0.8785316944122314, "num_tokens": 611116492.0, "step": 16095 }, { "epoch": 2.047576644192851, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.421220779418945, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8908330202102661, "num_tokens": 611155683.0, "step": 16096 }, { "epoch": 2.0477038544714414, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.943001747131348, "learning_rate": 5e-05, "loss": 0.6916, "mean_token_accuracy": 0.9114838242530823, "num_tokens": 611194352.0, "step": 16097 }, { "epoch": 2.047831064750032, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.337978363037109, "learning_rate": 5e-05, "loss": 0.691, "mean_token_accuracy": 0.8922533392906189, "num_tokens": 611229919.0, "step": 16098 }, { "epoch": 2.0479582750286225, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.043773651123047, "learning_rate": 5e-05, "loss": 0.7775, "mean_token_accuracy": 0.8849538564682007, "num_tokens": 611267394.0, "step": 16099 }, { "epoch": 2.048085485307213, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.130435466766357, "learning_rate": 5e-05, "loss": 0.7655, "mean_token_accuracy": 0.8621957302093506, "num_tokens": 611312009.0, "step": 16100 }, { "epoch": 2.0482126955858035, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.2669677734375, "learning_rate": 5e-05, "loss": 0.8314, "mean_token_accuracy": 0.8715769648551941, "num_tokens": 611346723.0, "step": 16101 }, { "epoch": 2.0483399058643936, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.892531394958496, "learning_rate": 5e-05, "loss": 0.6749, "mean_token_accuracy": 0.8887714147567749, "num_tokens": 611381803.0, "step": 16102 }, { "epoch": 2.048467116142984, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.573054313659668, "learning_rate": 5e-05, "loss": 0.8086, "mean_token_accuracy": 0.8871194124221802, "num_tokens": 611419579.0, "step": 16103 }, { "epoch": 2.0485943264215747, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.640816688537598, "learning_rate": 5e-05, "loss": 0.6591, "mean_token_accuracy": 0.8892459869384766, "num_tokens": 611456904.0, "step": 16104 }, { "epoch": 2.048721536700165, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.915535926818848, "learning_rate": 5e-05, "loss": 0.8102, "mean_token_accuracy": 0.8976923227310181, "num_tokens": 611495597.0, "step": 16105 }, { "epoch": 2.0488487469787557, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.594229698181152, "learning_rate": 5e-05, "loss": 0.6562, "mean_token_accuracy": 0.8924326300621033, "num_tokens": 611531307.0, "step": 16106 }, { "epoch": 2.0489759572573463, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.007829666137695, "learning_rate": 5e-05, "loss": 0.9291, "mean_token_accuracy": 0.873199999332428, "num_tokens": 611567657.0, "step": 16107 }, { "epoch": 2.049103167535937, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.97257137298584, "learning_rate": 5e-05, "loss": 0.7595, "mean_token_accuracy": 0.8837106227874756, "num_tokens": 611608107.0, "step": 16108 }, { "epoch": 2.0492303778145273, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.491650581359863, "learning_rate": 5e-05, "loss": 0.8761, "mean_token_accuracy": 0.8838386535644531, "num_tokens": 611644695.0, "step": 16109 }, { "epoch": 2.049357588093118, "ewc_loss": 0.5, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.67588472366333, "learning_rate": 5e-05, "loss": 0.7989, "mean_token_accuracy": 0.8913052082061768, "num_tokens": 611679330.0, "step": 16110 }, { "epoch": 2.0494847983717084, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.547003269195557, "learning_rate": 5e-05, "loss": 0.7888, "mean_token_accuracy": 0.8742198348045349, "num_tokens": 611718792.0, "step": 16111 }, { "epoch": 2.049612008650299, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.356193542480469, "learning_rate": 5e-05, "loss": 0.8567, "mean_token_accuracy": 0.878883421421051, "num_tokens": 611752559.0, "step": 16112 }, { "epoch": 2.0497392189288894, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.770948886871338, "learning_rate": 5e-05, "loss": 0.6597, "mean_token_accuracy": 0.8919953107833862, "num_tokens": 611791616.0, "step": 16113 }, { "epoch": 2.04986642920748, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.877681732177734, "learning_rate": 5e-05, "loss": 0.8944, "mean_token_accuracy": 0.8746888041496277, "num_tokens": 611826577.0, "step": 16114 }, { "epoch": 2.0499936394860705, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.43858003616333, "learning_rate": 5e-05, "loss": 0.6737, "mean_token_accuracy": 0.8860203623771667, "num_tokens": 611861777.0, "step": 16115 }, { "epoch": 2.050120849764661, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.224876403808594, "learning_rate": 5e-05, "loss": 0.9372, "mean_token_accuracy": 0.878844141960144, "num_tokens": 611897933.0, "step": 16116 }, { "epoch": 2.0502480600432516, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.586511611938477, "learning_rate": 5e-05, "loss": 0.7052, "mean_token_accuracy": 0.8863826990127563, "num_tokens": 611935285.0, "step": 16117 }, { "epoch": 2.050375270321842, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000644683837890625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.146309852600098, "learning_rate": 5e-05, "loss": 0.8955, "mean_token_accuracy": 0.8932466506958008, "num_tokens": 611969879.0, "step": 16118 }, { "epoch": 2.0505024806004326, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.185691833496094, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8888568878173828, "num_tokens": 612003608.0, "step": 16119 }, { "epoch": 2.050629690879023, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.516077995300293, "learning_rate": 5e-05, "loss": 0.9078, "mean_token_accuracy": 0.883510947227478, "num_tokens": 612044247.0, "step": 16120 }, { "epoch": 2.0507569011576137, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.859822750091553, "learning_rate": 5e-05, "loss": 0.8939, "mean_token_accuracy": 0.8778142929077148, "num_tokens": 612080776.0, "step": 16121 }, { "epoch": 2.050884111436204, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.6723737716674805, "learning_rate": 5e-05, "loss": 0.8256, "mean_token_accuracy": 0.8756657242774963, "num_tokens": 612119162.0, "step": 16122 }, { "epoch": 2.0510113217147947, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.168254852294922, "learning_rate": 5e-05, "loss": 0.8211, "mean_token_accuracy": 0.8940544724464417, "num_tokens": 612157774.0, "step": 16123 }, { "epoch": 2.0511385319933853, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.417553424835205, "learning_rate": 5e-05, "loss": 0.7448, "mean_token_accuracy": 0.8891086578369141, "num_tokens": 612192169.0, "step": 16124 }, { "epoch": 2.051265742271976, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.901409149169922, "learning_rate": 5e-05, "loss": 0.771, "mean_token_accuracy": 0.8894559144973755, "num_tokens": 612225936.0, "step": 16125 }, { "epoch": 2.0513929525505663, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.709100246429443, "learning_rate": 5e-05, "loss": 0.754, "mean_token_accuracy": 0.8913949728012085, "num_tokens": 612262509.0, "step": 16126 }, { "epoch": 2.0515201628291564, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.423303127288818, "learning_rate": 5e-05, "loss": 0.7241, "mean_token_accuracy": 0.8837951421737671, "num_tokens": 612297952.0, "step": 16127 }, { "epoch": 2.051647373107747, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.037152290344238, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8875445127487183, "num_tokens": 612332512.0, "step": 16128 }, { "epoch": 2.0517745833863374, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.061927795410156, "learning_rate": 5e-05, "loss": 0.6667, "mean_token_accuracy": 0.8932807445526123, "num_tokens": 612370350.0, "step": 16129 }, { "epoch": 2.051901793664928, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.326180458068848, "learning_rate": 5e-05, "loss": 0.8296, "mean_token_accuracy": 0.8750549554824829, "num_tokens": 612404775.0, "step": 16130 }, { "epoch": 2.0520290039435185, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.7975897789001465, "learning_rate": 5e-05, "loss": 0.6814, "mean_token_accuracy": 0.8874747157096863, "num_tokens": 612447226.0, "step": 16131 }, { "epoch": 2.052156214222109, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.71471881866455, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8768746852874756, "num_tokens": 612481358.0, "step": 16132 }, { "epoch": 2.0522834245006996, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.500272750854492, "learning_rate": 5e-05, "loss": 0.6416, "mean_token_accuracy": 0.8956872224807739, "num_tokens": 612519132.0, "step": 16133 }, { "epoch": 2.05241063477929, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.194693565368652, "learning_rate": 5e-05, "loss": 0.9249, "mean_token_accuracy": 0.8767571449279785, "num_tokens": 612559187.0, "step": 16134 }, { "epoch": 2.0525378450578806, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.413002967834473, "learning_rate": 5e-05, "loss": 0.6698, "mean_token_accuracy": 0.8906271457672119, "num_tokens": 612598208.0, "step": 16135 }, { "epoch": 2.052665055336471, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.450547218322754, "learning_rate": 5e-05, "loss": 0.9387, "mean_token_accuracy": 0.8877815008163452, "num_tokens": 612633951.0, "step": 16136 }, { "epoch": 2.0527922656150617, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 6.979898929595947, "learning_rate": 5e-05, "loss": 0.8099, "mean_token_accuracy": 0.8801642656326294, "num_tokens": 612673978.0, "step": 16137 }, { "epoch": 2.052919475893652, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.941606521606445, "learning_rate": 5e-05, "loss": 0.9147, "mean_token_accuracy": 0.8869651556015015, "num_tokens": 612720488.0, "step": 16138 }, { "epoch": 2.0530466861722427, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.8707194328308105, "learning_rate": 5e-05, "loss": 0.914, "mean_token_accuracy": 0.8826376795768738, "num_tokens": 612755009.0, "step": 16139 }, { "epoch": 2.0531738964508333, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.786236763000488, "learning_rate": 5e-05, "loss": 0.8503, "mean_token_accuracy": 0.8786695003509521, "num_tokens": 612790653.0, "step": 16140 }, { "epoch": 2.053301106729424, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.491541862487793, "learning_rate": 5e-05, "loss": 0.8785, "mean_token_accuracy": 0.8893738985061646, "num_tokens": 612829406.0, "step": 16141 }, { "epoch": 2.0534283170080143, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.029572486877441, "learning_rate": 5e-05, "loss": 0.7567, "mean_token_accuracy": 0.8759604096412659, "num_tokens": 612872304.0, "step": 16142 }, { "epoch": 2.053555527286605, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.651540756225586, "learning_rate": 5e-05, "loss": 0.8706, "mean_token_accuracy": 0.8862721920013428, "num_tokens": 612907909.0, "step": 16143 }, { "epoch": 2.0536827375651954, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.99606990814209, "learning_rate": 5e-05, "loss": 0.6946, "mean_token_accuracy": 0.895606517791748, "num_tokens": 612943794.0, "step": 16144 }, { "epoch": 2.053809947843786, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.455121994018555, "learning_rate": 5e-05, "loss": 0.8366, "mean_token_accuracy": 0.888127326965332, "num_tokens": 612987125.0, "step": 16145 }, { "epoch": 2.0539371581223764, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.195110321044922, "learning_rate": 5e-05, "loss": 0.7781, "mean_token_accuracy": 0.8771786689758301, "num_tokens": 613028232.0, "step": 16146 }, { "epoch": 2.054064368400967, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.273088455200195, "learning_rate": 5e-05, "loss": 0.8258, "mean_token_accuracy": 0.8866410255432129, "num_tokens": 613066651.0, "step": 16147 }, { "epoch": 2.0541915786795575, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.168514728546143, "learning_rate": 5e-05, "loss": 0.7267, "mean_token_accuracy": 0.894175112247467, "num_tokens": 613105153.0, "step": 16148 }, { "epoch": 2.054318788958148, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.407367706298828, "learning_rate": 5e-05, "loss": 0.8038, "mean_token_accuracy": 0.8938603401184082, "num_tokens": 613146817.0, "step": 16149 }, { "epoch": 2.0544459992367385, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.974318981170654, "learning_rate": 5e-05, "loss": 0.7414, "mean_token_accuracy": 0.879239559173584, "num_tokens": 613192644.0, "step": 16150 }, { "epoch": 2.0545732095153286, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.643577575683594, "learning_rate": 5e-05, "loss": 0.8956, "mean_token_accuracy": 0.8811061382293701, "num_tokens": 613232283.0, "step": 16151 }, { "epoch": 2.054700419793919, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.8638129234313965, "learning_rate": 5e-05, "loss": 0.7021, "mean_token_accuracy": 0.8905840516090393, "num_tokens": 613270916.0, "step": 16152 }, { "epoch": 2.0548276300725097, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.741926193237305, "learning_rate": 5e-05, "loss": 0.8539, "mean_token_accuracy": 0.8944450616836548, "num_tokens": 613303095.0, "step": 16153 }, { "epoch": 2.0549548403511, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 6.985205173492432, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8802176713943481, "num_tokens": 613341551.0, "step": 16154 }, { "epoch": 2.0550820506296907, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.611038208007812, "learning_rate": 5e-05, "loss": 0.8534, "mean_token_accuracy": 0.8929393291473389, "num_tokens": 613377511.0, "step": 16155 }, { "epoch": 2.0552092609082813, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.259232044219971, "learning_rate": 5e-05, "loss": 0.7655, "mean_token_accuracy": 0.89339679479599, "num_tokens": 613408534.0, "step": 16156 }, { "epoch": 2.055336471186872, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.249565124511719, "learning_rate": 5e-05, "loss": 0.8137, "mean_token_accuracy": 0.8969735503196716, "num_tokens": 613443461.0, "step": 16157 }, { "epoch": 2.0554636814654623, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.514341354370117, "learning_rate": 5e-05, "loss": 0.8246, "mean_token_accuracy": 0.8774003982543945, "num_tokens": 613482438.0, "step": 16158 }, { "epoch": 2.055590891744053, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.917639255523682, "learning_rate": 5e-05, "loss": 0.8335, "mean_token_accuracy": 0.8767690658569336, "num_tokens": 613521937.0, "step": 16159 }, { "epoch": 2.0557181020226434, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.596464157104492, "learning_rate": 5e-05, "loss": 0.8145, "mean_token_accuracy": 0.8732847571372986, "num_tokens": 613563471.0, "step": 16160 }, { "epoch": 2.055845312301234, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.818466663360596, "learning_rate": 5e-05, "loss": 0.7791, "mean_token_accuracy": 0.8850432634353638, "num_tokens": 613603753.0, "step": 16161 }, { "epoch": 2.0559725225798244, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.468926906585693, "learning_rate": 5e-05, "loss": 0.7249, "mean_token_accuracy": 0.8950486183166504, "num_tokens": 613637602.0, "step": 16162 }, { "epoch": 2.056099732858415, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.935426235198975, "learning_rate": 5e-05, "loss": 0.7748, "mean_token_accuracy": 0.8838832378387451, "num_tokens": 613673807.0, "step": 16163 }, { "epoch": 2.0562269431370055, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.244232654571533, "learning_rate": 5e-05, "loss": 0.7177, "mean_token_accuracy": 0.885408341884613, "num_tokens": 613710702.0, "step": 16164 }, { "epoch": 2.056354153415596, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.121673583984375, "learning_rate": 5e-05, "loss": 0.7767, "mean_token_accuracy": 0.8836580514907837, "num_tokens": 613748890.0, "step": 16165 }, { "epoch": 2.0564813636941865, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.039549827575684, "learning_rate": 5e-05, "loss": 0.6794, "mean_token_accuracy": 0.8869782090187073, "num_tokens": 613791039.0, "step": 16166 }, { "epoch": 2.056608573972777, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.313124656677246, "learning_rate": 5e-05, "loss": 0.7775, "mean_token_accuracy": 0.8928394317626953, "num_tokens": 613821345.0, "step": 16167 }, { "epoch": 2.0567357842513676, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.9005446434021, "learning_rate": 5e-05, "loss": 0.7458, "mean_token_accuracy": 0.8635552525520325, "num_tokens": 613857770.0, "step": 16168 }, { "epoch": 2.056862994529958, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.440428733825684, "learning_rate": 5e-05, "loss": 0.7877, "mean_token_accuracy": 0.8900251388549805, "num_tokens": 613898844.0, "step": 16169 }, { "epoch": 2.0569902048085487, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.903426647186279, "learning_rate": 5e-05, "loss": 0.705, "mean_token_accuracy": 0.8772417902946472, "num_tokens": 613940884.0, "step": 16170 }, { "epoch": 2.057117415087139, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.420796394348145, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8869341015815735, "num_tokens": 613981664.0, "step": 16171 }, { "epoch": 2.0572446253657297, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.065999507904053, "learning_rate": 5e-05, "loss": 0.7283, "mean_token_accuracy": 0.8793649673461914, "num_tokens": 614013965.0, "step": 16172 }, { "epoch": 2.0573718356443202, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.15587043762207, "learning_rate": 5e-05, "loss": 0.7906, "mean_token_accuracy": 0.8852335214614868, "num_tokens": 614048841.0, "step": 16173 }, { "epoch": 2.0574990459229108, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.367153167724609, "learning_rate": 5e-05, "loss": 0.7312, "mean_token_accuracy": 0.8851139545440674, "num_tokens": 614087677.0, "step": 16174 }, { "epoch": 2.057626256201501, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.73701810836792, "learning_rate": 5e-05, "loss": 0.7507, "mean_token_accuracy": 0.8837015628814697, "num_tokens": 614128546.0, "step": 16175 }, { "epoch": 2.0577534664800914, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.734982490539551, "learning_rate": 5e-05, "loss": 0.7378, "mean_token_accuracy": 0.8928430080413818, "num_tokens": 614163947.0, "step": 16176 }, { "epoch": 2.057880676758682, "ewc_loss": 0.375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.2805585861206055, "learning_rate": 5e-05, "loss": 0.7061, "mean_token_accuracy": 0.8829954862594604, "num_tokens": 614202353.0, "step": 16177 }, { "epoch": 2.0580078870372724, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.097810745239258, "learning_rate": 5e-05, "loss": 0.7805, "mean_token_accuracy": 0.8807438611984253, "num_tokens": 614244781.0, "step": 16178 }, { "epoch": 2.058135097315863, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.877884387969971, "learning_rate": 5e-05, "loss": 0.6581, "mean_token_accuracy": 0.8895618319511414, "num_tokens": 614278335.0, "step": 16179 }, { "epoch": 2.0582623075944535, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.488326072692871, "learning_rate": 5e-05, "loss": 0.8008, "mean_token_accuracy": 0.8841364979743958, "num_tokens": 614317726.0, "step": 16180 }, { "epoch": 2.058389517873044, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.506335258483887, "learning_rate": 5e-05, "loss": 0.6352, "mean_token_accuracy": 0.8909433484077454, "num_tokens": 614355403.0, "step": 16181 }, { "epoch": 2.0585167281516346, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.96933650970459, "learning_rate": 5e-05, "loss": 0.8855, "mean_token_accuracy": 0.873419463634491, "num_tokens": 614404370.0, "step": 16182 }, { "epoch": 2.058643938430225, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.244426727294922, "learning_rate": 5e-05, "loss": 0.6221, "mean_token_accuracy": 0.8938168287277222, "num_tokens": 614439613.0, "step": 16183 }, { "epoch": 2.0587711487088156, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.40588092803955, "learning_rate": 5e-05, "loss": 0.9446, "mean_token_accuracy": 0.8779764175415039, "num_tokens": 614481634.0, "step": 16184 }, { "epoch": 2.058898358987406, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.466245174407959, "learning_rate": 5e-05, "loss": 0.6854, "mean_token_accuracy": 0.8892785310745239, "num_tokens": 614517178.0, "step": 16185 }, { "epoch": 2.0590255692659967, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.283373832702637, "learning_rate": 5e-05, "loss": 0.9699, "mean_token_accuracy": 0.8725350499153137, "num_tokens": 614551531.0, "step": 16186 }, { "epoch": 2.059152779544587, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.391453266143799, "learning_rate": 5e-05, "loss": 0.8588, "mean_token_accuracy": 0.8804068565368652, "num_tokens": 614589173.0, "step": 16187 }, { "epoch": 2.0592799898231777, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.164397239685059, "learning_rate": 5e-05, "loss": 0.8513, "mean_token_accuracy": 0.8846186399459839, "num_tokens": 614633657.0, "step": 16188 }, { "epoch": 2.0594072001017683, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.415011405944824, "learning_rate": 5e-05, "loss": 0.9175, "mean_token_accuracy": 0.8809819221496582, "num_tokens": 614670013.0, "step": 16189 }, { "epoch": 2.059534410380359, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.83907413482666, "learning_rate": 5e-05, "loss": 0.7158, "mean_token_accuracy": 0.8799729347229004, "num_tokens": 614706414.0, "step": 16190 }, { "epoch": 2.0596616206589493, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000652313232421875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.191903114318848, "learning_rate": 5e-05, "loss": 0.932, "mean_token_accuracy": 0.8837160468101501, "num_tokens": 614746397.0, "step": 16191 }, { "epoch": 2.05978883093754, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.467994689941406, "learning_rate": 5e-05, "loss": 0.6825, "mean_token_accuracy": 0.89227294921875, "num_tokens": 614785587.0, "step": 16192 }, { "epoch": 2.0599160412161304, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.294856071472168, "learning_rate": 5e-05, "loss": 0.9326, "mean_token_accuracy": 0.8894073963165283, "num_tokens": 614823304.0, "step": 16193 }, { "epoch": 2.060043251494721, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.11306095123291, "learning_rate": 5e-05, "loss": 0.8085, "mean_token_accuracy": 0.8878511190414429, "num_tokens": 614857544.0, "step": 16194 }, { "epoch": 2.0601704617733114, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.44600772857666, "learning_rate": 5e-05, "loss": 0.9062, "mean_token_accuracy": 0.8818784952163696, "num_tokens": 614893095.0, "step": 16195 }, { "epoch": 2.060297672051902, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 7.935400009155273, "learning_rate": 5e-05, "loss": 0.8936, "mean_token_accuracy": 0.8787766098976135, "num_tokens": 614932634.0, "step": 16196 }, { "epoch": 2.0604248823304925, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.4749531745910645, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.88126140832901, "num_tokens": 614975039.0, "step": 16197 }, { "epoch": 2.060552092609083, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.32886791229248, "learning_rate": 5e-05, "loss": 0.8491, "mean_token_accuracy": 0.8879350423812866, "num_tokens": 615019092.0, "step": 16198 }, { "epoch": 2.0606793028876735, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.201015949249268, "learning_rate": 5e-05, "loss": 0.7068, "mean_token_accuracy": 0.8945072889328003, "num_tokens": 615060183.0, "step": 16199 }, { "epoch": 2.0608065131662636, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.147013664245605, "learning_rate": 5e-05, "loss": 0.8246, "mean_token_accuracy": 0.8824503421783447, "num_tokens": 615098376.0, "step": 16200 }, { "epoch": 2.060933723444854, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.427231788635254, "learning_rate": 5e-05, "loss": 0.7503, "mean_token_accuracy": 0.8889981508255005, "num_tokens": 615134401.0, "step": 16201 }, { "epoch": 2.0610609337234447, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.6966023445129395, "learning_rate": 5e-05, "loss": 0.7522, "mean_token_accuracy": 0.8869351148605347, "num_tokens": 615176145.0, "step": 16202 }, { "epoch": 2.061188144002035, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.749427795410156, "learning_rate": 5e-05, "loss": 0.7724, "mean_token_accuracy": 0.8806159496307373, "num_tokens": 615210281.0, "step": 16203 }, { "epoch": 2.0613153542806257, "ewc_loss": 0.390625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.3657050132751465, "learning_rate": 5e-05, "loss": 0.6649, "mean_token_accuracy": 0.9030634164810181, "num_tokens": 615246952.0, "step": 16204 }, { "epoch": 2.0614425645592163, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.895424842834473, "learning_rate": 5e-05, "loss": 0.7224, "mean_token_accuracy": 0.897746205329895, "num_tokens": 615285807.0, "step": 16205 }, { "epoch": 2.061569774837807, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.228307723999023, "learning_rate": 5e-05, "loss": 0.647, "mean_token_accuracy": 0.8965674638748169, "num_tokens": 615319641.0, "step": 16206 }, { "epoch": 2.0616969851163973, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.990869998931885, "learning_rate": 5e-05, "loss": 0.7993, "mean_token_accuracy": 0.8720096349716187, "num_tokens": 615358072.0, "step": 16207 }, { "epoch": 2.061824195394988, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.07668399810791, "learning_rate": 5e-05, "loss": 0.675, "mean_token_accuracy": 0.8861858248710632, "num_tokens": 615393924.0, "step": 16208 }, { "epoch": 2.0619514056735784, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.180459976196289, "learning_rate": 5e-05, "loss": 0.767, "mean_token_accuracy": 0.8848316669464111, "num_tokens": 615436771.0, "step": 16209 }, { "epoch": 2.062078615952169, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.829345226287842, "learning_rate": 5e-05, "loss": 0.6828, "mean_token_accuracy": 0.8797045946121216, "num_tokens": 615473577.0, "step": 16210 }, { "epoch": 2.0622058262307594, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.5424222946167, "learning_rate": 5e-05, "loss": 0.8344, "mean_token_accuracy": 0.8709554076194763, "num_tokens": 615507585.0, "step": 16211 }, { "epoch": 2.06233303650935, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.470545768737793, "learning_rate": 5e-05, "loss": 0.6348, "mean_token_accuracy": 0.8923288583755493, "num_tokens": 615542638.0, "step": 16212 }, { "epoch": 2.0624602467879405, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 9.083160400390625, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.8824523687362671, "num_tokens": 615579567.0, "step": 16213 }, { "epoch": 2.062587457066531, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.134894847869873, "learning_rate": 5e-05, "loss": 0.6351, "mean_token_accuracy": 0.8932962417602539, "num_tokens": 615616563.0, "step": 16214 }, { "epoch": 2.0627146673451215, "ewc_loss": 0.625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.722382545471191, "learning_rate": 5e-05, "loss": 0.9762, "mean_token_accuracy": 0.8756093382835388, "num_tokens": 615653243.0, "step": 16215 }, { "epoch": 2.062841877623712, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.343380928039551, "learning_rate": 5e-05, "loss": 0.6864, "mean_token_accuracy": 0.8920618295669556, "num_tokens": 615689552.0, "step": 16216 }, { "epoch": 2.0629690879023026, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.935523986816406, "learning_rate": 5e-05, "loss": 1.0515, "mean_token_accuracy": 0.8756667375564575, "num_tokens": 615723803.0, "step": 16217 }, { "epoch": 2.063096298180893, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 7.506565570831299, "learning_rate": 5e-05, "loss": 0.9223, "mean_token_accuracy": 0.8873474597930908, "num_tokens": 615761181.0, "step": 16218 }, { "epoch": 2.0632235084594837, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.950626373291016, "learning_rate": 5e-05, "loss": 0.9276, "mean_token_accuracy": 0.8921887874603271, "num_tokens": 615802200.0, "step": 16219 }, { "epoch": 2.063350718738074, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.571601867675781, "learning_rate": 5e-05, "loss": 0.9629, "mean_token_accuracy": 0.8897925019264221, "num_tokens": 615843299.0, "step": 16220 }, { "epoch": 2.0634779290166647, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.311110019683838, "learning_rate": 5e-05, "loss": 0.7891, "mean_token_accuracy": 0.8834363222122192, "num_tokens": 615886148.0, "step": 16221 }, { "epoch": 2.0636051392952552, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.310059547424316, "learning_rate": 5e-05, "loss": 0.9861, "mean_token_accuracy": 0.8775303363800049, "num_tokens": 615930595.0, "step": 16222 }, { "epoch": 2.0637323495738458, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0006561279296875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.720480918884277, "learning_rate": 5e-05, "loss": 0.7498, "mean_token_accuracy": 0.8850149512290955, "num_tokens": 615966873.0, "step": 16223 }, { "epoch": 2.0638595598524363, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.384119987487793, "learning_rate": 5e-05, "loss": 0.9161, "mean_token_accuracy": 0.8968827724456787, "num_tokens": 616002021.0, "step": 16224 }, { "epoch": 2.0639867701310264, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.2798237800598145, "learning_rate": 5e-05, "loss": 0.8561, "mean_token_accuracy": 0.8876034021377563, "num_tokens": 616043692.0, "step": 16225 }, { "epoch": 2.064113980409617, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.645790100097656, "learning_rate": 5e-05, "loss": 0.8795, "mean_token_accuracy": 0.895063579082489, "num_tokens": 616083811.0, "step": 16226 }, { "epoch": 2.0642411906882074, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.911593437194824, "learning_rate": 5e-05, "loss": 0.8728, "mean_token_accuracy": 0.8969678282737732, "num_tokens": 616121714.0, "step": 16227 }, { "epoch": 2.064368400966798, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.783476829528809, "learning_rate": 5e-05, "loss": 0.8247, "mean_token_accuracy": 0.8823641538619995, "num_tokens": 616156571.0, "step": 16228 }, { "epoch": 2.0644956112453885, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.152698516845703, "learning_rate": 5e-05, "loss": 0.908, "mean_token_accuracy": 0.8792004585266113, "num_tokens": 616197198.0, "step": 16229 }, { "epoch": 2.064622821523979, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.400606155395508, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.8755448460578918, "num_tokens": 616233495.0, "step": 16230 }, { "epoch": 2.0647500318025696, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.124507904052734, "learning_rate": 5e-05, "loss": 0.7991, "mean_token_accuracy": 0.8976284265518188, "num_tokens": 616268894.0, "step": 16231 }, { "epoch": 2.06487724208116, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.357967376708984, "learning_rate": 5e-05, "loss": 0.7567, "mean_token_accuracy": 0.8850830793380737, "num_tokens": 616302899.0, "step": 16232 }, { "epoch": 2.0650044523597506, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.030709266662598, "learning_rate": 5e-05, "loss": 0.855, "mean_token_accuracy": 0.865239679813385, "num_tokens": 616342937.0, "step": 16233 }, { "epoch": 2.065131662638341, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.320497989654541, "learning_rate": 5e-05, "loss": 0.7472, "mean_token_accuracy": 0.8849971890449524, "num_tokens": 616391581.0, "step": 16234 }, { "epoch": 2.0652588729169317, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.02750301361084, "learning_rate": 5e-05, "loss": 0.782, "mean_token_accuracy": 0.8863766193389893, "num_tokens": 616430303.0, "step": 16235 }, { "epoch": 2.065386083195522, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.206048965454102, "learning_rate": 5e-05, "loss": 0.6944, "mean_token_accuracy": 0.8943552374839783, "num_tokens": 616468294.0, "step": 16236 }, { "epoch": 2.0655132934741127, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.096294403076172, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8910000324249268, "num_tokens": 616503551.0, "step": 16237 }, { "epoch": 2.0656405037527032, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.08208703994751, "learning_rate": 5e-05, "loss": 0.6919, "mean_token_accuracy": 0.8924756050109863, "num_tokens": 616541374.0, "step": 16238 }, { "epoch": 2.065767714031294, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.19601821899414, "learning_rate": 5e-05, "loss": 0.7713, "mean_token_accuracy": 0.8904685974121094, "num_tokens": 616575480.0, "step": 16239 }, { "epoch": 2.0658949243098843, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.954018592834473, "learning_rate": 5e-05, "loss": 0.7073, "mean_token_accuracy": 0.8826984167098999, "num_tokens": 616620323.0, "step": 16240 }, { "epoch": 2.066022134588475, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.373003959655762, "learning_rate": 5e-05, "loss": 0.7669, "mean_token_accuracy": 0.89668208360672, "num_tokens": 616659843.0, "step": 16241 }, { "epoch": 2.0661493448670654, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.755197525024414, "learning_rate": 5e-05, "loss": 0.6704, "mean_token_accuracy": 0.8917863368988037, "num_tokens": 616696065.0, "step": 16242 }, { "epoch": 2.066276555145656, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.629443168640137, "learning_rate": 5e-05, "loss": 0.8158, "mean_token_accuracy": 0.8900840282440186, "num_tokens": 616729444.0, "step": 16243 }, { "epoch": 2.0664037654242464, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.5808281898498535, "learning_rate": 5e-05, "loss": 0.6816, "mean_token_accuracy": 0.8848724961280823, "num_tokens": 616773351.0, "step": 16244 }, { "epoch": 2.066530975702837, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.925667762756348, "learning_rate": 5e-05, "loss": 0.8731, "mean_token_accuracy": 0.881233811378479, "num_tokens": 616812161.0, "step": 16245 }, { "epoch": 2.0666581859814275, "ewc_loss": 0.375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.5924506187438965, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8713631629943848, "num_tokens": 616847623.0, "step": 16246 }, { "epoch": 2.066785396260018, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.93977165222168, "learning_rate": 5e-05, "loss": 0.9001, "mean_token_accuracy": 0.8824235796928406, "num_tokens": 616884078.0, "step": 16247 }, { "epoch": 2.0669126065386085, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 6.969292640686035, "learning_rate": 5e-05, "loss": 0.7661, "mean_token_accuracy": 0.8846803903579712, "num_tokens": 616924291.0, "step": 16248 }, { "epoch": 2.0670398168171986, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.49736499786377, "learning_rate": 5e-05, "loss": 0.8513, "mean_token_accuracy": 0.8918307423591614, "num_tokens": 616964580.0, "step": 16249 }, { "epoch": 2.067167027095789, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.512194633483887, "learning_rate": 5e-05, "loss": 0.7906, "mean_token_accuracy": 0.8956300616264343, "num_tokens": 617000574.0, "step": 16250 }, { "epoch": 2.0672942373743797, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.776679992675781, "learning_rate": 5e-05, "loss": 0.7959, "mean_token_accuracy": 0.8830325603485107, "num_tokens": 617043323.0, "step": 16251 }, { "epoch": 2.06742144765297, "ewc_loss": 0.5, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.967526912689209, "learning_rate": 5e-05, "loss": 0.8106, "mean_token_accuracy": 0.8890032172203064, "num_tokens": 617082585.0, "step": 16252 }, { "epoch": 2.0675486579315607, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.246918201446533, "learning_rate": 5e-05, "loss": 0.7109, "mean_token_accuracy": 0.8865604400634766, "num_tokens": 617118384.0, "step": 16253 }, { "epoch": 2.0676758682101513, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.20145320892334, "learning_rate": 5e-05, "loss": 0.7683, "mean_token_accuracy": 0.898716151714325, "num_tokens": 617153354.0, "step": 16254 }, { "epoch": 2.067803078488742, "ewc_loss": 0.375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.032973289489746, "learning_rate": 5e-05, "loss": 0.6922, "mean_token_accuracy": 0.8876161575317383, "num_tokens": 617192829.0, "step": 16255 }, { "epoch": 2.0679302887673323, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.23293399810791, "learning_rate": 5e-05, "loss": 0.7986, "mean_token_accuracy": 0.8866868019104004, "num_tokens": 617230824.0, "step": 16256 }, { "epoch": 2.068057499045923, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.029137134552002, "learning_rate": 5e-05, "loss": 0.7178, "mean_token_accuracy": 0.8774891495704651, "num_tokens": 617268791.0, "step": 16257 }, { "epoch": 2.0681847093245134, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.200787544250488, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.8843412399291992, "num_tokens": 617305366.0, "step": 16258 }, { "epoch": 2.068311919603104, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.057519435882568, "learning_rate": 5e-05, "loss": 0.7068, "mean_token_accuracy": 0.8866903185844421, "num_tokens": 617342294.0, "step": 16259 }, { "epoch": 2.0684391298816944, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.17000675201416, "learning_rate": 5e-05, "loss": 0.7853, "mean_token_accuracy": 0.8853066563606262, "num_tokens": 617380456.0, "step": 16260 }, { "epoch": 2.068566340160285, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.056667327880859, "learning_rate": 5e-05, "loss": 0.6809, "mean_token_accuracy": 0.8917059302330017, "num_tokens": 617420705.0, "step": 16261 }, { "epoch": 2.0686935504388755, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.191253662109375, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.889135479927063, "num_tokens": 617462740.0, "step": 16262 }, { "epoch": 2.068820760717466, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.01101541519165, "learning_rate": 5e-05, "loss": 0.6751, "mean_token_accuracy": 0.8923991918563843, "num_tokens": 617497496.0, "step": 16263 }, { "epoch": 2.0689479709960565, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.239005088806152, "learning_rate": 5e-05, "loss": 0.7544, "mean_token_accuracy": 0.8965948820114136, "num_tokens": 617535297.0, "step": 16264 }, { "epoch": 2.069075181274647, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.956454753875732, "learning_rate": 5e-05, "loss": 0.6936, "mean_token_accuracy": 0.8857212066650391, "num_tokens": 617568332.0, "step": 16265 }, { "epoch": 2.0692023915532376, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.323747634887695, "learning_rate": 5e-05, "loss": 0.8042, "mean_token_accuracy": 0.8807677030563354, "num_tokens": 617608934.0, "step": 16266 }, { "epoch": 2.069329601831828, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.886081218719482, "learning_rate": 5e-05, "loss": 0.6999, "mean_token_accuracy": 0.8776860237121582, "num_tokens": 617643794.0, "step": 16267 }, { "epoch": 2.0694568121104187, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.394185066223145, "learning_rate": 5e-05, "loss": 0.7983, "mean_token_accuracy": 0.8891065120697021, "num_tokens": 617676448.0, "step": 16268 }, { "epoch": 2.069584022389009, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.857524394989014, "learning_rate": 5e-05, "loss": 0.7543, "mean_token_accuracy": 0.8639993667602539, "num_tokens": 617713460.0, "step": 16269 }, { "epoch": 2.0697112326675997, "ewc_loss": 0.5, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.436196327209473, "learning_rate": 5e-05, "loss": 0.8353, "mean_token_accuracy": 0.8792111873626709, "num_tokens": 617749767.0, "step": 16270 }, { "epoch": 2.0698384429461902, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.906620979309082, "learning_rate": 5e-05, "loss": 0.7429, "mean_token_accuracy": 0.870690107345581, "num_tokens": 617784991.0, "step": 16271 }, { "epoch": 2.0699656532247808, "ewc_loss": 0.5, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.392424583435059, "learning_rate": 5e-05, "loss": 0.7905, "mean_token_accuracy": 0.8933374881744385, "num_tokens": 617820241.0, "step": 16272 }, { "epoch": 2.070092863503371, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.0359368324279785, "learning_rate": 5e-05, "loss": 0.7488, "mean_token_accuracy": 0.8760334253311157, "num_tokens": 617861838.0, "step": 16273 }, { "epoch": 2.0702200737819614, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.206289291381836, "learning_rate": 5e-05, "loss": 0.7995, "mean_token_accuracy": 0.8865495324134827, "num_tokens": 617893001.0, "step": 16274 }, { "epoch": 2.070347284060552, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.249741554260254, "learning_rate": 5e-05, "loss": 0.7416, "mean_token_accuracy": 0.8844910860061646, "num_tokens": 617936361.0, "step": 16275 }, { "epoch": 2.0704744943391424, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.910438060760498, "learning_rate": 5e-05, "loss": 0.7769, "mean_token_accuracy": 0.884546160697937, "num_tokens": 617968267.0, "step": 16276 }, { "epoch": 2.070601704617733, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.485731601715088, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.8834716081619263, "num_tokens": 618007710.0, "step": 16277 }, { "epoch": 2.0707289148963235, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.581465244293213, "learning_rate": 5e-05, "loss": 0.7305, "mean_token_accuracy": 0.886573851108551, "num_tokens": 618043185.0, "step": 16278 }, { "epoch": 2.070856125174914, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.675603866577148, "learning_rate": 5e-05, "loss": 0.713, "mean_token_accuracy": 0.8939554691314697, "num_tokens": 618085460.0, "step": 16279 }, { "epoch": 2.0709833354535045, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.378164768218994, "learning_rate": 5e-05, "loss": 0.6891, "mean_token_accuracy": 0.8944735527038574, "num_tokens": 618117155.0, "step": 16280 }, { "epoch": 2.071110545732095, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.774224758148193, "learning_rate": 5e-05, "loss": 0.7199, "mean_token_accuracy": 0.891110897064209, "num_tokens": 618153380.0, "step": 16281 }, { "epoch": 2.0712377560106856, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.243334770202637, "learning_rate": 5e-05, "loss": 0.6909, "mean_token_accuracy": 0.8857308030128479, "num_tokens": 618190125.0, "step": 16282 }, { "epoch": 2.071364966289276, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.828486442565918, "learning_rate": 5e-05, "loss": 0.7345, "mean_token_accuracy": 0.8842217922210693, "num_tokens": 618225281.0, "step": 16283 }, { "epoch": 2.0714921765678667, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.152583599090576, "learning_rate": 5e-05, "loss": 0.6727, "mean_token_accuracy": 0.8892960548400879, "num_tokens": 618262213.0, "step": 16284 }, { "epoch": 2.071619386846457, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.907553672790527, "learning_rate": 5e-05, "loss": 0.777, "mean_token_accuracy": 0.8701915144920349, "num_tokens": 618292780.0, "step": 16285 }, { "epoch": 2.0717465971250477, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.033078193664551, "learning_rate": 5e-05, "loss": 0.7131, "mean_token_accuracy": 0.8739013671875, "num_tokens": 618332968.0, "step": 16286 }, { "epoch": 2.0718738074036382, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.059014320373535, "learning_rate": 5e-05, "loss": 0.7065, "mean_token_accuracy": 0.8955961465835571, "num_tokens": 618369585.0, "step": 16287 }, { "epoch": 2.0720010176822288, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.865951061248779, "learning_rate": 5e-05, "loss": 0.6316, "mean_token_accuracy": 0.8989050388336182, "num_tokens": 618410388.0, "step": 16288 }, { "epoch": 2.0721282279608193, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.317696571350098, "learning_rate": 5e-05, "loss": 0.7206, "mean_token_accuracy": 0.8983515501022339, "num_tokens": 618453591.0, "step": 16289 }, { "epoch": 2.07225543823941, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.572909355163574, "learning_rate": 5e-05, "loss": 0.6455, "mean_token_accuracy": 0.8903940916061401, "num_tokens": 618489757.0, "step": 16290 }, { "epoch": 2.0723826485180004, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.740069389343262, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.8890730738639832, "num_tokens": 618524733.0, "step": 16291 }, { "epoch": 2.072509858796591, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.268614768981934, "learning_rate": 5e-05, "loss": 0.6572, "mean_token_accuracy": 0.8850628137588501, "num_tokens": 618564734.0, "step": 16292 }, { "epoch": 2.0726370690751814, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.254488945007324, "learning_rate": 5e-05, "loss": 0.8776, "mean_token_accuracy": 0.8880633115768433, "num_tokens": 618602503.0, "step": 16293 }, { "epoch": 2.072764279353772, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.227043628692627, "learning_rate": 5e-05, "loss": 0.6491, "mean_token_accuracy": 0.8950765132904053, "num_tokens": 618639907.0, "step": 16294 }, { "epoch": 2.0728914896323625, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.459348678588867, "learning_rate": 5e-05, "loss": 0.9367, "mean_token_accuracy": 0.8879081010818481, "num_tokens": 618681366.0, "step": 16295 }, { "epoch": 2.073018699910953, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 6.974757671356201, "learning_rate": 5e-05, "loss": 0.7611, "mean_token_accuracy": 0.8964518904685974, "num_tokens": 618718610.0, "step": 16296 }, { "epoch": 2.0731459101895435, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.64136791229248, "learning_rate": 5e-05, "loss": 0.8819, "mean_token_accuracy": 0.8914389610290527, "num_tokens": 618757741.0, "step": 16297 }, { "epoch": 2.0732731204681336, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.138221740722656, "learning_rate": 5e-05, "loss": 0.8945, "mean_token_accuracy": 0.8929197788238525, "num_tokens": 618798189.0, "step": 16298 }, { "epoch": 2.073400330746724, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.9805145263671875, "learning_rate": 5e-05, "loss": 0.7264, "mean_token_accuracy": 0.8842555284500122, "num_tokens": 618837308.0, "step": 16299 }, { "epoch": 2.0735275410253147, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.443511962890625, "learning_rate": 5e-05, "loss": 0.9946, "mean_token_accuracy": 0.87285315990448, "num_tokens": 618876058.0, "step": 16300 }, { "epoch": 2.073654751303905, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 5.9238505363464355, "learning_rate": 5e-05, "loss": 0.6977, "mean_token_accuracy": 0.8771529197692871, "num_tokens": 618916246.0, "step": 16301 }, { "epoch": 2.0737819615824957, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.672605514526367, "learning_rate": 5e-05, "loss": 1.1141, "mean_token_accuracy": 0.8781319856643677, "num_tokens": 618957639.0, "step": 16302 }, { "epoch": 2.0739091718610863, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 6.914508819580078, "learning_rate": 5e-05, "loss": 0.8624, "mean_token_accuracy": 0.8948431015014648, "num_tokens": 618999505.0, "step": 16303 }, { "epoch": 2.074036382139677, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.64655590057373, "learning_rate": 5e-05, "loss": 1.0945, "mean_token_accuracy": 0.8911265730857849, "num_tokens": 619034807.0, "step": 16304 }, { "epoch": 2.0741635924182673, "ewc_loss": 0.85546875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 8.829402923583984, "learning_rate": 5e-05, "loss": 1.1488, "mean_token_accuracy": 0.894449770450592, "num_tokens": 619073262.0, "step": 16305 }, { "epoch": 2.074290802696858, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.90955924987793, "learning_rate": 5e-05, "loss": 0.9272, "mean_token_accuracy": 0.8782773017883301, "num_tokens": 619112706.0, "step": 16306 }, { "epoch": 2.0744180129754484, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000732421875, "grad_norm": 9.44698715209961, "learning_rate": 5e-05, "loss": 1.146, "mean_token_accuracy": 0.8769879341125488, "num_tokens": 619144444.0, "step": 16307 }, { "epoch": 2.074545223254039, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.605432033538818, "learning_rate": 5e-05, "loss": 0.83, "mean_token_accuracy": 0.8925857543945312, "num_tokens": 619183291.0, "step": 16308 }, { "epoch": 2.0746724335326294, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.815592765808105, "learning_rate": 5e-05, "loss": 0.9905, "mean_token_accuracy": 0.8940919637680054, "num_tokens": 619223865.0, "step": 16309 }, { "epoch": 2.07479964381122, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 7.826781272888184, "learning_rate": 5e-05, "loss": 0.9249, "mean_token_accuracy": 0.8929569721221924, "num_tokens": 619257418.0, "step": 16310 }, { "epoch": 2.0749268540898105, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.03331470489502, "learning_rate": 5e-05, "loss": 0.953, "mean_token_accuracy": 0.8793951869010925, "num_tokens": 619298724.0, "step": 16311 }, { "epoch": 2.075054064368401, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.353066444396973, "learning_rate": 5e-05, "loss": 0.9721, "mean_token_accuracy": 0.8868011236190796, "num_tokens": 619332497.0, "step": 16312 }, { "epoch": 2.0751812746469915, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.827950954437256, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.885868489742279, "num_tokens": 619363371.0, "step": 16313 }, { "epoch": 2.075308484925582, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.790477752685547, "learning_rate": 5e-05, "loss": 0.9305, "mean_token_accuracy": 0.8869878053665161, "num_tokens": 619396068.0, "step": 16314 }, { "epoch": 2.0754356952041726, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.2491984367370605, "learning_rate": 5e-05, "loss": 0.7864, "mean_token_accuracy": 0.8845424652099609, "num_tokens": 619433728.0, "step": 16315 }, { "epoch": 2.075562905482763, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.940522193908691, "learning_rate": 5e-05, "loss": 0.9405, "mean_token_accuracy": 0.8757952451705933, "num_tokens": 619476185.0, "step": 16316 }, { "epoch": 2.0756901157613536, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.2892961502075195, "learning_rate": 5e-05, "loss": 0.7765, "mean_token_accuracy": 0.9027197360992432, "num_tokens": 619513184.0, "step": 16317 }, { "epoch": 2.075817326039944, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.658839225769043, "learning_rate": 5e-05, "loss": 0.8809, "mean_token_accuracy": 0.8792607188224792, "num_tokens": 619550168.0, "step": 16318 }, { "epoch": 2.0759445363185347, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.493499755859375, "learning_rate": 5e-05, "loss": 0.8604, "mean_token_accuracy": 0.883394718170166, "num_tokens": 619588960.0, "step": 16319 }, { "epoch": 2.0760717465971252, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.132857322692871, "learning_rate": 5e-05, "loss": 0.7871, "mean_token_accuracy": 0.8909687399864197, "num_tokens": 619624907.0, "step": 16320 }, { "epoch": 2.0761989568757158, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.689414978027344, "learning_rate": 5e-05, "loss": 0.8137, "mean_token_accuracy": 0.8939071893692017, "num_tokens": 619670576.0, "step": 16321 }, { "epoch": 2.0763261671543063, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.617692470550537, "learning_rate": 5e-05, "loss": 0.7328, "mean_token_accuracy": 0.8912687301635742, "num_tokens": 619710452.0, "step": 16322 }, { "epoch": 2.0764533774328964, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.977543354034424, "learning_rate": 5e-05, "loss": 0.8271, "mean_token_accuracy": 0.8878703713417053, "num_tokens": 619750104.0, "step": 16323 }, { "epoch": 2.076580587711487, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.1946306228637695, "learning_rate": 5e-05, "loss": 0.6927, "mean_token_accuracy": 0.8931041955947876, "num_tokens": 619791364.0, "step": 16324 }, { "epoch": 2.0767077979900774, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.295775413513184, "learning_rate": 5e-05, "loss": 0.8338, "mean_token_accuracy": 0.8808903694152832, "num_tokens": 619831138.0, "step": 16325 }, { "epoch": 2.076835008268668, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.9135518074035645, "learning_rate": 5e-05, "loss": 0.6907, "mean_token_accuracy": 0.8910613059997559, "num_tokens": 619862897.0, "step": 16326 }, { "epoch": 2.0769622185472585, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.555212020874023, "learning_rate": 5e-05, "loss": 0.8041, "mean_token_accuracy": 0.8950077295303345, "num_tokens": 619897096.0, "step": 16327 }, { "epoch": 2.077089428825849, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.745198726654053, "learning_rate": 5e-05, "loss": 0.6732, "mean_token_accuracy": 0.8987007141113281, "num_tokens": 619937458.0, "step": 16328 }, { "epoch": 2.0772166391044395, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.732857704162598, "learning_rate": 5e-05, "loss": 0.8778, "mean_token_accuracy": 0.8751973509788513, "num_tokens": 619974735.0, "step": 16329 }, { "epoch": 2.07734384938303, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.7509613037109375, "learning_rate": 5e-05, "loss": 0.7446, "mean_token_accuracy": 0.8818789720535278, "num_tokens": 620012497.0, "step": 16330 }, { "epoch": 2.0774710596616206, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.726655960083008, "learning_rate": 5e-05, "loss": 0.8496, "mean_token_accuracy": 0.8825734853744507, "num_tokens": 620051120.0, "step": 16331 }, { "epoch": 2.077598269940211, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 6.913036346435547, "learning_rate": 5e-05, "loss": 0.7511, "mean_token_accuracy": 0.8852821588516235, "num_tokens": 620091479.0, "step": 16332 }, { "epoch": 2.0777254802188017, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.541903495788574, "learning_rate": 5e-05, "loss": 0.8404, "mean_token_accuracy": 0.8869062066078186, "num_tokens": 620126509.0, "step": 16333 }, { "epoch": 2.077852690497392, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.092690467834473, "learning_rate": 5e-05, "loss": 0.7565, "mean_token_accuracy": 0.8888200521469116, "num_tokens": 620164278.0, "step": 16334 }, { "epoch": 2.0779799007759827, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.369762420654297, "learning_rate": 5e-05, "loss": 0.8438, "mean_token_accuracy": 0.879210889339447, "num_tokens": 620202388.0, "step": 16335 }, { "epoch": 2.0781071110545732, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.169966220855713, "learning_rate": 5e-05, "loss": 0.739, "mean_token_accuracy": 0.8910104036331177, "num_tokens": 620236981.0, "step": 16336 }, { "epoch": 2.0782343213331638, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.2532320022583, "learning_rate": 5e-05, "loss": 0.8115, "mean_token_accuracy": 0.8890209197998047, "num_tokens": 620275946.0, "step": 16337 }, { "epoch": 2.0783615316117543, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.149931907653809, "learning_rate": 5e-05, "loss": 0.7872, "mean_token_accuracy": 0.8732849359512329, "num_tokens": 620317385.0, "step": 16338 }, { "epoch": 2.078488741890345, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.243304252624512, "learning_rate": 5e-05, "loss": 0.7885, "mean_token_accuracy": 0.8921299576759338, "num_tokens": 620353625.0, "step": 16339 }, { "epoch": 2.0786159521689354, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000659942626953125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.146241188049316, "learning_rate": 5e-05, "loss": 0.7903, "mean_token_accuracy": 0.8685718774795532, "num_tokens": 620389421.0, "step": 16340 }, { "epoch": 2.078743162447526, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.169368743896484, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.8896634578704834, "num_tokens": 620428233.0, "step": 16341 }, { "epoch": 2.0788703727261164, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.144845008850098, "learning_rate": 5e-05, "loss": 0.8032, "mean_token_accuracy": 0.8656932711601257, "num_tokens": 620465156.0, "step": 16342 }, { "epoch": 2.078997583004707, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.14582633972168, "learning_rate": 5e-05, "loss": 0.7979, "mean_token_accuracy": 0.8855504989624023, "num_tokens": 620501731.0, "step": 16343 }, { "epoch": 2.0791247932832975, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.134271621704102, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.8829150795936584, "num_tokens": 620538507.0, "step": 16344 }, { "epoch": 2.079252003561888, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.101456642150879, "learning_rate": 5e-05, "loss": 0.7422, "mean_token_accuracy": 0.8988885283470154, "num_tokens": 620577355.0, "step": 16345 }, { "epoch": 2.0793792138404785, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.147608757019043, "learning_rate": 5e-05, "loss": 0.7158, "mean_token_accuracy": 0.8803682923316956, "num_tokens": 620613320.0, "step": 16346 }, { "epoch": 2.0795064241190686, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.032513618469238, "learning_rate": 5e-05, "loss": 0.7397, "mean_token_accuracy": 0.8961648941040039, "num_tokens": 620655167.0, "step": 16347 }, { "epoch": 2.079633634397659, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.201544761657715, "learning_rate": 5e-05, "loss": 0.7412, "mean_token_accuracy": 0.8707492351531982, "num_tokens": 620697898.0, "step": 16348 }, { "epoch": 2.0797608446762497, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.9386115074157715, "learning_rate": 5e-05, "loss": 0.7513, "mean_token_accuracy": 0.8897441029548645, "num_tokens": 620736548.0, "step": 16349 }, { "epoch": 2.07988805495484, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.237483501434326, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8745070099830627, "num_tokens": 620781368.0, "step": 16350 }, { "epoch": 2.0800152652334307, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.866183757781982, "learning_rate": 5e-05, "loss": 0.7548, "mean_token_accuracy": 0.8839223384857178, "num_tokens": 620821749.0, "step": 16351 }, { "epoch": 2.0801424755120212, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.266279220581055, "learning_rate": 5e-05, "loss": 0.7116, "mean_token_accuracy": 0.8838625550270081, "num_tokens": 620857906.0, "step": 16352 }, { "epoch": 2.0802696857906118, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.814647674560547, "learning_rate": 5e-05, "loss": 0.7386, "mean_token_accuracy": 0.8864903450012207, "num_tokens": 620897124.0, "step": 16353 }, { "epoch": 2.0803968960692023, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.261620044708252, "learning_rate": 5e-05, "loss": 0.7376, "mean_token_accuracy": 0.8751030564308167, "num_tokens": 620932123.0, "step": 16354 }, { "epoch": 2.080524106347793, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.787115097045898, "learning_rate": 5e-05, "loss": 0.7301, "mean_token_accuracy": 0.8861702680587769, "num_tokens": 620970906.0, "step": 16355 }, { "epoch": 2.0806513166263834, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.267009258270264, "learning_rate": 5e-05, "loss": 0.7123, "mean_token_accuracy": 0.8793758749961853, "num_tokens": 621007387.0, "step": 16356 }, { "epoch": 2.080778526904974, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.762960910797119, "learning_rate": 5e-05, "loss": 0.7149, "mean_token_accuracy": 0.8875466585159302, "num_tokens": 621043723.0, "step": 16357 }, { "epoch": 2.0809057371835644, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.2627177238464355, "learning_rate": 5e-05, "loss": 0.6627, "mean_token_accuracy": 0.8934615254402161, "num_tokens": 621083686.0, "step": 16358 }, { "epoch": 2.081032947462155, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.742023944854736, "learning_rate": 5e-05, "loss": 0.7137, "mean_token_accuracy": 0.8850151896476746, "num_tokens": 621125758.0, "step": 16359 }, { "epoch": 2.0811601577407455, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.246574878692627, "learning_rate": 5e-05, "loss": 0.6949, "mean_token_accuracy": 0.8846592903137207, "num_tokens": 621163407.0, "step": 16360 }, { "epoch": 2.081287368019336, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.752613544464111, "learning_rate": 5e-05, "loss": 0.7195, "mean_token_accuracy": 0.8822570443153381, "num_tokens": 621203321.0, "step": 16361 }, { "epoch": 2.0814145782979265, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.210742473602295, "learning_rate": 5e-05, "loss": 0.6723, "mean_token_accuracy": 0.8856927752494812, "num_tokens": 621239777.0, "step": 16362 }, { "epoch": 2.081541788576517, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.794103145599365, "learning_rate": 5e-05, "loss": 0.6962, "mean_token_accuracy": 0.8882551789283752, "num_tokens": 621271841.0, "step": 16363 }, { "epoch": 2.0816689988551076, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.120809078216553, "learning_rate": 5e-05, "loss": 0.67, "mean_token_accuracy": 0.8878298997879028, "num_tokens": 621309393.0, "step": 16364 }, { "epoch": 2.081796209133698, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.9182634353637695, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.8742074370384216, "num_tokens": 621344362.0, "step": 16365 }, { "epoch": 2.0819234194122886, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.956582546234131, "learning_rate": 5e-05, "loss": 0.7097, "mean_token_accuracy": 0.868554949760437, "num_tokens": 621383999.0, "step": 16366 }, { "epoch": 2.082050629690879, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.14031982421875, "learning_rate": 5e-05, "loss": 0.7193, "mean_token_accuracy": 0.8909543752670288, "num_tokens": 621424475.0, "step": 16367 }, { "epoch": 2.0821778399694697, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.722391605377197, "learning_rate": 5e-05, "loss": 0.674, "mean_token_accuracy": 0.8769757151603699, "num_tokens": 621458241.0, "step": 16368 }, { "epoch": 2.0823050502480602, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.470325469970703, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.8927422761917114, "num_tokens": 621495424.0, "step": 16369 }, { "epoch": 2.0824322605266508, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.422979831695557, "learning_rate": 5e-05, "loss": 0.6466, "mean_token_accuracy": 0.8854952454566956, "num_tokens": 621531256.0, "step": 16370 }, { "epoch": 2.082559470805241, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.905380249023438, "learning_rate": 5e-05, "loss": 0.8291, "mean_token_accuracy": 0.886642575263977, "num_tokens": 621571665.0, "step": 16371 }, { "epoch": 2.0826866810838314, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.1869049072265625, "learning_rate": 5e-05, "loss": 0.647, "mean_token_accuracy": 0.8889981508255005, "num_tokens": 621604552.0, "step": 16372 }, { "epoch": 2.082813891362422, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.286251068115234, "learning_rate": 5e-05, "loss": 0.9557, "mean_token_accuracy": 0.8640744686126709, "num_tokens": 621651677.0, "step": 16373 }, { "epoch": 2.0829411016410124, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.434529781341553, "learning_rate": 5e-05, "loss": 0.6752, "mean_token_accuracy": 0.8899383544921875, "num_tokens": 621695108.0, "step": 16374 }, { "epoch": 2.083068311919603, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.031189918518066, "learning_rate": 5e-05, "loss": 0.8966, "mean_token_accuracy": 0.89518141746521, "num_tokens": 621725768.0, "step": 16375 }, { "epoch": 2.0831955221981935, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.391016006469727, "learning_rate": 5e-05, "loss": 0.8572, "mean_token_accuracy": 0.8772916793823242, "num_tokens": 621768559.0, "step": 16376 }, { "epoch": 2.083322732476784, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.696265697479248, "learning_rate": 5e-05, "loss": 0.7964, "mean_token_accuracy": 0.8839125633239746, "num_tokens": 621805464.0, "step": 16377 }, { "epoch": 2.0834499427553745, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.69444465637207, "learning_rate": 5e-05, "loss": 0.9398, "mean_token_accuracy": 0.8752155303955078, "num_tokens": 621843129.0, "step": 16378 }, { "epoch": 2.083577153033965, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.089181423187256, "learning_rate": 5e-05, "loss": 0.6492, "mean_token_accuracy": 0.8841606974601746, "num_tokens": 621882450.0, "step": 16379 }, { "epoch": 2.0837043633125556, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.428544998168945, "learning_rate": 5e-05, "loss": 0.9942, "mean_token_accuracy": 0.8822981715202332, "num_tokens": 621915797.0, "step": 16380 }, { "epoch": 2.083831573591146, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 5.877673149108887, "learning_rate": 5e-05, "loss": 0.7108, "mean_token_accuracy": 0.8748672604560852, "num_tokens": 621953563.0, "step": 16381 }, { "epoch": 2.0839587838697367, "ewc_loss": 0.9765625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000911712646484375, "grad_norm": 11.95186710357666, "learning_rate": 5e-05, "loss": 1.2712, "mean_token_accuracy": 0.8968520164489746, "num_tokens": 621989579.0, "step": 16382 }, { "epoch": 2.084085994148327, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 8.88442611694336, "learning_rate": 5e-05, "loss": 1.1503, "mean_token_accuracy": 0.9001315236091614, "num_tokens": 622027446.0, "step": 16383 }, { "epoch": 2.0842132044269177, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.883903503417969, "learning_rate": 5e-05, "loss": 1.0743, "mean_token_accuracy": 0.8737501502037048, "num_tokens": 622064972.0, "step": 16384 }, { "epoch": 2.0843404147055082, "ewc_loss": 0.92578125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 9.567756652832031, "learning_rate": 5e-05, "loss": 1.2172, "mean_token_accuracy": 0.8945175409317017, "num_tokens": 622103260.0, "step": 16385 }, { "epoch": 2.0844676249840988, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.592148780822754, "learning_rate": 5e-05, "loss": 0.977, "mean_token_accuracy": 0.8886001110076904, "num_tokens": 622140894.0, "step": 16386 }, { "epoch": 2.0845948352626893, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 9.351072311401367, "learning_rate": 5e-05, "loss": 1.036, "mean_token_accuracy": 0.8906281590461731, "num_tokens": 622177537.0, "step": 16387 }, { "epoch": 2.08472204554128, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.433141708374023, "learning_rate": 5e-05, "loss": 0.9262, "mean_token_accuracy": 0.8884899020195007, "num_tokens": 622211172.0, "step": 16388 }, { "epoch": 2.0848492558198704, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.264223098754883, "learning_rate": 5e-05, "loss": 0.9735, "mean_token_accuracy": 0.8822538256645203, "num_tokens": 622244805.0, "step": 16389 }, { "epoch": 2.084976466098461, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.875063896179199, "learning_rate": 5e-05, "loss": 0.8564, "mean_token_accuracy": 0.8946553468704224, "num_tokens": 622279496.0, "step": 16390 }, { "epoch": 2.0851036763770514, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.87841796875, "learning_rate": 5e-05, "loss": 0.8898, "mean_token_accuracy": 0.8937106132507324, "num_tokens": 622318818.0, "step": 16391 }, { "epoch": 2.085230886655642, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.628157138824463, "learning_rate": 5e-05, "loss": 0.8101, "mean_token_accuracy": 0.8888028264045715, "num_tokens": 622359313.0, "step": 16392 }, { "epoch": 2.0853580969342325, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.584037780761719, "learning_rate": 5e-05, "loss": 0.8866, "mean_token_accuracy": 0.8819522857666016, "num_tokens": 622392840.0, "step": 16393 }, { "epoch": 2.085485307212823, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.392238616943359, "learning_rate": 5e-05, "loss": 0.8264, "mean_token_accuracy": 0.8711536526679993, "num_tokens": 622433522.0, "step": 16394 }, { "epoch": 2.0856125174914135, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.560317039489746, "learning_rate": 5e-05, "loss": 0.863, "mean_token_accuracy": 0.8876121044158936, "num_tokens": 622468805.0, "step": 16395 }, { "epoch": 2.0857397277700036, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.107122421264648, "learning_rate": 5e-05, "loss": 0.7207, "mean_token_accuracy": 0.8928250074386597, "num_tokens": 622511462.0, "step": 16396 }, { "epoch": 2.085866938048594, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.745160102844238, "learning_rate": 5e-05, "loss": 0.8715, "mean_token_accuracy": 0.8894203305244446, "num_tokens": 622546246.0, "step": 16397 }, { "epoch": 2.0859941483271847, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 6.9196882247924805, "learning_rate": 5e-05, "loss": 0.7177, "mean_token_accuracy": 0.8921545147895813, "num_tokens": 622583113.0, "step": 16398 }, { "epoch": 2.086121358605775, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.920510292053223, "learning_rate": 5e-05, "loss": 0.9003, "mean_token_accuracy": 0.8846449851989746, "num_tokens": 622621414.0, "step": 16399 }, { "epoch": 2.0862485688843657, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 6.963918685913086, "learning_rate": 5e-05, "loss": 0.8104, "mean_token_accuracy": 0.8696275353431702, "num_tokens": 622660079.0, "step": 16400 }, { "epoch": 2.0863757791629562, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.898987770080566, "learning_rate": 5e-05, "loss": 0.9224, "mean_token_accuracy": 0.8855396509170532, "num_tokens": 622700413.0, "step": 16401 }, { "epoch": 2.0865029894415468, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.267232894897461, "learning_rate": 5e-05, "loss": 0.8111, "mean_token_accuracy": 0.8894151449203491, "num_tokens": 622736448.0, "step": 16402 }, { "epoch": 2.0866301997201373, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.502641677856445, "learning_rate": 5e-05, "loss": 0.8823, "mean_token_accuracy": 0.8849299550056458, "num_tokens": 622777014.0, "step": 16403 }, { "epoch": 2.086757409998728, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.672579765319824, "learning_rate": 5e-05, "loss": 0.845, "mean_token_accuracy": 0.8869482278823853, "num_tokens": 622817608.0, "step": 16404 }, { "epoch": 2.0868846202773184, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.808290004730225, "learning_rate": 5e-05, "loss": 0.7963, "mean_token_accuracy": 0.8871808648109436, "num_tokens": 622853090.0, "step": 16405 }, { "epoch": 2.087011830555909, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.015178680419922, "learning_rate": 5e-05, "loss": 0.8086, "mean_token_accuracy": 0.8972182869911194, "num_tokens": 622891843.0, "step": 16406 }, { "epoch": 2.0871390408344994, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.21599006652832, "learning_rate": 5e-05, "loss": 0.7861, "mean_token_accuracy": 0.8643040657043457, "num_tokens": 622930364.0, "step": 16407 }, { "epoch": 2.08726625111309, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.291600227355957, "learning_rate": 5e-05, "loss": 0.8931, "mean_token_accuracy": 0.8675171136856079, "num_tokens": 622970199.0, "step": 16408 }, { "epoch": 2.0873934613916805, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.874413013458252, "learning_rate": 5e-05, "loss": 0.7073, "mean_token_accuracy": 0.8789758682250977, "num_tokens": 623003669.0, "step": 16409 }, { "epoch": 2.087520671670271, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.576817512512207, "learning_rate": 5e-05, "loss": 0.8821, "mean_token_accuracy": 0.870954155921936, "num_tokens": 623043110.0, "step": 16410 }, { "epoch": 2.0876478819488615, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.625882148742676, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8714190125465393, "num_tokens": 623082474.0, "step": 16411 }, { "epoch": 2.087775092227452, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.987857818603516, "learning_rate": 5e-05, "loss": 0.9367, "mean_token_accuracy": 0.8714509010314941, "num_tokens": 623119650.0, "step": 16412 }, { "epoch": 2.0879023025060426, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.567159175872803, "learning_rate": 5e-05, "loss": 0.7333, "mean_token_accuracy": 0.8741140365600586, "num_tokens": 623150359.0, "step": 16413 }, { "epoch": 2.088029512784633, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.235687255859375, "learning_rate": 5e-05, "loss": 0.9347, "mean_token_accuracy": 0.8861109018325806, "num_tokens": 623189251.0, "step": 16414 }, { "epoch": 2.0881567230632236, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 6.997818946838379, "learning_rate": 5e-05, "loss": 0.7564, "mean_token_accuracy": 0.8988574743270874, "num_tokens": 623231763.0, "step": 16415 }, { "epoch": 2.088283933341814, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.932397842407227, "learning_rate": 5e-05, "loss": 0.9109, "mean_token_accuracy": 0.8924040794372559, "num_tokens": 623269322.0, "step": 16416 }, { "epoch": 2.0884111436204047, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.674798965454102, "learning_rate": 5e-05, "loss": 0.8725, "mean_token_accuracy": 0.8903175592422485, "num_tokens": 623303765.0, "step": 16417 }, { "epoch": 2.0885383538989952, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.073381423950195, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.8795593976974487, "num_tokens": 623355484.0, "step": 16418 }, { "epoch": 2.0886655641775858, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.110496520996094, "learning_rate": 5e-05, "loss": 0.9093, "mean_token_accuracy": 0.8819557428359985, "num_tokens": 623399441.0, "step": 16419 }, { "epoch": 2.0887927744561763, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.38707160949707, "learning_rate": 5e-05, "loss": 0.8269, "mean_token_accuracy": 0.863635778427124, "num_tokens": 623438690.0, "step": 16420 }, { "epoch": 2.0889199847347664, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.129250526428223, "learning_rate": 5e-05, "loss": 0.8594, "mean_token_accuracy": 0.8782763481140137, "num_tokens": 623476291.0, "step": 16421 }, { "epoch": 2.089047195013357, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.365071773529053, "learning_rate": 5e-05, "loss": 0.73, "mean_token_accuracy": 0.8955525755882263, "num_tokens": 623513830.0, "step": 16422 }, { "epoch": 2.0891744052919474, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.727123737335205, "learning_rate": 5e-05, "loss": 0.7917, "mean_token_accuracy": 0.8787722587585449, "num_tokens": 623554789.0, "step": 16423 }, { "epoch": 2.089301615570538, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.694427490234375, "learning_rate": 5e-05, "loss": 0.7734, "mean_token_accuracy": 0.8865909576416016, "num_tokens": 623593646.0, "step": 16424 }, { "epoch": 2.0894288258491285, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.28781270980835, "learning_rate": 5e-05, "loss": 0.6912, "mean_token_accuracy": 0.8919153213500977, "num_tokens": 623630646.0, "step": 16425 }, { "epoch": 2.089556036127719, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.986977577209473, "learning_rate": 5e-05, "loss": 0.8331, "mean_token_accuracy": 0.8647913932800293, "num_tokens": 623675488.0, "step": 16426 }, { "epoch": 2.0896832464063095, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.013785362243652, "learning_rate": 5e-05, "loss": 0.7184, "mean_token_accuracy": 0.8754793405532837, "num_tokens": 623719717.0, "step": 16427 }, { "epoch": 2.0898104566849, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.197168350219727, "learning_rate": 5e-05, "loss": 0.7719, "mean_token_accuracy": 0.8901684284210205, "num_tokens": 623753248.0, "step": 16428 }, { "epoch": 2.0899376669634906, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.812093257904053, "learning_rate": 5e-05, "loss": 0.7135, "mean_token_accuracy": 0.8726975321769714, "num_tokens": 623791562.0, "step": 16429 }, { "epoch": 2.090064877242081, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.480753898620605, "learning_rate": 5e-05, "loss": 0.8493, "mean_token_accuracy": 0.8734964728355408, "num_tokens": 623834187.0, "step": 16430 }, { "epoch": 2.0901920875206716, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.579804420471191, "learning_rate": 5e-05, "loss": 0.7275, "mean_token_accuracy": 0.8716278672218323, "num_tokens": 623869664.0, "step": 16431 }, { "epoch": 2.090319297799262, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.844098091125488, "learning_rate": 5e-05, "loss": 0.8533, "mean_token_accuracy": 0.8833059072494507, "num_tokens": 623907461.0, "step": 16432 }, { "epoch": 2.0904465080778527, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.445603847503662, "learning_rate": 5e-05, "loss": 0.6841, "mean_token_accuracy": 0.8802618980407715, "num_tokens": 623947679.0, "step": 16433 }, { "epoch": 2.0905737183564432, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.119406700134277, "learning_rate": 5e-05, "loss": 0.8871, "mean_token_accuracy": 0.8907828330993652, "num_tokens": 623992659.0, "step": 16434 }, { "epoch": 2.0907009286350338, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.709387302398682, "learning_rate": 5e-05, "loss": 0.6989, "mean_token_accuracy": 0.8950542211532593, "num_tokens": 624028305.0, "step": 16435 }, { "epoch": 2.0908281389136243, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.960031509399414, "learning_rate": 5e-05, "loss": 0.9124, "mean_token_accuracy": 0.8849902749061584, "num_tokens": 624062585.0, "step": 16436 }, { "epoch": 2.090955349192215, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.357128143310547, "learning_rate": 5e-05, "loss": 0.8143, "mean_token_accuracy": 0.8872376084327698, "num_tokens": 624100525.0, "step": 16437 }, { "epoch": 2.0910825594708053, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.156354904174805, "learning_rate": 5e-05, "loss": 0.8155, "mean_token_accuracy": 0.8959762454032898, "num_tokens": 624135982.0, "step": 16438 }, { "epoch": 2.091209769749396, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.9802937507629395, "learning_rate": 5e-05, "loss": 0.8218, "mean_token_accuracy": 0.8951138257980347, "num_tokens": 624172867.0, "step": 16439 }, { "epoch": 2.0913369800279864, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.304200649261475, "learning_rate": 5e-05, "loss": 0.7051, "mean_token_accuracy": 0.8986110687255859, "num_tokens": 624208588.0, "step": 16440 }, { "epoch": 2.091464190306577, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.311262130737305, "learning_rate": 5e-05, "loss": 0.8516, "mean_token_accuracy": 0.8810528516769409, "num_tokens": 624241060.0, "step": 16441 }, { "epoch": 2.0915914005851675, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.966470241546631, "learning_rate": 5e-05, "loss": 0.7278, "mean_token_accuracy": 0.877549409866333, "num_tokens": 624284508.0, "step": 16442 }, { "epoch": 2.091718610863758, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.309136390686035, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.8885310888290405, "num_tokens": 624326787.0, "step": 16443 }, { "epoch": 2.0918458211423485, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.032852649688721, "learning_rate": 5e-05, "loss": 0.6867, "mean_token_accuracy": 0.8937209844589233, "num_tokens": 624363375.0, "step": 16444 }, { "epoch": 2.0919730314209386, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.167126655578613, "learning_rate": 5e-05, "loss": 0.8183, "mean_token_accuracy": 0.878856897354126, "num_tokens": 624405508.0, "step": 16445 }, { "epoch": 2.092100241699529, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.0904412269592285, "learning_rate": 5e-05, "loss": 0.6782, "mean_token_accuracy": 0.8937713503837585, "num_tokens": 624438505.0, "step": 16446 }, { "epoch": 2.0922274519781197, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.164427757263184, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.8847743272781372, "num_tokens": 624475463.0, "step": 16447 }, { "epoch": 2.09235466225671, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.012325286865234, "learning_rate": 5e-05, "loss": 0.6919, "mean_token_accuracy": 0.8887256383895874, "num_tokens": 624511577.0, "step": 16448 }, { "epoch": 2.0924818725353007, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.303817749023438, "learning_rate": 5e-05, "loss": 0.8125, "mean_token_accuracy": 0.8865050673484802, "num_tokens": 624552213.0, "step": 16449 }, { "epoch": 2.0926090828138912, "ewc_loss": 0.375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.87216854095459, "learning_rate": 5e-05, "loss": 0.681, "mean_token_accuracy": 0.8899972438812256, "num_tokens": 624589026.0, "step": 16450 }, { "epoch": 2.0927362930924818, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.473136901855469, "learning_rate": 5e-05, "loss": 0.8371, "mean_token_accuracy": 0.8828738331794739, "num_tokens": 624623245.0, "step": 16451 }, { "epoch": 2.0928635033710723, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.785429000854492, "learning_rate": 5e-05, "loss": 0.7177, "mean_token_accuracy": 0.8782004117965698, "num_tokens": 624659713.0, "step": 16452 }, { "epoch": 2.092990713649663, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.595355987548828, "learning_rate": 5e-05, "loss": 0.8383, "mean_token_accuracy": 0.887252688407898, "num_tokens": 624695539.0, "step": 16453 }, { "epoch": 2.0931179239282534, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.834653854370117, "learning_rate": 5e-05, "loss": 0.7209, "mean_token_accuracy": 0.8799034357070923, "num_tokens": 624735311.0, "step": 16454 }, { "epoch": 2.093245134206844, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.536338806152344, "learning_rate": 5e-05, "loss": 0.8496, "mean_token_accuracy": 0.8885568976402283, "num_tokens": 624771941.0, "step": 16455 }, { "epoch": 2.0933723444854344, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.04278039932251, "learning_rate": 5e-05, "loss": 0.7106, "mean_token_accuracy": 0.8971588611602783, "num_tokens": 624809963.0, "step": 16456 }, { "epoch": 2.093499554764025, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.288602828979492, "learning_rate": 5e-05, "loss": 0.8441, "mean_token_accuracy": 0.8831197619438171, "num_tokens": 624849839.0, "step": 16457 }, { "epoch": 2.0936267650426155, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.3193182945251465, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8822438716888428, "num_tokens": 624883920.0, "step": 16458 }, { "epoch": 2.093753975321206, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.91105318069458, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.892486572265625, "num_tokens": 624916024.0, "step": 16459 }, { "epoch": 2.0938811855997965, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.5201497077941895, "learning_rate": 5e-05, "loss": 0.7668, "mean_token_accuracy": 0.8869200348854065, "num_tokens": 624950978.0, "step": 16460 }, { "epoch": 2.094008395878387, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.6533732414245605, "learning_rate": 5e-05, "loss": 0.7206, "mean_token_accuracy": 0.8942044377326965, "num_tokens": 624992177.0, "step": 16461 }, { "epoch": 2.0941356061569776, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.550045013427734, "learning_rate": 5e-05, "loss": 0.7508, "mean_token_accuracy": 0.8822981715202332, "num_tokens": 625028211.0, "step": 16462 }, { "epoch": 2.094262816435568, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.617101192474365, "learning_rate": 5e-05, "loss": 0.6972, "mean_token_accuracy": 0.8985829949378967, "num_tokens": 625065895.0, "step": 16463 }, { "epoch": 2.0943900267141586, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.418712139129639, "learning_rate": 5e-05, "loss": 0.7046, "mean_token_accuracy": 0.8896306157112122, "num_tokens": 625100521.0, "step": 16464 }, { "epoch": 2.094517236992749, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.7163286209106445, "learning_rate": 5e-05, "loss": 0.7499, "mean_token_accuracy": 0.8829883337020874, "num_tokens": 625144810.0, "step": 16465 }, { "epoch": 2.0946444472713397, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.245401859283447, "learning_rate": 5e-05, "loss": 0.7104, "mean_token_accuracy": 0.8810487985610962, "num_tokens": 625178001.0, "step": 16466 }, { "epoch": 2.09477165754993, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.870197296142578, "learning_rate": 5e-05, "loss": 0.737, "mean_token_accuracy": 0.8839058876037598, "num_tokens": 625214857.0, "step": 16467 }, { "epoch": 2.0948988678285207, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.034508228302002, "learning_rate": 5e-05, "loss": 0.6902, "mean_token_accuracy": 0.8880081176757812, "num_tokens": 625253396.0, "step": 16468 }, { "epoch": 2.095026078107111, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.093445777893066, "learning_rate": 5e-05, "loss": 0.7924, "mean_token_accuracy": 0.8700662851333618, "num_tokens": 625291922.0, "step": 16469 }, { "epoch": 2.0951532883857014, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.814382076263428, "learning_rate": 5e-05, "loss": 0.6859, "mean_token_accuracy": 0.8841722011566162, "num_tokens": 625326548.0, "step": 16470 }, { "epoch": 2.095280498664292, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.351515769958496, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.8893844485282898, "num_tokens": 625361354.0, "step": 16471 }, { "epoch": 2.0954077089428824, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.634045600891113, "learning_rate": 5e-05, "loss": 0.697, "mean_token_accuracy": 0.8766337633132935, "num_tokens": 625400885.0, "step": 16472 }, { "epoch": 2.095534919221473, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.575724601745605, "learning_rate": 5e-05, "loss": 0.8107, "mean_token_accuracy": 0.8859791159629822, "num_tokens": 625438975.0, "step": 16473 }, { "epoch": 2.0956621295000635, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.569567680358887, "learning_rate": 5e-05, "loss": 0.7239, "mean_token_accuracy": 0.8680641651153564, "num_tokens": 625475384.0, "step": 16474 }, { "epoch": 2.095789339778654, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.66884708404541, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.8972140550613403, "num_tokens": 625508422.0, "step": 16475 }, { "epoch": 2.0959165500572445, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.783424377441406, "learning_rate": 5e-05, "loss": 0.6657, "mean_token_accuracy": 0.8927783370018005, "num_tokens": 625546434.0, "step": 16476 }, { "epoch": 2.096043760335835, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.355646133422852, "learning_rate": 5e-05, "loss": 0.815, "mean_token_accuracy": 0.8861197233200073, "num_tokens": 625584005.0, "step": 16477 }, { "epoch": 2.0961709706144256, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.3294782638549805, "learning_rate": 5e-05, "loss": 0.72, "mean_token_accuracy": 0.8963935375213623, "num_tokens": 625616713.0, "step": 16478 }, { "epoch": 2.096298180893016, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.566373825073242, "learning_rate": 5e-05, "loss": 0.7431, "mean_token_accuracy": 0.8875089883804321, "num_tokens": 625657738.0, "step": 16479 }, { "epoch": 2.0964253911716066, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.043975830078125, "learning_rate": 5e-05, "loss": 0.7873, "mean_token_accuracy": 0.8816525340080261, "num_tokens": 625693485.0, "step": 16480 }, { "epoch": 2.096552601450197, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.69929313659668, "learning_rate": 5e-05, "loss": 0.6412, "mean_token_accuracy": 0.8916381597518921, "num_tokens": 625727219.0, "step": 16481 }, { "epoch": 2.0966798117287877, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.859434127807617, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.895453691482544, "num_tokens": 625762858.0, "step": 16482 }, { "epoch": 2.0968070220073782, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.00336217880249, "learning_rate": 5e-05, "loss": 0.6506, "mean_token_accuracy": 0.8821179866790771, "num_tokens": 625799752.0, "step": 16483 }, { "epoch": 2.0969342322859688, "ewc_loss": 0.625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.787836074829102, "learning_rate": 5e-05, "loss": 0.9159, "mean_token_accuracy": 0.8937115669250488, "num_tokens": 625835453.0, "step": 16484 }, { "epoch": 2.0970614425645593, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.026017665863037, "learning_rate": 5e-05, "loss": 0.7281, "mean_token_accuracy": 0.8710398077964783, "num_tokens": 625875181.0, "step": 16485 }, { "epoch": 2.09718865284315, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.270560264587402, "learning_rate": 5e-05, "loss": 1.0274, "mean_token_accuracy": 0.8942698240280151, "num_tokens": 625918851.0, "step": 16486 }, { "epoch": 2.0973158631217403, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 7.513528823852539, "learning_rate": 5e-05, "loss": 0.9245, "mean_token_accuracy": 0.885250449180603, "num_tokens": 625954859.0, "step": 16487 }, { "epoch": 2.097443073400331, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.235950469970703, "learning_rate": 5e-05, "loss": 0.9447, "mean_token_accuracy": 0.8914982080459595, "num_tokens": 625986042.0, "step": 16488 }, { "epoch": 2.0975702836789214, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 8.827814102172852, "learning_rate": 5e-05, "loss": 1.0944, "mean_token_accuracy": 0.8632707595825195, "num_tokens": 626023604.0, "step": 16489 }, { "epoch": 2.097697493957512, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.254871368408203, "learning_rate": 5e-05, "loss": 0.7228, "mean_token_accuracy": 0.9008278846740723, "num_tokens": 626059193.0, "step": 16490 }, { "epoch": 2.0978247042361025, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.006510734558105, "learning_rate": 5e-05, "loss": 1.04, "mean_token_accuracy": 0.8774558305740356, "num_tokens": 626099994.0, "step": 16491 }, { "epoch": 2.097951914514693, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 6.802150249481201, "learning_rate": 5e-05, "loss": 0.8024, "mean_token_accuracy": 0.8851790428161621, "num_tokens": 626133105.0, "step": 16492 }, { "epoch": 2.0980791247932835, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.447905540466309, "learning_rate": 5e-05, "loss": 1.0459, "mean_token_accuracy": 0.8924144506454468, "num_tokens": 626171837.0, "step": 16493 }, { "epoch": 2.0982063350718736, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 8.07512378692627, "learning_rate": 5e-05, "loss": 1.0222, "mean_token_accuracy": 0.8778027296066284, "num_tokens": 626214057.0, "step": 16494 }, { "epoch": 2.098333545350464, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.375670433044434, "learning_rate": 5e-05, "loss": 0.9388, "mean_token_accuracy": 0.888522744178772, "num_tokens": 626249046.0, "step": 16495 }, { "epoch": 2.0984607556290547, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 8.613306045532227, "learning_rate": 5e-05, "loss": 1.0431, "mean_token_accuracy": 0.8734326362609863, "num_tokens": 626283991.0, "step": 16496 }, { "epoch": 2.098587965907645, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.224555015563965, "learning_rate": 5e-05, "loss": 0.8684, "mean_token_accuracy": 0.8819196224212646, "num_tokens": 626321919.0, "step": 16497 }, { "epoch": 2.0987151761862357, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.575861930847168, "learning_rate": 5e-05, "loss": 0.9049, "mean_token_accuracy": 0.8940966129302979, "num_tokens": 626360281.0, "step": 16498 }, { "epoch": 2.0988423864648262, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.869555473327637, "learning_rate": 5e-05, "loss": 0.8627, "mean_token_accuracy": 0.8860714435577393, "num_tokens": 626404314.0, "step": 16499 }, { "epoch": 2.0989695967434168, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.0474853515625, "learning_rate": 5e-05, "loss": 0.8788, "mean_token_accuracy": 0.8682801723480225, "num_tokens": 626452743.0, "step": 16500 }, { "epoch": 2.0990968070220073, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.016386032104492, "learning_rate": 5e-05, "loss": 0.8154, "mean_token_accuracy": 0.8993954658508301, "num_tokens": 626494310.0, "step": 16501 }, { "epoch": 2.099224017300598, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.379178047180176, "learning_rate": 5e-05, "loss": 0.7398, "mean_token_accuracy": 0.8866883516311646, "num_tokens": 626534995.0, "step": 16502 }, { "epoch": 2.0993512275791884, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.371366500854492, "learning_rate": 5e-05, "loss": 0.8187, "mean_token_accuracy": 0.894709587097168, "num_tokens": 626575084.0, "step": 16503 }, { "epoch": 2.099478437857779, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.908700942993164, "learning_rate": 5e-05, "loss": 0.6916, "mean_token_accuracy": 0.8910422921180725, "num_tokens": 626608922.0, "step": 16504 }, { "epoch": 2.0996056481363694, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.761857986450195, "learning_rate": 5e-05, "loss": 0.8402, "mean_token_accuracy": 0.8945427536964417, "num_tokens": 626645480.0, "step": 16505 }, { "epoch": 2.09973285841496, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.673314571380615, "learning_rate": 5e-05, "loss": 0.7334, "mean_token_accuracy": 0.8790688514709473, "num_tokens": 626687641.0, "step": 16506 }, { "epoch": 2.0998600686935505, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.053330421447754, "learning_rate": 5e-05, "loss": 0.8687, "mean_token_accuracy": 0.8958496451377869, "num_tokens": 626720624.0, "step": 16507 }, { "epoch": 2.099987278972141, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.723966121673584, "learning_rate": 5e-05, "loss": 0.7969, "mean_token_accuracy": 0.8612897396087646, "num_tokens": 626760174.0, "step": 16508 }, { "epoch": 2.1001144892507315, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.109491348266602, "learning_rate": 5e-05, "loss": 0.8574, "mean_token_accuracy": 0.8971721529960632, "num_tokens": 626792970.0, "step": 16509 }, { "epoch": 2.100241699529322, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.018553733825684, "learning_rate": 5e-05, "loss": 0.8153, "mean_token_accuracy": 0.8830606341362, "num_tokens": 626826793.0, "step": 16510 }, { "epoch": 2.1003689098079126, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.821134567260742, "learning_rate": 5e-05, "loss": 0.893, "mean_token_accuracy": 0.888702392578125, "num_tokens": 626862938.0, "step": 16511 }, { "epoch": 2.100496120086503, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.322892665863037, "learning_rate": 5e-05, "loss": 0.8288, "mean_token_accuracy": 0.8853394985198975, "num_tokens": 626900738.0, "step": 16512 }, { "epoch": 2.1006233303650936, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.452710151672363, "learning_rate": 5e-05, "loss": 0.8497, "mean_token_accuracy": 0.8920202851295471, "num_tokens": 626932670.0, "step": 16513 }, { "epoch": 2.100750540643684, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.405406951904297, "learning_rate": 5e-05, "loss": 0.831, "mean_token_accuracy": 0.8825238943099976, "num_tokens": 626971683.0, "step": 16514 }, { "epoch": 2.1008777509222747, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.239686012268066, "learning_rate": 5e-05, "loss": 0.8621, "mean_token_accuracy": 0.8795457482337952, "num_tokens": 627009955.0, "step": 16515 }, { "epoch": 2.101004961200865, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.332340717315674, "learning_rate": 5e-05, "loss": 0.8172, "mean_token_accuracy": 0.8732022643089294, "num_tokens": 627052469.0, "step": 16516 }, { "epoch": 2.1011321714794557, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.18386459350586, "learning_rate": 5e-05, "loss": 0.8457, "mean_token_accuracy": 0.8842234015464783, "num_tokens": 627088314.0, "step": 16517 }, { "epoch": 2.1012593817580463, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.2107062339782715, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.884132981300354, "num_tokens": 627132030.0, "step": 16518 }, { "epoch": 2.1013865920366364, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.22468376159668, "learning_rate": 5e-05, "loss": 0.8557, "mean_token_accuracy": 0.8808620572090149, "num_tokens": 627171135.0, "step": 16519 }, { "epoch": 2.101513802315227, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.110511302947998, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.8879439830780029, "num_tokens": 627206303.0, "step": 16520 }, { "epoch": 2.1016410125938174, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.259659767150879, "learning_rate": 5e-05, "loss": 0.8391, "mean_token_accuracy": 0.8826552629470825, "num_tokens": 627242859.0, "step": 16521 }, { "epoch": 2.101768222872408, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.0958380699157715, "learning_rate": 5e-05, "loss": 0.7397, "mean_token_accuracy": 0.879927396774292, "num_tokens": 627284659.0, "step": 16522 }, { "epoch": 2.1018954331509985, "ewc_loss": 0.5, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.220812797546387, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8913270235061646, "num_tokens": 627320012.0, "step": 16523 }, { "epoch": 2.102022643429589, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.1669535636901855, "learning_rate": 5e-05, "loss": 0.7147, "mean_token_accuracy": 0.8892126083374023, "num_tokens": 627358117.0, "step": 16524 }, { "epoch": 2.1021498537081795, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.08548355102539, "learning_rate": 5e-05, "loss": 0.8375, "mean_token_accuracy": 0.8753286600112915, "num_tokens": 627393331.0, "step": 16525 }, { "epoch": 2.10227706398677, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.268253326416016, "learning_rate": 5e-05, "loss": 0.7666, "mean_token_accuracy": 0.8754024505615234, "num_tokens": 627433905.0, "step": 16526 }, { "epoch": 2.1024042742653606, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.900634765625, "learning_rate": 5e-05, "loss": 0.7462, "mean_token_accuracy": 0.8975670337677002, "num_tokens": 627469673.0, "step": 16527 }, { "epoch": 2.102531484543951, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.383909225463867, "learning_rate": 5e-05, "loss": 0.7526, "mean_token_accuracy": 0.8808872699737549, "num_tokens": 627502917.0, "step": 16528 }, { "epoch": 2.1026586948225416, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.698291301727295, "learning_rate": 5e-05, "loss": 0.7205, "mean_token_accuracy": 0.8967866897583008, "num_tokens": 627536632.0, "step": 16529 }, { "epoch": 2.102785905101132, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.48960542678833, "learning_rate": 5e-05, "loss": 0.711, "mean_token_accuracy": 0.8921696543693542, "num_tokens": 627580820.0, "step": 16530 }, { "epoch": 2.1029131153797227, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.549488544464111, "learning_rate": 5e-05, "loss": 0.6872, "mean_token_accuracy": 0.9013108611106873, "num_tokens": 627616895.0, "step": 16531 }, { "epoch": 2.1030403256583132, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.547982692718506, "learning_rate": 5e-05, "loss": 0.7232, "mean_token_accuracy": 0.8817169070243835, "num_tokens": 627655648.0, "step": 16532 }, { "epoch": 2.1031675359369038, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.436039924621582, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.8833039999008179, "num_tokens": 627692259.0, "step": 16533 }, { "epoch": 2.1032947462154943, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.593052387237549, "learning_rate": 5e-05, "loss": 0.7042, "mean_token_accuracy": 0.8888328075408936, "num_tokens": 627731601.0, "step": 16534 }, { "epoch": 2.103421956494085, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.363067150115967, "learning_rate": 5e-05, "loss": 0.6981, "mean_token_accuracy": 0.8850957155227661, "num_tokens": 627768381.0, "step": 16535 }, { "epoch": 2.1035491667726753, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.641991138458252, "learning_rate": 5e-05, "loss": 0.7003, "mean_token_accuracy": 0.8897621631622314, "num_tokens": 627808327.0, "step": 16536 }, { "epoch": 2.103676377051266, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.282293796539307, "learning_rate": 5e-05, "loss": 0.6742, "mean_token_accuracy": 0.8869437575340271, "num_tokens": 627838779.0, "step": 16537 }, { "epoch": 2.1038035873298564, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.700263023376465, "learning_rate": 5e-05, "loss": 0.7076, "mean_token_accuracy": 0.8858729600906372, "num_tokens": 627875634.0, "step": 16538 }, { "epoch": 2.103930797608447, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.165077209472656, "learning_rate": 5e-05, "loss": 0.6827, "mean_token_accuracy": 0.8827872276306152, "num_tokens": 627918350.0, "step": 16539 }, { "epoch": 2.1040580078870375, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.852569103240967, "learning_rate": 5e-05, "loss": 0.7221, "mean_token_accuracy": 0.880678117275238, "num_tokens": 627955649.0, "step": 16540 }, { "epoch": 2.104185218165628, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.979767322540283, "learning_rate": 5e-05, "loss": 0.6695, "mean_token_accuracy": 0.8860881328582764, "num_tokens": 627989374.0, "step": 16541 }, { "epoch": 2.1043124284442185, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.139362335205078, "learning_rate": 5e-05, "loss": 0.7807, "mean_token_accuracy": 0.8750401735305786, "num_tokens": 628026338.0, "step": 16542 }, { "epoch": 2.1044396387228086, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.5998969078063965, "learning_rate": 5e-05, "loss": 0.6335, "mean_token_accuracy": 0.8915033340454102, "num_tokens": 628062550.0, "step": 16543 }, { "epoch": 2.104566849001399, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.685107231140137, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.8799071907997131, "num_tokens": 628104868.0, "step": 16544 }, { "epoch": 2.1046940592799896, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.081141948699951, "learning_rate": 5e-05, "loss": 0.6467, "mean_token_accuracy": 0.8803214430809021, "num_tokens": 628142009.0, "step": 16545 }, { "epoch": 2.10482126955858, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.530370712280273, "learning_rate": 5e-05, "loss": 0.9068, "mean_token_accuracy": 0.8815294504165649, "num_tokens": 628180992.0, "step": 16546 }, { "epoch": 2.1049484798371707, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 5.77340030670166, "learning_rate": 5e-05, "loss": 0.6529, "mean_token_accuracy": 0.8855113983154297, "num_tokens": 628215636.0, "step": 16547 }, { "epoch": 2.1050756901157612, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.347112655639648, "learning_rate": 5e-05, "loss": 0.9948, "mean_token_accuracy": 0.8902244567871094, "num_tokens": 628256993.0, "step": 16548 }, { "epoch": 2.1052029003943518, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 6.819523334503174, "learning_rate": 5e-05, "loss": 0.8561, "mean_token_accuracy": 0.8813269734382629, "num_tokens": 628293408.0, "step": 16549 }, { "epoch": 2.1053301106729423, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.967432975769043, "learning_rate": 5e-05, "loss": 1.0185, "mean_token_accuracy": 0.8858238458633423, "num_tokens": 628329353.0, "step": 16550 }, { "epoch": 2.105457320951533, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 8.58515739440918, "learning_rate": 5e-05, "loss": 1.0841, "mean_token_accuracy": 0.8768515586853027, "num_tokens": 628364312.0, "step": 16551 }, { "epoch": 2.1055845312301233, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.779597759246826, "learning_rate": 5e-05, "loss": 0.7738, "mean_token_accuracy": 0.8876386880874634, "num_tokens": 628401354.0, "step": 16552 }, { "epoch": 2.105711741508714, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000732421875, "grad_norm": 10.190571784973145, "learning_rate": 5e-05, "loss": 1.1031, "mean_token_accuracy": 0.8923920392990112, "num_tokens": 628432840.0, "step": 16553 }, { "epoch": 2.1058389517873044, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.4391703605651855, "learning_rate": 5e-05, "loss": 0.7416, "mean_token_accuracy": 0.8813870549201965, "num_tokens": 628476300.0, "step": 16554 }, { "epoch": 2.105966162065895, "ewc_loss": 1.0234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000957489013671875, "grad_norm": 12.372917175292969, "learning_rate": 5e-05, "loss": 1.3493, "mean_token_accuracy": 0.8866159915924072, "num_tokens": 628514248.0, "step": 16555 }, { "epoch": 2.1060933723444855, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 9.364770889282227, "learning_rate": 5e-05, "loss": 1.2114, "mean_token_accuracy": 0.8878722190856934, "num_tokens": 628557161.0, "step": 16556 }, { "epoch": 2.106220582623076, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 11.593793869018555, "learning_rate": 5e-05, "loss": 1.0806, "mean_token_accuracy": 0.8888351917266846, "num_tokens": 628591073.0, "step": 16557 }, { "epoch": 2.1063477929016665, "ewc_loss": 0.92578125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 9.739367485046387, "learning_rate": 5e-05, "loss": 1.234, "mean_token_accuracy": 0.8886942863464355, "num_tokens": 628627215.0, "step": 16558 }, { "epoch": 2.106475003180257, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 11.1738862991333, "learning_rate": 5e-05, "loss": 1.1055, "mean_token_accuracy": 0.8811586499214172, "num_tokens": 628669858.0, "step": 16559 }, { "epoch": 2.1066022134588476, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.79783821105957, "learning_rate": 5e-05, "loss": 0.9427, "mean_token_accuracy": 0.8989889621734619, "num_tokens": 628704504.0, "step": 16560 }, { "epoch": 2.106729423737438, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 11.11903190612793, "learning_rate": 5e-05, "loss": 1.1722, "mean_token_accuracy": 0.8698931932449341, "num_tokens": 628743444.0, "step": 16561 }, { "epoch": 2.1068566340160286, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.586761474609375, "learning_rate": 5e-05, "loss": 0.9765, "mean_token_accuracy": 0.8732693195343018, "num_tokens": 628782256.0, "step": 16562 }, { "epoch": 2.106983844294619, "ewc_loss": 0.828125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 11.319293022155762, "learning_rate": 5e-05, "loss": 1.1474, "mean_token_accuracy": 0.8875839710235596, "num_tokens": 628822150.0, "step": 16563 }, { "epoch": 2.1071110545732097, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 9.211638450622559, "learning_rate": 5e-05, "loss": 1.003, "mean_token_accuracy": 0.8951109647750854, "num_tokens": 628862445.0, "step": 16564 }, { "epoch": 2.1072382648518, "ewc_loss": 0.765625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.931245803833008, "learning_rate": 5e-05, "loss": 1.0808, "mean_token_accuracy": 0.8862767815589905, "num_tokens": 628897166.0, "step": 16565 }, { "epoch": 2.1073654751303907, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 9.40461540222168, "learning_rate": 5e-05, "loss": 1.0045, "mean_token_accuracy": 0.8915361762046814, "num_tokens": 628933443.0, "step": 16566 }, { "epoch": 2.107492685408981, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.302828788757324, "learning_rate": 5e-05, "loss": 0.9912, "mean_token_accuracy": 0.8918235301971436, "num_tokens": 628969614.0, "step": 16567 }, { "epoch": 2.1076198956875714, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.026557922363281, "learning_rate": 5e-05, "loss": 0.9472, "mean_token_accuracy": 0.8869767189025879, "num_tokens": 629010386.0, "step": 16568 }, { "epoch": 2.107747105966162, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.897468566894531, "learning_rate": 5e-05, "loss": 1.0139, "mean_token_accuracy": 0.8761777281761169, "num_tokens": 629041388.0, "step": 16569 }, { "epoch": 2.1078743162447524, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.524638175964355, "learning_rate": 5e-05, "loss": 0.8584, "mean_token_accuracy": 0.8961902260780334, "num_tokens": 629077306.0, "step": 16570 }, { "epoch": 2.108001526523343, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.841900825500488, "learning_rate": 5e-05, "loss": 0.9517, "mean_token_accuracy": 0.8948855400085449, "num_tokens": 629110636.0, "step": 16571 }, { "epoch": 2.1081287368019335, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.167149543762207, "learning_rate": 5e-05, "loss": 0.8027, "mean_token_accuracy": 0.900687575340271, "num_tokens": 629145834.0, "step": 16572 }, { "epoch": 2.108255947080524, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.042827606201172, "learning_rate": 5e-05, "loss": 1.0084, "mean_token_accuracy": 0.8860980868339539, "num_tokens": 629187148.0, "step": 16573 }, { "epoch": 2.1083831573591145, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.19062328338623, "learning_rate": 5e-05, "loss": 0.8276, "mean_token_accuracy": 0.8937357068061829, "num_tokens": 629224391.0, "step": 16574 }, { "epoch": 2.108510367637705, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.209545135498047, "learning_rate": 5e-05, "loss": 1.033, "mean_token_accuracy": 0.8877037763595581, "num_tokens": 629261074.0, "step": 16575 }, { "epoch": 2.1086375779162956, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.561726570129395, "learning_rate": 5e-05, "loss": 0.8976, "mean_token_accuracy": 0.8873746395111084, "num_tokens": 629296568.0, "step": 16576 }, { "epoch": 2.108764788194886, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.84438705444336, "learning_rate": 5e-05, "loss": 0.9752, "mean_token_accuracy": 0.8883194923400879, "num_tokens": 629342784.0, "step": 16577 }, { "epoch": 2.1088919984734766, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.738113403320312, "learning_rate": 5e-05, "loss": 0.8849, "mean_token_accuracy": 0.8909728527069092, "num_tokens": 629377705.0, "step": 16578 }, { "epoch": 2.109019208752067, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.175702095031738, "learning_rate": 5e-05, "loss": 0.9097, "mean_token_accuracy": 0.8897570967674255, "num_tokens": 629414823.0, "step": 16579 }, { "epoch": 2.1091464190306577, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.4269437789917, "learning_rate": 5e-05, "loss": 0.867, "mean_token_accuracy": 0.876546323299408, "num_tokens": 629452404.0, "step": 16580 }, { "epoch": 2.109273629309248, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.832575798034668, "learning_rate": 5e-05, "loss": 0.8728, "mean_token_accuracy": 0.8898056745529175, "num_tokens": 629486147.0, "step": 16581 }, { "epoch": 2.1094008395878387, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.863960266113281, "learning_rate": 5e-05, "loss": 0.7681, "mean_token_accuracy": 0.8912501335144043, "num_tokens": 629523581.0, "step": 16582 }, { "epoch": 2.1095280498664293, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.990565299987793, "learning_rate": 5e-05, "loss": 0.8909, "mean_token_accuracy": 0.8864292502403259, "num_tokens": 629562125.0, "step": 16583 }, { "epoch": 2.10965526014502, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.481845378875732, "learning_rate": 5e-05, "loss": 0.7259, "mean_token_accuracy": 0.8940610885620117, "num_tokens": 629601786.0, "step": 16584 }, { "epoch": 2.1097824704236103, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.327462196350098, "learning_rate": 5e-05, "loss": 0.9158, "mean_token_accuracy": 0.893185019493103, "num_tokens": 629645022.0, "step": 16585 }, { "epoch": 2.109909680702201, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.552812099456787, "learning_rate": 5e-05, "loss": 0.7179, "mean_token_accuracy": 0.8985070586204529, "num_tokens": 629675604.0, "step": 16586 }, { "epoch": 2.1100368909807914, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.416155815124512, "learning_rate": 5e-05, "loss": 0.9601, "mean_token_accuracy": 0.8815643191337585, "num_tokens": 629718508.0, "step": 16587 }, { "epoch": 2.110164101259382, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.000078201293945, "learning_rate": 5e-05, "loss": 0.865, "mean_token_accuracy": 0.8713443875312805, "num_tokens": 629758476.0, "step": 16588 }, { "epoch": 2.1102913115379724, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.911388397216797, "learning_rate": 5e-05, "loss": 0.9278, "mean_token_accuracy": 0.8785426020622253, "num_tokens": 629793692.0, "step": 16589 }, { "epoch": 2.110418521816563, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.217606544494629, "learning_rate": 5e-05, "loss": 0.8533, "mean_token_accuracy": 0.876014232635498, "num_tokens": 629829750.0, "step": 16590 }, { "epoch": 2.1105457320951535, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.201458930969238, "learning_rate": 5e-05, "loss": 0.8875, "mean_token_accuracy": 0.8671455979347229, "num_tokens": 629865832.0, "step": 16591 }, { "epoch": 2.1106729423737436, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.992544174194336, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8750617504119873, "num_tokens": 629906450.0, "step": 16592 }, { "epoch": 2.110800152652334, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.9631805419921875, "learning_rate": 5e-05, "loss": 0.8404, "mean_token_accuracy": 0.8707122802734375, "num_tokens": 629939650.0, "step": 16593 }, { "epoch": 2.1109273629309246, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.642920970916748, "learning_rate": 5e-05, "loss": 0.7332, "mean_token_accuracy": 0.8865334987640381, "num_tokens": 629977692.0, "step": 16594 }, { "epoch": 2.111054573209515, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00066375732421875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.980521202087402, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.8875831365585327, "num_tokens": 630019111.0, "step": 16595 }, { "epoch": 2.1111817834881057, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.4093852043151855, "learning_rate": 5e-05, "loss": 0.6788, "mean_token_accuracy": 0.892795979976654, "num_tokens": 630054924.0, "step": 16596 }, { "epoch": 2.1113089937666962, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.036223411560059, "learning_rate": 5e-05, "loss": 0.7922, "mean_token_accuracy": 0.8785852193832397, "num_tokens": 630100344.0, "step": 16597 }, { "epoch": 2.1114362040452868, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.27198600769043, "learning_rate": 5e-05, "loss": 0.6866, "mean_token_accuracy": 0.8854458332061768, "num_tokens": 630137296.0, "step": 16598 }, { "epoch": 2.1115634143238773, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.06679630279541, "learning_rate": 5e-05, "loss": 0.7624, "mean_token_accuracy": 0.8868069648742676, "num_tokens": 630177477.0, "step": 16599 }, { "epoch": 2.111690624602468, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.156949996948242, "learning_rate": 5e-05, "loss": 0.6233, "mean_token_accuracy": 0.9042136073112488, "num_tokens": 630205830.0, "step": 16600 }, { "epoch": 2.1118178348810583, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.200736999511719, "learning_rate": 5e-05, "loss": 0.7684, "mean_token_accuracy": 0.8861703276634216, "num_tokens": 630247114.0, "step": 16601 }, { "epoch": 2.111945045159649, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.026288032531738, "learning_rate": 5e-05, "loss": 0.6471, "mean_token_accuracy": 0.8972342014312744, "num_tokens": 630279848.0, "step": 16602 }, { "epoch": 2.1120722554382394, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.3695068359375, "learning_rate": 5e-05, "loss": 0.7724, "mean_token_accuracy": 0.8925309181213379, "num_tokens": 630320993.0, "step": 16603 }, { "epoch": 2.11219946571683, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.973458290100098, "learning_rate": 5e-05, "loss": 0.7318, "mean_token_accuracy": 0.865894079208374, "num_tokens": 630354703.0, "step": 16604 }, { "epoch": 2.1123266759954205, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.363606452941895, "learning_rate": 5e-05, "loss": 0.8142, "mean_token_accuracy": 0.8798747658729553, "num_tokens": 630388902.0, "step": 16605 }, { "epoch": 2.112453886274011, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.123432636260986, "learning_rate": 5e-05, "loss": 0.6797, "mean_token_accuracy": 0.8880813121795654, "num_tokens": 630431446.0, "step": 16606 }, { "epoch": 2.1125810965526015, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.145081520080566, "learning_rate": 5e-05, "loss": 0.7815, "mean_token_accuracy": 0.8784487247467041, "num_tokens": 630468334.0, "step": 16607 }, { "epoch": 2.112708306831192, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.398597240447998, "learning_rate": 5e-05, "loss": 0.6746, "mean_token_accuracy": 0.8905841708183289, "num_tokens": 630498607.0, "step": 16608 }, { "epoch": 2.1128355171097826, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.690786361694336, "learning_rate": 5e-05, "loss": 0.709, "mean_token_accuracy": 0.8874063491821289, "num_tokens": 630534900.0, "step": 16609 }, { "epoch": 2.112962727388373, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.7833099365234375, "learning_rate": 5e-05, "loss": 0.7494, "mean_token_accuracy": 0.8759581446647644, "num_tokens": 630577045.0, "step": 16610 }, { "epoch": 2.1130899376669636, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.145554065704346, "learning_rate": 5e-05, "loss": 0.6753, "mean_token_accuracy": 0.8873803615570068, "num_tokens": 630613477.0, "step": 16611 }, { "epoch": 2.113217147945554, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.253661155700684, "learning_rate": 5e-05, "loss": 0.7546, "mean_token_accuracy": 0.8852893114089966, "num_tokens": 630651868.0, "step": 16612 }, { "epoch": 2.1133443582241447, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.656236171722412, "learning_rate": 5e-05, "loss": 0.6543, "mean_token_accuracy": 0.8874357342720032, "num_tokens": 630691483.0, "step": 16613 }, { "epoch": 2.113471568502735, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.791741371154785, "learning_rate": 5e-05, "loss": 0.8018, "mean_token_accuracy": 0.8891257643699646, "num_tokens": 630730392.0, "step": 16614 }, { "epoch": 2.1135987787813257, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.291454792022705, "learning_rate": 5e-05, "loss": 0.6261, "mean_token_accuracy": 0.892454981803894, "num_tokens": 630764336.0, "step": 16615 }, { "epoch": 2.1137259890599163, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.31209659576416, "learning_rate": 5e-05, "loss": 0.9038, "mean_token_accuracy": 0.8783513903617859, "num_tokens": 630805912.0, "step": 16616 }, { "epoch": 2.1138531993385064, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.414990425109863, "learning_rate": 5e-05, "loss": 0.6955, "mean_token_accuracy": 0.8790676593780518, "num_tokens": 630844667.0, "step": 16617 }, { "epoch": 2.113980409617097, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.272754669189453, "learning_rate": 5e-05, "loss": 0.9229, "mean_token_accuracy": 0.8852424025535583, "num_tokens": 630888217.0, "step": 16618 }, { "epoch": 2.1141076198956874, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.34181547164917, "learning_rate": 5e-05, "loss": 0.7238, "mean_token_accuracy": 0.8939133882522583, "num_tokens": 630922415.0, "step": 16619 }, { "epoch": 2.114234830174278, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.173358917236328, "learning_rate": 5e-05, "loss": 0.8154, "mean_token_accuracy": 0.8877780437469482, "num_tokens": 630955469.0, "step": 16620 }, { "epoch": 2.1143620404528685, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.33917236328125, "learning_rate": 5e-05, "loss": 0.8186, "mean_token_accuracy": 0.8900130987167358, "num_tokens": 630994641.0, "step": 16621 }, { "epoch": 2.114489250731459, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.909211158752441, "learning_rate": 5e-05, "loss": 0.7157, "mean_token_accuracy": 0.8760181069374084, "num_tokens": 631034671.0, "step": 16622 }, { "epoch": 2.1146164610100495, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 9.013288497924805, "learning_rate": 5e-05, "loss": 0.8863, "mean_token_accuracy": 0.8812950253486633, "num_tokens": 631073518.0, "step": 16623 }, { "epoch": 2.11474367128864, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.619096755981445, "learning_rate": 5e-05, "loss": 0.7461, "mean_token_accuracy": 0.8691973686218262, "num_tokens": 631111631.0, "step": 16624 }, { "epoch": 2.1148708815672306, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.981846809387207, "learning_rate": 5e-05, "loss": 0.879, "mean_token_accuracy": 0.8838768005371094, "num_tokens": 631148081.0, "step": 16625 }, { "epoch": 2.114998091845821, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.20781135559082, "learning_rate": 5e-05, "loss": 0.763, "mean_token_accuracy": 0.880110502243042, "num_tokens": 631183176.0, "step": 16626 }, { "epoch": 2.1151253021244116, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.159460067749023, "learning_rate": 5e-05, "loss": 0.7892, "mean_token_accuracy": 0.8888380527496338, "num_tokens": 631213841.0, "step": 16627 }, { "epoch": 2.115252512403002, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.887920379638672, "learning_rate": 5e-05, "loss": 0.8244, "mean_token_accuracy": 0.8707249164581299, "num_tokens": 631249410.0, "step": 16628 }, { "epoch": 2.1153797226815927, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.381946563720703, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.8834430575370789, "num_tokens": 631292153.0, "step": 16629 }, { "epoch": 2.115506932960183, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.161863327026367, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8820475339889526, "num_tokens": 631330461.0, "step": 16630 }, { "epoch": 2.1156341432387737, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.247147560119629, "learning_rate": 5e-05, "loss": 0.733, "mean_token_accuracy": 0.8797183036804199, "num_tokens": 631367384.0, "step": 16631 }, { "epoch": 2.1157613535173643, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.939306735992432, "learning_rate": 5e-05, "loss": 0.7629, "mean_token_accuracy": 0.8776503801345825, "num_tokens": 631402558.0, "step": 16632 }, { "epoch": 2.115888563795955, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.537119388580322, "learning_rate": 5e-05, "loss": 0.7737, "mean_token_accuracy": 0.8653067946434021, "num_tokens": 631444040.0, "step": 16633 }, { "epoch": 2.1160157740745453, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.470582485198975, "learning_rate": 5e-05, "loss": 0.6926, "mean_token_accuracy": 0.8926428556442261, "num_tokens": 631486350.0, "step": 16634 }, { "epoch": 2.116142984353136, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.935245990753174, "learning_rate": 5e-05, "loss": 0.7127, "mean_token_accuracy": 0.8929846882820129, "num_tokens": 631519145.0, "step": 16635 }, { "epoch": 2.1162701946317264, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.035538673400879, "learning_rate": 5e-05, "loss": 0.6923, "mean_token_accuracy": 0.8775151968002319, "num_tokens": 631554637.0, "step": 16636 }, { "epoch": 2.116397404910317, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.28077220916748, "learning_rate": 5e-05, "loss": 0.737, "mean_token_accuracy": 0.8916220664978027, "num_tokens": 631593580.0, "step": 16637 }, { "epoch": 2.1165246151889074, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.758542537689209, "learning_rate": 5e-05, "loss": 0.6659, "mean_token_accuracy": 0.8841094374656677, "num_tokens": 631630410.0, "step": 16638 }, { "epoch": 2.116651825467498, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.595236778259277, "learning_rate": 5e-05, "loss": 0.7954, "mean_token_accuracy": 0.8891006708145142, "num_tokens": 631661894.0, "step": 16639 }, { "epoch": 2.116779035746088, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.542978763580322, "learning_rate": 5e-05, "loss": 0.6457, "mean_token_accuracy": 0.8874009251594543, "num_tokens": 631694514.0, "step": 16640 }, { "epoch": 2.1169062460246786, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.902863502502441, "learning_rate": 5e-05, "loss": 0.8601, "mean_token_accuracy": 0.8761720657348633, "num_tokens": 631737715.0, "step": 16641 }, { "epoch": 2.117033456303269, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.42901611328125, "learning_rate": 5e-05, "loss": 0.7039, "mean_token_accuracy": 0.8738380670547485, "num_tokens": 631773972.0, "step": 16642 }, { "epoch": 2.1171606665818596, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.195695877075195, "learning_rate": 5e-05, "loss": 0.901, "mean_token_accuracy": 0.8829094171524048, "num_tokens": 631812292.0, "step": 16643 }, { "epoch": 2.11728787686045, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.495167255401611, "learning_rate": 5e-05, "loss": 0.6565, "mean_token_accuracy": 0.8948062658309937, "num_tokens": 631849518.0, "step": 16644 }, { "epoch": 2.1174150871390407, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.369162559509277, "learning_rate": 5e-05, "loss": 0.9305, "mean_token_accuracy": 0.8851703405380249, "num_tokens": 631889821.0, "step": 16645 }, { "epoch": 2.1175422974176312, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 6.896076679229736, "learning_rate": 5e-05, "loss": 0.7236, "mean_token_accuracy": 0.8952945470809937, "num_tokens": 631925192.0, "step": 16646 }, { "epoch": 2.1176695076962218, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.197531700134277, "learning_rate": 5e-05, "loss": 0.9447, "mean_token_accuracy": 0.8827638626098633, "num_tokens": 631967876.0, "step": 16647 }, { "epoch": 2.1177967179748123, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.510272979736328, "learning_rate": 5e-05, "loss": 0.8328, "mean_token_accuracy": 0.8875998258590698, "num_tokens": 632006026.0, "step": 16648 }, { "epoch": 2.117923928253403, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.545235633850098, "learning_rate": 5e-05, "loss": 0.8732, "mean_token_accuracy": 0.8956514596939087, "num_tokens": 632045886.0, "step": 16649 }, { "epoch": 2.1180511385319933, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.912302494049072, "learning_rate": 5e-05, "loss": 0.8479, "mean_token_accuracy": 0.8900082111358643, "num_tokens": 632083583.0, "step": 16650 }, { "epoch": 2.118178348810584, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.922667026519775, "learning_rate": 5e-05, "loss": 0.8221, "mean_token_accuracy": 0.8883569240570068, "num_tokens": 632124427.0, "step": 16651 }, { "epoch": 2.1183055590891744, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.850633144378662, "learning_rate": 5e-05, "loss": 0.85, "mean_token_accuracy": 0.8804630041122437, "num_tokens": 632159787.0, "step": 16652 }, { "epoch": 2.118432769367765, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.872209072113037, "learning_rate": 5e-05, "loss": 0.7877, "mean_token_accuracy": 0.8909555077552795, "num_tokens": 632194737.0, "step": 16653 }, { "epoch": 2.1185599796463555, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.4182963371276855, "learning_rate": 5e-05, "loss": 0.7404, "mean_token_accuracy": 0.8899046182632446, "num_tokens": 632230737.0, "step": 16654 }, { "epoch": 2.118687189924946, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.199033737182617, "learning_rate": 5e-05, "loss": 0.8221, "mean_token_accuracy": 0.8812881708145142, "num_tokens": 632271374.0, "step": 16655 }, { "epoch": 2.1188144002035365, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.060544967651367, "learning_rate": 5e-05, "loss": 0.6816, "mean_token_accuracy": 0.8939447402954102, "num_tokens": 632305326.0, "step": 16656 }, { "epoch": 2.118941610482127, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.403203010559082, "learning_rate": 5e-05, "loss": 0.8477, "mean_token_accuracy": 0.8791795969009399, "num_tokens": 632338807.0, "step": 16657 }, { "epoch": 2.1190688207607176, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.012977123260498, "learning_rate": 5e-05, "loss": 0.7451, "mean_token_accuracy": 0.8722129464149475, "num_tokens": 632373933.0, "step": 16658 }, { "epoch": 2.119196031039308, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.31340217590332, "learning_rate": 5e-05, "loss": 0.8187, "mean_token_accuracy": 0.8831766247749329, "num_tokens": 632414149.0, "step": 16659 }, { "epoch": 2.1193232413178986, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.164004802703857, "learning_rate": 5e-05, "loss": 0.6836, "mean_token_accuracy": 0.8937658071517944, "num_tokens": 632453622.0, "step": 16660 }, { "epoch": 2.119450451596489, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.139152526855469, "learning_rate": 5e-05, "loss": 0.8391, "mean_token_accuracy": 0.8726909756660461, "num_tokens": 632488398.0, "step": 16661 }, { "epoch": 2.1195776618750797, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.292619705200195, "learning_rate": 5e-05, "loss": 0.7378, "mean_token_accuracy": 0.8795337677001953, "num_tokens": 632524013.0, "step": 16662 }, { "epoch": 2.11970487215367, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.013106346130371, "learning_rate": 5e-05, "loss": 0.8085, "mean_token_accuracy": 0.8788048028945923, "num_tokens": 632562247.0, "step": 16663 }, { "epoch": 2.1198320824322607, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.328709602355957, "learning_rate": 5e-05, "loss": 0.7628, "mean_token_accuracy": 0.8750730752944946, "num_tokens": 632600066.0, "step": 16664 }, { "epoch": 2.119959292710851, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.916026592254639, "learning_rate": 5e-05, "loss": 0.7806, "mean_token_accuracy": 0.8817128539085388, "num_tokens": 632638157.0, "step": 16665 }, { "epoch": 2.1200865029894413, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.345729827880859, "learning_rate": 5e-05, "loss": 0.7821, "mean_token_accuracy": 0.8636583089828491, "num_tokens": 632682534.0, "step": 16666 }, { "epoch": 2.120213713268032, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.847628116607666, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.8906872272491455, "num_tokens": 632717574.0, "step": 16667 }, { "epoch": 2.1203409235466224, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.351545333862305, "learning_rate": 5e-05, "loss": 0.6936, "mean_token_accuracy": 0.8868412971496582, "num_tokens": 632754113.0, "step": 16668 }, { "epoch": 2.120468133825213, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.771488189697266, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8730856776237488, "num_tokens": 632797776.0, "step": 16669 }, { "epoch": 2.1205953441038035, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.398380279541016, "learning_rate": 5e-05, "loss": 0.6924, "mean_token_accuracy": 0.8861663341522217, "num_tokens": 632834841.0, "step": 16670 }, { "epoch": 2.120722554382394, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.687786102294922, "learning_rate": 5e-05, "loss": 0.7096, "mean_token_accuracy": 0.8904927968978882, "num_tokens": 632872698.0, "step": 16671 }, { "epoch": 2.1208497646609845, "ewc_loss": 0.375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.434162139892578, "learning_rate": 5e-05, "loss": 0.7081, "mean_token_accuracy": 0.8841162919998169, "num_tokens": 632911067.0, "step": 16672 }, { "epoch": 2.120976974939575, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.602655410766602, "learning_rate": 5e-05, "loss": 0.698, "mean_token_accuracy": 0.8896122574806213, "num_tokens": 632952037.0, "step": 16673 }, { "epoch": 2.1211041852181656, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.490311622619629, "learning_rate": 5e-05, "loss": 0.7029, "mean_token_accuracy": 0.8811212778091431, "num_tokens": 632984667.0, "step": 16674 }, { "epoch": 2.121231395496756, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.504031181335449, "learning_rate": 5e-05, "loss": 0.7274, "mean_token_accuracy": 0.8769088983535767, "num_tokens": 633027165.0, "step": 16675 }, { "epoch": 2.1213586057753466, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.564149379730225, "learning_rate": 5e-05, "loss": 0.67, "mean_token_accuracy": 0.895736813545227, "num_tokens": 633068636.0, "step": 16676 }, { "epoch": 2.121485816053937, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.394923210144043, "learning_rate": 5e-05, "loss": 0.6656, "mean_token_accuracy": 0.8918488025665283, "num_tokens": 633103546.0, "step": 16677 }, { "epoch": 2.1216130263325277, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.649044990539551, "learning_rate": 5e-05, "loss": 0.7344, "mean_token_accuracy": 0.8742341995239258, "num_tokens": 633141529.0, "step": 16678 }, { "epoch": 2.121740236611118, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.281434059143066, "learning_rate": 5e-05, "loss": 0.6528, "mean_token_accuracy": 0.8958781957626343, "num_tokens": 633177324.0, "step": 16679 }, { "epoch": 2.1218674468897087, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.761581897735596, "learning_rate": 5e-05, "loss": 0.7139, "mean_token_accuracy": 0.8828022480010986, "num_tokens": 633216726.0, "step": 16680 }, { "epoch": 2.1219946571682993, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.139903545379639, "learning_rate": 5e-05, "loss": 0.683, "mean_token_accuracy": 0.879820704460144, "num_tokens": 633255583.0, "step": 16681 }, { "epoch": 2.12212186744689, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.939087867736816, "learning_rate": 5e-05, "loss": 0.7084, "mean_token_accuracy": 0.8868489265441895, "num_tokens": 633295167.0, "step": 16682 }, { "epoch": 2.1222490777254803, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.906978607177734, "learning_rate": 5e-05, "loss": 0.6942, "mean_token_accuracy": 0.8700649738311768, "num_tokens": 633336662.0, "step": 16683 }, { "epoch": 2.122376288004071, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.256181716918945, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.8867695331573486, "num_tokens": 633379826.0, "step": 16684 }, { "epoch": 2.1225034982826614, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.542189121246338, "learning_rate": 5e-05, "loss": 0.6698, "mean_token_accuracy": 0.8744627833366394, "num_tokens": 633421447.0, "step": 16685 }, { "epoch": 2.122630708561252, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.827300071716309, "learning_rate": 5e-05, "loss": 0.7816, "mean_token_accuracy": 0.8924851417541504, "num_tokens": 633462704.0, "step": 16686 }, { "epoch": 2.1227579188398424, "ewc_loss": 0.294921875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00022792816162109375, "grad_norm": 5.980068683624268, "learning_rate": 5e-05, "loss": 0.6184, "mean_token_accuracy": 0.8835811018943787, "num_tokens": 633496452.0, "step": 16687 }, { "epoch": 2.122885129118433, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.776130676269531, "learning_rate": 5e-05, "loss": 0.8738, "mean_token_accuracy": 0.9020097255706787, "num_tokens": 633534560.0, "step": 16688 }, { "epoch": 2.1230123393970235, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.658022880554199, "learning_rate": 5e-05, "loss": 0.6356, "mean_token_accuracy": 0.8879443407058716, "num_tokens": 633572429.0, "step": 16689 }, { "epoch": 2.1231395496756136, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.740317344665527, "learning_rate": 5e-05, "loss": 1.106, "mean_token_accuracy": 0.8768559098243713, "num_tokens": 633603588.0, "step": 16690 }, { "epoch": 2.123266759954204, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.074304103851318, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.887998104095459, "num_tokens": 633640975.0, "step": 16691 }, { "epoch": 2.1233939702327946, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.277579307556152, "learning_rate": 5e-05, "loss": 1.0295, "mean_token_accuracy": 0.8886922597885132, "num_tokens": 633684340.0, "step": 16692 }, { "epoch": 2.123521180511385, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 8.994317054748535, "learning_rate": 5e-05, "loss": 1.1292, "mean_token_accuracy": 0.8867175579071045, "num_tokens": 633720969.0, "step": 16693 }, { "epoch": 2.1236483907899757, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.870052337646484, "learning_rate": 5e-05, "loss": 0.7465, "mean_token_accuracy": 0.8951401114463806, "num_tokens": 633751726.0, "step": 16694 }, { "epoch": 2.123775601068566, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000827789306640625, "grad_norm": 10.813641548156738, "learning_rate": 5e-05, "loss": 1.2031, "mean_token_accuracy": 0.8902394771575928, "num_tokens": 633784645.0, "step": 16695 }, { "epoch": 2.1239028113471567, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 6.995171546936035, "learning_rate": 5e-05, "loss": 0.7475, "mean_token_accuracy": 0.8977768421173096, "num_tokens": 633816028.0, "step": 16696 }, { "epoch": 2.1240300216257473, "ewc_loss": 1.078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.001007080078125, "grad_norm": 13.002273559570312, "learning_rate": 5e-05, "loss": 1.4269, "mean_token_accuracy": 0.8753585815429688, "num_tokens": 633859018.0, "step": 16697 }, { "epoch": 2.124157231904338, "ewc_loss": 1.078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.001007080078125, "grad_norm": 10.468096733093262, "learning_rate": 5e-05, "loss": 1.4095, "mean_token_accuracy": 0.8791171908378601, "num_tokens": 633897549.0, "step": 16698 }, { "epoch": 2.1242844421829283, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 11.67869758605957, "learning_rate": 5e-05, "loss": 1.057, "mean_token_accuracy": 0.8930690288543701, "num_tokens": 633933111.0, "step": 16699 }, { "epoch": 2.124411652461519, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00077056884765625, "grad_norm": 9.690824508666992, "learning_rate": 5e-05, "loss": 1.1882, "mean_token_accuracy": 0.872489333152771, "num_tokens": 633972409.0, "step": 16700 }, { "epoch": 2.1245388627401094, "ewc_loss": 0.87890625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 11.983345985412598, "learning_rate": 5e-05, "loss": 1.2568, "mean_token_accuracy": 0.8671018481254578, "num_tokens": 634007112.0, "step": 16701 }, { "epoch": 2.1246660730187, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.716349601745605, "learning_rate": 5e-05, "loss": 0.9038, "mean_token_accuracy": 0.899724006652832, "num_tokens": 634043509.0, "step": 16702 }, { "epoch": 2.1247932832972904, "ewc_loss": 0.98046875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 12.69379997253418, "learning_rate": 5e-05, "loss": 1.3084, "mean_token_accuracy": 0.8847084045410156, "num_tokens": 634081883.0, "step": 16703 }, { "epoch": 2.124920493575881, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 10.02548599243164, "learning_rate": 5e-05, "loss": 1.0745, "mean_token_accuracy": 0.8939294815063477, "num_tokens": 634116203.0, "step": 16704 }, { "epoch": 2.1250477038544715, "ewc_loss": 0.86328125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 12.226435661315918, "learning_rate": 5e-05, "loss": 1.1877, "mean_token_accuracy": 0.8853777647018433, "num_tokens": 634152538.0, "step": 16705 }, { "epoch": 2.125174914133062, "ewc_loss": 0.83203125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000766754150390625, "grad_norm": 10.748374938964844, "learning_rate": 5e-05, "loss": 1.1524, "mean_token_accuracy": 0.8866513967514038, "num_tokens": 634188563.0, "step": 16706 }, { "epoch": 2.1253021244116526, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 11.234570503234863, "learning_rate": 5e-05, "loss": 1.115, "mean_token_accuracy": 0.8713885545730591, "num_tokens": 634235164.0, "step": 16707 }, { "epoch": 2.125429334690243, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 9.984478950500488, "learning_rate": 5e-05, "loss": 1.0164, "mean_token_accuracy": 0.8911123871803284, "num_tokens": 634275136.0, "step": 16708 }, { "epoch": 2.1255565449688336, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.852283477783203, "learning_rate": 5e-05, "loss": 1.0289, "mean_token_accuracy": 0.8888541460037231, "num_tokens": 634312056.0, "step": 16709 }, { "epoch": 2.125683755247424, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.354297637939453, "learning_rate": 5e-05, "loss": 1.0046, "mean_token_accuracy": 0.872668981552124, "num_tokens": 634345737.0, "step": 16710 }, { "epoch": 2.1258109655260147, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.647602081298828, "learning_rate": 5e-05, "loss": 1.0148, "mean_token_accuracy": 0.8840609788894653, "num_tokens": 634381122.0, "step": 16711 }, { "epoch": 2.125938175804605, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.888056755065918, "learning_rate": 5e-05, "loss": 0.8924, "mean_token_accuracy": 0.8839371204376221, "num_tokens": 634420529.0, "step": 16712 }, { "epoch": 2.1260653860831957, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.825074195861816, "learning_rate": 5e-05, "loss": 1.0057, "mean_token_accuracy": 0.8920387029647827, "num_tokens": 634460540.0, "step": 16713 }, { "epoch": 2.1261925963617863, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.73534870147705, "learning_rate": 5e-05, "loss": 0.9177, "mean_token_accuracy": 0.8750331401824951, "num_tokens": 634497818.0, "step": 16714 }, { "epoch": 2.1263198066403763, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 11.178279876708984, "learning_rate": 5e-05, "loss": 1.081, "mean_token_accuracy": 0.8806283473968506, "num_tokens": 634532262.0, "step": 16715 }, { "epoch": 2.126447016918967, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.165267944335938, "learning_rate": 5e-05, "loss": 0.9419, "mean_token_accuracy": 0.8830612897872925, "num_tokens": 634567316.0, "step": 16716 }, { "epoch": 2.1265742271975574, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 11.05555534362793, "learning_rate": 5e-05, "loss": 1.0746, "mean_token_accuracy": 0.8797339200973511, "num_tokens": 634608603.0, "step": 16717 }, { "epoch": 2.126701437476148, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.594947814941406, "learning_rate": 5e-05, "loss": 0.9644, "mean_token_accuracy": 0.8837263584136963, "num_tokens": 634646539.0, "step": 16718 }, { "epoch": 2.1268286477547385, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.308828353881836, "learning_rate": 5e-05, "loss": 0.9667, "mean_token_accuracy": 0.8899677991867065, "num_tokens": 634673998.0, "step": 16719 }, { "epoch": 2.126955858033329, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.330449104309082, "learning_rate": 5e-05, "loss": 0.9118, "mean_token_accuracy": 0.887647271156311, "num_tokens": 634707609.0, "step": 16720 }, { "epoch": 2.1270830683119195, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.668804168701172, "learning_rate": 5e-05, "loss": 0.9127, "mean_token_accuracy": 0.8890200853347778, "num_tokens": 634742893.0, "step": 16721 }, { "epoch": 2.12721027859051, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.693611145019531, "learning_rate": 5e-05, "loss": 0.8324, "mean_token_accuracy": 0.8886511325836182, "num_tokens": 634780887.0, "step": 16722 }, { "epoch": 2.1273374888691006, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.509268760681152, "learning_rate": 5e-05, "loss": 0.9064, "mean_token_accuracy": 0.8895996809005737, "num_tokens": 634819512.0, "step": 16723 }, { "epoch": 2.127464699147691, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.065131187438965, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.8729501962661743, "num_tokens": 634857213.0, "step": 16724 }, { "epoch": 2.1275919094262816, "ewc_loss": 0.625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.847044944763184, "learning_rate": 5e-05, "loss": 0.9781, "mean_token_accuracy": 0.8771605491638184, "num_tokens": 634899221.0, "step": 16725 }, { "epoch": 2.127719119704872, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.874882698059082, "learning_rate": 5e-05, "loss": 0.7125, "mean_token_accuracy": 0.9002069234848022, "num_tokens": 634943486.0, "step": 16726 }, { "epoch": 2.1278463299834627, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.315380096435547, "learning_rate": 5e-05, "loss": 1.0266, "mean_token_accuracy": 0.8789137601852417, "num_tokens": 634981303.0, "step": 16727 }, { "epoch": 2.127973540262053, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.483963966369629, "learning_rate": 5e-05, "loss": 0.8483, "mean_token_accuracy": 0.8815706968307495, "num_tokens": 635021411.0, "step": 16728 }, { "epoch": 2.1281007505406437, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.983126640319824, "learning_rate": 5e-05, "loss": 0.95, "mean_token_accuracy": 0.8925077319145203, "num_tokens": 635063449.0, "step": 16729 }, { "epoch": 2.1282279608192343, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.950163841247559, "learning_rate": 5e-05, "loss": 0.9048, "mean_token_accuracy": 0.8734265565872192, "num_tokens": 635107052.0, "step": 16730 }, { "epoch": 2.128355171097825, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.974024772644043, "learning_rate": 5e-05, "loss": 0.8468, "mean_token_accuracy": 0.8950121998786926, "num_tokens": 635141036.0, "step": 16731 }, { "epoch": 2.1284823813764153, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.633560180664062, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8762118220329285, "num_tokens": 635181060.0, "step": 16732 }, { "epoch": 2.128609591655006, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.337166786193848, "learning_rate": 5e-05, "loss": 0.8047, "mean_token_accuracy": 0.8834856748580933, "num_tokens": 635222679.0, "step": 16733 }, { "epoch": 2.1287368019335964, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.066329002380371, "learning_rate": 5e-05, "loss": 0.7445, "mean_token_accuracy": 0.8891225457191467, "num_tokens": 635258351.0, "step": 16734 }, { "epoch": 2.128864012212187, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.21190071105957, "learning_rate": 5e-05, "loss": 0.8446, "mean_token_accuracy": 0.8682433366775513, "num_tokens": 635300647.0, "step": 16735 }, { "epoch": 2.1289912224907774, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.603417873382568, "learning_rate": 5e-05, "loss": 0.7195, "mean_token_accuracy": 0.882876992225647, "num_tokens": 635335239.0, "step": 16736 }, { "epoch": 2.129118432769368, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.19452953338623, "learning_rate": 5e-05, "loss": 0.773, "mean_token_accuracy": 0.8866023421287537, "num_tokens": 635370786.0, "step": 16737 }, { "epoch": 2.129245643047958, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.392698764801025, "learning_rate": 5e-05, "loss": 0.6987, "mean_token_accuracy": 0.8846357464790344, "num_tokens": 635410033.0, "step": 16738 }, { "epoch": 2.129372853326549, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.203734397888184, "learning_rate": 5e-05, "loss": 0.7357, "mean_token_accuracy": 0.8962599039077759, "num_tokens": 635447157.0, "step": 16739 }, { "epoch": 2.129500063605139, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.209603309631348, "learning_rate": 5e-05, "loss": 0.6598, "mean_token_accuracy": 0.8915543556213379, "num_tokens": 635485580.0, "step": 16740 }, { "epoch": 2.1296272738837296, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.389081954956055, "learning_rate": 5e-05, "loss": 0.7207, "mean_token_accuracy": 0.8994200229644775, "num_tokens": 635520289.0, "step": 16741 }, { "epoch": 2.12975448416232, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.981654167175293, "learning_rate": 5e-05, "loss": 0.64, "mean_token_accuracy": 0.8926554322242737, "num_tokens": 635552390.0, "step": 16742 }, { "epoch": 2.1298816944409107, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.75015640258789, "learning_rate": 5e-05, "loss": 0.8111, "mean_token_accuracy": 0.8833134174346924, "num_tokens": 635588281.0, "step": 16743 }, { "epoch": 2.130008904719501, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.852060794830322, "learning_rate": 5e-05, "loss": 0.7016, "mean_token_accuracy": 0.8742126226425171, "num_tokens": 635632229.0, "step": 16744 }, { "epoch": 2.1301361149980917, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.988929748535156, "learning_rate": 5e-05, "loss": 0.8322, "mean_token_accuracy": 0.890264093875885, "num_tokens": 635671315.0, "step": 16745 }, { "epoch": 2.1302633252766823, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.201545715332031, "learning_rate": 5e-05, "loss": 0.7266, "mean_token_accuracy": 0.8792175054550171, "num_tokens": 635711765.0, "step": 16746 }, { "epoch": 2.130390535555273, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.614890098571777, "learning_rate": 5e-05, "loss": 0.8556, "mean_token_accuracy": 0.8677314519882202, "num_tokens": 635747710.0, "step": 16747 }, { "epoch": 2.1305177458338633, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.78145170211792, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8732951879501343, "num_tokens": 635781355.0, "step": 16748 }, { "epoch": 2.130644956112454, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.707245349884033, "learning_rate": 5e-05, "loss": 0.7291, "mean_token_accuracy": 0.8859025835990906, "num_tokens": 635821340.0, "step": 16749 }, { "epoch": 2.1307721663910444, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.280733108520508, "learning_rate": 5e-05, "loss": 0.79, "mean_token_accuracy": 0.8718478679656982, "num_tokens": 635857943.0, "step": 16750 }, { "epoch": 2.130899376669635, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.896469593048096, "learning_rate": 5e-05, "loss": 0.7031, "mean_token_accuracy": 0.8781793117523193, "num_tokens": 635896808.0, "step": 16751 }, { "epoch": 2.1310265869482254, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.93237590789795, "learning_rate": 5e-05, "loss": 0.7999, "mean_token_accuracy": 0.8896278738975525, "num_tokens": 635934952.0, "step": 16752 }, { "epoch": 2.131153797226816, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.370258808135986, "learning_rate": 5e-05, "loss": 0.6524, "mean_token_accuracy": 0.8843531012535095, "num_tokens": 635972968.0, "step": 16753 }, { "epoch": 2.1312810075054065, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.708338737487793, "learning_rate": 5e-05, "loss": 0.8737, "mean_token_accuracy": 0.8971611261367798, "num_tokens": 636004761.0, "step": 16754 }, { "epoch": 2.131408217783997, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.73805046081543, "learning_rate": 5e-05, "loss": 0.6999, "mean_token_accuracy": 0.8794282674789429, "num_tokens": 636043003.0, "step": 16755 }, { "epoch": 2.1315354280625876, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.604825019836426, "learning_rate": 5e-05, "loss": 0.9335, "mean_token_accuracy": 0.8854154348373413, "num_tokens": 636083580.0, "step": 16756 }, { "epoch": 2.131662638341178, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.097962379455566, "learning_rate": 5e-05, "loss": 0.8177, "mean_token_accuracy": 0.8853437900543213, "num_tokens": 636124248.0, "step": 16757 }, { "epoch": 2.1317898486197686, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.19036865234375, "learning_rate": 5e-05, "loss": 0.8312, "mean_token_accuracy": 0.874731183052063, "num_tokens": 636164665.0, "step": 16758 }, { "epoch": 2.131917058898359, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.78015422821045, "learning_rate": 5e-05, "loss": 0.8515, "mean_token_accuracy": 0.8826435804367065, "num_tokens": 636201526.0, "step": 16759 }, { "epoch": 2.1320442691769497, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.264683246612549, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.882691502571106, "num_tokens": 636247793.0, "step": 16760 }, { "epoch": 2.13217147945554, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.597481727600098, "learning_rate": 5e-05, "loss": 0.7988, "mean_token_accuracy": 0.892265260219574, "num_tokens": 636289989.0, "step": 16761 }, { "epoch": 2.1322986897341307, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.666428565979004, "learning_rate": 5e-05, "loss": 0.7892, "mean_token_accuracy": 0.8699619770050049, "num_tokens": 636328729.0, "step": 16762 }, { "epoch": 2.132425900012721, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.568160057067871, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.8827347755432129, "num_tokens": 636362593.0, "step": 16763 }, { "epoch": 2.1325531102913113, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.345254898071289, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.8906158804893494, "num_tokens": 636402790.0, "step": 16764 }, { "epoch": 2.132680320569902, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.875692367553711, "learning_rate": 5e-05, "loss": 0.684, "mean_token_accuracy": 0.8812171816825867, "num_tokens": 636443052.0, "step": 16765 }, { "epoch": 2.1328075308484924, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.770157814025879, "learning_rate": 5e-05, "loss": 0.8095, "mean_token_accuracy": 0.883674144744873, "num_tokens": 636481345.0, "step": 16766 }, { "epoch": 2.132934741127083, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.744660377502441, "learning_rate": 5e-05, "loss": 0.6706, "mean_token_accuracy": 0.8822895884513855, "num_tokens": 636516638.0, "step": 16767 }, { "epoch": 2.1330619514056735, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.804557800292969, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.8977484107017517, "num_tokens": 636551606.0, "step": 16768 }, { "epoch": 2.133189161684264, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.99852180480957, "learning_rate": 5e-05, "loss": 0.702, "mean_token_accuracy": 0.8810042142868042, "num_tokens": 636589298.0, "step": 16769 }, { "epoch": 2.1333163719628545, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.57494831085205, "learning_rate": 5e-05, "loss": 0.808, "mean_token_accuracy": 0.8846254348754883, "num_tokens": 636622010.0, "step": 16770 }, { "epoch": 2.133443582241445, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.228835105895996, "learning_rate": 5e-05, "loss": 0.6944, "mean_token_accuracy": 0.8867863416671753, "num_tokens": 636665355.0, "step": 16771 }, { "epoch": 2.1335707925200356, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.355584144592285, "learning_rate": 5e-05, "loss": 0.7983, "mean_token_accuracy": 0.8913999795913696, "num_tokens": 636706135.0, "step": 16772 }, { "epoch": 2.133698002798626, "ewc_loss": 0.375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.191425323486328, "learning_rate": 5e-05, "loss": 0.6629, "mean_token_accuracy": 0.8955326080322266, "num_tokens": 636744587.0, "step": 16773 }, { "epoch": 2.1338252130772166, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.518341064453125, "learning_rate": 5e-05, "loss": 0.7683, "mean_token_accuracy": 0.8965048789978027, "num_tokens": 636784175.0, "step": 16774 }, { "epoch": 2.133952423355807, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.912875652313232, "learning_rate": 5e-05, "loss": 0.7054, "mean_token_accuracy": 0.8748023509979248, "num_tokens": 636821752.0, "step": 16775 }, { "epoch": 2.1340796336343977, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.850629806518555, "learning_rate": 5e-05, "loss": 0.91, "mean_token_accuracy": 0.8721412420272827, "num_tokens": 636855507.0, "step": 16776 }, { "epoch": 2.134206843912988, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.816402435302734, "learning_rate": 5e-05, "loss": 0.716, "mean_token_accuracy": 0.8778476715087891, "num_tokens": 636895401.0, "step": 16777 }, { "epoch": 2.1343340541915787, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.862752914428711, "learning_rate": 5e-05, "loss": 0.9191, "mean_token_accuracy": 0.865878701210022, "num_tokens": 636940225.0, "step": 16778 }, { "epoch": 2.1344612644701693, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.217487335205078, "learning_rate": 5e-05, "loss": 0.7447, "mean_token_accuracy": 0.8776988983154297, "num_tokens": 636976685.0, "step": 16779 }, { "epoch": 2.13458847474876, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.27072525024414, "learning_rate": 5e-05, "loss": 0.8225, "mean_token_accuracy": 0.882822573184967, "num_tokens": 637018170.0, "step": 16780 }, { "epoch": 2.1347156850273503, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.772424221038818, "learning_rate": 5e-05, "loss": 0.7597, "mean_token_accuracy": 0.885971188545227, "num_tokens": 637055115.0, "step": 16781 }, { "epoch": 2.134842895305941, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.575908184051514, "learning_rate": 5e-05, "loss": 0.7754, "mean_token_accuracy": 0.8702386617660522, "num_tokens": 637093964.0, "step": 16782 }, { "epoch": 2.1349701055845314, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.080044746398926, "learning_rate": 5e-05, "loss": 0.7202, "mean_token_accuracy": 0.903239369392395, "num_tokens": 637135242.0, "step": 16783 }, { "epoch": 2.135097315863122, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.212666988372803, "learning_rate": 5e-05, "loss": 0.6991, "mean_token_accuracy": 0.8842936754226685, "num_tokens": 637173407.0, "step": 16784 }, { "epoch": 2.1352245261417124, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.264321327209473, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.8822706341743469, "num_tokens": 637207613.0, "step": 16785 }, { "epoch": 2.135351736420303, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.983098983764648, "learning_rate": 5e-05, "loss": 0.6341, "mean_token_accuracy": 0.898921012878418, "num_tokens": 637242010.0, "step": 16786 }, { "epoch": 2.1354789466988935, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.510700225830078, "learning_rate": 5e-05, "loss": 0.7739, "mean_token_accuracy": 0.8905676007270813, "num_tokens": 637284034.0, "step": 16787 }, { "epoch": 2.1356061569774836, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.729799270629883, "learning_rate": 5e-05, "loss": 0.6547, "mean_token_accuracy": 0.8890959024429321, "num_tokens": 637325657.0, "step": 16788 }, { "epoch": 2.135733367256074, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.86323356628418, "learning_rate": 5e-05, "loss": 0.8145, "mean_token_accuracy": 0.892292857170105, "num_tokens": 637360946.0, "step": 16789 }, { "epoch": 2.1358605775346646, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.604284286499023, "learning_rate": 5e-05, "loss": 0.6859, "mean_token_accuracy": 0.8797245025634766, "num_tokens": 637397540.0, "step": 16790 }, { "epoch": 2.135987787813255, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.076166152954102, "learning_rate": 5e-05, "loss": 0.8719, "mean_token_accuracy": 0.8891464471817017, "num_tokens": 637443148.0, "step": 16791 }, { "epoch": 2.1361149980918457, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.90897274017334, "learning_rate": 5e-05, "loss": 0.6815, "mean_token_accuracy": 0.8916403651237488, "num_tokens": 637480796.0, "step": 16792 }, { "epoch": 2.136242208370436, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.696327209472656, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8814269304275513, "num_tokens": 637525465.0, "step": 16793 }, { "epoch": 2.1363694186490267, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.528943061828613, "learning_rate": 5e-05, "loss": 0.7587, "mean_token_accuracy": 0.884961724281311, "num_tokens": 637563235.0, "step": 16794 }, { "epoch": 2.1364966289276173, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.957202434539795, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8926587104797363, "num_tokens": 637600061.0, "step": 16795 }, { "epoch": 2.136623839206208, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.939105987548828, "learning_rate": 5e-05, "loss": 0.7481, "mean_token_accuracy": 0.8987283110618591, "num_tokens": 637640877.0, "step": 16796 }, { "epoch": 2.1367510494847983, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.5019211769104, "learning_rate": 5e-05, "loss": 0.7329, "mean_token_accuracy": 0.8828927278518677, "num_tokens": 637679248.0, "step": 16797 }, { "epoch": 2.136878259763389, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.915955543518066, "learning_rate": 5e-05, "loss": 0.7693, "mean_token_accuracy": 0.8790670037269592, "num_tokens": 637715973.0, "step": 16798 }, { "epoch": 2.1370054700419794, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.618951797485352, "learning_rate": 5e-05, "loss": 0.7224, "mean_token_accuracy": 0.8897361159324646, "num_tokens": 637753358.0, "step": 16799 }, { "epoch": 2.13713268032057, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.4879374504089355, "learning_rate": 5e-05, "loss": 0.7237, "mean_token_accuracy": 0.8835240602493286, "num_tokens": 637798608.0, "step": 16800 }, { "epoch": 2.1372598905991604, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.040244102478027, "learning_rate": 5e-05, "loss": 0.7519, "mean_token_accuracy": 0.8866803646087646, "num_tokens": 637837173.0, "step": 16801 }, { "epoch": 2.137387100877751, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.9889936447143555, "learning_rate": 5e-05, "loss": 0.6639, "mean_token_accuracy": 0.889667272567749, "num_tokens": 637874072.0, "step": 16802 }, { "epoch": 2.1375143111563415, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.532304763793945, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.873725175857544, "num_tokens": 637909369.0, "step": 16803 }, { "epoch": 2.137641521434932, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.573838233947754, "learning_rate": 5e-05, "loss": 0.6397, "mean_token_accuracy": 0.8928655385971069, "num_tokens": 637946526.0, "step": 16804 }, { "epoch": 2.1377687317135226, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 9.00043773651123, "learning_rate": 5e-05, "loss": 0.8591, "mean_token_accuracy": 0.8802053928375244, "num_tokens": 637984476.0, "step": 16805 }, { "epoch": 2.137895941992113, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.439460277557373, "learning_rate": 5e-05, "loss": 0.6762, "mean_token_accuracy": 0.8789363503456116, "num_tokens": 638019098.0, "step": 16806 }, { "epoch": 2.1380231522707036, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.232169151306152, "learning_rate": 5e-05, "loss": 0.8914, "mean_token_accuracy": 0.8855595588684082, "num_tokens": 638059049.0, "step": 16807 }, { "epoch": 2.138150362549294, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.799448490142822, "learning_rate": 5e-05, "loss": 0.7138, "mean_token_accuracy": 0.8838058710098267, "num_tokens": 638095887.0, "step": 16808 }, { "epoch": 2.1382775728278847, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.820521354675293, "learning_rate": 5e-05, "loss": 0.9432, "mean_token_accuracy": 0.8678867816925049, "num_tokens": 638141855.0, "step": 16809 }, { "epoch": 2.138404783106475, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.61020040512085, "learning_rate": 5e-05, "loss": 0.7768, "mean_token_accuracy": 0.8954702615737915, "num_tokens": 638184491.0, "step": 16810 }, { "epoch": 2.1385319933850653, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.795559883117676, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.895845890045166, "num_tokens": 638218341.0, "step": 16811 }, { "epoch": 2.1386592036636562, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.314920425415039, "learning_rate": 5e-05, "loss": 0.8602, "mean_token_accuracy": 0.8768960237503052, "num_tokens": 638257422.0, "step": 16812 }, { "epoch": 2.1387864139422463, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.9698944091796875, "learning_rate": 5e-05, "loss": 0.6812, "mean_token_accuracy": 0.8872880935668945, "num_tokens": 638298053.0, "step": 16813 }, { "epoch": 2.138913624220837, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.76991081237793, "learning_rate": 5e-05, "loss": 0.8997, "mean_token_accuracy": 0.8682161569595337, "num_tokens": 638337870.0, "step": 16814 }, { "epoch": 2.1390408344994274, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.667529582977295, "learning_rate": 5e-05, "loss": 0.683, "mean_token_accuracy": 0.8836677074432373, "num_tokens": 638371587.0, "step": 16815 }, { "epoch": 2.139168044778018, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.958465576171875, "learning_rate": 5e-05, "loss": 0.8839, "mean_token_accuracy": 0.8839976191520691, "num_tokens": 638416879.0, "step": 16816 }, { "epoch": 2.1392952550566084, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.737947940826416, "learning_rate": 5e-05, "loss": 0.6653, "mean_token_accuracy": 0.8935918211936951, "num_tokens": 638452809.0, "step": 16817 }, { "epoch": 2.139422465335199, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.992341995239258, "learning_rate": 5e-05, "loss": 0.9251, "mean_token_accuracy": 0.8782947659492493, "num_tokens": 638487027.0, "step": 16818 }, { "epoch": 2.1395496756137895, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.031623363494873, "learning_rate": 5e-05, "loss": 0.7448, "mean_token_accuracy": 0.8842344284057617, "num_tokens": 638517922.0, "step": 16819 }, { "epoch": 2.13967688589238, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.772422790527344, "learning_rate": 5e-05, "loss": 0.9289, "mean_token_accuracy": 0.8743422031402588, "num_tokens": 638562360.0, "step": 16820 }, { "epoch": 2.1398040961709706, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.38529109954834, "learning_rate": 5e-05, "loss": 0.7878, "mean_token_accuracy": 0.881305456161499, "num_tokens": 638595649.0, "step": 16821 }, { "epoch": 2.139931306449561, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.39244270324707, "learning_rate": 5e-05, "loss": 0.8685, "mean_token_accuracy": 0.884256899356842, "num_tokens": 638632588.0, "step": 16822 }, { "epoch": 2.1400585167281516, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.5393171310424805, "learning_rate": 5e-05, "loss": 0.7938, "mean_token_accuracy": 0.8854326009750366, "num_tokens": 638663000.0, "step": 16823 }, { "epoch": 2.140185727006742, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.192985534667969, "learning_rate": 5e-05, "loss": 0.8433, "mean_token_accuracy": 0.8817065954208374, "num_tokens": 638703737.0, "step": 16824 }, { "epoch": 2.1403129372853327, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.385409355163574, "learning_rate": 5e-05, "loss": 0.8044, "mean_token_accuracy": 0.8715835809707642, "num_tokens": 638745526.0, "step": 16825 }, { "epoch": 2.140440147563923, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000667572021484375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.350903511047363, "learning_rate": 5e-05, "loss": 0.8388, "mean_token_accuracy": 0.8875920176506042, "num_tokens": 638780865.0, "step": 16826 }, { "epoch": 2.1405673578425137, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.047662734985352, "learning_rate": 5e-05, "loss": 0.7268, "mean_token_accuracy": 0.8809564113616943, "num_tokens": 638822274.0, "step": 16827 }, { "epoch": 2.1406945681211043, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.665863990783691, "learning_rate": 5e-05, "loss": 0.9027, "mean_token_accuracy": 0.8741479516029358, "num_tokens": 638859077.0, "step": 16828 }, { "epoch": 2.140821778399695, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.861870288848877, "learning_rate": 5e-05, "loss": 0.7053, "mean_token_accuracy": 0.8843182325363159, "num_tokens": 638898055.0, "step": 16829 }, { "epoch": 2.1409489886782853, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.851814270019531, "learning_rate": 5e-05, "loss": 0.9281, "mean_token_accuracy": 0.8753695487976074, "num_tokens": 638939597.0, "step": 16830 }, { "epoch": 2.141076198956876, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.339157581329346, "learning_rate": 5e-05, "loss": 0.7846, "mean_token_accuracy": 0.8719706535339355, "num_tokens": 638978268.0, "step": 16831 }, { "epoch": 2.1412034092354664, "ewc_loss": 0.5, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.051115036010742, "learning_rate": 5e-05, "loss": 0.7922, "mean_token_accuracy": 0.8955177068710327, "num_tokens": 639010554.0, "step": 16832 }, { "epoch": 2.141330619514057, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.811980724334717, "learning_rate": 5e-05, "loss": 0.7284, "mean_token_accuracy": 0.9051692485809326, "num_tokens": 639046721.0, "step": 16833 }, { "epoch": 2.1414578297926474, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.6366705894470215, "learning_rate": 5e-05, "loss": 0.7819, "mean_token_accuracy": 0.8780364990234375, "num_tokens": 639080078.0, "step": 16834 }, { "epoch": 2.141585040071238, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.757868766784668, "learning_rate": 5e-05, "loss": 0.7739, "mean_token_accuracy": 0.8850723505020142, "num_tokens": 639119469.0, "step": 16835 }, { "epoch": 2.141712250349828, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.7448859214782715, "learning_rate": 5e-05, "loss": 0.7667, "mean_token_accuracy": 0.8810306787490845, "num_tokens": 639159954.0, "step": 16836 }, { "epoch": 2.141839460628419, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.457397937774658, "learning_rate": 5e-05, "loss": 0.7345, "mean_token_accuracy": 0.8848662376403809, "num_tokens": 639202455.0, "step": 16837 }, { "epoch": 2.141966670907009, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00067138671875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.984625339508057, "learning_rate": 5e-05, "loss": 0.7357, "mean_token_accuracy": 0.8928017616271973, "num_tokens": 639240889.0, "step": 16838 }, { "epoch": 2.1420938811855996, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.215228080749512, "learning_rate": 5e-05, "loss": 0.7019, "mean_token_accuracy": 0.8852939009666443, "num_tokens": 639275079.0, "step": 16839 }, { "epoch": 2.14222109146419, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.131710052490234, "learning_rate": 5e-05, "loss": 0.7997, "mean_token_accuracy": 0.872061014175415, "num_tokens": 639308166.0, "step": 16840 }, { "epoch": 2.1423483017427807, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.11007833480835, "learning_rate": 5e-05, "loss": 0.6871, "mean_token_accuracy": 0.8920310735702515, "num_tokens": 639341862.0, "step": 16841 }, { "epoch": 2.142475512021371, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.203079223632812, "learning_rate": 5e-05, "loss": 0.7662, "mean_token_accuracy": 0.8939640522003174, "num_tokens": 639381065.0, "step": 16842 }, { "epoch": 2.1426027222999617, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.050653457641602, "learning_rate": 5e-05, "loss": 0.6732, "mean_token_accuracy": 0.8905690908432007, "num_tokens": 639421232.0, "step": 16843 }, { "epoch": 2.1427299325785523, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.265592575073242, "learning_rate": 5e-05, "loss": 0.7993, "mean_token_accuracy": 0.8792601823806763, "num_tokens": 639459462.0, "step": 16844 }, { "epoch": 2.142857142857143, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.001901149749756, "learning_rate": 5e-05, "loss": 0.7044, "mean_token_accuracy": 0.881999135017395, "num_tokens": 639499759.0, "step": 16845 }, { "epoch": 2.1429843531357333, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.31476879119873, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.8857828974723816, "num_tokens": 639534285.0, "step": 16846 }, { "epoch": 2.143111563414324, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.032054424285889, "learning_rate": 5e-05, "loss": 0.7216, "mean_token_accuracy": 0.8798647522926331, "num_tokens": 639576058.0, "step": 16847 }, { "epoch": 2.1432387736929144, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.245048522949219, "learning_rate": 5e-05, "loss": 0.7891, "mean_token_accuracy": 0.8858307600021362, "num_tokens": 639610464.0, "step": 16848 }, { "epoch": 2.143365983971505, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.166055679321289, "learning_rate": 5e-05, "loss": 0.6684, "mean_token_accuracy": 0.8959288597106934, "num_tokens": 639645327.0, "step": 16849 }, { "epoch": 2.1434931942500954, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.02463436126709, "learning_rate": 5e-05, "loss": 0.8044, "mean_token_accuracy": 0.8736478090286255, "num_tokens": 639684501.0, "step": 16850 }, { "epoch": 2.143620404528686, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.3825860023498535, "learning_rate": 5e-05, "loss": 0.7409, "mean_token_accuracy": 0.8791473507881165, "num_tokens": 639721687.0, "step": 16851 }, { "epoch": 2.1437476148072765, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.719281196594238, "learning_rate": 5e-05, "loss": 0.7889, "mean_token_accuracy": 0.864923894405365, "num_tokens": 639761503.0, "step": 16852 }, { "epoch": 2.143874825085867, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.662389278411865, "learning_rate": 5e-05, "loss": 0.7196, "mean_token_accuracy": 0.8864376544952393, "num_tokens": 639800938.0, "step": 16853 }, { "epoch": 2.1440020353644575, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.3197150230407715, "learning_rate": 5e-05, "loss": 0.6497, "mean_token_accuracy": 0.8968318700790405, "num_tokens": 639835318.0, "step": 16854 }, { "epoch": 2.144129245643048, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.035482406616211, "learning_rate": 5e-05, "loss": 0.7409, "mean_token_accuracy": 0.8886153101921082, "num_tokens": 639878155.0, "step": 16855 }, { "epoch": 2.1442564559216386, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.886372089385986, "learning_rate": 5e-05, "loss": 0.6669, "mean_token_accuracy": 0.8848782777786255, "num_tokens": 639920369.0, "step": 16856 }, { "epoch": 2.144383666200229, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.507180213928223, "learning_rate": 5e-05, "loss": 0.7764, "mean_token_accuracy": 0.8882107734680176, "num_tokens": 639964690.0, "step": 16857 }, { "epoch": 2.1445108764788197, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.412845611572266, "learning_rate": 5e-05, "loss": 0.6917, "mean_token_accuracy": 0.8699934482574463, "num_tokens": 640002969.0, "step": 16858 }, { "epoch": 2.14463808675741, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 9.182785034179688, "learning_rate": 5e-05, "loss": 0.8246, "mean_token_accuracy": 0.8939533829689026, "num_tokens": 640044061.0, "step": 16859 }, { "epoch": 2.1447652970360007, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 5.953473091125488, "learning_rate": 5e-05, "loss": 0.6275, "mean_token_accuracy": 0.8880003690719604, "num_tokens": 640082803.0, "step": 16860 }, { "epoch": 2.144892507314591, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.974384307861328, "learning_rate": 5e-05, "loss": 0.9255, "mean_token_accuracy": 0.8988361954689026, "num_tokens": 640123223.0, "step": 16861 }, { "epoch": 2.1450197175931813, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.214606285095215, "learning_rate": 5e-05, "loss": 0.6637, "mean_token_accuracy": 0.8992289304733276, "num_tokens": 640160674.0, "step": 16862 }, { "epoch": 2.145146927871772, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.150862693786621, "learning_rate": 5e-05, "loss": 1.0317, "mean_token_accuracy": 0.8902628421783447, "num_tokens": 640197347.0, "step": 16863 }, { "epoch": 2.1452741381503624, "ewc_loss": 0.625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 7.7738189697265625, "learning_rate": 5e-05, "loss": 0.9544, "mean_token_accuracy": 0.8841800689697266, "num_tokens": 640235599.0, "step": 16864 }, { "epoch": 2.145401348428953, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.654915809631348, "learning_rate": 5e-05, "loss": 0.9224, "mean_token_accuracy": 0.8804141283035278, "num_tokens": 640277753.0, "step": 16865 }, { "epoch": 2.1455285587075434, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.202082633972168, "learning_rate": 5e-05, "loss": 1.0016, "mean_token_accuracy": 0.8896654844284058, "num_tokens": 640317486.0, "step": 16866 }, { "epoch": 2.145655768986134, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.565732479095459, "learning_rate": 5e-05, "loss": 0.7501, "mean_token_accuracy": 0.8690769672393799, "num_tokens": 640357240.0, "step": 16867 }, { "epoch": 2.1457829792647245, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000743865966796875, "grad_norm": 10.990463256835938, "learning_rate": 5e-05, "loss": 1.1279, "mean_token_accuracy": 0.8889290690422058, "num_tokens": 640392429.0, "step": 16868 }, { "epoch": 2.145910189543315, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 6.987257480621338, "learning_rate": 5e-05, "loss": 0.8886, "mean_token_accuracy": 0.8717109560966492, "num_tokens": 640432529.0, "step": 16869 }, { "epoch": 2.1460373998219056, "ewc_loss": 0.90625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 11.788553237915039, "learning_rate": 5e-05, "loss": 1.2246, "mean_token_accuracy": 0.8864402770996094, "num_tokens": 640468443.0, "step": 16870 }, { "epoch": 2.146164610100496, "ewc_loss": 0.9296875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 9.35678768157959, "learning_rate": 5e-05, "loss": 1.2878, "mean_token_accuracy": 0.8733935952186584, "num_tokens": 640505647.0, "step": 16871 }, { "epoch": 2.1462918203790866, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.513997077941895, "learning_rate": 5e-05, "loss": 1.0562, "mean_token_accuracy": 0.8724284172058105, "num_tokens": 640551180.0, "step": 16872 }, { "epoch": 2.146419030657677, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000732421875, "grad_norm": 9.231241226196289, "learning_rate": 5e-05, "loss": 1.1187, "mean_token_accuracy": 0.8887608647346497, "num_tokens": 640590939.0, "step": 16873 }, { "epoch": 2.1465462409362677, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.627442359924316, "learning_rate": 5e-05, "loss": 1.0471, "mean_token_accuracy": 0.8691220879554749, "num_tokens": 640625688.0, "step": 16874 }, { "epoch": 2.146673451214858, "ewc_loss": 0.625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.62841510772705, "learning_rate": 5e-05, "loss": 1.0234, "mean_token_accuracy": 0.8642077445983887, "num_tokens": 640669142.0, "step": 16875 }, { "epoch": 2.1468006614934487, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.15864372253418, "learning_rate": 5e-05, "loss": 1.0038, "mean_token_accuracy": 0.8762811422348022, "num_tokens": 640713345.0, "step": 16876 }, { "epoch": 2.1469278717720393, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.86893367767334, "learning_rate": 5e-05, "loss": 0.9038, "mean_token_accuracy": 0.8617911338806152, "num_tokens": 640750930.0, "step": 16877 }, { "epoch": 2.14705508205063, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.355013847351074, "learning_rate": 5e-05, "loss": 0.9578, "mean_token_accuracy": 0.8897668123245239, "num_tokens": 640792029.0, "step": 16878 }, { "epoch": 2.1471822923292203, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.3966383934021, "learning_rate": 5e-05, "loss": 0.7978, "mean_token_accuracy": 0.8883464336395264, "num_tokens": 640833952.0, "step": 16879 }, { "epoch": 2.147309502607811, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.761858940124512, "learning_rate": 5e-05, "loss": 0.991, "mean_token_accuracy": 0.882748544216156, "num_tokens": 640867980.0, "step": 16880 }, { "epoch": 2.1474367128864014, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.503695964813232, "learning_rate": 5e-05, "loss": 0.9035, "mean_token_accuracy": 0.8693461418151855, "num_tokens": 640913241.0, "step": 16881 }, { "epoch": 2.147563923164992, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 9.85811710357666, "learning_rate": 5e-05, "loss": 0.978, "mean_token_accuracy": 0.8988142013549805, "num_tokens": 640946705.0, "step": 16882 }, { "epoch": 2.1476911334435824, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.9084577560424805, "learning_rate": 5e-05, "loss": 0.8948, "mean_token_accuracy": 0.8905640244483948, "num_tokens": 640985658.0, "step": 16883 }, { "epoch": 2.147818343722173, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.604482650756836, "learning_rate": 5e-05, "loss": 1.0344, "mean_token_accuracy": 0.8762843608856201, "num_tokens": 641025963.0, "step": 16884 }, { "epoch": 2.1479455540007635, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.1308012008667, "learning_rate": 5e-05, "loss": 0.911, "mean_token_accuracy": 0.8900176882743835, "num_tokens": 641070853.0, "step": 16885 }, { "epoch": 2.1480727642793536, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.143630981445312, "learning_rate": 5e-05, "loss": 0.9495, "mean_token_accuracy": 0.8872450590133667, "num_tokens": 641106336.0, "step": 16886 }, { "epoch": 2.148199974557944, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.131932258605957, "learning_rate": 5e-05, "loss": 0.8419, "mean_token_accuracy": 0.9005424976348877, "num_tokens": 641147842.0, "step": 16887 }, { "epoch": 2.1483271848365346, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.724722862243652, "learning_rate": 5e-05, "loss": 0.9072, "mean_token_accuracy": 0.8943406343460083, "num_tokens": 641186329.0, "step": 16888 }, { "epoch": 2.148454395115125, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.830286502838135, "learning_rate": 5e-05, "loss": 0.8564, "mean_token_accuracy": 0.8768183588981628, "num_tokens": 641227892.0, "step": 16889 }, { "epoch": 2.1485816053937157, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.796161651611328, "learning_rate": 5e-05, "loss": 0.9178, "mean_token_accuracy": 0.8909971117973328, "num_tokens": 641265611.0, "step": 16890 }, { "epoch": 2.148708815672306, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.43588924407959, "learning_rate": 5e-05, "loss": 0.8445, "mean_token_accuracy": 0.864648699760437, "num_tokens": 641299564.0, "step": 16891 }, { "epoch": 2.1488360259508967, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.13914680480957, "learning_rate": 5e-05, "loss": 0.9573, "mean_token_accuracy": 0.8871593475341797, "num_tokens": 641335892.0, "step": 16892 }, { "epoch": 2.1489632362294873, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.380459785461426, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.8826388120651245, "num_tokens": 641375470.0, "step": 16893 }, { "epoch": 2.149090446508078, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.259608268737793, "learning_rate": 5e-05, "loss": 0.9896, "mean_token_accuracy": 0.884120762348175, "num_tokens": 641408793.0, "step": 16894 }, { "epoch": 2.1492176567866683, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.758176803588867, "learning_rate": 5e-05, "loss": 0.8695, "mean_token_accuracy": 0.8807550072669983, "num_tokens": 641444949.0, "step": 16895 }, { "epoch": 2.149344867065259, "ewc_loss": 0.625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.82154369354248, "learning_rate": 5e-05, "loss": 0.9459, "mean_token_accuracy": 0.8845722079277039, "num_tokens": 641482775.0, "step": 16896 }, { "epoch": 2.1494720773438494, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.07668399810791, "learning_rate": 5e-05, "loss": 0.8671, "mean_token_accuracy": 0.8874854445457458, "num_tokens": 641520385.0, "step": 16897 }, { "epoch": 2.14959928762244, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.111956596374512, "learning_rate": 5e-05, "loss": 0.8989, "mean_token_accuracy": 0.8713293075561523, "num_tokens": 641563885.0, "step": 16898 }, { "epoch": 2.1497264979010304, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.067732810974121, "learning_rate": 5e-05, "loss": 0.8473, "mean_token_accuracy": 0.8827208876609802, "num_tokens": 641600868.0, "step": 16899 }, { "epoch": 2.149853708179621, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.790022373199463, "learning_rate": 5e-05, "loss": 0.8125, "mean_token_accuracy": 0.8814849257469177, "num_tokens": 641643936.0, "step": 16900 }, { "epoch": 2.1499809184582115, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.748928546905518, "learning_rate": 5e-05, "loss": 0.797, "mean_token_accuracy": 0.8794550895690918, "num_tokens": 641680248.0, "step": 16901 }, { "epoch": 2.150108128736802, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.015804290771484, "learning_rate": 5e-05, "loss": 0.7943, "mean_token_accuracy": 0.8902292847633362, "num_tokens": 641721280.0, "step": 16902 }, { "epoch": 2.1502353390153925, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.305984020233154, "learning_rate": 5e-05, "loss": 0.735, "mean_token_accuracy": 0.8817087411880493, "num_tokens": 641760168.0, "step": 16903 }, { "epoch": 2.150362549293983, "ewc_loss": 0.5, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.346454620361328, "learning_rate": 5e-05, "loss": 0.8729, "mean_token_accuracy": 0.8721944093704224, "num_tokens": 641798517.0, "step": 16904 }, { "epoch": 2.1504897595725736, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.066577434539795, "learning_rate": 5e-05, "loss": 0.7148, "mean_token_accuracy": 0.879951536655426, "num_tokens": 641841486.0, "step": 16905 }, { "epoch": 2.150616969851164, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.490460395812988, "learning_rate": 5e-05, "loss": 0.8683, "mean_token_accuracy": 0.8763061761856079, "num_tokens": 641880042.0, "step": 16906 }, { "epoch": 2.1507441801297547, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.043388843536377, "learning_rate": 5e-05, "loss": 0.6956, "mean_token_accuracy": 0.8865293264389038, "num_tokens": 641916069.0, "step": 16907 }, { "epoch": 2.150871390408345, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.551725387573242, "learning_rate": 5e-05, "loss": 0.8551, "mean_token_accuracy": 0.8842101097106934, "num_tokens": 641951298.0, "step": 16908 }, { "epoch": 2.1509986006869353, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.114595413208008, "learning_rate": 5e-05, "loss": 0.734, "mean_token_accuracy": 0.8855872750282288, "num_tokens": 641987577.0, "step": 16909 }, { "epoch": 2.1511258109655262, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.445480346679688, "learning_rate": 5e-05, "loss": 0.8972, "mean_token_accuracy": 0.8682805299758911, "num_tokens": 642031914.0, "step": 16910 }, { "epoch": 2.1512530212441163, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.349330425262451, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.8787616491317749, "num_tokens": 642069729.0, "step": 16911 }, { "epoch": 2.151380231522707, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.109485626220703, "learning_rate": 5e-05, "loss": 0.8111, "mean_token_accuracy": 0.8817447423934937, "num_tokens": 642106831.0, "step": 16912 }, { "epoch": 2.1515074418012974, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.6161041259765625, "learning_rate": 5e-05, "loss": 0.7381, "mean_token_accuracy": 0.8920141458511353, "num_tokens": 642142408.0, "step": 16913 }, { "epoch": 2.151634652079888, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.728046894073486, "learning_rate": 5e-05, "loss": 0.7787, "mean_token_accuracy": 0.8789971470832825, "num_tokens": 642181860.0, "step": 16914 }, { "epoch": 2.1517618623584784, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.801351547241211, "learning_rate": 5e-05, "loss": 0.728, "mean_token_accuracy": 0.8958183526992798, "num_tokens": 642219073.0, "step": 16915 }, { "epoch": 2.151889072637069, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.44317626953125, "learning_rate": 5e-05, "loss": 0.7249, "mean_token_accuracy": 0.8832287192344666, "num_tokens": 642254872.0, "step": 16916 }, { "epoch": 2.1520162829156595, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.918417930603027, "learning_rate": 5e-05, "loss": 0.7312, "mean_token_accuracy": 0.8918843865394592, "num_tokens": 642295379.0, "step": 16917 }, { "epoch": 2.15214349319425, "ewc_loss": 0.375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.2601094245910645, "learning_rate": 5e-05, "loss": 0.749, "mean_token_accuracy": 0.8690338134765625, "num_tokens": 642333316.0, "step": 16918 }, { "epoch": 2.1522707034728406, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.04910659790039, "learning_rate": 5e-05, "loss": 0.7814, "mean_token_accuracy": 0.8785631656646729, "num_tokens": 642373677.0, "step": 16919 }, { "epoch": 2.152397913751431, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.10083532333374, "learning_rate": 5e-05, "loss": 0.7187, "mean_token_accuracy": 0.875194787979126, "num_tokens": 642416066.0, "step": 16920 }, { "epoch": 2.1525251240300216, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.20746898651123, "learning_rate": 5e-05, "loss": 0.8124, "mean_token_accuracy": 0.8723207712173462, "num_tokens": 642456818.0, "step": 16921 }, { "epoch": 2.152652334308612, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.923422336578369, "learning_rate": 5e-05, "loss": 0.6914, "mean_token_accuracy": 0.8795609474182129, "num_tokens": 642498021.0, "step": 16922 }, { "epoch": 2.1527795445872027, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.423356056213379, "learning_rate": 5e-05, "loss": 0.8091, "mean_token_accuracy": 0.8753900527954102, "num_tokens": 642533999.0, "step": 16923 }, { "epoch": 2.152906754865793, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.752453804016113, "learning_rate": 5e-05, "loss": 0.6791, "mean_token_accuracy": 0.8836540579795837, "num_tokens": 642571943.0, "step": 16924 }, { "epoch": 2.1530339651443837, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.65590763092041, "learning_rate": 5e-05, "loss": 0.7931, "mean_token_accuracy": 0.8879051208496094, "num_tokens": 642605145.0, "step": 16925 }, { "epoch": 2.1531611754229742, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.698465824127197, "learning_rate": 5e-05, "loss": 0.6911, "mean_token_accuracy": 0.8787527084350586, "num_tokens": 642644686.0, "step": 16926 }, { "epoch": 2.1532883857015648, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.69920539855957, "learning_rate": 5e-05, "loss": 0.8128, "mean_token_accuracy": 0.8904314041137695, "num_tokens": 642677815.0, "step": 16927 }, { "epoch": 2.1534155959801553, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.902559280395508, "learning_rate": 5e-05, "loss": 0.6667, "mean_token_accuracy": 0.895615816116333, "num_tokens": 642716245.0, "step": 16928 }, { "epoch": 2.153542806258746, "ewc_loss": 0.5, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.429573059082031, "learning_rate": 5e-05, "loss": 0.8408, "mean_token_accuracy": 0.881926417350769, "num_tokens": 642750428.0, "step": 16929 }, { "epoch": 2.1536700165373364, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.295400142669678, "learning_rate": 5e-05, "loss": 0.7654, "mean_token_accuracy": 0.876194953918457, "num_tokens": 642790349.0, "step": 16930 }, { "epoch": 2.153797226815927, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.904334545135498, "learning_rate": 5e-05, "loss": 0.7686, "mean_token_accuracy": 0.8880307674407959, "num_tokens": 642835063.0, "step": 16931 }, { "epoch": 2.1539244370945174, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.736205577850342, "learning_rate": 5e-05, "loss": 0.7673, "mean_token_accuracy": 0.8807004690170288, "num_tokens": 642875013.0, "step": 16932 }, { "epoch": 2.154051647373108, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.3679304122924805, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8735122084617615, "num_tokens": 642914934.0, "step": 16933 }, { "epoch": 2.154178857651698, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.067939758300781, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8848699927330017, "num_tokens": 642949368.0, "step": 16934 }, { "epoch": 2.1543060679302886, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.088418006896973, "learning_rate": 5e-05, "loss": 0.7197, "mean_token_accuracy": 0.8743959665298462, "num_tokens": 642988083.0, "step": 16935 }, { "epoch": 2.154433278208879, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.163199424743652, "learning_rate": 5e-05, "loss": 0.8223, "mean_token_accuracy": 0.86855149269104, "num_tokens": 643028868.0, "step": 16936 }, { "epoch": 2.1545604884874696, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.0850043296813965, "learning_rate": 5e-05, "loss": 0.702, "mean_token_accuracy": 0.881306529045105, "num_tokens": 643072251.0, "step": 16937 }, { "epoch": 2.15468769876606, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.039777755737305, "learning_rate": 5e-05, "loss": 0.745, "mean_token_accuracy": 0.8869419097900391, "num_tokens": 643109864.0, "step": 16938 }, { "epoch": 2.1548149090446507, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.252037525177002, "learning_rate": 5e-05, "loss": 0.6741, "mean_token_accuracy": 0.8949263691902161, "num_tokens": 643147868.0, "step": 16939 }, { "epoch": 2.154942119323241, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.769533634185791, "learning_rate": 5e-05, "loss": 0.6552, "mean_token_accuracy": 0.9053528904914856, "num_tokens": 643182173.0, "step": 16940 }, { "epoch": 2.1550693296018317, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.537243843078613, "learning_rate": 5e-05, "loss": 0.7284, "mean_token_accuracy": 0.8774799108505249, "num_tokens": 643217006.0, "step": 16941 }, { "epoch": 2.1551965398804223, "ewc_loss": 0.375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.458783149719238, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8679913282394409, "num_tokens": 643253724.0, "step": 16942 }, { "epoch": 2.155323750159013, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.807102680206299, "learning_rate": 5e-05, "loss": 0.7635, "mean_token_accuracy": 0.8715595006942749, "num_tokens": 643294633.0, "step": 16943 }, { "epoch": 2.1554509604376033, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.135496616363525, "learning_rate": 5e-05, "loss": 0.6956, "mean_token_accuracy": 0.8752783536911011, "num_tokens": 643333269.0, "step": 16944 }, { "epoch": 2.155578170716194, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.156086921691895, "learning_rate": 5e-05, "loss": 0.7293, "mean_token_accuracy": 0.8915534615516663, "num_tokens": 643374892.0, "step": 16945 }, { "epoch": 2.1557053809947844, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.756180763244629, "learning_rate": 5e-05, "loss": 0.6303, "mean_token_accuracy": 0.8955941200256348, "num_tokens": 643411774.0, "step": 16946 }, { "epoch": 2.155832591273375, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.637496948242188, "learning_rate": 5e-05, "loss": 0.7651, "mean_token_accuracy": 0.8892408013343811, "num_tokens": 643450034.0, "step": 16947 }, { "epoch": 2.1559598015519654, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00024127960205078125, "grad_norm": 6.300283432006836, "learning_rate": 5e-05, "loss": 0.641, "mean_token_accuracy": 0.8822488188743591, "num_tokens": 643482901.0, "step": 16948 }, { "epoch": 2.156087011830556, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 9.322879791259766, "learning_rate": 5e-05, "loss": 0.8845, "mean_token_accuracy": 0.8800047636032104, "num_tokens": 643524918.0, "step": 16949 }, { "epoch": 2.1562142221091465, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.91778039932251, "learning_rate": 5e-05, "loss": 0.6719, "mean_token_accuracy": 0.8736640810966492, "num_tokens": 643563931.0, "step": 16950 }, { "epoch": 2.156341432387737, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 10.055209159851074, "learning_rate": 5e-05, "loss": 0.9769, "mean_token_accuracy": 0.881939172744751, "num_tokens": 643602838.0, "step": 16951 }, { "epoch": 2.1564686426663275, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.364201068878174, "learning_rate": 5e-05, "loss": 0.7303, "mean_token_accuracy": 0.8828462362289429, "num_tokens": 643641128.0, "step": 16952 }, { "epoch": 2.156595852944918, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.060136795043945, "learning_rate": 5e-05, "loss": 1.015, "mean_token_accuracy": 0.8871885538101196, "num_tokens": 643674853.0, "step": 16953 }, { "epoch": 2.1567230632235086, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 7.902853965759277, "learning_rate": 5e-05, "loss": 0.9669, "mean_token_accuracy": 0.8799059391021729, "num_tokens": 643709924.0, "step": 16954 }, { "epoch": 2.156850273502099, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.501310348510742, "learning_rate": 5e-05, "loss": 0.8888, "mean_token_accuracy": 0.8866846561431885, "num_tokens": 643744899.0, "step": 16955 }, { "epoch": 2.1569774837806897, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.163990020751953, "learning_rate": 5e-05, "loss": 0.9632, "mean_token_accuracy": 0.8957316875457764, "num_tokens": 643781316.0, "step": 16956 }, { "epoch": 2.15710469405928, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.655036926269531, "learning_rate": 5e-05, "loss": 0.7201, "mean_token_accuracy": 0.8839479684829712, "num_tokens": 643824267.0, "step": 16957 }, { "epoch": 2.1572319043378707, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.498754501342773, "learning_rate": 5e-05, "loss": 1.065, "mean_token_accuracy": 0.8820232152938843, "num_tokens": 643867155.0, "step": 16958 }, { "epoch": 2.157359114616461, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 6.8395304679870605, "learning_rate": 5e-05, "loss": 0.8196, "mean_token_accuracy": 0.8903729319572449, "num_tokens": 643907813.0, "step": 16959 }, { "epoch": 2.1574863248950513, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 10.7827730178833, "learning_rate": 5e-05, "loss": 1.0964, "mean_token_accuracy": 0.8886481523513794, "num_tokens": 643940387.0, "step": 16960 }, { "epoch": 2.157613535173642, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 8.46340274810791, "learning_rate": 5e-05, "loss": 1.0918, "mean_token_accuracy": 0.8749856352806091, "num_tokens": 643980387.0, "step": 16961 }, { "epoch": 2.1577407454522324, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.549040794372559, "learning_rate": 5e-05, "loss": 0.9581, "mean_token_accuracy": 0.8864352703094482, "num_tokens": 644020883.0, "step": 16962 }, { "epoch": 2.157867955730823, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 8.895039558410645, "learning_rate": 5e-05, "loss": 1.0945, "mean_token_accuracy": 0.8687509298324585, "num_tokens": 644054492.0, "step": 16963 }, { "epoch": 2.1579951660094134, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.409093856811523, "learning_rate": 5e-05, "loss": 0.9245, "mean_token_accuracy": 0.8719747066497803, "num_tokens": 644095009.0, "step": 16964 }, { "epoch": 2.158122376288004, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.704495429992676, "learning_rate": 5e-05, "loss": 0.9066, "mean_token_accuracy": 0.892987847328186, "num_tokens": 644128469.0, "step": 16965 }, { "epoch": 2.1582495865665945, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.14372730255127, "learning_rate": 5e-05, "loss": 0.9363, "mean_token_accuracy": 0.8670743703842163, "num_tokens": 644167102.0, "step": 16966 }, { "epoch": 2.158376796845185, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.963078498840332, "learning_rate": 5e-05, "loss": 0.8575, "mean_token_accuracy": 0.8800021409988403, "num_tokens": 644204319.0, "step": 16967 }, { "epoch": 2.1585040071237755, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.397693634033203, "learning_rate": 5e-05, "loss": 0.934, "mean_token_accuracy": 0.878392219543457, "num_tokens": 644246950.0, "step": 16968 }, { "epoch": 2.158631217402366, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.259236812591553, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8848927617073059, "num_tokens": 644288744.0, "step": 16969 }, { "epoch": 2.1587584276809566, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.746567726135254, "learning_rate": 5e-05, "loss": 0.899, "mean_token_accuracy": 0.8821038603782654, "num_tokens": 644326719.0, "step": 16970 }, { "epoch": 2.158885637959547, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.981130599975586, "learning_rate": 5e-05, "loss": 0.7412, "mean_token_accuracy": 0.8811131715774536, "num_tokens": 644362027.0, "step": 16971 }, { "epoch": 2.1590128482381377, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.939743041992188, "learning_rate": 5e-05, "loss": 0.9195, "mean_token_accuracy": 0.8823820352554321, "num_tokens": 644401917.0, "step": 16972 }, { "epoch": 2.159140058516728, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 6.921217441558838, "learning_rate": 5e-05, "loss": 0.7649, "mean_token_accuracy": 0.8774963617324829, "num_tokens": 644441735.0, "step": 16973 }, { "epoch": 2.1592672687953187, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.074982643127441, "learning_rate": 5e-05, "loss": 0.8646, "mean_token_accuracy": 0.897552490234375, "num_tokens": 644475833.0, "step": 16974 }, { "epoch": 2.1593944790739092, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.006099700927734, "learning_rate": 5e-05, "loss": 0.7901, "mean_token_accuracy": 0.8811553716659546, "num_tokens": 644516506.0, "step": 16975 }, { "epoch": 2.1595216893524998, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.003547668457031, "learning_rate": 5e-05, "loss": 0.8962, "mean_token_accuracy": 0.8895650506019592, "num_tokens": 644554081.0, "step": 16976 }, { "epoch": 2.1596488996310903, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.247626781463623, "learning_rate": 5e-05, "loss": 0.8652, "mean_token_accuracy": 0.8703734874725342, "num_tokens": 644599514.0, "step": 16977 }, { "epoch": 2.159776109909681, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.683358192443848, "learning_rate": 5e-05, "loss": 0.9208, "mean_token_accuracy": 0.8739107847213745, "num_tokens": 644639569.0, "step": 16978 }, { "epoch": 2.1599033201882714, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.433403491973877, "learning_rate": 5e-05, "loss": 0.7989, "mean_token_accuracy": 0.8969253301620483, "num_tokens": 644685721.0, "step": 16979 }, { "epoch": 2.160030530466862, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.42268180847168, "learning_rate": 5e-05, "loss": 0.9274, "mean_token_accuracy": 0.8675576448440552, "num_tokens": 644729939.0, "step": 16980 }, { "epoch": 2.1601577407454524, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.422468185424805, "learning_rate": 5e-05, "loss": 0.8399, "mean_token_accuracy": 0.8763333559036255, "num_tokens": 644768181.0, "step": 16981 }, { "epoch": 2.160284951024043, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.302367210388184, "learning_rate": 5e-05, "loss": 0.8906, "mean_token_accuracy": 0.8751277923583984, "num_tokens": 644806288.0, "step": 16982 }, { "epoch": 2.1604121613026335, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.361743927001953, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.8891481757164001, "num_tokens": 644846116.0, "step": 16983 }, { "epoch": 2.1605393715812236, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.271686553955078, "learning_rate": 5e-05, "loss": 0.873, "mean_token_accuracy": 0.8766037821769714, "num_tokens": 644885330.0, "step": 16984 }, { "epoch": 2.160666581859814, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.256762981414795, "learning_rate": 5e-05, "loss": 0.7861, "mean_token_accuracy": 0.8746567964553833, "num_tokens": 644923341.0, "step": 16985 }, { "epoch": 2.1607937921384046, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.360932350158691, "learning_rate": 5e-05, "loss": 0.8785, "mean_token_accuracy": 0.876859188079834, "num_tokens": 644964165.0, "step": 16986 }, { "epoch": 2.160921002416995, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.1294074058532715, "learning_rate": 5e-05, "loss": 0.6945, "mean_token_accuracy": 0.8960413932800293, "num_tokens": 644999319.0, "step": 16987 }, { "epoch": 2.1610482126955857, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.47566032409668, "learning_rate": 5e-05, "loss": 0.836, "mean_token_accuracy": 0.894291877746582, "num_tokens": 645040687.0, "step": 16988 }, { "epoch": 2.161175422974176, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.136474132537842, "learning_rate": 5e-05, "loss": 0.7483, "mean_token_accuracy": 0.8810679316520691, "num_tokens": 645075772.0, "step": 16989 }, { "epoch": 2.1613026332527667, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.364152908325195, "learning_rate": 5e-05, "loss": 0.869, "mean_token_accuracy": 0.8808304071426392, "num_tokens": 645111856.0, "step": 16990 }, { "epoch": 2.1614298435313573, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.328537940979004, "learning_rate": 5e-05, "loss": 0.783, "mean_token_accuracy": 0.8766309022903442, "num_tokens": 645155995.0, "step": 16991 }, { "epoch": 2.161557053809948, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.094082832336426, "learning_rate": 5e-05, "loss": 0.8702, "mean_token_accuracy": 0.8675757646560669, "num_tokens": 645200101.0, "step": 16992 }, { "epoch": 2.1616842640885383, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.5197248458862305, "learning_rate": 5e-05, "loss": 0.7408, "mean_token_accuracy": 0.8944146037101746, "num_tokens": 645235243.0, "step": 16993 }, { "epoch": 2.161811474367129, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.789815425872803, "learning_rate": 5e-05, "loss": 0.7503, "mean_token_accuracy": 0.8936188220977783, "num_tokens": 645270369.0, "step": 16994 }, { "epoch": 2.1619386846457194, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.67805814743042, "learning_rate": 5e-05, "loss": 0.7321, "mean_token_accuracy": 0.8942680358886719, "num_tokens": 645304111.0, "step": 16995 }, { "epoch": 2.16206589492431, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.548770904541016, "learning_rate": 5e-05, "loss": 0.7672, "mean_token_accuracy": 0.871620237827301, "num_tokens": 645339884.0, "step": 16996 }, { "epoch": 2.1621931052029004, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.792111396789551, "learning_rate": 5e-05, "loss": 0.704, "mean_token_accuracy": 0.8988049030303955, "num_tokens": 645372465.0, "step": 16997 }, { "epoch": 2.162320315481491, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.412877082824707, "learning_rate": 5e-05, "loss": 0.7068, "mean_token_accuracy": 0.8888250589370728, "num_tokens": 645413022.0, "step": 16998 }, { "epoch": 2.1624475257600815, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.790647506713867, "learning_rate": 5e-05, "loss": 0.7093, "mean_token_accuracy": 0.8965137004852295, "num_tokens": 645451344.0, "step": 16999 }, { "epoch": 2.162574736038672, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.382534027099609, "learning_rate": 5e-05, "loss": 0.6462, "mean_token_accuracy": 0.9008259773254395, "num_tokens": 645493463.0, "step": 17000 }, { "epoch": 2.1627019463172625, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.776546001434326, "learning_rate": 5e-05, "loss": 0.7148, "mean_token_accuracy": 0.8910195827484131, "num_tokens": 645533508.0, "step": 17001 }, { "epoch": 2.162829156595853, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.355379104614258, "learning_rate": 5e-05, "loss": 0.697, "mean_token_accuracy": 0.8872336149215698, "num_tokens": 645573866.0, "step": 17002 }, { "epoch": 2.1629563668744436, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.79123067855835, "learning_rate": 5e-05, "loss": 0.7024, "mean_token_accuracy": 0.890483558177948, "num_tokens": 645609835.0, "step": 17003 }, { "epoch": 2.163083577153034, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.309945583343506, "learning_rate": 5e-05, "loss": 0.6861, "mean_token_accuracy": 0.8868488669395447, "num_tokens": 645644228.0, "step": 17004 }, { "epoch": 2.1632107874316246, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.810060501098633, "learning_rate": 5e-05, "loss": 0.7208, "mean_token_accuracy": 0.8876620531082153, "num_tokens": 645682619.0, "step": 17005 }, { "epoch": 2.163337997710215, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.24912691116333, "learning_rate": 5e-05, "loss": 0.6684, "mean_token_accuracy": 0.8871558308601379, "num_tokens": 645719506.0, "step": 17006 }, { "epoch": 2.1634652079888053, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.917311668395996, "learning_rate": 5e-05, "loss": 0.741, "mean_token_accuracy": 0.8773482441902161, "num_tokens": 645761930.0, "step": 17007 }, { "epoch": 2.1635924182673962, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.0826568603515625, "learning_rate": 5e-05, "loss": 0.7134, "mean_token_accuracy": 0.8726381063461304, "num_tokens": 645800322.0, "step": 17008 }, { "epoch": 2.1637196285459863, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.131818771362305, "learning_rate": 5e-05, "loss": 0.7545, "mean_token_accuracy": 0.8811442852020264, "num_tokens": 645842914.0, "step": 17009 }, { "epoch": 2.163846838824577, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000675201416015625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.822597026824951, "learning_rate": 5e-05, "loss": 0.6869, "mean_token_accuracy": 0.876460075378418, "num_tokens": 645885798.0, "step": 17010 }, { "epoch": 2.1639740491031674, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.524794578552246, "learning_rate": 5e-05, "loss": 0.7847, "mean_token_accuracy": 0.8825886249542236, "num_tokens": 645922187.0, "step": 17011 }, { "epoch": 2.164101259381758, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.411032676696777, "learning_rate": 5e-05, "loss": 0.6085, "mean_token_accuracy": 0.8945342898368835, "num_tokens": 645955594.0, "step": 17012 }, { "epoch": 2.1642284696603484, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 9.170817375183105, "learning_rate": 5e-05, "loss": 0.8334, "mean_token_accuracy": 0.8917362093925476, "num_tokens": 645992579.0, "step": 17013 }, { "epoch": 2.164355679938939, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 5.973034858703613, "learning_rate": 5e-05, "loss": 0.6456, "mean_token_accuracy": 0.8801864385604858, "num_tokens": 646033243.0, "step": 17014 }, { "epoch": 2.1644828902175295, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.964414596557617, "learning_rate": 5e-05, "loss": 0.939, "mean_token_accuracy": 0.8930668234825134, "num_tokens": 646069309.0, "step": 17015 }, { "epoch": 2.16461010049612, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.248264312744141, "learning_rate": 5e-05, "loss": 0.7077, "mean_token_accuracy": 0.8847712278366089, "num_tokens": 646100720.0, "step": 17016 }, { "epoch": 2.1647373107747105, "ewc_loss": 0.703125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.089531898498535, "learning_rate": 5e-05, "loss": 0.973, "mean_token_accuracy": 0.9017065763473511, "num_tokens": 646135917.0, "step": 17017 }, { "epoch": 2.164864521053301, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 7.826255798339844, "learning_rate": 5e-05, "loss": 0.9148, "mean_token_accuracy": 0.8972241282463074, "num_tokens": 646173227.0, "step": 17018 }, { "epoch": 2.1649917313318916, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.466726303100586, "learning_rate": 5e-05, "loss": 0.9167, "mean_token_accuracy": 0.8751362562179565, "num_tokens": 646210023.0, "step": 17019 }, { "epoch": 2.165118941610482, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 9.35805606842041, "learning_rate": 5e-05, "loss": 1.0179, "mean_token_accuracy": 0.8880187273025513, "num_tokens": 646241124.0, "step": 17020 }, { "epoch": 2.1652461518890727, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.28557825088501, "learning_rate": 5e-05, "loss": 0.7306, "mean_token_accuracy": 0.8698403239250183, "num_tokens": 646276892.0, "step": 17021 }, { "epoch": 2.165373362167663, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 11.658167839050293, "learning_rate": 5e-05, "loss": 1.1971, "mean_token_accuracy": 0.8955488204956055, "num_tokens": 646308959.0, "step": 17022 }, { "epoch": 2.1655005724462537, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.376002788543701, "learning_rate": 5e-05, "loss": 0.9016, "mean_token_accuracy": 0.8885968923568726, "num_tokens": 646342614.0, "step": 17023 }, { "epoch": 2.1656277827248442, "ewc_loss": 1.0078125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00093841552734375, "grad_norm": 12.839699745178223, "learning_rate": 5e-05, "loss": 1.3355, "mean_token_accuracy": 0.8822213411331177, "num_tokens": 646380395.0, "step": 17024 }, { "epoch": 2.1657549930034348, "ewc_loss": 1.1171875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00104522705078125, "grad_norm": 10.543761253356934, "learning_rate": 5e-05, "loss": 1.4836, "mean_token_accuracy": 0.8741070032119751, "num_tokens": 646417068.0, "step": 17025 }, { "epoch": 2.1658822032820253, "ewc_loss": 0.765625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 11.39857292175293, "learning_rate": 5e-05, "loss": 1.098, "mean_token_accuracy": 0.879852831363678, "num_tokens": 646456655.0, "step": 17026 }, { "epoch": 2.166009413560616, "ewc_loss": 0.83203125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 9.668174743652344, "learning_rate": 5e-05, "loss": 1.1073, "mean_token_accuracy": 0.9022667407989502, "num_tokens": 646495035.0, "step": 17027 }, { "epoch": 2.1661366238392064, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000766754150390625, "grad_norm": 11.118066787719727, "learning_rate": 5e-05, "loss": 1.1192, "mean_token_accuracy": 0.8959518671035767, "num_tokens": 646529695.0, "step": 17028 }, { "epoch": 2.166263834117797, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.887389183044434, "learning_rate": 5e-05, "loss": 1.0155, "mean_token_accuracy": 0.86799556016922, "num_tokens": 646566563.0, "step": 17029 }, { "epoch": 2.1663910443963874, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 10.785804748535156, "learning_rate": 5e-05, "loss": 1.1362, "mean_token_accuracy": 0.8792357444763184, "num_tokens": 646609515.0, "step": 17030 }, { "epoch": 2.166518254674978, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.815048217773438, "learning_rate": 5e-05, "loss": 0.9486, "mean_token_accuracy": 0.891779899597168, "num_tokens": 646650024.0, "step": 17031 }, { "epoch": 2.166645464953568, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.522241592407227, "learning_rate": 5e-05, "loss": 1.0328, "mean_token_accuracy": 0.8916569352149963, "num_tokens": 646690547.0, "step": 17032 }, { "epoch": 2.1667726752321586, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.852611541748047, "learning_rate": 5e-05, "loss": 0.9717, "mean_token_accuracy": 0.891947865486145, "num_tokens": 646722060.0, "step": 17033 }, { "epoch": 2.166899885510749, "ewc_loss": 0.6875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.144699096679688, "learning_rate": 5e-05, "loss": 1.0284, "mean_token_accuracy": 0.8800451755523682, "num_tokens": 646761972.0, "step": 17034 }, { "epoch": 2.1670270957893396, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.604127883911133, "learning_rate": 5e-05, "loss": 0.9908, "mean_token_accuracy": 0.8716869950294495, "num_tokens": 646794847.0, "step": 17035 }, { "epoch": 2.16715430606793, "ewc_loss": 0.6875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.958268165588379, "learning_rate": 5e-05, "loss": 0.9932, "mean_token_accuracy": 0.8915345668792725, "num_tokens": 646833744.0, "step": 17036 }, { "epoch": 2.1672815163465207, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.200584411621094, "learning_rate": 5e-05, "loss": 0.8626, "mean_token_accuracy": 0.8880423307418823, "num_tokens": 646874257.0, "step": 17037 }, { "epoch": 2.167408726625111, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.2000093460083, "learning_rate": 5e-05, "loss": 1.0199, "mean_token_accuracy": 0.8951182961463928, "num_tokens": 646912439.0, "step": 17038 }, { "epoch": 2.1675359369037017, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.194283485412598, "learning_rate": 5e-05, "loss": 0.8513, "mean_token_accuracy": 0.8939348459243774, "num_tokens": 646946311.0, "step": 17039 }, { "epoch": 2.1676631471822922, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 10.38525104522705, "learning_rate": 5e-05, "loss": 1.1038, "mean_token_accuracy": 0.8717127442359924, "num_tokens": 646982426.0, "step": 17040 }, { "epoch": 2.1677903574608828, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.684916496276855, "learning_rate": 5e-05, "loss": 0.8989, "mean_token_accuracy": 0.8955145478248596, "num_tokens": 647023244.0, "step": 17041 }, { "epoch": 2.1679175677394733, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.89929485321045, "learning_rate": 5e-05, "loss": 1.0613, "mean_token_accuracy": 0.8683309555053711, "num_tokens": 647062434.0, "step": 17042 }, { "epoch": 2.168044778018064, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.852425575256348, "learning_rate": 5e-05, "loss": 0.9488, "mean_token_accuracy": 0.8787826895713806, "num_tokens": 647098021.0, "step": 17043 }, { "epoch": 2.1681719882966544, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.158170700073242, "learning_rate": 5e-05, "loss": 0.9554, "mean_token_accuracy": 0.8801298141479492, "num_tokens": 647134709.0, "step": 17044 }, { "epoch": 2.168299198575245, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.478516578674316, "learning_rate": 5e-05, "loss": 0.9013, "mean_token_accuracy": 0.8709883093833923, "num_tokens": 647169267.0, "step": 17045 }, { "epoch": 2.1684264088538354, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.842437744140625, "learning_rate": 5e-05, "loss": 0.9168, "mean_token_accuracy": 0.8815947771072388, "num_tokens": 647207399.0, "step": 17046 }, { "epoch": 2.168553619132426, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.872565746307373, "learning_rate": 5e-05, "loss": 0.7987, "mean_token_accuracy": 0.8813124895095825, "num_tokens": 647244722.0, "step": 17047 }, { "epoch": 2.1686808294110165, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.069070816040039, "learning_rate": 5e-05, "loss": 0.9068, "mean_token_accuracy": 0.8897807002067566, "num_tokens": 647279051.0, "step": 17048 }, { "epoch": 2.168808039689607, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.415731906890869, "learning_rate": 5e-05, "loss": 0.7586, "mean_token_accuracy": 0.8817055225372314, "num_tokens": 647319201.0, "step": 17049 }, { "epoch": 2.1689352499681975, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.560617446899414, "learning_rate": 5e-05, "loss": 0.9842, "mean_token_accuracy": 0.8814956545829773, "num_tokens": 647358306.0, "step": 17050 }, { "epoch": 2.169062460246788, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.600905895233154, "learning_rate": 5e-05, "loss": 0.8114, "mean_token_accuracy": 0.8765906095504761, "num_tokens": 647393879.0, "step": 17051 }, { "epoch": 2.1691896705253786, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.61618423461914, "learning_rate": 5e-05, "loss": 1.0091, "mean_token_accuracy": 0.8770171403884888, "num_tokens": 647434848.0, "step": 17052 }, { "epoch": 2.169316880803969, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.2302885055542, "learning_rate": 5e-05, "loss": 0.8488, "mean_token_accuracy": 0.891042947769165, "num_tokens": 647469711.0, "step": 17053 }, { "epoch": 2.1694440910825596, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.880537033081055, "learning_rate": 5e-05, "loss": 0.9719, "mean_token_accuracy": 0.8685522675514221, "num_tokens": 647507788.0, "step": 17054 }, { "epoch": 2.16957130136115, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.40197467803955, "learning_rate": 5e-05, "loss": 0.8764, "mean_token_accuracy": 0.8831984996795654, "num_tokens": 647542193.0, "step": 17055 }, { "epoch": 2.1696985116397407, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.126690864562988, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.8977993130683899, "num_tokens": 647578692.0, "step": 17056 }, { "epoch": 2.169825721918331, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.08060073852539, "learning_rate": 5e-05, "loss": 0.7874, "mean_token_accuracy": 0.8889892101287842, "num_tokens": 647618158.0, "step": 17057 }, { "epoch": 2.1699529321969213, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.998289108276367, "learning_rate": 5e-05, "loss": 0.8032, "mean_token_accuracy": 0.884280264377594, "num_tokens": 647660788.0, "step": 17058 }, { "epoch": 2.170080142475512, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.651481628417969, "learning_rate": 5e-05, "loss": 0.7611, "mean_token_accuracy": 0.8743683695793152, "num_tokens": 647700418.0, "step": 17059 }, { "epoch": 2.1702073527541024, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.134735107421875, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.8885276317596436, "num_tokens": 647738248.0, "step": 17060 }, { "epoch": 2.170334563032693, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.340187072753906, "learning_rate": 5e-05, "loss": 0.6895, "mean_token_accuracy": 0.8921905159950256, "num_tokens": 647777287.0, "step": 17061 }, { "epoch": 2.1704617733112834, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.295042991638184, "learning_rate": 5e-05, "loss": 0.8045, "mean_token_accuracy": 0.8790751099586487, "num_tokens": 647823213.0, "step": 17062 }, { "epoch": 2.170588983589874, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.136333465576172, "learning_rate": 5e-05, "loss": 0.6612, "mean_token_accuracy": 0.8969008922576904, "num_tokens": 647861506.0, "step": 17063 }, { "epoch": 2.1707161938684645, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.476305961608887, "learning_rate": 5e-05, "loss": 0.7785, "mean_token_accuracy": 0.8928260207176208, "num_tokens": 647900039.0, "step": 17064 }, { "epoch": 2.170843404147055, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.999269008636475, "learning_rate": 5e-05, "loss": 0.7028, "mean_token_accuracy": 0.8800053596496582, "num_tokens": 647938368.0, "step": 17065 }, { "epoch": 2.1709706144256455, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.689963340759277, "learning_rate": 5e-05, "loss": 0.8739, "mean_token_accuracy": 0.8719974756240845, "num_tokens": 647972549.0, "step": 17066 }, { "epoch": 2.171097824704236, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.994380950927734, "learning_rate": 5e-05, "loss": 0.7013, "mean_token_accuracy": 0.8807438015937805, "num_tokens": 648016923.0, "step": 17067 }, { "epoch": 2.1712250349828266, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.70232105255127, "learning_rate": 5e-05, "loss": 0.8383, "mean_token_accuracy": 0.8880933523178101, "num_tokens": 648059880.0, "step": 17068 }, { "epoch": 2.171352245261417, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.299610137939453, "learning_rate": 5e-05, "loss": 0.7397, "mean_token_accuracy": 0.8806856274604797, "num_tokens": 648093657.0, "step": 17069 }, { "epoch": 2.1714794555400077, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.258051872253418, "learning_rate": 5e-05, "loss": 0.792, "mean_token_accuracy": 0.8866575956344604, "num_tokens": 648130723.0, "step": 17070 }, { "epoch": 2.171606665818598, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.723742961883545, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.8746593594551086, "num_tokens": 648176657.0, "step": 17071 }, { "epoch": 2.1717338760971887, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.628136157989502, "learning_rate": 5e-05, "loss": 0.7369, "mean_token_accuracy": 0.8879219889640808, "num_tokens": 648214505.0, "step": 17072 }, { "epoch": 2.1718610863757792, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.078922271728516, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.891181230545044, "num_tokens": 648247816.0, "step": 17073 }, { "epoch": 2.1719882966543698, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.116710662841797, "learning_rate": 5e-05, "loss": 0.685, "mean_token_accuracy": 0.8857721090316772, "num_tokens": 648276418.0, "step": 17074 }, { "epoch": 2.1721155069329603, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.443531036376953, "learning_rate": 5e-05, "loss": 0.8169, "mean_token_accuracy": 0.8761937618255615, "num_tokens": 648314722.0, "step": 17075 }, { "epoch": 2.172242717211551, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.746884346008301, "learning_rate": 5e-05, "loss": 0.6872, "mean_token_accuracy": 0.8759258985519409, "num_tokens": 648354733.0, "step": 17076 }, { "epoch": 2.1723699274901414, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.843852996826172, "learning_rate": 5e-05, "loss": 0.84, "mean_token_accuracy": 0.881603479385376, "num_tokens": 648392164.0, "step": 17077 }, { "epoch": 2.172497137768732, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.5795674324035645, "learning_rate": 5e-05, "loss": 0.6587, "mean_token_accuracy": 0.8872635364532471, "num_tokens": 648431728.0, "step": 17078 }, { "epoch": 2.1726243480473224, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.03469467163086, "learning_rate": 5e-05, "loss": 0.8938, "mean_token_accuracy": 0.8779594898223877, "num_tokens": 648471225.0, "step": 17079 }, { "epoch": 2.172751558325913, "ewc_loss": 0.375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.921749114990234, "learning_rate": 5e-05, "loss": 0.708, "mean_token_accuracy": 0.8809946775436401, "num_tokens": 648507742.0, "step": 17080 }, { "epoch": 2.1728787686045035, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.611477851867676, "learning_rate": 5e-05, "loss": 0.8691, "mean_token_accuracy": 0.8813562393188477, "num_tokens": 648548641.0, "step": 17081 }, { "epoch": 2.1730059788830935, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.6176323890686035, "learning_rate": 5e-05, "loss": 0.7315, "mean_token_accuracy": 0.8933073282241821, "num_tokens": 648583540.0, "step": 17082 }, { "epoch": 2.173133189161684, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.680621147155762, "learning_rate": 5e-05, "loss": 0.7298, "mean_token_accuracy": 0.8946782350540161, "num_tokens": 648627225.0, "step": 17083 }, { "epoch": 2.1732603994402746, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.214949607849121, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8984148502349854, "num_tokens": 648662760.0, "step": 17084 }, { "epoch": 2.173387609718865, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.105795860290527, "learning_rate": 5e-05, "loss": 0.7002, "mean_token_accuracy": 0.8783243894577026, "num_tokens": 648697748.0, "step": 17085 }, { "epoch": 2.1735148199974557, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.41858959197998, "learning_rate": 5e-05, "loss": 0.8103, "mean_token_accuracy": 0.8771674633026123, "num_tokens": 648732981.0, "step": 17086 }, { "epoch": 2.173642030276046, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.085062026977539, "learning_rate": 5e-05, "loss": 0.6818, "mean_token_accuracy": 0.8886339068412781, "num_tokens": 648764356.0, "step": 17087 }, { "epoch": 2.1737692405546367, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.192886352539062, "learning_rate": 5e-05, "loss": 0.7806, "mean_token_accuracy": 0.8813220262527466, "num_tokens": 648809064.0, "step": 17088 }, { "epoch": 2.1738964508332272, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.377209663391113, "learning_rate": 5e-05, "loss": 0.6898, "mean_token_accuracy": 0.8899509310722351, "num_tokens": 648843299.0, "step": 17089 }, { "epoch": 2.1740236611118178, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.7718505859375, "learning_rate": 5e-05, "loss": 0.7564, "mean_token_accuracy": 0.8772621154785156, "num_tokens": 648881251.0, "step": 17090 }, { "epoch": 2.1741508713904083, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.704769134521484, "learning_rate": 5e-05, "loss": 0.7448, "mean_token_accuracy": 0.8796208500862122, "num_tokens": 648915837.0, "step": 17091 }, { "epoch": 2.174278081668999, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.41791296005249, "learning_rate": 5e-05, "loss": 0.7264, "mean_token_accuracy": 0.8749170303344727, "num_tokens": 648953354.0, "step": 17092 }, { "epoch": 2.1744052919475894, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.938446521759033, "learning_rate": 5e-05, "loss": 0.724, "mean_token_accuracy": 0.8879190683364868, "num_tokens": 648993224.0, "step": 17093 }, { "epoch": 2.17453250222618, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.230644226074219, "learning_rate": 5e-05, "loss": 0.6846, "mean_token_accuracy": 0.8818367719650269, "num_tokens": 649026429.0, "step": 17094 }, { "epoch": 2.1746597125047704, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.043591499328613, "learning_rate": 5e-05, "loss": 0.7238, "mean_token_accuracy": 0.8873702883720398, "num_tokens": 649064672.0, "step": 17095 }, { "epoch": 2.174786922783361, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.1291327476501465, "learning_rate": 5e-05, "loss": 0.6795, "mean_token_accuracy": 0.8861715793609619, "num_tokens": 649097926.0, "step": 17096 }, { "epoch": 2.1749141330619515, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.13744068145752, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8802176713943481, "num_tokens": 649138502.0, "step": 17097 }, { "epoch": 2.175041343340542, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.0601911544799805, "learning_rate": 5e-05, "loss": 0.708, "mean_token_accuracy": 0.8737602233886719, "num_tokens": 649175867.0, "step": 17098 }, { "epoch": 2.1751685536191325, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.217649459838867, "learning_rate": 5e-05, "loss": 0.7228, "mean_token_accuracy": 0.896214485168457, "num_tokens": 649214132.0, "step": 17099 }, { "epoch": 2.175295763897723, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.943530559539795, "learning_rate": 5e-05, "loss": 0.6848, "mean_token_accuracy": 0.879994809627533, "num_tokens": 649262397.0, "step": 17100 }, { "epoch": 2.1754229741763136, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.418915748596191, "learning_rate": 5e-05, "loss": 0.8309, "mean_token_accuracy": 0.868352472782135, "num_tokens": 649298165.0, "step": 17101 }, { "epoch": 2.175550184454904, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.6738996505737305, "learning_rate": 5e-05, "loss": 0.6654, "mean_token_accuracy": 0.8792359828948975, "num_tokens": 649337802.0, "step": 17102 }, { "epoch": 2.1756773947334946, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.854217529296875, "learning_rate": 5e-05, "loss": 0.7968, "mean_token_accuracy": 0.8923993706703186, "num_tokens": 649378922.0, "step": 17103 }, { "epoch": 2.175804605012085, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.299814701080322, "learning_rate": 5e-05, "loss": 0.6195, "mean_token_accuracy": 0.8936161398887634, "num_tokens": 649414332.0, "step": 17104 }, { "epoch": 2.1759318152906753, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.46007251739502, "learning_rate": 5e-05, "loss": 0.8749, "mean_token_accuracy": 0.8928583264350891, "num_tokens": 649448965.0, "step": 17105 }, { "epoch": 2.1760590255692662, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.176414489746094, "learning_rate": 5e-05, "loss": 0.6756, "mean_token_accuracy": 0.8814184665679932, "num_tokens": 649487322.0, "step": 17106 }, { "epoch": 2.1761862358478563, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.864255905151367, "learning_rate": 5e-05, "loss": 1.0478, "mean_token_accuracy": 0.8577336668968201, "num_tokens": 649526524.0, "step": 17107 }, { "epoch": 2.176313446126447, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 6.986186504364014, "learning_rate": 5e-05, "loss": 0.8253, "mean_token_accuracy": 0.877561628818512, "num_tokens": 649566169.0, "step": 17108 }, { "epoch": 2.1764406564050374, "ewc_loss": 0.625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.222782135009766, "learning_rate": 5e-05, "loss": 0.9175, "mean_token_accuracy": 0.8931370973587036, "num_tokens": 649603725.0, "step": 17109 }, { "epoch": 2.176567866683628, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.255115509033203, "learning_rate": 5e-05, "loss": 0.9515, "mean_token_accuracy": 0.8779430389404297, "num_tokens": 649644789.0, "step": 17110 }, { "epoch": 2.1766950769622184, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.5099029541015625, "learning_rate": 5e-05, "loss": 0.8024, "mean_token_accuracy": 0.8768005967140198, "num_tokens": 649682051.0, "step": 17111 }, { "epoch": 2.176822287240809, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.281759262084961, "learning_rate": 5e-05, "loss": 0.9147, "mean_token_accuracy": 0.8953179717063904, "num_tokens": 649715730.0, "step": 17112 }, { "epoch": 2.1769494975193995, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.421454906463623, "learning_rate": 5e-05, "loss": 0.7079, "mean_token_accuracy": 0.8802825212478638, "num_tokens": 649750237.0, "step": 17113 }, { "epoch": 2.17707670779799, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.908159255981445, "learning_rate": 5e-05, "loss": 0.9507, "mean_token_accuracy": 0.8935670852661133, "num_tokens": 649791250.0, "step": 17114 }, { "epoch": 2.1772039180765805, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 6.871740818023682, "learning_rate": 5e-05, "loss": 0.798, "mean_token_accuracy": 0.8841229677200317, "num_tokens": 649831750.0, "step": 17115 }, { "epoch": 2.177331128355171, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.478063583374023, "learning_rate": 5e-05, "loss": 0.9879, "mean_token_accuracy": 0.8765976428985596, "num_tokens": 649869245.0, "step": 17116 }, { "epoch": 2.1774583386337616, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.000567436218262, "learning_rate": 5e-05, "loss": 0.9473, "mean_token_accuracy": 0.8799972534179688, "num_tokens": 649903763.0, "step": 17117 }, { "epoch": 2.177585548912352, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.154518127441406, "learning_rate": 5e-05, "loss": 0.8092, "mean_token_accuracy": 0.893966019153595, "num_tokens": 649946601.0, "step": 17118 }, { "epoch": 2.1777127591909426, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.658799171447754, "learning_rate": 5e-05, "loss": 0.9063, "mean_token_accuracy": 0.8922147154808044, "num_tokens": 649979308.0, "step": 17119 }, { "epoch": 2.177839969469533, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.313640117645264, "learning_rate": 5e-05, "loss": 0.7371, "mean_token_accuracy": 0.8947774171829224, "num_tokens": 650018068.0, "step": 17120 }, { "epoch": 2.1779671797481237, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.72064208984375, "learning_rate": 5e-05, "loss": 0.9285, "mean_token_accuracy": 0.8773694634437561, "num_tokens": 650059559.0, "step": 17121 }, { "epoch": 2.1780943900267142, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.389165878295898, "learning_rate": 5e-05, "loss": 0.7811, "mean_token_accuracy": 0.8858190774917603, "num_tokens": 650103266.0, "step": 17122 }, { "epoch": 2.1782216003053048, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.226293563842773, "learning_rate": 5e-05, "loss": 0.8593, "mean_token_accuracy": 0.8798348307609558, "num_tokens": 650139525.0, "step": 17123 }, { "epoch": 2.1783488105838953, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.732461929321289, "learning_rate": 5e-05, "loss": 0.7619, "mean_token_accuracy": 0.8985377550125122, "num_tokens": 650179249.0, "step": 17124 }, { "epoch": 2.178476020862486, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.712512969970703, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8733607530593872, "num_tokens": 650217258.0, "step": 17125 }, { "epoch": 2.1786032311410763, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.893468379974365, "learning_rate": 5e-05, "loss": 0.7424, "mean_token_accuracy": 0.9002439379692078, "num_tokens": 650261923.0, "step": 17126 }, { "epoch": 2.178730441419667, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.460211277008057, "learning_rate": 5e-05, "loss": 0.7153, "mean_token_accuracy": 0.890500009059906, "num_tokens": 650295562.0, "step": 17127 }, { "epoch": 2.1788576516982574, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.999407768249512, "learning_rate": 5e-05, "loss": 0.781, "mean_token_accuracy": 0.8841896057128906, "num_tokens": 650331988.0, "step": 17128 }, { "epoch": 2.178984861976848, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.283956050872803, "learning_rate": 5e-05, "loss": 0.6913, "mean_token_accuracy": 0.8900173902511597, "num_tokens": 650370572.0, "step": 17129 }, { "epoch": 2.179112072255438, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.153197288513184, "learning_rate": 5e-05, "loss": 0.7567, "mean_token_accuracy": 0.8920621275901794, "num_tokens": 650412838.0, "step": 17130 }, { "epoch": 2.1792392825340285, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.114869117736816, "learning_rate": 5e-05, "loss": 0.6944, "mean_token_accuracy": 0.881901741027832, "num_tokens": 650455588.0, "step": 17131 }, { "epoch": 2.179366492812619, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.249805450439453, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8893693089485168, "num_tokens": 650496369.0, "step": 17132 }, { "epoch": 2.1794937030912096, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.003676414489746, "learning_rate": 5e-05, "loss": 0.7155, "mean_token_accuracy": 0.8733570575714111, "num_tokens": 650535366.0, "step": 17133 }, { "epoch": 2.1796209133698, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.365880012512207, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8938814401626587, "num_tokens": 650574171.0, "step": 17134 }, { "epoch": 2.1797481236483907, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.843434810638428, "learning_rate": 5e-05, "loss": 0.6403, "mean_token_accuracy": 0.897361159324646, "num_tokens": 650612132.0, "step": 17135 }, { "epoch": 2.179875333926981, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.621593475341797, "learning_rate": 5e-05, "loss": 0.7894, "mean_token_accuracy": 0.8923722505569458, "num_tokens": 650648813.0, "step": 17136 }, { "epoch": 2.1800025442055717, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.608736991882324, "learning_rate": 5e-05, "loss": 0.6621, "mean_token_accuracy": 0.8912200927734375, "num_tokens": 650678807.0, "step": 17137 }, { "epoch": 2.1801297544841622, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.982155799865723, "learning_rate": 5e-05, "loss": 0.8721, "mean_token_accuracy": 0.8746105432510376, "num_tokens": 650716011.0, "step": 17138 }, { "epoch": 2.1802569647627528, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.43709659576416, "learning_rate": 5e-05, "loss": 0.6375, "mean_token_accuracy": 0.8967550992965698, "num_tokens": 650749763.0, "step": 17139 }, { "epoch": 2.1803841750413433, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.29228687286377, "learning_rate": 5e-05, "loss": 0.9221, "mean_token_accuracy": 0.8840063810348511, "num_tokens": 650786977.0, "step": 17140 }, { "epoch": 2.180511385319934, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.641113758087158, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8707983493804932, "num_tokens": 650823293.0, "step": 17141 }, { "epoch": 2.1806385955985244, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.211305618286133, "learning_rate": 5e-05, "loss": 0.9444, "mean_token_accuracy": 0.8814245462417603, "num_tokens": 650865037.0, "step": 17142 }, { "epoch": 2.180765805877115, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.288614749908447, "learning_rate": 5e-05, "loss": 0.8143, "mean_token_accuracy": 0.8875346183776855, "num_tokens": 650901803.0, "step": 17143 }, { "epoch": 2.1808930161557054, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.480788230895996, "learning_rate": 5e-05, "loss": 0.8624, "mean_token_accuracy": 0.8890082836151123, "num_tokens": 650942163.0, "step": 17144 }, { "epoch": 2.181020226434296, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 7.995389938354492, "learning_rate": 5e-05, "loss": 0.8787, "mean_token_accuracy": 0.8802059888839722, "num_tokens": 650977989.0, "step": 17145 }, { "epoch": 2.1811474367128865, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.482875823974609, "learning_rate": 5e-05, "loss": 0.7669, "mean_token_accuracy": 0.8874553442001343, "num_tokens": 651016831.0, "step": 17146 }, { "epoch": 2.181274646991477, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.439986228942871, "learning_rate": 5e-05, "loss": 0.8855, "mean_token_accuracy": 0.8775655627250671, "num_tokens": 651052493.0, "step": 17147 }, { "epoch": 2.1814018572700675, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.052662372589111, "learning_rate": 5e-05, "loss": 0.7691, "mean_token_accuracy": 0.869242787361145, "num_tokens": 651093140.0, "step": 17148 }, { "epoch": 2.181529067548658, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.445070266723633, "learning_rate": 5e-05, "loss": 0.8518, "mean_token_accuracy": 0.8770177364349365, "num_tokens": 651129367.0, "step": 17149 }, { "epoch": 2.1816562778272486, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.156822204589844, "learning_rate": 5e-05, "loss": 0.7182, "mean_token_accuracy": 0.8912091255187988, "num_tokens": 651169573.0, "step": 17150 }, { "epoch": 2.181783488105839, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.178421020507812, "learning_rate": 5e-05, "loss": 0.784, "mean_token_accuracy": 0.8921006917953491, "num_tokens": 651206771.0, "step": 17151 }, { "epoch": 2.1819106983844296, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.344104290008545, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.8924049139022827, "num_tokens": 651251322.0, "step": 17152 }, { "epoch": 2.18203790866302, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.946994781494141, "learning_rate": 5e-05, "loss": 0.7612, "mean_token_accuracy": 0.8902292847633362, "num_tokens": 651287495.0, "step": 17153 }, { "epoch": 2.1821651189416107, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.433248996734619, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.8730424046516418, "num_tokens": 651326005.0, "step": 17154 }, { "epoch": 2.1822923292202008, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.835823059082031, "learning_rate": 5e-05, "loss": 0.7921, "mean_token_accuracy": 0.8727846741676331, "num_tokens": 651364843.0, "step": 17155 }, { "epoch": 2.1824195394987913, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.44413948059082, "learning_rate": 5e-05, "loss": 0.7033, "mean_token_accuracy": 0.8903034925460815, "num_tokens": 651403981.0, "step": 17156 }, { "epoch": 2.182546749777382, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.774989604949951, "learning_rate": 5e-05, "loss": 0.7148, "mean_token_accuracy": 0.8899258971214294, "num_tokens": 651443751.0, "step": 17157 }, { "epoch": 2.1826739600559724, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.414350986480713, "learning_rate": 5e-05, "loss": 0.746, "mean_token_accuracy": 0.8791640400886536, "num_tokens": 651487248.0, "step": 17158 }, { "epoch": 2.182801170334563, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.817324638366699, "learning_rate": 5e-05, "loss": 0.7969, "mean_token_accuracy": 0.873073935508728, "num_tokens": 651523055.0, "step": 17159 }, { "epoch": 2.1829283806131534, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.297129154205322, "learning_rate": 5e-05, "loss": 0.6979, "mean_token_accuracy": 0.8830298781394958, "num_tokens": 651554749.0, "step": 17160 }, { "epoch": 2.183055590891744, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.93610954284668, "learning_rate": 5e-05, "loss": 0.8066, "mean_token_accuracy": 0.865446925163269, "num_tokens": 651597342.0, "step": 17161 }, { "epoch": 2.1831828011703345, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.097727298736572, "learning_rate": 5e-05, "loss": 0.6889, "mean_token_accuracy": 0.8841350078582764, "num_tokens": 651634578.0, "step": 17162 }, { "epoch": 2.183310011448925, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.186755180358887, "learning_rate": 5e-05, "loss": 0.732, "mean_token_accuracy": 0.8929034471511841, "num_tokens": 651674561.0, "step": 17163 }, { "epoch": 2.1834372217275155, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.808140277862549, "learning_rate": 5e-05, "loss": 0.6472, "mean_token_accuracy": 0.8954839110374451, "num_tokens": 651713426.0, "step": 17164 }, { "epoch": 2.183564432006106, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.612255096435547, "learning_rate": 5e-05, "loss": 0.7923, "mean_token_accuracy": 0.8873195648193359, "num_tokens": 651753267.0, "step": 17165 }, { "epoch": 2.1836916422846966, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.380195140838623, "learning_rate": 5e-05, "loss": 0.6383, "mean_token_accuracy": 0.8891387581825256, "num_tokens": 651797580.0, "step": 17166 }, { "epoch": 2.183818852563287, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 9.276065826416016, "learning_rate": 5e-05, "loss": 0.888, "mean_token_accuracy": 0.8765292167663574, "num_tokens": 651828294.0, "step": 17167 }, { "epoch": 2.1839460628418776, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.033867835998535, "learning_rate": 5e-05, "loss": 0.7178, "mean_token_accuracy": 0.8668068647384644, "num_tokens": 651865572.0, "step": 17168 }, { "epoch": 2.184073273120468, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.93430233001709, "learning_rate": 5e-05, "loss": 0.9608, "mean_token_accuracy": 0.8866259455680847, "num_tokens": 651897453.0, "step": 17169 }, { "epoch": 2.1842004833990587, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.532570838928223, "learning_rate": 5e-05, "loss": 0.7646, "mean_token_accuracy": 0.8801016807556152, "num_tokens": 651933096.0, "step": 17170 }, { "epoch": 2.1843276936776492, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.763982772827148, "learning_rate": 5e-05, "loss": 0.9805, "mean_token_accuracy": 0.890612781047821, "num_tokens": 651970042.0, "step": 17171 }, { "epoch": 2.1844549039562398, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 7.9746928215026855, "learning_rate": 5e-05, "loss": 0.9761, "mean_token_accuracy": 0.8797225952148438, "num_tokens": 652006513.0, "step": 17172 }, { "epoch": 2.1845821142348303, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.08802604675293, "learning_rate": 5e-05, "loss": 0.8267, "mean_token_accuracy": 0.8922629356384277, "num_tokens": 652042706.0, "step": 17173 }, { "epoch": 2.184709324513421, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.283469200134277, "learning_rate": 5e-05, "loss": 0.9756, "mean_token_accuracy": 0.8907735347747803, "num_tokens": 652084219.0, "step": 17174 }, { "epoch": 2.1848365347920113, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.351405620574951, "learning_rate": 5e-05, "loss": 0.6736, "mean_token_accuracy": 0.8905135989189148, "num_tokens": 652130640.0, "step": 17175 }, { "epoch": 2.184963745070602, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 10.82143783569336, "learning_rate": 5e-05, "loss": 1.1347, "mean_token_accuracy": 0.8779455423355103, "num_tokens": 652169812.0, "step": 17176 }, { "epoch": 2.1850909553491924, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 6.922276973724365, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.8987070918083191, "num_tokens": 652206317.0, "step": 17177 }, { "epoch": 2.185218165627783, "ewc_loss": 0.85546875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 11.30554485321045, "learning_rate": 5e-05, "loss": 1.187, "mean_token_accuracy": 0.8822836875915527, "num_tokens": 652241420.0, "step": 17178 }, { "epoch": 2.1853453759063735, "ewc_loss": 0.90625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000835418701171875, "grad_norm": 9.133914947509766, "learning_rate": 5e-05, "loss": 1.2257, "mean_token_accuracy": 0.8849570155143738, "num_tokens": 652283951.0, "step": 17179 }, { "epoch": 2.1854725861849635, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.743216514587402, "learning_rate": 5e-05, "loss": 0.9613, "mean_token_accuracy": 0.8870190382003784, "num_tokens": 652320729.0, "step": 17180 }, { "epoch": 2.185599796463554, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 9.42577838897705, "learning_rate": 5e-05, "loss": 1.1654, "mean_token_accuracy": 0.8764873743057251, "num_tokens": 652362969.0, "step": 17181 }, { "epoch": 2.1857270067421446, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.446051597595215, "learning_rate": 5e-05, "loss": 0.9305, "mean_token_accuracy": 0.8801680207252502, "num_tokens": 652403120.0, "step": 17182 }, { "epoch": 2.185854217020735, "ewc_loss": 0.65625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.39041519165039, "learning_rate": 5e-05, "loss": 1.0425, "mean_token_accuracy": 0.8686106204986572, "num_tokens": 652436904.0, "step": 17183 }, { "epoch": 2.1859814272993257, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.128351211547852, "learning_rate": 5e-05, "loss": 0.9163, "mean_token_accuracy": 0.8901419639587402, "num_tokens": 652474938.0, "step": 17184 }, { "epoch": 2.186108637577916, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.653502464294434, "learning_rate": 5e-05, "loss": 0.8953, "mean_token_accuracy": 0.8867170214653015, "num_tokens": 652503718.0, "step": 17185 }, { "epoch": 2.1862358478565067, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.366153717041016, "learning_rate": 5e-05, "loss": 0.9251, "mean_token_accuracy": 0.8854108452796936, "num_tokens": 652542235.0, "step": 17186 }, { "epoch": 2.1863630581350972, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.620173931121826, "learning_rate": 5e-05, "loss": 0.8148, "mean_token_accuracy": 0.8683573007583618, "num_tokens": 652584779.0, "step": 17187 }, { "epoch": 2.1864902684136878, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.876029014587402, "learning_rate": 5e-05, "loss": 0.9438, "mean_token_accuracy": 0.8763090968132019, "num_tokens": 652627132.0, "step": 17188 }, { "epoch": 2.1866174786922783, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.008665084838867, "learning_rate": 5e-05, "loss": 0.774, "mean_token_accuracy": 0.8758859038352966, "num_tokens": 652670156.0, "step": 17189 }, { "epoch": 2.186744688970869, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.36559772491455, "learning_rate": 5e-05, "loss": 0.9167, "mean_token_accuracy": 0.8934766054153442, "num_tokens": 652705246.0, "step": 17190 }, { "epoch": 2.1868718992494594, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.003448009490967, "learning_rate": 5e-05, "loss": 0.8037, "mean_token_accuracy": 0.8851085901260376, "num_tokens": 652742234.0, "step": 17191 }, { "epoch": 2.18699910952805, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.0006866455078125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.496541976928711, "learning_rate": 5e-05, "loss": 0.9359, "mean_token_accuracy": 0.8904399871826172, "num_tokens": 652779100.0, "step": 17192 }, { "epoch": 2.1871263198066404, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 7.451864719390869, "learning_rate": 5e-05, "loss": 0.876, "mean_token_accuracy": 0.8867358565330505, "num_tokens": 652821174.0, "step": 17193 }, { "epoch": 2.187253530085231, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.060212135314941, "learning_rate": 5e-05, "loss": 0.8855, "mean_token_accuracy": 0.8959006071090698, "num_tokens": 652856271.0, "step": 17194 }, { "epoch": 2.1873807403638215, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.791738033294678, "learning_rate": 5e-05, "loss": 0.9146, "mean_token_accuracy": 0.8814325928688049, "num_tokens": 652898737.0, "step": 17195 }, { "epoch": 2.187507950642412, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.441773414611816, "learning_rate": 5e-05, "loss": 0.8525, "mean_token_accuracy": 0.8862630128860474, "num_tokens": 652938778.0, "step": 17196 }, { "epoch": 2.1876351609210025, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.79032564163208, "learning_rate": 5e-05, "loss": 0.8721, "mean_token_accuracy": 0.8824577331542969, "num_tokens": 652982202.0, "step": 17197 }, { "epoch": 2.187762371199593, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.185565948486328, "learning_rate": 5e-05, "loss": 0.8651, "mean_token_accuracy": 0.8736675381660461, "num_tokens": 653024990.0, "step": 17198 }, { "epoch": 2.1878895814781836, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.471129894256592, "learning_rate": 5e-05, "loss": 0.8224, "mean_token_accuracy": 0.8737648129463196, "num_tokens": 653061913.0, "step": 17199 }, { "epoch": 2.188016791756774, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.360042572021484, "learning_rate": 5e-05, "loss": 0.8552, "mean_token_accuracy": 0.8811120390892029, "num_tokens": 653105944.0, "step": 17200 }, { "epoch": 2.1881440020353646, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.044073104858398, "learning_rate": 5e-05, "loss": 0.6862, "mean_token_accuracy": 0.8954268097877502, "num_tokens": 653139819.0, "step": 17201 }, { "epoch": 2.188271212313955, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.78276252746582, "learning_rate": 5e-05, "loss": 0.855, "mean_token_accuracy": 0.8936818242073059, "num_tokens": 653184569.0, "step": 17202 }, { "epoch": 2.1883984225925452, "ewc_loss": 0.375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.723864555358887, "learning_rate": 5e-05, "loss": 0.6455, "mean_token_accuracy": 0.9034976959228516, "num_tokens": 653220346.0, "step": 17203 }, { "epoch": 2.188525632871136, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.327173233032227, "learning_rate": 5e-05, "loss": 0.933, "mean_token_accuracy": 0.8832433223724365, "num_tokens": 653257288.0, "step": 17204 }, { "epoch": 2.1886528431497263, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.776723384857178, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8804553747177124, "num_tokens": 653293055.0, "step": 17205 }, { "epoch": 2.188780053428317, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.548730850219727, "learning_rate": 5e-05, "loss": 1.006, "mean_token_accuracy": 0.8748548626899719, "num_tokens": 653329988.0, "step": 17206 }, { "epoch": 2.1889072637069074, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.53908634185791, "learning_rate": 5e-05, "loss": 0.8299, "mean_token_accuracy": 0.8960511088371277, "num_tokens": 653367704.0, "step": 17207 }, { "epoch": 2.189034473985498, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.840740203857422, "learning_rate": 5e-05, "loss": 0.9211, "mean_token_accuracy": 0.8889874219894409, "num_tokens": 653403915.0, "step": 17208 }, { "epoch": 2.1891616842640884, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.289023399353027, "learning_rate": 5e-05, "loss": 0.9161, "mean_token_accuracy": 0.8907203674316406, "num_tokens": 653438045.0, "step": 17209 }, { "epoch": 2.189288894542679, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.6392621994018555, "learning_rate": 5e-05, "loss": 0.7855, "mean_token_accuracy": 0.8822240829467773, "num_tokens": 653469552.0, "step": 17210 }, { "epoch": 2.1894161048212695, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.681386947631836, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.887033224105835, "num_tokens": 653511809.0, "step": 17211 }, { "epoch": 2.18954331509986, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.98622989654541, "learning_rate": 5e-05, "loss": 0.6826, "mean_token_accuracy": 0.8974800705909729, "num_tokens": 653543812.0, "step": 17212 }, { "epoch": 2.1896705253784505, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.89218521118164, "learning_rate": 5e-05, "loss": 0.9264, "mean_token_accuracy": 0.8751602172851562, "num_tokens": 653586232.0, "step": 17213 }, { "epoch": 2.189797735657041, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.904299259185791, "learning_rate": 5e-05, "loss": 0.732, "mean_token_accuracy": 0.8806523084640503, "num_tokens": 653626700.0, "step": 17214 }, { "epoch": 2.1899249459356316, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.954062461853027, "learning_rate": 5e-05, "loss": 0.9067, "mean_token_accuracy": 0.8784594535827637, "num_tokens": 653666648.0, "step": 17215 }, { "epoch": 2.190052156214222, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.057114124298096, "learning_rate": 5e-05, "loss": 0.7439, "mean_token_accuracy": 0.890854001045227, "num_tokens": 653698991.0, "step": 17216 }, { "epoch": 2.1901793664928126, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.873130798339844, "learning_rate": 5e-05, "loss": 0.8749, "mean_token_accuracy": 0.8917348980903625, "num_tokens": 653737696.0, "step": 17217 }, { "epoch": 2.190306576771403, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.272522449493408, "learning_rate": 5e-05, "loss": 0.7757, "mean_token_accuracy": 0.8908441066741943, "num_tokens": 653776386.0, "step": 17218 }, { "epoch": 2.1904337870499937, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.6582612991333, "learning_rate": 5e-05, "loss": 0.8542, "mean_token_accuracy": 0.8930971026420593, "num_tokens": 653811346.0, "step": 17219 }, { "epoch": 2.1905609973285842, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.444748401641846, "learning_rate": 5e-05, "loss": 0.8402, "mean_token_accuracy": 0.8766984343528748, "num_tokens": 653847966.0, "step": 17220 }, { "epoch": 2.1906882076071748, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00067901611328125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.412177085876465, "learning_rate": 5e-05, "loss": 0.8203, "mean_token_accuracy": 0.8950487375259399, "num_tokens": 653882381.0, "step": 17221 }, { "epoch": 2.1908154178857653, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.44480562210083, "learning_rate": 5e-05, "loss": 0.8091, "mean_token_accuracy": 0.8886316418647766, "num_tokens": 653925026.0, "step": 17222 }, { "epoch": 2.190942628164356, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.329716682434082, "learning_rate": 5e-05, "loss": 0.8379, "mean_token_accuracy": 0.8849813938140869, "num_tokens": 653962210.0, "step": 17223 }, { "epoch": 2.1910698384429463, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.310817718505859, "learning_rate": 5e-05, "loss": 0.748, "mean_token_accuracy": 0.8872798681259155, "num_tokens": 653998713.0, "step": 17224 }, { "epoch": 2.191197048721537, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.371095657348633, "learning_rate": 5e-05, "loss": 0.8481, "mean_token_accuracy": 0.8804339170455933, "num_tokens": 654033383.0, "step": 17225 }, { "epoch": 2.1913242590001274, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.17744255065918, "learning_rate": 5e-05, "loss": 0.7641, "mean_token_accuracy": 0.8808873891830444, "num_tokens": 654074712.0, "step": 17226 }, { "epoch": 2.191451469278718, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.440844535827637, "learning_rate": 5e-05, "loss": 0.8824, "mean_token_accuracy": 0.879332959651947, "num_tokens": 654110650.0, "step": 17227 }, { "epoch": 2.191578679557308, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.138190269470215, "learning_rate": 5e-05, "loss": 0.6976, "mean_token_accuracy": 0.8962529897689819, "num_tokens": 654146628.0, "step": 17228 }, { "epoch": 2.1917058898358985, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.44742202758789, "learning_rate": 5e-05, "loss": 0.8773, "mean_token_accuracy": 0.8755637407302856, "num_tokens": 654192239.0, "step": 17229 }, { "epoch": 2.191833100114489, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.170355796813965, "learning_rate": 5e-05, "loss": 0.7982, "mean_token_accuracy": 0.8651329278945923, "num_tokens": 654232458.0, "step": 17230 }, { "epoch": 2.1919603103930796, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.358587265014648, "learning_rate": 5e-05, "loss": 0.902, "mean_token_accuracy": 0.8652104139328003, "num_tokens": 654271266.0, "step": 17231 }, { "epoch": 2.19208752067167, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.295589923858643, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8863030672073364, "num_tokens": 654307801.0, "step": 17232 }, { "epoch": 2.1922147309502606, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.165026664733887, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8986334204673767, "num_tokens": 654348903.0, "step": 17233 }, { "epoch": 2.192341941228851, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.44227409362793, "learning_rate": 5e-05, "loss": 0.7451, "mean_token_accuracy": 0.8873823881149292, "num_tokens": 654388659.0, "step": 17234 }, { "epoch": 2.1924691515074417, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.929110050201416, "learning_rate": 5e-05, "loss": 0.7677, "mean_token_accuracy": 0.8867253065109253, "num_tokens": 654425231.0, "step": 17235 }, { "epoch": 2.1925963617860322, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.546927452087402, "learning_rate": 5e-05, "loss": 0.7394, "mean_token_accuracy": 0.889031708240509, "num_tokens": 654457560.0, "step": 17236 }, { "epoch": 2.1927235720646228, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.763752460479736, "learning_rate": 5e-05, "loss": 0.7612, "mean_token_accuracy": 0.8819358348846436, "num_tokens": 654491000.0, "step": 17237 }, { "epoch": 2.1928507823432133, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.5839619636535645, "learning_rate": 5e-05, "loss": 0.754, "mean_token_accuracy": 0.8811072707176208, "num_tokens": 654531142.0, "step": 17238 }, { "epoch": 2.192977992621804, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.678572177886963, "learning_rate": 5e-05, "loss": 0.7493, "mean_token_accuracy": 0.8798545002937317, "num_tokens": 654574133.0, "step": 17239 }, { "epoch": 2.1931052029003943, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.578241348266602, "learning_rate": 5e-05, "loss": 0.7439, "mean_token_accuracy": 0.8765976428985596, "num_tokens": 654612745.0, "step": 17240 }, { "epoch": 2.193232413178985, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.657121658325195, "learning_rate": 5e-05, "loss": 0.72, "mean_token_accuracy": 0.8858375549316406, "num_tokens": 654646097.0, "step": 17241 }, { "epoch": 2.1933596234575754, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.5240349769592285, "learning_rate": 5e-05, "loss": 0.716, "mean_token_accuracy": 0.8809946179389954, "num_tokens": 654688216.0, "step": 17242 }, { "epoch": 2.193486833736166, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.687732696533203, "learning_rate": 5e-05, "loss": 0.6679, "mean_token_accuracy": 0.9034072160720825, "num_tokens": 654725697.0, "step": 17243 }, { "epoch": 2.1936140440147565, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.446640491485596, "learning_rate": 5e-05, "loss": 0.7139, "mean_token_accuracy": 0.8831618428230286, "num_tokens": 654762440.0, "step": 17244 }, { "epoch": 2.193741254293347, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.733447551727295, "learning_rate": 5e-05, "loss": 0.7397, "mean_token_accuracy": 0.8761725425720215, "num_tokens": 654795843.0, "step": 17245 }, { "epoch": 2.1938684645719375, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.344813823699951, "learning_rate": 5e-05, "loss": 0.6749, "mean_token_accuracy": 0.8876051902770996, "num_tokens": 654835212.0, "step": 17246 }, { "epoch": 2.193995674850528, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.85511589050293, "learning_rate": 5e-05, "loss": 0.7491, "mean_token_accuracy": 0.8750160932540894, "num_tokens": 654869092.0, "step": 17247 }, { "epoch": 2.1941228851291186, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.203042984008789, "learning_rate": 5e-05, "loss": 0.6643, "mean_token_accuracy": 0.8883723020553589, "num_tokens": 654907202.0, "step": 17248 }, { "epoch": 2.194250095407709, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.03079605102539, "learning_rate": 5e-05, "loss": 0.7312, "mean_token_accuracy": 0.8850457668304443, "num_tokens": 654938867.0, "step": 17249 }, { "epoch": 2.1943773056862996, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.968545913696289, "learning_rate": 5e-05, "loss": 0.6822, "mean_token_accuracy": 0.8796093463897705, "num_tokens": 654978838.0, "step": 17250 }, { "epoch": 2.19450451596489, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.312653541564941, "learning_rate": 5e-05, "loss": 0.7458, "mean_token_accuracy": 0.8896700739860535, "num_tokens": 655016987.0, "step": 17251 }, { "epoch": 2.1946317262434807, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.672811985015869, "learning_rate": 5e-05, "loss": 0.6581, "mean_token_accuracy": 0.8828502297401428, "num_tokens": 655056797.0, "step": 17252 }, { "epoch": 2.1947589365220708, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.749074935913086, "learning_rate": 5e-05, "loss": 0.802, "mean_token_accuracy": 0.8873660564422607, "num_tokens": 655094959.0, "step": 17253 }, { "epoch": 2.1948861468006613, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.306756019592285, "learning_rate": 5e-05, "loss": 0.6766, "mean_token_accuracy": 0.8733362555503845, "num_tokens": 655136490.0, "step": 17254 }, { "epoch": 2.195013357079252, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.301115989685059, "learning_rate": 5e-05, "loss": 0.8758, "mean_token_accuracy": 0.8871952295303345, "num_tokens": 655175174.0, "step": 17255 }, { "epoch": 2.1951405673578424, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.201394081115723, "learning_rate": 5e-05, "loss": 0.6835, "mean_token_accuracy": 0.8760108351707458, "num_tokens": 655214753.0, "step": 17256 }, { "epoch": 2.195267777636433, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.536323547363281, "learning_rate": 5e-05, "loss": 0.9093, "mean_token_accuracy": 0.8949902653694153, "num_tokens": 655249737.0, "step": 17257 }, { "epoch": 2.1953949879150234, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.916208267211914, "learning_rate": 5e-05, "loss": 0.739, "mean_token_accuracy": 0.8919638395309448, "num_tokens": 655286553.0, "step": 17258 }, { "epoch": 2.195522198193614, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.69226360321045, "learning_rate": 5e-05, "loss": 0.9041, "mean_token_accuracy": 0.8850029706954956, "num_tokens": 655320409.0, "step": 17259 }, { "epoch": 2.1956494084722045, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.174023628234863, "learning_rate": 5e-05, "loss": 0.8901, "mean_token_accuracy": 0.8823621273040771, "num_tokens": 655362436.0, "step": 17260 }, { "epoch": 2.195776618750795, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.925552845001221, "learning_rate": 5e-05, "loss": 0.715, "mean_token_accuracy": 0.8790535926818848, "num_tokens": 655405059.0, "step": 17261 }, { "epoch": 2.1959038290293855, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.654284477233887, "learning_rate": 5e-05, "loss": 0.964, "mean_token_accuracy": 0.8805820941925049, "num_tokens": 655448688.0, "step": 17262 }, { "epoch": 2.196031039307976, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 5.808992385864258, "learning_rate": 5e-05, "loss": 0.6422, "mean_token_accuracy": 0.8879402875900269, "num_tokens": 655486361.0, "step": 17263 }, { "epoch": 2.1961582495865666, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000732421875, "grad_norm": 11.196895599365234, "learning_rate": 5e-05, "loss": 1.1193, "mean_token_accuracy": 0.8839649558067322, "num_tokens": 655518610.0, "step": 17264 }, { "epoch": 2.196285459865157, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.220494747161865, "learning_rate": 5e-05, "loss": 0.9131, "mean_token_accuracy": 0.8795729875564575, "num_tokens": 655557860.0, "step": 17265 }, { "epoch": 2.1964126701437476, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0007781982421875, "grad_norm": 11.325243949890137, "learning_rate": 5e-05, "loss": 1.1853, "mean_token_accuracy": 0.8821943998336792, "num_tokens": 655597885.0, "step": 17266 }, { "epoch": 2.196539880422338, "ewc_loss": 0.91796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00084686279296875, "grad_norm": 9.449055671691895, "learning_rate": 5e-05, "loss": 1.2158, "mean_token_accuracy": 0.8902246952056885, "num_tokens": 655632847.0, "step": 17267 }, { "epoch": 2.1966670907009287, "ewc_loss": 0.625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.539843559265137, "learning_rate": 5e-05, "loss": 0.9785, "mean_token_accuracy": 0.8783242702484131, "num_tokens": 655665374.0, "step": 17268 }, { "epoch": 2.196794300979519, "ewc_loss": 0.796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 9.709802627563477, "learning_rate": 5e-05, "loss": 1.1146, "mean_token_accuracy": 0.8861624002456665, "num_tokens": 655705926.0, "step": 17269 }, { "epoch": 2.1969215112581097, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.210236549377441, "learning_rate": 5e-05, "loss": 0.8652, "mean_token_accuracy": 0.8949692845344543, "num_tokens": 655743500.0, "step": 17270 }, { "epoch": 2.1970487215367003, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.901697158813477, "learning_rate": 5e-05, "loss": 1.0242, "mean_token_accuracy": 0.8765870332717896, "num_tokens": 655788063.0, "step": 17271 }, { "epoch": 2.197175931815291, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.029369354248047, "learning_rate": 5e-05, "loss": 0.91, "mean_token_accuracy": 0.8945547938346863, "num_tokens": 655828393.0, "step": 17272 }, { "epoch": 2.1973031420938813, "ewc_loss": 0.625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.312352180480957, "learning_rate": 5e-05, "loss": 0.9346, "mean_token_accuracy": 0.8894240856170654, "num_tokens": 655871666.0, "step": 17273 }, { "epoch": 2.197430352372472, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.209189414978027, "learning_rate": 5e-05, "loss": 0.8884, "mean_token_accuracy": 0.8974379301071167, "num_tokens": 655905396.0, "step": 17274 }, { "epoch": 2.1975575626510624, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.333305358886719, "learning_rate": 5e-05, "loss": 0.8615, "mean_token_accuracy": 0.8793714046478271, "num_tokens": 655944515.0, "step": 17275 }, { "epoch": 2.197684772929653, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.3659029006958, "learning_rate": 5e-05, "loss": 0.933, "mean_token_accuracy": 0.8727545142173767, "num_tokens": 655979773.0, "step": 17276 }, { "epoch": 2.1978119832082434, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.7168450355529785, "learning_rate": 5e-05, "loss": 0.787, "mean_token_accuracy": 0.8840096592903137, "num_tokens": 656017786.0, "step": 17277 }, { "epoch": 2.1979391934868335, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.325029373168945, "learning_rate": 5e-05, "loss": 0.9041, "mean_token_accuracy": 0.862912654876709, "num_tokens": 656058776.0, "step": 17278 }, { "epoch": 2.198066403765424, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.5100908279418945, "learning_rate": 5e-05, "loss": 0.7568, "mean_token_accuracy": 0.8901579976081848, "num_tokens": 656100231.0, "step": 17279 }, { "epoch": 2.1981936140440146, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.180246353149414, "learning_rate": 5e-05, "loss": 0.7894, "mean_token_accuracy": 0.8847805261611938, "num_tokens": 656131786.0, "step": 17280 }, { "epoch": 2.198320824322605, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.506008625030518, "learning_rate": 5e-05, "loss": 0.7473, "mean_token_accuracy": 0.8867276310920715, "num_tokens": 656172041.0, "step": 17281 }, { "epoch": 2.1984480346011956, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.916018962860107, "learning_rate": 5e-05, "loss": 0.7363, "mean_token_accuracy": 0.8898735046386719, "num_tokens": 656203152.0, "step": 17282 }, { "epoch": 2.198575244879786, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.65886116027832, "learning_rate": 5e-05, "loss": 0.7897, "mean_token_accuracy": 0.8789728879928589, "num_tokens": 656241568.0, "step": 17283 }, { "epoch": 2.1987024551583767, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.537103652954102, "learning_rate": 5e-05, "loss": 0.7551, "mean_token_accuracy": 0.8694015741348267, "num_tokens": 656275221.0, "step": 17284 }, { "epoch": 2.1988296654369672, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.004178047180176, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.877051591873169, "num_tokens": 656317535.0, "step": 17285 }, { "epoch": 2.1989568757155578, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.0825371742248535, "learning_rate": 5e-05, "loss": 0.6743, "mean_token_accuracy": 0.8889731168746948, "num_tokens": 656351440.0, "step": 17286 }, { "epoch": 2.1990840859941483, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.502364158630371, "learning_rate": 5e-05, "loss": 0.803, "mean_token_accuracy": 0.880744218826294, "num_tokens": 656390942.0, "step": 17287 }, { "epoch": 2.199211296272739, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.5939812660217285, "learning_rate": 5e-05, "loss": 0.6562, "mean_token_accuracy": 0.8898443579673767, "num_tokens": 656427314.0, "step": 17288 }, { "epoch": 2.1993385065513293, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 9.112975120544434, "learning_rate": 5e-05, "loss": 0.8671, "mean_token_accuracy": 0.8776637315750122, "num_tokens": 656464014.0, "step": 17289 }, { "epoch": 2.19946571682992, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.244368076324463, "learning_rate": 5e-05, "loss": 0.6684, "mean_token_accuracy": 0.8825815916061401, "num_tokens": 656497783.0, "step": 17290 }, { "epoch": 2.1995929271085104, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.754085540771484, "learning_rate": 5e-05, "loss": 0.9556, "mean_token_accuracy": 0.8814523220062256, "num_tokens": 656538160.0, "step": 17291 }, { "epoch": 2.199720137387101, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.470355987548828, "learning_rate": 5e-05, "loss": 0.7107, "mean_token_accuracy": 0.8896663784980774, "num_tokens": 656578296.0, "step": 17292 }, { "epoch": 2.1998473476656915, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.8990478515625, "learning_rate": 5e-05, "loss": 0.9969, "mean_token_accuracy": 0.8833208680152893, "num_tokens": 656620662.0, "step": 17293 }, { "epoch": 2.199974557944282, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.478591442108154, "learning_rate": 5e-05, "loss": 0.8961, "mean_token_accuracy": 0.8846535682678223, "num_tokens": 656665511.0, "step": 17294 }, { "epoch": 2.2001017682228725, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.08317756652832, "learning_rate": 5e-05, "loss": 0.9534, "mean_token_accuracy": 0.8838796615600586, "num_tokens": 656703079.0, "step": 17295 }, { "epoch": 2.200228978501463, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000682830810546875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.279938697814941, "learning_rate": 5e-05, "loss": 0.9669, "mean_token_accuracy": 0.8754622936248779, "num_tokens": 656741430.0, "step": 17296 }, { "epoch": 2.2003561887800536, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.884193420410156, "learning_rate": 5e-05, "loss": 0.7836, "mean_token_accuracy": 0.8906664848327637, "num_tokens": 656778836.0, "step": 17297 }, { "epoch": 2.200483399058644, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.57630443572998, "learning_rate": 5e-05, "loss": 0.9125, "mean_token_accuracy": 0.884175717830658, "num_tokens": 656811331.0, "step": 17298 }, { "epoch": 2.2006106093372346, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.4202165603637695, "learning_rate": 5e-05, "loss": 0.7572, "mean_token_accuracy": 0.8917183876037598, "num_tokens": 656845579.0, "step": 17299 }, { "epoch": 2.200737819615825, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.291376113891602, "learning_rate": 5e-05, "loss": 0.8498, "mean_token_accuracy": 0.8799245953559875, "num_tokens": 656880718.0, "step": 17300 }, { "epoch": 2.2008650298944152, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.630460262298584, "learning_rate": 5e-05, "loss": 0.7962, "mean_token_accuracy": 0.8834271430969238, "num_tokens": 656918350.0, "step": 17301 }, { "epoch": 2.200992240173006, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.6795268058776855, "learning_rate": 5e-05, "loss": 0.7706, "mean_token_accuracy": 0.8804437518119812, "num_tokens": 656954661.0, "step": 17302 }, { "epoch": 2.2011194504515963, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.041818618774414, "learning_rate": 5e-05, "loss": 0.799, "mean_token_accuracy": 0.8856645822525024, "num_tokens": 656993233.0, "step": 17303 }, { "epoch": 2.201246660730187, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.193368911743164, "learning_rate": 5e-05, "loss": 0.7164, "mean_token_accuracy": 0.8803688287734985, "num_tokens": 657032843.0, "step": 17304 }, { "epoch": 2.2013738710087773, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.335086822509766, "learning_rate": 5e-05, "loss": 0.8327, "mean_token_accuracy": 0.880936861038208, "num_tokens": 657078625.0, "step": 17305 }, { "epoch": 2.201501081287368, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.936787128448486, "learning_rate": 5e-05, "loss": 0.6967, "mean_token_accuracy": 0.8859018683433533, "num_tokens": 657113696.0, "step": 17306 }, { "epoch": 2.2016282915659584, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.606752395629883, "learning_rate": 5e-05, "loss": 0.8424, "mean_token_accuracy": 0.8837029933929443, "num_tokens": 657149224.0, "step": 17307 }, { "epoch": 2.201755501844549, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.780269145965576, "learning_rate": 5e-05, "loss": 0.6963, "mean_token_accuracy": 0.8798689246177673, "num_tokens": 657181946.0, "step": 17308 }, { "epoch": 2.2018827121231395, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.818981170654297, "learning_rate": 5e-05, "loss": 0.9329, "mean_token_accuracy": 0.8612138628959656, "num_tokens": 657214372.0, "step": 17309 }, { "epoch": 2.20200992240173, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.776928424835205, "learning_rate": 5e-05, "loss": 0.7088, "mean_token_accuracy": 0.883258581161499, "num_tokens": 657255572.0, "step": 17310 }, { "epoch": 2.2021371326803205, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.843043327331543, "learning_rate": 5e-05, "loss": 0.8615, "mean_token_accuracy": 0.8861945271492004, "num_tokens": 657288336.0, "step": 17311 }, { "epoch": 2.202264342958911, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.9762282371521, "learning_rate": 5e-05, "loss": 0.7604, "mean_token_accuracy": 0.8825604915618896, "num_tokens": 657327954.0, "step": 17312 }, { "epoch": 2.2023915532375016, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.66322135925293, "learning_rate": 5e-05, "loss": 0.8934, "mean_token_accuracy": 0.8765217065811157, "num_tokens": 657366691.0, "step": 17313 }, { "epoch": 2.202518763516092, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.243656635284424, "learning_rate": 5e-05, "loss": 0.7807, "mean_token_accuracy": 0.881074070930481, "num_tokens": 657405694.0, "step": 17314 }, { "epoch": 2.2026459737946826, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.36385726928711, "learning_rate": 5e-05, "loss": 0.8071, "mean_token_accuracy": 0.897068977355957, "num_tokens": 657442461.0, "step": 17315 }, { "epoch": 2.202773184073273, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.39333438873291, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8866369724273682, "num_tokens": 657476384.0, "step": 17316 }, { "epoch": 2.2029003943518637, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.148320198059082, "learning_rate": 5e-05, "loss": 0.8526, "mean_token_accuracy": 0.8750401139259338, "num_tokens": 657511291.0, "step": 17317 }, { "epoch": 2.203027604630454, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.409420013427734, "learning_rate": 5e-05, "loss": 0.7798, "mean_token_accuracy": 0.876013994216919, "num_tokens": 657546820.0, "step": 17318 }, { "epoch": 2.2031548149090447, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.092571258544922, "learning_rate": 5e-05, "loss": 0.7845, "mean_token_accuracy": 0.8875491619110107, "num_tokens": 657582513.0, "step": 17319 }, { "epoch": 2.2032820251876353, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.327475547790527, "learning_rate": 5e-05, "loss": 0.7331, "mean_token_accuracy": 0.885087788105011, "num_tokens": 657617108.0, "step": 17320 }, { "epoch": 2.203409235466226, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.085599899291992, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8909820318222046, "num_tokens": 657656220.0, "step": 17321 }, { "epoch": 2.2035364457448163, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.265377044677734, "learning_rate": 5e-05, "loss": 0.6968, "mean_token_accuracy": 0.8947445154190063, "num_tokens": 657691386.0, "step": 17322 }, { "epoch": 2.203663656023407, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.068735122680664, "learning_rate": 5e-05, "loss": 0.8285, "mean_token_accuracy": 0.8703351616859436, "num_tokens": 657728118.0, "step": 17323 }, { "epoch": 2.2037908663019974, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.279241561889648, "learning_rate": 5e-05, "loss": 0.6859, "mean_token_accuracy": 0.8959907293319702, "num_tokens": 657768832.0, "step": 17324 }, { "epoch": 2.203918076580588, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.985658168792725, "learning_rate": 5e-05, "loss": 0.7583, "mean_token_accuracy": 0.8892992734909058, "num_tokens": 657805511.0, "step": 17325 }, { "epoch": 2.204045286859178, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.326394557952881, "learning_rate": 5e-05, "loss": 0.7259, "mean_token_accuracy": 0.8828837871551514, "num_tokens": 657843250.0, "step": 17326 }, { "epoch": 2.2041724971377685, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.88098669052124, "learning_rate": 5e-05, "loss": 0.72, "mean_token_accuracy": 0.8946670293807983, "num_tokens": 657878882.0, "step": 17327 }, { "epoch": 2.204299707416359, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.401859760284424, "learning_rate": 5e-05, "loss": 0.6586, "mean_token_accuracy": 0.9040122032165527, "num_tokens": 657917007.0, "step": 17328 }, { "epoch": 2.2044269176949496, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.765603065490723, "learning_rate": 5e-05, "loss": 0.7683, "mean_token_accuracy": 0.8740313649177551, "num_tokens": 657956741.0, "step": 17329 }, { "epoch": 2.20455412797354, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.459611415863037, "learning_rate": 5e-05, "loss": 0.6954, "mean_token_accuracy": 0.8838260769844055, "num_tokens": 657995050.0, "step": 17330 }, { "epoch": 2.2046813382521306, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.689579486846924, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8796851634979248, "num_tokens": 658032708.0, "step": 17331 }, { "epoch": 2.204808548530721, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.470680236816406, "learning_rate": 5e-05, "loss": 0.6601, "mean_token_accuracy": 0.8988329172134399, "num_tokens": 658065725.0, "step": 17332 }, { "epoch": 2.2049357588093117, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.632329940795898, "learning_rate": 5e-05, "loss": 0.6866, "mean_token_accuracy": 0.8925401568412781, "num_tokens": 658107290.0, "step": 17333 }, { "epoch": 2.2050629690879022, "ewc_loss": 0.375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.4930739402771, "learning_rate": 5e-05, "loss": 0.6892, "mean_token_accuracy": 0.8871569037437439, "num_tokens": 658147516.0, "step": 17334 }, { "epoch": 2.2051901793664928, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.582571983337402, "learning_rate": 5e-05, "loss": 0.6936, "mean_token_accuracy": 0.8886010646820068, "num_tokens": 658188477.0, "step": 17335 }, { "epoch": 2.2053173896450833, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.511290550231934, "learning_rate": 5e-05, "loss": 0.6892, "mean_token_accuracy": 0.8868228197097778, "num_tokens": 658223517.0, "step": 17336 }, { "epoch": 2.205444599923674, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.53652811050415, "learning_rate": 5e-05, "loss": 0.6816, "mean_token_accuracy": 0.8885335922241211, "num_tokens": 658259167.0, "step": 17337 }, { "epoch": 2.2055718102022643, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.536279201507568, "learning_rate": 5e-05, "loss": 0.6839, "mean_token_accuracy": 0.8866836428642273, "num_tokens": 658294825.0, "step": 17338 }, { "epoch": 2.205699020480855, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.488687038421631, "learning_rate": 5e-05, "loss": 0.7138, "mean_token_accuracy": 0.8753147125244141, "num_tokens": 658336898.0, "step": 17339 }, { "epoch": 2.2058262307594454, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.5663886070251465, "learning_rate": 5e-05, "loss": 0.651, "mean_token_accuracy": 0.8979976773262024, "num_tokens": 658373201.0, "step": 17340 }, { "epoch": 2.205953441038036, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.433125019073486, "learning_rate": 5e-05, "loss": 0.7099, "mean_token_accuracy": 0.8789008855819702, "num_tokens": 658413147.0, "step": 17341 }, { "epoch": 2.2060806513166265, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.609861850738525, "learning_rate": 5e-05, "loss": 0.7291, "mean_token_accuracy": 0.8724744915962219, "num_tokens": 658453832.0, "step": 17342 }, { "epoch": 2.206207861595217, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.388854503631592, "learning_rate": 5e-05, "loss": 0.7056, "mean_token_accuracy": 0.87593013048172, "num_tokens": 658498302.0, "step": 17343 }, { "epoch": 2.2063350718738075, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.672055721282959, "learning_rate": 5e-05, "loss": 0.692, "mean_token_accuracy": 0.8886145949363708, "num_tokens": 658535017.0, "step": 17344 }, { "epoch": 2.206462282152398, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.25710916519165, "learning_rate": 5e-05, "loss": 0.7203, "mean_token_accuracy": 0.8675004243850708, "num_tokens": 658575495.0, "step": 17345 }, { "epoch": 2.2065894924309886, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.82977294921875, "learning_rate": 5e-05, "loss": 0.6947, "mean_token_accuracy": 0.8900738954544067, "num_tokens": 658619493.0, "step": 17346 }, { "epoch": 2.206716702709579, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 7.0484089851379395, "learning_rate": 5e-05, "loss": 0.6618, "mean_token_accuracy": 0.8822579383850098, "num_tokens": 658654371.0, "step": 17347 }, { "epoch": 2.2068439129881696, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.117846488952637, "learning_rate": 5e-05, "loss": 0.7635, "mean_token_accuracy": 0.8750089406967163, "num_tokens": 658694543.0, "step": 17348 }, { "epoch": 2.20697112326676, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.656437397003174, "learning_rate": 5e-05, "loss": 0.644, "mean_token_accuracy": 0.8800396919250488, "num_tokens": 658724859.0, "step": 17349 }, { "epoch": 2.2070983335453507, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.733687400817871, "learning_rate": 5e-05, "loss": 0.7508, "mean_token_accuracy": 0.895280122756958, "num_tokens": 658759941.0, "step": 17350 }, { "epoch": 2.2072255438239408, "ewc_loss": 0.28125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002117156982421875, "grad_norm": 5.950638294219971, "learning_rate": 5e-05, "loss": 0.5788, "mean_token_accuracy": 0.8910864591598511, "num_tokens": 658801254.0, "step": 17351 }, { "epoch": 2.2073527541025313, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.988203048706055, "learning_rate": 5e-05, "loss": 0.9449, "mean_token_accuracy": 0.8778430223464966, "num_tokens": 658835177.0, "step": 17352 }, { "epoch": 2.207479964381122, "ewc_loss": 0.279296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000209808349609375, "grad_norm": 5.16961145401001, "learning_rate": 5e-05, "loss": 0.5908, "mean_token_accuracy": 0.8905662894248962, "num_tokens": 658874636.0, "step": 17353 }, { "epoch": 2.2076071746597123, "ewc_loss": 0.93359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00086212158203125, "grad_norm": 12.001914024353027, "learning_rate": 5e-05, "loss": 1.2724, "mean_token_accuracy": 0.8789573907852173, "num_tokens": 658911388.0, "step": 17354 }, { "epoch": 2.207734384938303, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 7.166467189788818, "learning_rate": 5e-05, "loss": 0.9192, "mean_token_accuracy": 0.8893898725509644, "num_tokens": 658954685.0, "step": 17355 }, { "epoch": 2.2078615952168934, "ewc_loss": 1.0625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0009918212890625, "grad_norm": 13.04753303527832, "learning_rate": 5e-05, "loss": 1.3721, "mean_token_accuracy": 0.8877288103103638, "num_tokens": 658992573.0, "step": 17356 }, { "epoch": 2.207988805495484, "ewc_loss": 1.1875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0011138916015625, "grad_norm": 10.55970573425293, "learning_rate": 5e-05, "loss": 1.5341, "mean_token_accuracy": 0.876502275466919, "num_tokens": 659024460.0, "step": 17357 }, { "epoch": 2.2081160157740745, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 11.637807846069336, "learning_rate": 5e-05, "loss": 1.1655, "mean_token_accuracy": 0.8627955317497253, "num_tokens": 659066103.0, "step": 17358 }, { "epoch": 2.208243226052665, "ewc_loss": 0.93359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00086212158203125, "grad_norm": 9.77381420135498, "learning_rate": 5e-05, "loss": 1.2409, "mean_token_accuracy": 0.889509379863739, "num_tokens": 659097293.0, "step": 17359 }, { "epoch": 2.2083704363312555, "ewc_loss": 0.82421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00075531005859375, "grad_norm": 11.14684772491455, "learning_rate": 5e-05, "loss": 1.1719, "mean_token_accuracy": 0.8753958344459534, "num_tokens": 659141221.0, "step": 17360 }, { "epoch": 2.208497646609846, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 8.787114143371582, "learning_rate": 5e-05, "loss": 0.98, "mean_token_accuracy": 0.8858586549758911, "num_tokens": 659181026.0, "step": 17361 }, { "epoch": 2.2086248568884366, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.508021354675293, "learning_rate": 5e-05, "loss": 1.0851, "mean_token_accuracy": 0.8763411641120911, "num_tokens": 659218503.0, "step": 17362 }, { "epoch": 2.208752067167027, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.384285926818848, "learning_rate": 5e-05, "loss": 0.938, "mean_token_accuracy": 0.887567400932312, "num_tokens": 659255940.0, "step": 17363 }, { "epoch": 2.2088792774456176, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.23538875579834, "learning_rate": 5e-05, "loss": 1.0986, "mean_token_accuracy": 0.8642040491104126, "num_tokens": 659296994.0, "step": 17364 }, { "epoch": 2.209006487724208, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.28722858428955, "learning_rate": 5e-05, "loss": 0.9667, "mean_token_accuracy": 0.875579833984375, "num_tokens": 659335372.0, "step": 17365 }, { "epoch": 2.2091336980027987, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.942651748657227, "learning_rate": 5e-05, "loss": 1.017, "mean_token_accuracy": 0.8802676200866699, "num_tokens": 659375954.0, "step": 17366 }, { "epoch": 2.209260908281389, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.252521514892578, "learning_rate": 5e-05, "loss": 0.9423, "mean_token_accuracy": 0.8807665109634399, "num_tokens": 659409995.0, "step": 17367 }, { "epoch": 2.2093881185599797, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.664101600646973, "learning_rate": 5e-05, "loss": 0.9678, "mean_token_accuracy": 0.8895626664161682, "num_tokens": 659445640.0, "step": 17368 }, { "epoch": 2.2095153288385703, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.05134391784668, "learning_rate": 5e-05, "loss": 0.9013, "mean_token_accuracy": 0.8817179203033447, "num_tokens": 659488609.0, "step": 17369 }, { "epoch": 2.209642539117161, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.594850540161133, "learning_rate": 5e-05, "loss": 0.949, "mean_token_accuracy": 0.8942871689796448, "num_tokens": 659533077.0, "step": 17370 }, { "epoch": 2.2097697493957513, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.82673454284668, "learning_rate": 5e-05, "loss": 0.8606, "mean_token_accuracy": 0.8838789463043213, "num_tokens": 659574080.0, "step": 17371 }, { "epoch": 2.209896959674342, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.795088768005371, "learning_rate": 5e-05, "loss": 1.0166, "mean_token_accuracy": 0.8830131888389587, "num_tokens": 659610508.0, "step": 17372 }, { "epoch": 2.2100241699529324, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.822299003601074, "learning_rate": 5e-05, "loss": 0.8727, "mean_token_accuracy": 0.8831444978713989, "num_tokens": 659648378.0, "step": 17373 }, { "epoch": 2.210151380231523, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 9.975492477416992, "learning_rate": 5e-05, "loss": 1.0788, "mean_token_accuracy": 0.8755596280097961, "num_tokens": 659681639.0, "step": 17374 }, { "epoch": 2.2102785905101134, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.226995468139648, "learning_rate": 5e-05, "loss": 0.949, "mean_token_accuracy": 0.8756668567657471, "num_tokens": 659721579.0, "step": 17375 }, { "epoch": 2.2104058007887035, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.605666160583496, "learning_rate": 5e-05, "loss": 1.0298, "mean_token_accuracy": 0.8799827694892883, "num_tokens": 659761577.0, "step": 17376 }, { "epoch": 2.210533011067294, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.539172172546387, "learning_rate": 5e-05, "loss": 0.9518, "mean_token_accuracy": 0.8816925287246704, "num_tokens": 659799393.0, "step": 17377 }, { "epoch": 2.2106602213458846, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.81938362121582, "learning_rate": 5e-05, "loss": 0.9898, "mean_token_accuracy": 0.8642922639846802, "num_tokens": 659839118.0, "step": 17378 }, { "epoch": 2.210787431624475, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.376775741577148, "learning_rate": 5e-05, "loss": 0.8681, "mean_token_accuracy": 0.8949509263038635, "num_tokens": 659874955.0, "step": 17379 }, { "epoch": 2.2109146419030656, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.365586280822754, "learning_rate": 5e-05, "loss": 0.9096, "mean_token_accuracy": 0.8775419592857361, "num_tokens": 659908590.0, "step": 17380 }, { "epoch": 2.211041852181656, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.824862003326416, "learning_rate": 5e-05, "loss": 0.8192, "mean_token_accuracy": 0.8809618353843689, "num_tokens": 659945753.0, "step": 17381 }, { "epoch": 2.2111690624602467, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.614879608154297, "learning_rate": 5e-05, "loss": 0.8688, "mean_token_accuracy": 0.8910824656486511, "num_tokens": 659981284.0, "step": 17382 }, { "epoch": 2.211296272738837, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.192849159240723, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8819041848182678, "num_tokens": 660016160.0, "step": 17383 }, { "epoch": 2.2114234830174277, "ewc_loss": 0.625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.200484275817871, "learning_rate": 5e-05, "loss": 0.9443, "mean_token_accuracy": 0.8874469995498657, "num_tokens": 660053755.0, "step": 17384 }, { "epoch": 2.2115506932960183, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.124635696411133, "learning_rate": 5e-05, "loss": 0.6936, "mean_token_accuracy": 0.9048753380775452, "num_tokens": 660092808.0, "step": 17385 }, { "epoch": 2.211677903574609, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.478498458862305, "learning_rate": 5e-05, "loss": 1.0153, "mean_token_accuracy": 0.8779295682907104, "num_tokens": 660128996.0, "step": 17386 }, { "epoch": 2.2118051138531993, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.710378646850586, "learning_rate": 5e-05, "loss": 0.8004, "mean_token_accuracy": 0.8979979753494263, "num_tokens": 660163345.0, "step": 17387 }, { "epoch": 2.21193232413179, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.046732902526855, "learning_rate": 5e-05, "loss": 0.9614, "mean_token_accuracy": 0.8862560987472534, "num_tokens": 660196483.0, "step": 17388 }, { "epoch": 2.2120595344103804, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.284136772155762, "learning_rate": 5e-05, "loss": 0.9312, "mean_token_accuracy": 0.878753125667572, "num_tokens": 660226039.0, "step": 17389 }, { "epoch": 2.212186744688971, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.034505844116211, "learning_rate": 5e-05, "loss": 0.8667, "mean_token_accuracy": 0.8798174858093262, "num_tokens": 660274301.0, "step": 17390 }, { "epoch": 2.2123139549675614, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.339151382446289, "learning_rate": 5e-05, "loss": 0.9009, "mean_token_accuracy": 0.8736536502838135, "num_tokens": 660312893.0, "step": 17391 }, { "epoch": 2.212441165246152, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.495728015899658, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.8916771411895752, "num_tokens": 660349396.0, "step": 17392 }, { "epoch": 2.2125683755247425, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.077643394470215, "learning_rate": 5e-05, "loss": 0.7803, "mean_token_accuracy": 0.8886089324951172, "num_tokens": 660385902.0, "step": 17393 }, { "epoch": 2.212695585803333, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.600149631500244, "learning_rate": 5e-05, "loss": 0.7762, "mean_token_accuracy": 0.8800334930419922, "num_tokens": 660421527.0, "step": 17394 }, { "epoch": 2.2128227960819236, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.641092300415039, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.8889703750610352, "num_tokens": 660459687.0, "step": 17395 }, { "epoch": 2.212950006360514, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.829329490661621, "learning_rate": 5e-05, "loss": 0.8077, "mean_token_accuracy": 0.8685803413391113, "num_tokens": 660502415.0, "step": 17396 }, { "epoch": 2.2130772166391046, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.401351451873779, "learning_rate": 5e-05, "loss": 0.7151, "mean_token_accuracy": 0.8846664428710938, "num_tokens": 660542503.0, "step": 17397 }, { "epoch": 2.213204426917695, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.893558979034424, "learning_rate": 5e-05, "loss": 0.766, "mean_token_accuracy": 0.8791517019271851, "num_tokens": 660576482.0, "step": 17398 }, { "epoch": 2.2133316371962852, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.267543792724609, "learning_rate": 5e-05, "loss": 0.6494, "mean_token_accuracy": 0.8999382853507996, "num_tokens": 660615151.0, "step": 17399 }, { "epoch": 2.213458847474876, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.05450439453125, "learning_rate": 5e-05, "loss": 0.7828, "mean_token_accuracy": 0.8784720301628113, "num_tokens": 660656555.0, "step": 17400 }, { "epoch": 2.2135860577534663, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.024784564971924, "learning_rate": 5e-05, "loss": 0.7137, "mean_token_accuracy": 0.8720129132270813, "num_tokens": 660695991.0, "step": 17401 }, { "epoch": 2.213713268032057, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.355341911315918, "learning_rate": 5e-05, "loss": 0.784, "mean_token_accuracy": 0.8847607374191284, "num_tokens": 660727840.0, "step": 17402 }, { "epoch": 2.2138404783106473, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.734879493713379, "learning_rate": 5e-05, "loss": 0.6593, "mean_token_accuracy": 0.8887572884559631, "num_tokens": 660770107.0, "step": 17403 }, { "epoch": 2.213967688589238, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.752087593078613, "learning_rate": 5e-05, "loss": 0.8933, "mean_token_accuracy": 0.8666350245475769, "num_tokens": 660809813.0, "step": 17404 }, { "epoch": 2.2140948988678284, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.567330837249756, "learning_rate": 5e-05, "loss": 0.701, "mean_token_accuracy": 0.8751527070999146, "num_tokens": 660849427.0, "step": 17405 }, { "epoch": 2.214222109146419, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.065630912780762, "learning_rate": 5e-05, "loss": 0.8622, "mean_token_accuracy": 0.8911643624305725, "num_tokens": 660888456.0, "step": 17406 }, { "epoch": 2.2143493194250095, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.697768688201904, "learning_rate": 5e-05, "loss": 0.6747, "mean_token_accuracy": 0.8934667110443115, "num_tokens": 660931889.0, "step": 17407 }, { "epoch": 2.2144765297036, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.027325630187988, "learning_rate": 5e-05, "loss": 0.9164, "mean_token_accuracy": 0.8854873776435852, "num_tokens": 660972965.0, "step": 17408 }, { "epoch": 2.2146037399821905, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.326910495758057, "learning_rate": 5e-05, "loss": 0.7869, "mean_token_accuracy": 0.8804895281791687, "num_tokens": 661007352.0, "step": 17409 }, { "epoch": 2.214730950260781, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.24038314819336, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.8853419423103333, "num_tokens": 661044202.0, "step": 17410 }, { "epoch": 2.2148581605393716, "ewc_loss": 0.5, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.046748161315918, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.8839730620384216, "num_tokens": 661088542.0, "step": 17411 }, { "epoch": 2.214985370817962, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.234498023986816, "learning_rate": 5e-05, "loss": 0.7637, "mean_token_accuracy": 0.873832106590271, "num_tokens": 661130060.0, "step": 17412 }, { "epoch": 2.2151125810965526, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.57963752746582, "learning_rate": 5e-05, "loss": 0.8749, "mean_token_accuracy": 0.8754419088363647, "num_tokens": 661170325.0, "step": 17413 }, { "epoch": 2.215239791375143, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.690036296844482, "learning_rate": 5e-05, "loss": 0.6966, "mean_token_accuracy": 0.8794620037078857, "num_tokens": 661216151.0, "step": 17414 }, { "epoch": 2.2153670016537337, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.924870491027832, "learning_rate": 5e-05, "loss": 0.8598, "mean_token_accuracy": 0.8880332708358765, "num_tokens": 661257857.0, "step": 17415 }, { "epoch": 2.215494211932324, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.6244964599609375, "learning_rate": 5e-05, "loss": 0.6813, "mean_token_accuracy": 0.8840816617012024, "num_tokens": 661294113.0, "step": 17416 }, { "epoch": 2.2156214222109147, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.04973316192627, "learning_rate": 5e-05, "loss": 0.922, "mean_token_accuracy": 0.8807231187820435, "num_tokens": 661332681.0, "step": 17417 }, { "epoch": 2.2157486324895053, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.926087856292725, "learning_rate": 5e-05, "loss": 0.7439, "mean_token_accuracy": 0.8781530857086182, "num_tokens": 661373567.0, "step": 17418 }, { "epoch": 2.215875842768096, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.793599128723145, "learning_rate": 5e-05, "loss": 0.9512, "mean_token_accuracy": 0.8746743202209473, "num_tokens": 661415727.0, "step": 17419 }, { "epoch": 2.2160030530466863, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.4466400146484375, "learning_rate": 5e-05, "loss": 0.7841, "mean_token_accuracy": 0.8865300416946411, "num_tokens": 661454934.0, "step": 17420 }, { "epoch": 2.216130263325277, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.2188720703125, "learning_rate": 5e-05, "loss": 0.8985, "mean_token_accuracy": 0.8700491189956665, "num_tokens": 661499070.0, "step": 17421 }, { "epoch": 2.2162574736038674, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.763062000274658, "learning_rate": 5e-05, "loss": 0.8031, "mean_token_accuracy": 0.885753333568573, "num_tokens": 661538595.0, "step": 17422 }, { "epoch": 2.216384683882458, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.840651035308838, "learning_rate": 5e-05, "loss": 0.8187, "mean_token_accuracy": 0.875648021697998, "num_tokens": 661578472.0, "step": 17423 }, { "epoch": 2.216511894161048, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.631565570831299, "learning_rate": 5e-05, "loss": 0.7796, "mean_token_accuracy": 0.875665545463562, "num_tokens": 661615127.0, "step": 17424 }, { "epoch": 2.2166391044396385, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.010272979736328, "learning_rate": 5e-05, "loss": 0.8366, "mean_token_accuracy": 0.8709204196929932, "num_tokens": 661655514.0, "step": 17425 }, { "epoch": 2.216766314718229, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.135528087615967, "learning_rate": 5e-05, "loss": 0.7469, "mean_token_accuracy": 0.8718359470367432, "num_tokens": 661690838.0, "step": 17426 }, { "epoch": 2.2168935249968196, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.505565643310547, "learning_rate": 5e-05, "loss": 0.8336, "mean_token_accuracy": 0.8833380937576294, "num_tokens": 661735697.0, "step": 17427 }, { "epoch": 2.21702073527541, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.6832075119018555, "learning_rate": 5e-05, "loss": 0.7425, "mean_token_accuracy": 0.8682860136032104, "num_tokens": 661766875.0, "step": 17428 }, { "epoch": 2.2171479455540006, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.989754676818848, "learning_rate": 5e-05, "loss": 0.8869, "mean_token_accuracy": 0.8817348480224609, "num_tokens": 661798059.0, "step": 17429 }, { "epoch": 2.217275155832591, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.565794467926025, "learning_rate": 5e-05, "loss": 0.6619, "mean_token_accuracy": 0.8918496370315552, "num_tokens": 661833128.0, "step": 17430 }, { "epoch": 2.2174023661111817, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.172101020812988, "learning_rate": 5e-05, "loss": 0.9571, "mean_token_accuracy": 0.8740427494049072, "num_tokens": 661872672.0, "step": 17431 }, { "epoch": 2.217529576389772, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.018311977386475, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8731088042259216, "num_tokens": 661919611.0, "step": 17432 }, { "epoch": 2.2176567866683627, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.696900367736816, "learning_rate": 5e-05, "loss": 0.9758, "mean_token_accuracy": 0.8635793924331665, "num_tokens": 661955349.0, "step": 17433 }, { "epoch": 2.2177839969469533, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.7777934074401855, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.888727068901062, "num_tokens": 661990178.0, "step": 17434 }, { "epoch": 2.217911207225544, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.673470973968506, "learning_rate": 5e-05, "loss": 0.7875, "mean_token_accuracy": 0.8862438201904297, "num_tokens": 662035855.0, "step": 17435 }, { "epoch": 2.2180384175041343, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.349188804626465, "learning_rate": 5e-05, "loss": 0.8658, "mean_token_accuracy": 0.8742538094520569, "num_tokens": 662073959.0, "step": 17436 }, { "epoch": 2.218165627782725, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.973838806152344, "learning_rate": 5e-05, "loss": 0.6587, "mean_token_accuracy": 0.9011951684951782, "num_tokens": 662108598.0, "step": 17437 }, { "epoch": 2.2182928380613154, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.558279037475586, "learning_rate": 5e-05, "loss": 0.854, "mean_token_accuracy": 0.8801808953285217, "num_tokens": 662136316.0, "step": 17438 }, { "epoch": 2.218420048339906, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.8798627853393555, "learning_rate": 5e-05, "loss": 0.6954, "mean_token_accuracy": 0.8841522932052612, "num_tokens": 662175771.0, "step": 17439 }, { "epoch": 2.2185472586184964, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.474186897277832, "learning_rate": 5e-05, "loss": 0.8469, "mean_token_accuracy": 0.8790150284767151, "num_tokens": 662219031.0, "step": 17440 }, { "epoch": 2.218674468897087, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.997104167938232, "learning_rate": 5e-05, "loss": 0.69, "mean_token_accuracy": 0.8878767490386963, "num_tokens": 662257734.0, "step": 17441 }, { "epoch": 2.2188016791756775, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.37989330291748, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.88304603099823, "num_tokens": 662297329.0, "step": 17442 }, { "epoch": 2.218928889454268, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.071939468383789, "learning_rate": 5e-05, "loss": 0.7268, "mean_token_accuracy": 0.8761164546012878, "num_tokens": 662334597.0, "step": 17443 }, { "epoch": 2.2190560997328586, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.329367637634277, "learning_rate": 5e-05, "loss": 0.8269, "mean_token_accuracy": 0.8775948286056519, "num_tokens": 662370369.0, "step": 17444 }, { "epoch": 2.219183310011449, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.1459221839904785, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8762181997299194, "num_tokens": 662406193.0, "step": 17445 }, { "epoch": 2.2193105202900396, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.189729690551758, "learning_rate": 5e-05, "loss": 0.7666, "mean_token_accuracy": 0.8949247598648071, "num_tokens": 662439692.0, "step": 17446 }, { "epoch": 2.21943773056863, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.248890399932861, "learning_rate": 5e-05, "loss": 0.7228, "mean_token_accuracy": 0.8853739500045776, "num_tokens": 662479768.0, "step": 17447 }, { "epoch": 2.2195649408472207, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.01861572265625, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.8847610950469971, "num_tokens": 662523591.0, "step": 17448 }, { "epoch": 2.2196921511258108, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.371147155761719, "learning_rate": 5e-05, "loss": 0.7255, "mean_token_accuracy": 0.8873154520988464, "num_tokens": 662557693.0, "step": 17449 }, { "epoch": 2.2198193614044013, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.855953216552734, "learning_rate": 5e-05, "loss": 0.7733, "mean_token_accuracy": 0.8802249431610107, "num_tokens": 662595404.0, "step": 17450 }, { "epoch": 2.219946571682992, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.41262149810791, "learning_rate": 5e-05, "loss": 0.727, "mean_token_accuracy": 0.8813961744308472, "num_tokens": 662637033.0, "step": 17451 }, { "epoch": 2.2200737819615823, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.8067545890808105, "learning_rate": 5e-05, "loss": 0.7661, "mean_token_accuracy": 0.8812922239303589, "num_tokens": 662675199.0, "step": 17452 }, { "epoch": 2.220200992240173, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.334721088409424, "learning_rate": 5e-05, "loss": 0.7172, "mean_token_accuracy": 0.8811163306236267, "num_tokens": 662720809.0, "step": 17453 }, { "epoch": 2.2203282025187634, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.893350601196289, "learning_rate": 5e-05, "loss": 0.7164, "mean_token_accuracy": 0.8952329158782959, "num_tokens": 662759894.0, "step": 17454 }, { "epoch": 2.220455412797354, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.184420585632324, "learning_rate": 5e-05, "loss": 0.6705, "mean_token_accuracy": 0.8920034170150757, "num_tokens": 662795373.0, "step": 17455 }, { "epoch": 2.2205826230759445, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.070276260375977, "learning_rate": 5e-05, "loss": 0.7459, "mean_token_accuracy": 0.8905968070030212, "num_tokens": 662832764.0, "step": 17456 }, { "epoch": 2.220709833354535, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.963521957397461, "learning_rate": 5e-05, "loss": 0.7331, "mean_token_accuracy": 0.8673444986343384, "num_tokens": 662872796.0, "step": 17457 }, { "epoch": 2.2208370436331255, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.355951309204102, "learning_rate": 5e-05, "loss": 0.8439, "mean_token_accuracy": 0.871396541595459, "num_tokens": 662914027.0, "step": 17458 }, { "epoch": 2.220964253911716, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.665503025054932, "learning_rate": 5e-05, "loss": 0.7059, "mean_token_accuracy": 0.8722427487373352, "num_tokens": 662953924.0, "step": 17459 }, { "epoch": 2.2210914641903066, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.780588150024414, "learning_rate": 5e-05, "loss": 0.8598, "mean_token_accuracy": 0.8783711194992065, "num_tokens": 662982792.0, "step": 17460 }, { "epoch": 2.221218674468897, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.38927698135376, "learning_rate": 5e-05, "loss": 0.6545, "mean_token_accuracy": 0.8888236284255981, "num_tokens": 663023674.0, "step": 17461 }, { "epoch": 2.2213458847474876, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.190899848937988, "learning_rate": 5e-05, "loss": 0.9251, "mean_token_accuracy": 0.8740923404693604, "num_tokens": 663064187.0, "step": 17462 }, { "epoch": 2.221473095026078, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.47163724899292, "learning_rate": 5e-05, "loss": 0.6602, "mean_token_accuracy": 0.8948817849159241, "num_tokens": 663109146.0, "step": 17463 }, { "epoch": 2.2216003053046687, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.128130912780762, "learning_rate": 5e-05, "loss": 0.9133, "mean_token_accuracy": 0.8876246809959412, "num_tokens": 663147988.0, "step": 17464 }, { "epoch": 2.221727515583259, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.232512474060059, "learning_rate": 5e-05, "loss": 0.797, "mean_token_accuracy": 0.8872833251953125, "num_tokens": 663190072.0, "step": 17465 }, { "epoch": 2.2218547258618497, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.135030746459961, "learning_rate": 5e-05, "loss": 0.8416, "mean_token_accuracy": 0.8815783262252808, "num_tokens": 663222896.0, "step": 17466 }, { "epoch": 2.2219819361404403, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.250956535339355, "learning_rate": 5e-05, "loss": 0.8787, "mean_token_accuracy": 0.8814399838447571, "num_tokens": 663261846.0, "step": 17467 }, { "epoch": 2.222109146419031, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.766304016113281, "learning_rate": 5e-05, "loss": 0.6427, "mean_token_accuracy": 0.8994749784469604, "num_tokens": 663295766.0, "step": 17468 }, { "epoch": 2.2222363566976213, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.32423210144043, "learning_rate": 5e-05, "loss": 0.8947, "mean_token_accuracy": 0.8927288055419922, "num_tokens": 663333218.0, "step": 17469 }, { "epoch": 2.222363566976212, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 5.990816116333008, "learning_rate": 5e-05, "loss": 0.6682, "mean_token_accuracy": 0.8836714625358582, "num_tokens": 663371491.0, "step": 17470 }, { "epoch": 2.2224907772548024, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.254216194152832, "learning_rate": 5e-05, "loss": 0.9712, "mean_token_accuracy": 0.8960497379302979, "num_tokens": 663410824.0, "step": 17471 }, { "epoch": 2.222617987533393, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 6.634702205657959, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.8839517831802368, "num_tokens": 663445624.0, "step": 17472 }, { "epoch": 2.2227451978119834, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.24431324005127, "learning_rate": 5e-05, "loss": 1.1041, "mean_token_accuracy": 0.8718201518058777, "num_tokens": 663483223.0, "step": 17473 }, { "epoch": 2.2228724080905735, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 8.36640739440918, "learning_rate": 5e-05, "loss": 1.0415, "mean_token_accuracy": 0.8824617862701416, "num_tokens": 663520577.0, "step": 17474 }, { "epoch": 2.222999618369164, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.580057144165039, "learning_rate": 5e-05, "loss": 0.8994, "mean_token_accuracy": 0.8806921243667603, "num_tokens": 663554667.0, "step": 17475 }, { "epoch": 2.2231268286477546, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 9.29742431640625, "learning_rate": 5e-05, "loss": 1.0283, "mean_token_accuracy": 0.8892166614532471, "num_tokens": 663593475.0, "step": 17476 }, { "epoch": 2.223254038926345, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.092981338500977, "learning_rate": 5e-05, "loss": 0.7635, "mean_token_accuracy": 0.8890103101730347, "num_tokens": 663635957.0, "step": 17477 }, { "epoch": 2.2233812492049356, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.905984878540039, "learning_rate": 5e-05, "loss": 0.9948, "mean_token_accuracy": 0.8924700617790222, "num_tokens": 663678610.0, "step": 17478 }, { "epoch": 2.223508459483526, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.417090892791748, "learning_rate": 5e-05, "loss": 0.886, "mean_token_accuracy": 0.8829314708709717, "num_tokens": 663718204.0, "step": 17479 }, { "epoch": 2.2236356697621167, "ewc_loss": 0.625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.235045433044434, "learning_rate": 5e-05, "loss": 0.9526, "mean_token_accuracy": 0.8847380876541138, "num_tokens": 663760501.0, "step": 17480 }, { "epoch": 2.223762880040707, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.268964767456055, "learning_rate": 5e-05, "loss": 0.9448, "mean_token_accuracy": 0.8918743133544922, "num_tokens": 663802515.0, "step": 17481 }, { "epoch": 2.2238900903192977, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.897753715515137, "learning_rate": 5e-05, "loss": 0.8168, "mean_token_accuracy": 0.8761340379714966, "num_tokens": 663839095.0, "step": 17482 }, { "epoch": 2.2240173005978883, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.809602737426758, "learning_rate": 5e-05, "loss": 0.9889, "mean_token_accuracy": 0.8701612949371338, "num_tokens": 663875734.0, "step": 17483 }, { "epoch": 2.224144510876479, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.244133472442627, "learning_rate": 5e-05, "loss": 0.741, "mean_token_accuracy": 0.8895204067230225, "num_tokens": 663911321.0, "step": 17484 }, { "epoch": 2.2242717211550693, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.957435607910156, "learning_rate": 5e-05, "loss": 0.9334, "mean_token_accuracy": 0.8788910508155823, "num_tokens": 663942334.0, "step": 17485 }, { "epoch": 2.22439893143366, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.345132827758789, "learning_rate": 5e-05, "loss": 0.8188, "mean_token_accuracy": 0.8812843561172485, "num_tokens": 663976306.0, "step": 17486 }, { "epoch": 2.2245261417122504, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.591024398803711, "learning_rate": 5e-05, "loss": 0.8706, "mean_token_accuracy": 0.8857848644256592, "num_tokens": 664016234.0, "step": 17487 }, { "epoch": 2.224653351990841, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.589206695556641, "learning_rate": 5e-05, "loss": 0.8595, "mean_token_accuracy": 0.8799057602882385, "num_tokens": 664057012.0, "step": 17488 }, { "epoch": 2.2247805622694314, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.107422828674316, "learning_rate": 5e-05, "loss": 0.8129, "mean_token_accuracy": 0.8825922608375549, "num_tokens": 664092911.0, "step": 17489 }, { "epoch": 2.224907772548022, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.685414791107178, "learning_rate": 5e-05, "loss": 0.8511, "mean_token_accuracy": 0.8765920400619507, "num_tokens": 664133069.0, "step": 17490 }, { "epoch": 2.2250349828266125, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.799861907958984, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8837704658508301, "num_tokens": 664166059.0, "step": 17491 }, { "epoch": 2.225162193105203, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.724385738372803, "learning_rate": 5e-05, "loss": 0.8195, "mean_token_accuracy": 0.874642550945282, "num_tokens": 664204918.0, "step": 17492 }, { "epoch": 2.2252894033837936, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.609622001647949, "learning_rate": 5e-05, "loss": 0.7435, "mean_token_accuracy": 0.8843042850494385, "num_tokens": 664240561.0, "step": 17493 }, { "epoch": 2.225416613662384, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.78924036026001, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8770629167556763, "num_tokens": 664274905.0, "step": 17494 }, { "epoch": 2.2255438239409746, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.452542781829834, "learning_rate": 5e-05, "loss": 0.7396, "mean_token_accuracy": 0.878028929233551, "num_tokens": 664316666.0, "step": 17495 }, { "epoch": 2.225671034219565, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.842247009277344, "learning_rate": 5e-05, "loss": 0.7353, "mean_token_accuracy": 0.8894516229629517, "num_tokens": 664358590.0, "step": 17496 }, { "epoch": 2.225798244498155, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.342453956604004, "learning_rate": 5e-05, "loss": 0.7181, "mean_token_accuracy": 0.8821455240249634, "num_tokens": 664394764.0, "step": 17497 }, { "epoch": 2.225925454776746, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.891883373260498, "learning_rate": 5e-05, "loss": 0.7374, "mean_token_accuracy": 0.8866978883743286, "num_tokens": 664433730.0, "step": 17498 }, { "epoch": 2.2260526650553363, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.25846004486084, "learning_rate": 5e-05, "loss": 0.6994, "mean_token_accuracy": 0.8834111094474792, "num_tokens": 664469926.0, "step": 17499 }, { "epoch": 2.226179875333927, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.958599090576172, "learning_rate": 5e-05, "loss": 0.7609, "mean_token_accuracy": 0.8749833106994629, "num_tokens": 664511204.0, "step": 17500 }, { "epoch": 2.2263070856125173, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.148481845855713, "learning_rate": 5e-05, "loss": 0.7483, "mean_token_accuracy": 0.8688860535621643, "num_tokens": 664547474.0, "step": 17501 }, { "epoch": 2.226434295891108, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.101696968078613, "learning_rate": 5e-05, "loss": 0.7495, "mean_token_accuracy": 0.8814458847045898, "num_tokens": 664579252.0, "step": 17502 }, { "epoch": 2.2265615061696984, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.937838077545166, "learning_rate": 5e-05, "loss": 0.7115, "mean_token_accuracy": 0.8727078437805176, "num_tokens": 664612129.0, "step": 17503 }, { "epoch": 2.226688716448289, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.350680351257324, "learning_rate": 5e-05, "loss": 0.7294, "mean_token_accuracy": 0.897254228591919, "num_tokens": 664650847.0, "step": 17504 }, { "epoch": 2.2268159267268794, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.644679069519043, "learning_rate": 5e-05, "loss": 0.6919, "mean_token_accuracy": 0.8721553683280945, "num_tokens": 664694056.0, "step": 17505 }, { "epoch": 2.22694313700547, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.796037673950195, "learning_rate": 5e-05, "loss": 0.8379, "mean_token_accuracy": 0.8767355680465698, "num_tokens": 664732868.0, "step": 17506 }, { "epoch": 2.2270703472840605, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.27398157119751, "learning_rate": 5e-05, "loss": 0.6398, "mean_token_accuracy": 0.8876768350601196, "num_tokens": 664772345.0, "step": 17507 }, { "epoch": 2.227197557562651, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.420373916625977, "learning_rate": 5e-05, "loss": 0.9179, "mean_token_accuracy": 0.8751893639564514, "num_tokens": 664811398.0, "step": 17508 }, { "epoch": 2.2273247678412416, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.075170040130615, "learning_rate": 5e-05, "loss": 0.695, "mean_token_accuracy": 0.8787418603897095, "num_tokens": 664846773.0, "step": 17509 }, { "epoch": 2.227451978119832, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.947466850280762, "learning_rate": 5e-05, "loss": 0.9708, "mean_token_accuracy": 0.8854024410247803, "num_tokens": 664887027.0, "step": 17510 }, { "epoch": 2.2275791883984226, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 6.7726664543151855, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8948996663093567, "num_tokens": 664922798.0, "step": 17511 }, { "epoch": 2.227706398677013, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.606817245483398, "learning_rate": 5e-05, "loss": 0.9773, "mean_token_accuracy": 0.8847070932388306, "num_tokens": 664965862.0, "step": 17512 }, { "epoch": 2.2278336089556037, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.056614875793457, "learning_rate": 5e-05, "loss": 1.023, "mean_token_accuracy": 0.863743782043457, "num_tokens": 665006053.0, "step": 17513 }, { "epoch": 2.227960819234194, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.033910751342773, "learning_rate": 5e-05, "loss": 0.8376, "mean_token_accuracy": 0.8813751935958862, "num_tokens": 665040003.0, "step": 17514 }, { "epoch": 2.2280880295127847, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.108321189880371, "learning_rate": 5e-05, "loss": 0.9904, "mean_token_accuracy": 0.8760635852813721, "num_tokens": 665079989.0, "step": 17515 }, { "epoch": 2.2282152397913753, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.663915634155273, "learning_rate": 5e-05, "loss": 0.7056, "mean_token_accuracy": 0.8882719874382019, "num_tokens": 665114009.0, "step": 17516 }, { "epoch": 2.228342450069966, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.007226943969727, "learning_rate": 5e-05, "loss": 1.0152, "mean_token_accuracy": 0.8829039335250854, "num_tokens": 665153259.0, "step": 17517 }, { "epoch": 2.2284696603485563, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 6.96140193939209, "learning_rate": 5e-05, "loss": 0.7987, "mean_token_accuracy": 0.8916328549385071, "num_tokens": 665195024.0, "step": 17518 }, { "epoch": 2.228596870627147, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.869933128356934, "learning_rate": 5e-05, "loss": 1.0129, "mean_token_accuracy": 0.8827160000801086, "num_tokens": 665236000.0, "step": 17519 }, { "epoch": 2.2287240809057374, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.201959609985352, "learning_rate": 5e-05, "loss": 1.0323, "mean_token_accuracy": 0.8621735572814941, "num_tokens": 665277837.0, "step": 17520 }, { "epoch": 2.228851291184328, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.549245834350586, "learning_rate": 5e-05, "loss": 0.8588, "mean_token_accuracy": 0.8870954513549805, "num_tokens": 665312927.0, "step": 17521 }, { "epoch": 2.228978501462918, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.71530818939209, "learning_rate": 5e-05, "loss": 0.9628, "mean_token_accuracy": 0.8880559206008911, "num_tokens": 665346962.0, "step": 17522 }, { "epoch": 2.2291057117415085, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.645139217376709, "learning_rate": 5e-05, "loss": 0.7953, "mean_token_accuracy": 0.884524941444397, "num_tokens": 665385684.0, "step": 17523 }, { "epoch": 2.229232922020099, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.600342750549316, "learning_rate": 5e-05, "loss": 0.8606, "mean_token_accuracy": 0.9007322788238525, "num_tokens": 665421194.0, "step": 17524 }, { "epoch": 2.2293601322986896, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.70306921005249, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.8797807097434998, "num_tokens": 665459731.0, "step": 17525 }, { "epoch": 2.22948734257728, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.9156317710876465, "learning_rate": 5e-05, "loss": 0.8047, "mean_token_accuracy": 0.8838053345680237, "num_tokens": 665490310.0, "step": 17526 }, { "epoch": 2.2296145528558706, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.03569507598877, "learning_rate": 5e-05, "loss": 0.8652, "mean_token_accuracy": 0.8774072527885437, "num_tokens": 665527543.0, "step": 17527 }, { "epoch": 2.229741763134461, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.357797145843506, "learning_rate": 5e-05, "loss": 0.6902, "mean_token_accuracy": 0.8928024768829346, "num_tokens": 665562555.0, "step": 17528 }, { "epoch": 2.2298689734130517, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.312234878540039, "learning_rate": 5e-05, "loss": 0.8403, "mean_token_accuracy": 0.885269045829773, "num_tokens": 665598165.0, "step": 17529 }, { "epoch": 2.229996183691642, "ewc_loss": 0.375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.999925136566162, "learning_rate": 5e-05, "loss": 0.7375, "mean_token_accuracy": 0.868890106678009, "num_tokens": 665637225.0, "step": 17530 }, { "epoch": 2.2301233939702327, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.650026321411133, "learning_rate": 5e-05, "loss": 0.883, "mean_token_accuracy": 0.877443790435791, "num_tokens": 665676126.0, "step": 17531 }, { "epoch": 2.2302506042488233, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.683982849121094, "learning_rate": 5e-05, "loss": 0.6937, "mean_token_accuracy": 0.8865243792533875, "num_tokens": 665713570.0, "step": 17532 }, { "epoch": 2.230377814527414, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.142350196838379, "learning_rate": 5e-05, "loss": 0.8722, "mean_token_accuracy": 0.8921986818313599, "num_tokens": 665743648.0, "step": 17533 }, { "epoch": 2.2305050248060043, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.558177947998047, "learning_rate": 5e-05, "loss": 0.7073, "mean_token_accuracy": 0.8837838172912598, "num_tokens": 665780249.0, "step": 17534 }, { "epoch": 2.230632235084595, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.416383743286133, "learning_rate": 5e-05, "loss": 0.9092, "mean_token_accuracy": 0.8954179286956787, "num_tokens": 665821175.0, "step": 17535 }, { "epoch": 2.2307594453631854, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 6.957800388336182, "learning_rate": 5e-05, "loss": 0.763, "mean_token_accuracy": 0.8931016325950623, "num_tokens": 665854598.0, "step": 17536 }, { "epoch": 2.230886655641776, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.189846992492676, "learning_rate": 5e-05, "loss": 0.9179, "mean_token_accuracy": 0.891366720199585, "num_tokens": 665892294.0, "step": 17537 }, { "epoch": 2.2310138659203664, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.464026927947998, "learning_rate": 5e-05, "loss": 0.8868, "mean_token_accuracy": 0.8738498687744141, "num_tokens": 665928575.0, "step": 17538 }, { "epoch": 2.231141076198957, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.720978736877441, "learning_rate": 5e-05, "loss": 0.8689, "mean_token_accuracy": 0.8955038785934448, "num_tokens": 665963136.0, "step": 17539 }, { "epoch": 2.2312682864775475, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.605607509613037, "learning_rate": 5e-05, "loss": 0.864, "mean_token_accuracy": 0.8837554454803467, "num_tokens": 666008556.0, "step": 17540 }, { "epoch": 2.231395496756138, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.474282264709473, "learning_rate": 5e-05, "loss": 0.8837, "mean_token_accuracy": 0.8818475604057312, "num_tokens": 666047066.0, "step": 17541 }, { "epoch": 2.2315227070347285, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.406682014465332, "learning_rate": 5e-05, "loss": 0.8041, "mean_token_accuracy": 0.8868584632873535, "num_tokens": 666084998.0, "step": 17542 }, { "epoch": 2.231649917313319, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.56747055053711, "learning_rate": 5e-05, "loss": 0.9013, "mean_token_accuracy": 0.8809945583343506, "num_tokens": 666124935.0, "step": 17543 }, { "epoch": 2.2317771275919096, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.148805618286133, "learning_rate": 5e-05, "loss": 0.8176, "mean_token_accuracy": 0.8681992888450623, "num_tokens": 666165580.0, "step": 17544 }, { "epoch": 2.2319043378705, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.759893417358398, "learning_rate": 5e-05, "loss": 0.9309, "mean_token_accuracy": 0.874753475189209, "num_tokens": 666206020.0, "step": 17545 }, { "epoch": 2.2320315481490907, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.133044719696045, "learning_rate": 5e-05, "loss": 0.8155, "mean_token_accuracy": 0.8726847171783447, "num_tokens": 666239757.0, "step": 17546 }, { "epoch": 2.2321587584276807, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.644636154174805, "learning_rate": 5e-05, "loss": 0.8628, "mean_token_accuracy": 0.8977491855621338, "num_tokens": 666278426.0, "step": 17547 }, { "epoch": 2.2322859687062713, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.413064479827881, "learning_rate": 5e-05, "loss": 0.7994, "mean_token_accuracy": 0.8854637742042542, "num_tokens": 666315784.0, "step": 17548 }, { "epoch": 2.232413178984862, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.196619987487793, "learning_rate": 5e-05, "loss": 0.8685, "mean_token_accuracy": 0.874639630317688, "num_tokens": 666353851.0, "step": 17549 }, { "epoch": 2.2325403892634523, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.69349479675293, "learning_rate": 5e-05, "loss": 0.8378, "mean_token_accuracy": 0.8793061971664429, "num_tokens": 666391208.0, "step": 17550 }, { "epoch": 2.232667599542043, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.762328147888184, "learning_rate": 5e-05, "loss": 0.7999, "mean_token_accuracy": 0.882575511932373, "num_tokens": 666434477.0, "step": 17551 }, { "epoch": 2.2327948098206334, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.786541938781738, "learning_rate": 5e-05, "loss": 0.8166, "mean_token_accuracy": 0.8803842663764954, "num_tokens": 666472481.0, "step": 17552 }, { "epoch": 2.232922020099224, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.586174964904785, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8788537979125977, "num_tokens": 666502858.0, "step": 17553 }, { "epoch": 2.2330492303778144, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.708502769470215, "learning_rate": 5e-05, "loss": 0.8143, "mean_token_accuracy": 0.8668222427368164, "num_tokens": 666547311.0, "step": 17554 }, { "epoch": 2.233176440656405, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.6008195877075195, "learning_rate": 5e-05, "loss": 0.7233, "mean_token_accuracy": 0.889137864112854, "num_tokens": 666587818.0, "step": 17555 }, { "epoch": 2.2333036509349955, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.583280563354492, "learning_rate": 5e-05, "loss": 0.7271, "mean_token_accuracy": 0.8833553194999695, "num_tokens": 666626071.0, "step": 17556 }, { "epoch": 2.233430861213586, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.667208194732666, "learning_rate": 5e-05, "loss": 0.6912, "mean_token_accuracy": 0.8966687321662903, "num_tokens": 666659221.0, "step": 17557 }, { "epoch": 2.2335580714921766, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.463853359222412, "learning_rate": 5e-05, "loss": 0.73, "mean_token_accuracy": 0.8762836456298828, "num_tokens": 666696189.0, "step": 17558 }, { "epoch": 2.233685281770767, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.699631690979004, "learning_rate": 5e-05, "loss": 0.7646, "mean_token_accuracy": 0.8759024739265442, "num_tokens": 666738495.0, "step": 17559 }, { "epoch": 2.2338124920493576, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.430624008178711, "learning_rate": 5e-05, "loss": 0.6905, "mean_token_accuracy": 0.8867523074150085, "num_tokens": 666772848.0, "step": 17560 }, { "epoch": 2.233939702327948, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.68690824508667, "learning_rate": 5e-05, "loss": 0.708, "mean_token_accuracy": 0.8850324153900146, "num_tokens": 666813849.0, "step": 17561 }, { "epoch": 2.2340669126065387, "ewc_loss": 0.375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.406117916107178, "learning_rate": 5e-05, "loss": 0.7141, "mean_token_accuracy": 0.882293164730072, "num_tokens": 666848564.0, "step": 17562 }, { "epoch": 2.234194122885129, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.729939937591553, "learning_rate": 5e-05, "loss": 0.7201, "mean_token_accuracy": 0.8828012943267822, "num_tokens": 666877381.0, "step": 17563 }, { "epoch": 2.2343213331637197, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.310518264770508, "learning_rate": 5e-05, "loss": 0.7173, "mean_token_accuracy": 0.8744223713874817, "num_tokens": 666920452.0, "step": 17564 }, { "epoch": 2.2344485434423103, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.839671611785889, "learning_rate": 5e-05, "loss": 0.7054, "mean_token_accuracy": 0.8885586857795715, "num_tokens": 666958799.0, "step": 17565 }, { "epoch": 2.234575753720901, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.1727705001831055, "learning_rate": 5e-05, "loss": 0.6647, "mean_token_accuracy": 0.8856667876243591, "num_tokens": 666993899.0, "step": 17566 }, { "epoch": 2.2347029639994913, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.990227222442627, "learning_rate": 5e-05, "loss": 0.744, "mean_token_accuracy": 0.8802638053894043, "num_tokens": 667035154.0, "step": 17567 }, { "epoch": 2.234830174278082, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.967248916625977, "learning_rate": 5e-05, "loss": 0.6936, "mean_token_accuracy": 0.8750784993171692, "num_tokens": 667077047.0, "step": 17568 }, { "epoch": 2.2349573845566724, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.292394638061523, "learning_rate": 5e-05, "loss": 0.7509, "mean_token_accuracy": 0.8840214014053345, "num_tokens": 667114298.0, "step": 17569 }, { "epoch": 2.235084594835263, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.601273059844971, "learning_rate": 5e-05, "loss": 0.6434, "mean_token_accuracy": 0.8865953683853149, "num_tokens": 667157942.0, "step": 17570 }, { "epoch": 2.2352118051138534, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.84328842163086, "learning_rate": 5e-05, "loss": 0.8443, "mean_token_accuracy": 0.8752319812774658, "num_tokens": 667195547.0, "step": 17571 }, { "epoch": 2.2353390153924435, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.114084720611572, "learning_rate": 5e-05, "loss": 0.6092, "mean_token_accuracy": 0.8969443440437317, "num_tokens": 667232484.0, "step": 17572 }, { "epoch": 2.235466225671034, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.625176429748535, "learning_rate": 5e-05, "loss": 0.9549, "mean_token_accuracy": 0.8720383048057556, "num_tokens": 667267661.0, "step": 17573 }, { "epoch": 2.2355934359496246, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 5.98130464553833, "learning_rate": 5e-05, "loss": 0.6416, "mean_token_accuracy": 0.8925943374633789, "num_tokens": 667303963.0, "step": 17574 }, { "epoch": 2.235720646228215, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.112750053405762, "learning_rate": 5e-05, "loss": 1.0021, "mean_token_accuracy": 0.8809967041015625, "num_tokens": 667344103.0, "step": 17575 }, { "epoch": 2.2358478565068056, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.127476215362549, "learning_rate": 5e-05, "loss": 0.8492, "mean_token_accuracy": 0.8830965161323547, "num_tokens": 667377782.0, "step": 17576 }, { "epoch": 2.235975066785396, "ewc_loss": 0.625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.196099281311035, "learning_rate": 5e-05, "loss": 0.9565, "mean_token_accuracy": 0.883794903755188, "num_tokens": 667419015.0, "step": 17577 }, { "epoch": 2.2361022770639867, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.671825408935547, "learning_rate": 5e-05, "loss": 0.9873, "mean_token_accuracy": 0.8848716020584106, "num_tokens": 667457465.0, "step": 17578 }, { "epoch": 2.236229487342577, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.940821170806885, "learning_rate": 5e-05, "loss": 0.7495, "mean_token_accuracy": 0.8731705546379089, "num_tokens": 667497671.0, "step": 17579 }, { "epoch": 2.2363566976211677, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.530073165893555, "learning_rate": 5e-05, "loss": 1.0428, "mean_token_accuracy": 0.8926898241043091, "num_tokens": 667534258.0, "step": 17580 }, { "epoch": 2.2364839078997583, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.20523738861084, "learning_rate": 5e-05, "loss": 0.7208, "mean_token_accuracy": 0.883834183216095, "num_tokens": 667572689.0, "step": 17581 }, { "epoch": 2.236611118178349, "ewc_loss": 1.0078125, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.000942230224609375, "grad_norm": 12.447293281555176, "learning_rate": 5e-05, "loss": 1.3535, "mean_token_accuracy": 0.8787194490432739, "num_tokens": 667612425.0, "step": 17582 }, { "epoch": 2.2367383284569393, "ewc_loss": 0.921875, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 9.468521118164062, "learning_rate": 5e-05, "loss": 1.2713, "mean_token_accuracy": 0.8758993744850159, "num_tokens": 667653309.0, "step": 17583 }, { "epoch": 2.23686553873553, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 11.314229011535645, "learning_rate": 5e-05, "loss": 1.0824, "mean_token_accuracy": 0.8824806809425354, "num_tokens": 667699174.0, "step": 17584 }, { "epoch": 2.2369927490141204, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000823974609375, "grad_norm": 9.726221084594727, "learning_rate": 5e-05, "loss": 1.2226, "mean_token_accuracy": 0.8836051821708679, "num_tokens": 667738987.0, "step": 17585 }, { "epoch": 2.237119959292711, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.586875915527344, "learning_rate": 5e-05, "loss": 1.0576, "mean_token_accuracy": 0.8858264088630676, "num_tokens": 667780678.0, "step": 17586 }, { "epoch": 2.2372471695713014, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.919793128967285, "learning_rate": 5e-05, "loss": 0.9812, "mean_token_accuracy": 0.8861978054046631, "num_tokens": 667816298.0, "step": 17587 }, { "epoch": 2.237374379849892, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 9.904328346252441, "learning_rate": 5e-05, "loss": 1.079, "mean_token_accuracy": 0.8715100884437561, "num_tokens": 667855723.0, "step": 17588 }, { "epoch": 2.2375015901284825, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.35889720916748, "learning_rate": 5e-05, "loss": 0.9268, "mean_token_accuracy": 0.8750843405723572, "num_tokens": 667893895.0, "step": 17589 }, { "epoch": 2.237628800407073, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.707404136657715, "learning_rate": 5e-05, "loss": 0.9631, "mean_token_accuracy": 0.8967472314834595, "num_tokens": 667932061.0, "step": 17590 }, { "epoch": 2.2377560106856635, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.9927754402160645, "learning_rate": 5e-05, "loss": 0.8412, "mean_token_accuracy": 0.8899217844009399, "num_tokens": 667965723.0, "step": 17591 }, { "epoch": 2.237883220964254, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.743636131286621, "learning_rate": 5e-05, "loss": 0.9987, "mean_token_accuracy": 0.8781065940856934, "num_tokens": 668002310.0, "step": 17592 }, { "epoch": 2.2380104312428446, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.953047275543213, "learning_rate": 5e-05, "loss": 0.8958, "mean_token_accuracy": 0.8707349300384521, "num_tokens": 668044970.0, "step": 17593 }, { "epoch": 2.238137641521435, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.67178726196289, "learning_rate": 5e-05, "loss": 1.0144, "mean_token_accuracy": 0.8735909461975098, "num_tokens": 668081898.0, "step": 17594 }, { "epoch": 2.238264851800025, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.935123443603516, "learning_rate": 5e-05, "loss": 0.8645, "mean_token_accuracy": 0.8798060417175293, "num_tokens": 668119072.0, "step": 17595 }, { "epoch": 2.238392062078616, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.662444114685059, "learning_rate": 5e-05, "loss": 0.9627, "mean_token_accuracy": 0.8866925835609436, "num_tokens": 668157063.0, "step": 17596 }, { "epoch": 2.2385192723572063, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.885560035705566, "learning_rate": 5e-05, "loss": 0.916, "mean_token_accuracy": 0.8620216846466064, "num_tokens": 668196985.0, "step": 17597 }, { "epoch": 2.238646482635797, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.736327171325684, "learning_rate": 5e-05, "loss": 0.984, "mean_token_accuracy": 0.8896781802177429, "num_tokens": 668235954.0, "step": 17598 }, { "epoch": 2.2387736929143873, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.045191764831543, "learning_rate": 5e-05, "loss": 0.8709, "mean_token_accuracy": 0.8816884756088257, "num_tokens": 668275442.0, "step": 17599 }, { "epoch": 2.238900903192978, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.505509376525879, "learning_rate": 5e-05, "loss": 0.9997, "mean_token_accuracy": 0.8847250938415527, "num_tokens": 668311413.0, "step": 17600 }, { "epoch": 2.2390281134715684, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.220282554626465, "learning_rate": 5e-05, "loss": 0.834, "mean_token_accuracy": 0.8961454629898071, "num_tokens": 668346928.0, "step": 17601 }, { "epoch": 2.239155323750159, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.07644271850586, "learning_rate": 5e-05, "loss": 0.9301, "mean_token_accuracy": 0.892427921295166, "num_tokens": 668388566.0, "step": 17602 }, { "epoch": 2.2392825340287494, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.096985816955566, "learning_rate": 5e-05, "loss": 0.827, "mean_token_accuracy": 0.894792914390564, "num_tokens": 668430253.0, "step": 17603 }, { "epoch": 2.23940974430734, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.838388442993164, "learning_rate": 5e-05, "loss": 0.9431, "mean_token_accuracy": 0.8737467527389526, "num_tokens": 668470561.0, "step": 17604 }, { "epoch": 2.2395369545859305, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.826705455780029, "learning_rate": 5e-05, "loss": 0.8127, "mean_token_accuracy": 0.8831969499588013, "num_tokens": 668509718.0, "step": 17605 }, { "epoch": 2.239664164864521, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.842729568481445, "learning_rate": 5e-05, "loss": 0.9277, "mean_token_accuracy": 0.8788878917694092, "num_tokens": 668548367.0, "step": 17606 }, { "epoch": 2.2397913751431116, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.567334175109863, "learning_rate": 5e-05, "loss": 0.8356, "mean_token_accuracy": 0.8647306561470032, "num_tokens": 668586977.0, "step": 17607 }, { "epoch": 2.239918585421702, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.032215118408203, "learning_rate": 5e-05, "loss": 0.945, "mean_token_accuracy": 0.8825443983078003, "num_tokens": 668617801.0, "step": 17608 }, { "epoch": 2.2400457957002926, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.523558616638184, "learning_rate": 5e-05, "loss": 0.8094, "mean_token_accuracy": 0.8726174831390381, "num_tokens": 668656074.0, "step": 17609 }, { "epoch": 2.240173005978883, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.04533576965332, "learning_rate": 5e-05, "loss": 0.9205, "mean_token_accuracy": 0.8928409814834595, "num_tokens": 668698354.0, "step": 17610 }, { "epoch": 2.2403002162574737, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.787224769592285, "learning_rate": 5e-05, "loss": 0.8118, "mean_token_accuracy": 0.8895388245582581, "num_tokens": 668734822.0, "step": 17611 }, { "epoch": 2.240427426536064, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.595985412597656, "learning_rate": 5e-05, "loss": 0.8569, "mean_token_accuracy": 0.9020510911941528, "num_tokens": 668773136.0, "step": 17612 }, { "epoch": 2.2405546368146547, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.927975654602051, "learning_rate": 5e-05, "loss": 0.8254, "mean_token_accuracy": 0.8828461170196533, "num_tokens": 668811580.0, "step": 17613 }, { "epoch": 2.2406818470932452, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.204520225524902, "learning_rate": 5e-05, "loss": 0.8393, "mean_token_accuracy": 0.885661244392395, "num_tokens": 668846743.0, "step": 17614 }, { "epoch": 2.2408090573718358, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.718994140625, "learning_rate": 5e-05, "loss": 0.7586, "mean_token_accuracy": 0.8888846635818481, "num_tokens": 668885634.0, "step": 17615 }, { "epoch": 2.2409362676504263, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.216619491577148, "learning_rate": 5e-05, "loss": 0.8272, "mean_token_accuracy": 0.8831017017364502, "num_tokens": 668920319.0, "step": 17616 }, { "epoch": 2.241063477929017, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.4332275390625, "learning_rate": 5e-05, "loss": 0.7407, "mean_token_accuracy": 0.8855612277984619, "num_tokens": 668959741.0, "step": 17617 }, { "epoch": 2.2411906882076074, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.3512601852417, "learning_rate": 5e-05, "loss": 0.9287, "mean_token_accuracy": 0.8589084148406982, "num_tokens": 668998469.0, "step": 17618 }, { "epoch": 2.241317898486198, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.29597806930542, "learning_rate": 5e-05, "loss": 0.7575, "mean_token_accuracy": 0.8716245889663696, "num_tokens": 669030018.0, "step": 17619 }, { "epoch": 2.241445108764788, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.362735748291016, "learning_rate": 5e-05, "loss": 0.8153, "mean_token_accuracy": 0.8913449048995972, "num_tokens": 669066393.0, "step": 17620 }, { "epoch": 2.2415723190433785, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.292568683624268, "learning_rate": 5e-05, "loss": 0.7124, "mean_token_accuracy": 0.8897697925567627, "num_tokens": 669108321.0, "step": 17621 }, { "epoch": 2.241699529321969, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.34705924987793, "learning_rate": 5e-05, "loss": 0.8308, "mean_token_accuracy": 0.8852173089981079, "num_tokens": 669147878.0, "step": 17622 }, { "epoch": 2.2418267396005596, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.324796676635742, "learning_rate": 5e-05, "loss": 0.7612, "mean_token_accuracy": 0.872492253780365, "num_tokens": 669192348.0, "step": 17623 }, { "epoch": 2.24195394987915, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.192632675170898, "learning_rate": 5e-05, "loss": 0.8543, "mean_token_accuracy": 0.8739647269248962, "num_tokens": 669233763.0, "step": 17624 }, { "epoch": 2.2420811601577406, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.450850486755371, "learning_rate": 5e-05, "loss": 0.7151, "mean_token_accuracy": 0.8882099390029907, "num_tokens": 669269609.0, "step": 17625 }, { "epoch": 2.242208370436331, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.962834358215332, "learning_rate": 5e-05, "loss": 0.7821, "mean_token_accuracy": 0.8828603029251099, "num_tokens": 669308649.0, "step": 17626 }, { "epoch": 2.2423355807149217, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.55998420715332, "learning_rate": 5e-05, "loss": 0.734, "mean_token_accuracy": 0.8821372389793396, "num_tokens": 669350550.0, "step": 17627 }, { "epoch": 2.242462790993512, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.759103298187256, "learning_rate": 5e-05, "loss": 0.7579, "mean_token_accuracy": 0.8810392618179321, "num_tokens": 669392795.0, "step": 17628 }, { "epoch": 2.2425900012721027, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.629231929779053, "learning_rate": 5e-05, "loss": 0.733, "mean_token_accuracy": 0.881138265132904, "num_tokens": 669426928.0, "step": 17629 }, { "epoch": 2.2427172115506933, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.616988182067871, "learning_rate": 5e-05, "loss": 0.7187, "mean_token_accuracy": 0.8873347043991089, "num_tokens": 669467064.0, "step": 17630 }, { "epoch": 2.242844421829284, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.681279182434082, "learning_rate": 5e-05, "loss": 0.751, "mean_token_accuracy": 0.8734893798828125, "num_tokens": 669506656.0, "step": 17631 }, { "epoch": 2.2429716321078743, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.5066375732421875, "learning_rate": 5e-05, "loss": 0.8019, "mean_token_accuracy": 0.8541884422302246, "num_tokens": 669545787.0, "step": 17632 }, { "epoch": 2.243098842386465, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.696930408477783, "learning_rate": 5e-05, "loss": 0.7246, "mean_token_accuracy": 0.8809425830841064, "num_tokens": 669581242.0, "step": 17633 }, { "epoch": 2.2432260526650554, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.459105491638184, "learning_rate": 5e-05, "loss": 0.6607, "mean_token_accuracy": 0.8960474729537964, "num_tokens": 669621586.0, "step": 17634 }, { "epoch": 2.243353262943646, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.6934967041015625, "learning_rate": 5e-05, "loss": 0.7236, "mean_token_accuracy": 0.8831396102905273, "num_tokens": 669662903.0, "step": 17635 }, { "epoch": 2.2434804732222364, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.433019638061523, "learning_rate": 5e-05, "loss": 0.6917, "mean_token_accuracy": 0.8836544752120972, "num_tokens": 669698409.0, "step": 17636 }, { "epoch": 2.243607683500827, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.712098598480225, "learning_rate": 5e-05, "loss": 0.7588, "mean_token_accuracy": 0.8690892457962036, "num_tokens": 669741344.0, "step": 17637 }, { "epoch": 2.2437348937794175, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.373133659362793, "learning_rate": 5e-05, "loss": 0.669, "mean_token_accuracy": 0.8902809619903564, "num_tokens": 669780042.0, "step": 17638 }, { "epoch": 2.243862104058008, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.765810489654541, "learning_rate": 5e-05, "loss": 0.7699, "mean_token_accuracy": 0.865697979927063, "num_tokens": 669819340.0, "step": 17639 }, { "epoch": 2.2439893143365985, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.298789024353027, "learning_rate": 5e-05, "loss": 0.6537, "mean_token_accuracy": 0.8922802209854126, "num_tokens": 669859269.0, "step": 17640 }, { "epoch": 2.244116524615189, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.8256096839904785, "learning_rate": 5e-05, "loss": 0.7223, "mean_token_accuracy": 0.8831701278686523, "num_tokens": 669903085.0, "step": 17641 }, { "epoch": 2.2442437348937796, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.219101428985596, "learning_rate": 5e-05, "loss": 0.6653, "mean_token_accuracy": 0.8897061944007874, "num_tokens": 669945329.0, "step": 17642 }, { "epoch": 2.24437094517237, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.907603740692139, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.8874423503875732, "num_tokens": 669985141.0, "step": 17643 }, { "epoch": 2.2444981554509607, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 7.120667934417725, "learning_rate": 5e-05, "loss": 0.711, "mean_token_accuracy": 0.8718650341033936, "num_tokens": 670024034.0, "step": 17644 }, { "epoch": 2.2446253657295507, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 8.047636032104492, "learning_rate": 5e-05, "loss": 0.7363, "mean_token_accuracy": 0.8847524523735046, "num_tokens": 670062098.0, "step": 17645 }, { "epoch": 2.2447525760081413, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.956729888916016, "learning_rate": 5e-05, "loss": 0.6796, "mean_token_accuracy": 0.878930926322937, "num_tokens": 670100801.0, "step": 17646 }, { "epoch": 2.244879786286732, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.266682624816895, "learning_rate": 5e-05, "loss": 0.766, "mean_token_accuracy": 0.8781051635742188, "num_tokens": 670142221.0, "step": 17647 }, { "epoch": 2.2450069965653223, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.670374393463135, "learning_rate": 5e-05, "loss": 0.6327, "mean_token_accuracy": 0.8871428370475769, "num_tokens": 670183131.0, "step": 17648 }, { "epoch": 2.245134206843913, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.723047256469727, "learning_rate": 5e-05, "loss": 0.7439, "mean_token_accuracy": 0.8967870473861694, "num_tokens": 670219207.0, "step": 17649 }, { "epoch": 2.2452614171225034, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002346038818359375, "grad_norm": 6.254748821258545, "learning_rate": 5e-05, "loss": 0.6388, "mean_token_accuracy": 0.8820829391479492, "num_tokens": 670254085.0, "step": 17650 }, { "epoch": 2.245388627401094, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 9.342761039733887, "learning_rate": 5e-05, "loss": 0.9105, "mean_token_accuracy": 0.8733850717544556, "num_tokens": 670299669.0, "step": 17651 }, { "epoch": 2.2455158376796844, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.006009578704834, "learning_rate": 5e-05, "loss": 0.6087, "mean_token_accuracy": 0.8971586227416992, "num_tokens": 670332898.0, "step": 17652 }, { "epoch": 2.245643047958275, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.8562650680542, "learning_rate": 5e-05, "loss": 0.9743, "mean_token_accuracy": 0.8810244798660278, "num_tokens": 670371635.0, "step": 17653 }, { "epoch": 2.2457702582368655, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.676017761230469, "learning_rate": 5e-05, "loss": 0.7524, "mean_token_accuracy": 0.8816330432891846, "num_tokens": 670412568.0, "step": 17654 }, { "epoch": 2.245897468515456, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.25637149810791, "learning_rate": 5e-05, "loss": 0.925, "mean_token_accuracy": 0.8924837112426758, "num_tokens": 670454338.0, "step": 17655 }, { "epoch": 2.2460246787940465, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.14470100402832, "learning_rate": 5e-05, "loss": 0.9023, "mean_token_accuracy": 0.8883670568466187, "num_tokens": 670490826.0, "step": 17656 }, { "epoch": 2.246151889072637, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.227777481079102, "learning_rate": 5e-05, "loss": 0.7614, "mean_token_accuracy": 0.8739944696426392, "num_tokens": 670528806.0, "step": 17657 }, { "epoch": 2.2462790993512276, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.877809524536133, "learning_rate": 5e-05, "loss": 0.9872, "mean_token_accuracy": 0.8810047507286072, "num_tokens": 670563349.0, "step": 17658 }, { "epoch": 2.246406309629818, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 5.745224475860596, "learning_rate": 5e-05, "loss": 0.6728, "mean_token_accuracy": 0.877800703048706, "num_tokens": 670604682.0, "step": 17659 }, { "epoch": 2.2465335199084087, "ewc_loss": 0.9140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000843048095703125, "grad_norm": 12.028414726257324, "learning_rate": 5e-05, "loss": 1.2747, "mean_token_accuracy": 0.8728403449058533, "num_tokens": 670644807.0, "step": 17660 }, { "epoch": 2.246660730186999, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.008942604064941, "learning_rate": 5e-05, "loss": 0.9879, "mean_token_accuracy": 0.8858161568641663, "num_tokens": 670683788.0, "step": 17661 }, { "epoch": 2.2467879404655897, "ewc_loss": 0.88671875, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.00081634521484375, "grad_norm": 12.198305130004883, "learning_rate": 5e-05, "loss": 1.2148, "mean_token_accuracy": 0.8886457085609436, "num_tokens": 670728270.0, "step": 17662 }, { "epoch": 2.2469151507441802, "ewc_loss": 1.0, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000934600830078125, "grad_norm": 10.140995025634766, "learning_rate": 5e-05, "loss": 1.3255, "mean_token_accuracy": 0.8833786249160767, "num_tokens": 670763203.0, "step": 17663 }, { "epoch": 2.2470423610227708, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.966409683227539, "learning_rate": 5e-05, "loss": 1.0821, "mean_token_accuracy": 0.8750754594802856, "num_tokens": 670804548.0, "step": 17664 }, { "epoch": 2.2471695713013613, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 9.43128776550293, "learning_rate": 5e-05, "loss": 1.118, "mean_token_accuracy": 0.8777436017990112, "num_tokens": 670844922.0, "step": 17665 }, { "epoch": 2.247296781579952, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.134379386901855, "learning_rate": 5e-05, "loss": 1.0132, "mean_token_accuracy": 0.891582190990448, "num_tokens": 670882369.0, "step": 17666 }, { "epoch": 2.2474239918585424, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.855722427368164, "learning_rate": 5e-05, "loss": 0.9451, "mean_token_accuracy": 0.8940015435218811, "num_tokens": 670922716.0, "step": 17667 }, { "epoch": 2.247551202137133, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.421813011169434, "learning_rate": 5e-05, "loss": 0.9992, "mean_token_accuracy": 0.8762741088867188, "num_tokens": 670964618.0, "step": 17668 }, { "epoch": 2.2476784124157234, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.360212326049805, "learning_rate": 5e-05, "loss": 0.8554, "mean_token_accuracy": 0.8980135321617126, "num_tokens": 670997915.0, "step": 17669 }, { "epoch": 2.2478056226943135, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.067792892456055, "learning_rate": 5e-05, "loss": 0.9417, "mean_token_accuracy": 0.8748672008514404, "num_tokens": 671031675.0, "step": 17670 }, { "epoch": 2.247932832972904, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.037846565246582, "learning_rate": 5e-05, "loss": 0.8923, "mean_token_accuracy": 0.8738435506820679, "num_tokens": 671068060.0, "step": 17671 }, { "epoch": 2.2480600432514946, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.77191162109375, "learning_rate": 5e-05, "loss": 0.8821, "mean_token_accuracy": 0.8850561380386353, "num_tokens": 671109014.0, "step": 17672 }, { "epoch": 2.248187253530085, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.751343250274658, "learning_rate": 5e-05, "loss": 0.8008, "mean_token_accuracy": 0.8857293128967285, "num_tokens": 671140173.0, "step": 17673 }, { "epoch": 2.2483144638086756, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.679006576538086, "learning_rate": 5e-05, "loss": 0.8797, "mean_token_accuracy": 0.8787156343460083, "num_tokens": 671177644.0, "step": 17674 }, { "epoch": 2.248441674087266, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.339870929718018, "learning_rate": 5e-05, "loss": 0.7477, "mean_token_accuracy": 0.8841179609298706, "num_tokens": 671221860.0, "step": 17675 }, { "epoch": 2.2485688843658567, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.985145568847656, "learning_rate": 5e-05, "loss": 0.9011, "mean_token_accuracy": 0.8841158747673035, "num_tokens": 671262567.0, "step": 17676 }, { "epoch": 2.248696094644447, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.8871965408325195, "learning_rate": 5e-05, "loss": 0.7345, "mean_token_accuracy": 0.876975953578949, "num_tokens": 671301195.0, "step": 17677 }, { "epoch": 2.2488233049230377, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.72751235961914, "learning_rate": 5e-05, "loss": 0.9908, "mean_token_accuracy": 0.8805909156799316, "num_tokens": 671346028.0, "step": 17678 }, { "epoch": 2.2489505152016283, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.900359153747559, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8881083726882935, "num_tokens": 671386927.0, "step": 17679 }, { "epoch": 2.249077725480219, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 10.33701229095459, "learning_rate": 5e-05, "loss": 1.1015, "mean_token_accuracy": 0.878024697303772, "num_tokens": 671422447.0, "step": 17680 }, { "epoch": 2.2492049357588093, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.077326774597168, "learning_rate": 5e-05, "loss": 0.9402, "mean_token_accuracy": 0.8795797228813171, "num_tokens": 671461390.0, "step": 17681 }, { "epoch": 2.2493321460374, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.637107849121094, "learning_rate": 5e-05, "loss": 1.0026, "mean_token_accuracy": 0.8823213577270508, "num_tokens": 671499150.0, "step": 17682 }, { "epoch": 2.2494593563159904, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 8.901019096374512, "learning_rate": 5e-05, "loss": 1.0125, "mean_token_accuracy": 0.880535364151001, "num_tokens": 671537529.0, "step": 17683 }, { "epoch": 2.249586566594581, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.09953498840332, "learning_rate": 5e-05, "loss": 0.8633, "mean_token_accuracy": 0.876488208770752, "num_tokens": 671577661.0, "step": 17684 }, { "epoch": 2.2497137768731714, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.131512641906738, "learning_rate": 5e-05, "loss": 0.9442, "mean_token_accuracy": 0.8885878324508667, "num_tokens": 671611375.0, "step": 17685 }, { "epoch": 2.249840987151762, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.19309139251709, "learning_rate": 5e-05, "loss": 0.7966, "mean_token_accuracy": 0.8706507682800293, "num_tokens": 671655430.0, "step": 17686 }, { "epoch": 2.2499681974303525, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.195985794067383, "learning_rate": 5e-05, "loss": 0.9126, "mean_token_accuracy": 0.8864760994911194, "num_tokens": 671696451.0, "step": 17687 }, { "epoch": 2.250095407708943, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.259322643280029, "learning_rate": 5e-05, "loss": 0.7869, "mean_token_accuracy": 0.874942421913147, "num_tokens": 671736576.0, "step": 17688 }, { "epoch": 2.2502226179875335, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.78541088104248, "learning_rate": 5e-05, "loss": 0.8879, "mean_token_accuracy": 0.8852760791778564, "num_tokens": 671774638.0, "step": 17689 }, { "epoch": 2.250349828266124, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.542739391326904, "learning_rate": 5e-05, "loss": 0.7918, "mean_token_accuracy": 0.8788814544677734, "num_tokens": 671812626.0, "step": 17690 }, { "epoch": 2.2504770385447146, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.372237205505371, "learning_rate": 5e-05, "loss": 0.8334, "mean_token_accuracy": 0.8878494501113892, "num_tokens": 671844752.0, "step": 17691 }, { "epoch": 2.250604248823305, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.551364898681641, "learning_rate": 5e-05, "loss": 0.7879, "mean_token_accuracy": 0.8806231021881104, "num_tokens": 671883161.0, "step": 17692 }, { "epoch": 2.250731459101895, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.180612564086914, "learning_rate": 5e-05, "loss": 0.7969, "mean_token_accuracy": 0.8878394365310669, "num_tokens": 671924386.0, "step": 17693 }, { "epoch": 2.250858669380486, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.498661041259766, "learning_rate": 5e-05, "loss": 0.7607, "mean_token_accuracy": 0.8818028569221497, "num_tokens": 671958452.0, "step": 17694 }, { "epoch": 2.2509858796590763, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.020105361938477, "learning_rate": 5e-05, "loss": 0.7876, "mean_token_accuracy": 0.8802106380462646, "num_tokens": 671995335.0, "step": 17695 }, { "epoch": 2.251113089937667, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.480050086975098, "learning_rate": 5e-05, "loss": 0.7402, "mean_token_accuracy": 0.8771812319755554, "num_tokens": 672025640.0, "step": 17696 }, { "epoch": 2.2512403002162573, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.97094202041626, "learning_rate": 5e-05, "loss": 0.789, "mean_token_accuracy": 0.8755684494972229, "num_tokens": 672066013.0, "step": 17697 }, { "epoch": 2.251367510494848, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.329561710357666, "learning_rate": 5e-05, "loss": 0.7154, "mean_token_accuracy": 0.8824639320373535, "num_tokens": 672107243.0, "step": 17698 }, { "epoch": 2.2514947207734384, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.093757629394531, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8825069665908813, "num_tokens": 672145753.0, "step": 17699 }, { "epoch": 2.251621931052029, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.199679851531982, "learning_rate": 5e-05, "loss": 0.6957, "mean_token_accuracy": 0.8820577263832092, "num_tokens": 672185489.0, "step": 17700 }, { "epoch": 2.2517491413306194, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.136669158935547, "learning_rate": 5e-05, "loss": 0.7494, "mean_token_accuracy": 0.8862643837928772, "num_tokens": 672225702.0, "step": 17701 }, { "epoch": 2.25187635160921, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.147339344024658, "learning_rate": 5e-05, "loss": 0.7134, "mean_token_accuracy": 0.8762950897216797, "num_tokens": 672265901.0, "step": 17702 }, { "epoch": 2.2520035618878005, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.190457344055176, "learning_rate": 5e-05, "loss": 0.8014, "mean_token_accuracy": 0.8721251487731934, "num_tokens": 672304081.0, "step": 17703 }, { "epoch": 2.252130772166391, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.028072357177734, "learning_rate": 5e-05, "loss": 0.6837, "mean_token_accuracy": 0.8862340450286865, "num_tokens": 672343425.0, "step": 17704 }, { "epoch": 2.2522579824449815, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.358628273010254, "learning_rate": 5e-05, "loss": 0.7745, "mean_token_accuracy": 0.8877920508384705, "num_tokens": 672377745.0, "step": 17705 }, { "epoch": 2.252385192723572, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.864090442657471, "learning_rate": 5e-05, "loss": 0.6581, "mean_token_accuracy": 0.8902734518051147, "num_tokens": 672413811.0, "step": 17706 }, { "epoch": 2.2525124030021626, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.540273666381836, "learning_rate": 5e-05, "loss": 0.8225, "mean_token_accuracy": 0.879662036895752, "num_tokens": 672455083.0, "step": 17707 }, { "epoch": 2.252639613280753, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.766600608825684, "learning_rate": 5e-05, "loss": 0.6663, "mean_token_accuracy": 0.8879210948944092, "num_tokens": 672502401.0, "step": 17708 }, { "epoch": 2.2527668235593437, "ewc_loss": 0.5, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.691764831542969, "learning_rate": 5e-05, "loss": 0.8238, "mean_token_accuracy": 0.8841263651847839, "num_tokens": 672540504.0, "step": 17709 }, { "epoch": 2.252894033837934, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.76255989074707, "learning_rate": 5e-05, "loss": 0.6833, "mean_token_accuracy": 0.8839758038520813, "num_tokens": 672578157.0, "step": 17710 }, { "epoch": 2.2530212441165247, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.70954704284668, "learning_rate": 5e-05, "loss": 0.8529, "mean_token_accuracy": 0.8808310627937317, "num_tokens": 672611859.0, "step": 17711 }, { "epoch": 2.2531484543951152, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.973387718200684, "learning_rate": 5e-05, "loss": 0.7312, "mean_token_accuracy": 0.8764150142669678, "num_tokens": 672651779.0, "step": 17712 }, { "epoch": 2.2532756646737058, "ewc_loss": 0.5, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.372812271118164, "learning_rate": 5e-05, "loss": 0.8468, "mean_token_accuracy": 0.8787118792533875, "num_tokens": 672691559.0, "step": 17713 }, { "epoch": 2.2534028749522963, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.431978702545166, "learning_rate": 5e-05, "loss": 0.7538, "mean_token_accuracy": 0.8831999897956848, "num_tokens": 672731353.0, "step": 17714 }, { "epoch": 2.253530085230887, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.730625152587891, "learning_rate": 5e-05, "loss": 0.748, "mean_token_accuracy": 0.8874059319496155, "num_tokens": 672770427.0, "step": 17715 }, { "epoch": 2.2536572955094774, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.9654340744018555, "learning_rate": 5e-05, "loss": 0.7998, "mean_token_accuracy": 0.8823489546775818, "num_tokens": 672808817.0, "step": 17716 }, { "epoch": 2.253784505788068, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.07582426071167, "learning_rate": 5e-05, "loss": 0.6931, "mean_token_accuracy": 0.8822633624076843, "num_tokens": 672848368.0, "step": 17717 }, { "epoch": 2.253911716066658, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.48778247833252, "learning_rate": 5e-05, "loss": 0.7835, "mean_token_accuracy": 0.894224226474762, "num_tokens": 672885074.0, "step": 17718 }, { "epoch": 2.254038926345249, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.607149600982666, "learning_rate": 5e-05, "loss": 0.6679, "mean_token_accuracy": 0.882400631904602, "num_tokens": 672924866.0, "step": 17719 }, { "epoch": 2.254166136623839, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.989175796508789, "learning_rate": 5e-05, "loss": 0.844, "mean_token_accuracy": 0.8868631720542908, "num_tokens": 672962051.0, "step": 17720 }, { "epoch": 2.2542933469024296, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.341128826141357, "learning_rate": 5e-05, "loss": 0.6613, "mean_token_accuracy": 0.8858664035797119, "num_tokens": 672999954.0, "step": 17721 }, { "epoch": 2.25442055718102, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.424620628356934, "learning_rate": 5e-05, "loss": 0.9821, "mean_token_accuracy": 0.8677332997322083, "num_tokens": 673037825.0, "step": 17722 }, { "epoch": 2.2545477674596106, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.580245018005371, "learning_rate": 5e-05, "loss": 0.684, "mean_token_accuracy": 0.8940072059631348, "num_tokens": 673072036.0, "step": 17723 }, { "epoch": 2.254674977738201, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.29716968536377, "learning_rate": 5e-05, "loss": 0.975, "mean_token_accuracy": 0.8711661100387573, "num_tokens": 673110733.0, "step": 17724 }, { "epoch": 2.2548021880167917, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.4511637687683105, "learning_rate": 5e-05, "loss": 0.8269, "mean_token_accuracy": 0.8913120031356812, "num_tokens": 673146611.0, "step": 17725 }, { "epoch": 2.254929398295382, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.321452140808105, "learning_rate": 5e-05, "loss": 0.8858, "mean_token_accuracy": 0.8809134364128113, "num_tokens": 673189965.0, "step": 17726 }, { "epoch": 2.2550566085739727, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.23493766784668, "learning_rate": 5e-05, "loss": 0.9301, "mean_token_accuracy": 0.8732385635375977, "num_tokens": 673226990.0, "step": 17727 }, { "epoch": 2.2551838188525632, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.33428955078125, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.8793057203292847, "num_tokens": 673263109.0, "step": 17728 }, { "epoch": 2.2553110291311538, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.478540420532227, "learning_rate": 5e-05, "loss": 0.844, "mean_token_accuracy": 0.8892030715942383, "num_tokens": 673304081.0, "step": 17729 }, { "epoch": 2.2554382394097443, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.241419315338135, "learning_rate": 5e-05, "loss": 0.7309, "mean_token_accuracy": 0.889208972454071, "num_tokens": 673345510.0, "step": 17730 }, { "epoch": 2.255565449688335, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.971513748168945, "learning_rate": 5e-05, "loss": 0.8173, "mean_token_accuracy": 0.8762547969818115, "num_tokens": 673383607.0, "step": 17731 }, { "epoch": 2.2556926599669254, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.800136089324951, "learning_rate": 5e-05, "loss": 0.7891, "mean_token_accuracy": 0.8777180910110474, "num_tokens": 673422676.0, "step": 17732 }, { "epoch": 2.255819870245516, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.171862602233887, "learning_rate": 5e-05, "loss": 0.7048, "mean_token_accuracy": 0.8848221302032471, "num_tokens": 673466135.0, "step": 17733 }, { "epoch": 2.2559470805241064, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.449020385742188, "learning_rate": 5e-05, "loss": 0.8406, "mean_token_accuracy": 0.8756446242332458, "num_tokens": 673509003.0, "step": 17734 }, { "epoch": 2.256074290802697, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.6241536140441895, "learning_rate": 5e-05, "loss": 0.7004, "mean_token_accuracy": 0.8810742497444153, "num_tokens": 673546231.0, "step": 17735 }, { "epoch": 2.2562015010812875, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.971951484680176, "learning_rate": 5e-05, "loss": 0.8772, "mean_token_accuracy": 0.8771966099739075, "num_tokens": 673585812.0, "step": 17736 }, { "epoch": 2.256328711359878, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.418699264526367, "learning_rate": 5e-05, "loss": 0.6549, "mean_token_accuracy": 0.8916022777557373, "num_tokens": 673621566.0, "step": 17737 }, { "epoch": 2.2564559216384685, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.331459045410156, "learning_rate": 5e-05, "loss": 0.8817, "mean_token_accuracy": 0.9019219875335693, "num_tokens": 673665659.0, "step": 17738 }, { "epoch": 2.256583131917059, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.659450054168701, "learning_rate": 5e-05, "loss": 0.7069, "mean_token_accuracy": 0.8905163407325745, "num_tokens": 673705936.0, "step": 17739 }, { "epoch": 2.2567103421956496, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.249488830566406, "learning_rate": 5e-05, "loss": 0.962, "mean_token_accuracy": 0.8762743473052979, "num_tokens": 673740741.0, "step": 17740 }, { "epoch": 2.2568375524742397, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.3208394050598145, "learning_rate": 5e-05, "loss": 0.8314, "mean_token_accuracy": 0.8861669301986694, "num_tokens": 673778304.0, "step": 17741 }, { "epoch": 2.2569647627528306, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.625085830688477, "learning_rate": 5e-05, "loss": 0.8997, "mean_token_accuracy": 0.8848557472229004, "num_tokens": 673815136.0, "step": 17742 }, { "epoch": 2.2570919730314207, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.849048137664795, "learning_rate": 5e-05, "loss": 0.8705, "mean_token_accuracy": 0.8871190547943115, "num_tokens": 673847227.0, "step": 17743 }, { "epoch": 2.2572191833100113, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.963168144226074, "learning_rate": 5e-05, "loss": 0.8133, "mean_token_accuracy": 0.8920725584030151, "num_tokens": 673885110.0, "step": 17744 }, { "epoch": 2.257346393588602, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.8558783531188965, "learning_rate": 5e-05, "loss": 0.863, "mean_token_accuracy": 0.8731490969657898, "num_tokens": 673929534.0, "step": 17745 }, { "epoch": 2.2574736038671923, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.927257537841797, "learning_rate": 5e-05, "loss": 0.8431, "mean_token_accuracy": 0.8742277026176453, "num_tokens": 673972353.0, "step": 17746 }, { "epoch": 2.257600814145783, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.433535099029541, "learning_rate": 5e-05, "loss": 0.7313, "mean_token_accuracy": 0.8909999132156372, "num_tokens": 674009441.0, "step": 17747 }, { "epoch": 2.2577280244243734, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.221534729003906, "learning_rate": 5e-05, "loss": 0.7949, "mean_token_accuracy": 0.8933638334274292, "num_tokens": 674045397.0, "step": 17748 }, { "epoch": 2.257855234702964, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.117187023162842, "learning_rate": 5e-05, "loss": 0.7271, "mean_token_accuracy": 0.8781183958053589, "num_tokens": 674084896.0, "step": 17749 }, { "epoch": 2.2579824449815544, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.301335334777832, "learning_rate": 5e-05, "loss": 0.8569, "mean_token_accuracy": 0.8722786903381348, "num_tokens": 674121223.0, "step": 17750 }, { "epoch": 2.258109655260145, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.166857719421387, "learning_rate": 5e-05, "loss": 0.7351, "mean_token_accuracy": 0.8803082704544067, "num_tokens": 674157395.0, "step": 17751 }, { "epoch": 2.2582368655387355, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.096149444580078, "learning_rate": 5e-05, "loss": 0.8599, "mean_token_accuracy": 0.8652498722076416, "num_tokens": 674200157.0, "step": 17752 }, { "epoch": 2.258364075817326, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.37287712097168, "learning_rate": 5e-05, "loss": 0.772, "mean_token_accuracy": 0.8696054220199585, "num_tokens": 674234696.0, "step": 17753 }, { "epoch": 2.2584912860959165, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.858484268188477, "learning_rate": 5e-05, "loss": 0.7803, "mean_token_accuracy": 0.8810280561447144, "num_tokens": 674277456.0, "step": 17754 }, { "epoch": 2.258618496374507, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.452655792236328, "learning_rate": 5e-05, "loss": 0.7053, "mean_token_accuracy": 0.8933096528053284, "num_tokens": 674316913.0, "step": 17755 }, { "epoch": 2.2587457066530976, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.785399913787842, "learning_rate": 5e-05, "loss": 0.756, "mean_token_accuracy": 0.8841003179550171, "num_tokens": 674355539.0, "step": 17756 }, { "epoch": 2.258872916931688, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.434080123901367, "learning_rate": 5e-05, "loss": 0.7411, "mean_token_accuracy": 0.8753793239593506, "num_tokens": 674386759.0, "step": 17757 }, { "epoch": 2.2590001272102787, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.796986103057861, "learning_rate": 5e-05, "loss": 0.7246, "mean_token_accuracy": 0.8883694410324097, "num_tokens": 674427098.0, "step": 17758 }, { "epoch": 2.259127337488869, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.333989143371582, "learning_rate": 5e-05, "loss": 0.696, "mean_token_accuracy": 0.8883670568466187, "num_tokens": 674464766.0, "step": 17759 }, { "epoch": 2.2592545477674597, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.880123138427734, "learning_rate": 5e-05, "loss": 0.7606, "mean_token_accuracy": 0.8787072896957397, "num_tokens": 674506260.0, "step": 17760 }, { "epoch": 2.2593817580460502, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.2244181632995605, "learning_rate": 5e-05, "loss": 0.701, "mean_token_accuracy": 0.8850053548812866, "num_tokens": 674544255.0, "step": 17761 }, { "epoch": 2.2595089683246408, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.9835100173950195, "learning_rate": 5e-05, "loss": 0.7128, "mean_token_accuracy": 0.8917102813720703, "num_tokens": 674579806.0, "step": 17762 }, { "epoch": 2.2596361786032313, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.125889301300049, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8680946826934814, "num_tokens": 674620734.0, "step": 17763 }, { "epoch": 2.259763388881822, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.087040901184082, "learning_rate": 5e-05, "loss": 0.7562, "mean_token_accuracy": 0.8870624303817749, "num_tokens": 674661213.0, "step": 17764 }, { "epoch": 2.2598905991604123, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.997001647949219, "learning_rate": 5e-05, "loss": 0.6604, "mean_token_accuracy": 0.8891263604164124, "num_tokens": 674696323.0, "step": 17765 }, { "epoch": 2.2600178094390024, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.258970260620117, "learning_rate": 5e-05, "loss": 0.7743, "mean_token_accuracy": 0.8799746036529541, "num_tokens": 674728041.0, "step": 17766 }, { "epoch": 2.2601450197175934, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.822532653808594, "learning_rate": 5e-05, "loss": 0.6656, "mean_token_accuracy": 0.8880763053894043, "num_tokens": 674767518.0, "step": 17767 }, { "epoch": 2.2602722299961835, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.499287605285645, "learning_rate": 5e-05, "loss": 0.7901, "mean_token_accuracy": 0.8852897882461548, "num_tokens": 674807510.0, "step": 17768 }, { "epoch": 2.260399440274774, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.6687750816345215, "learning_rate": 5e-05, "loss": 0.6693, "mean_token_accuracy": 0.8823352456092834, "num_tokens": 674844851.0, "step": 17769 }, { "epoch": 2.2605266505533645, "ewc_loss": 0.5, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.685248374938965, "learning_rate": 5e-05, "loss": 0.7996, "mean_token_accuracy": 0.893704891204834, "num_tokens": 674881839.0, "step": 17770 }, { "epoch": 2.260653860831955, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.689379692077637, "learning_rate": 5e-05, "loss": 0.7046, "mean_token_accuracy": 0.8784278631210327, "num_tokens": 674920450.0, "step": 17771 }, { "epoch": 2.2607810711105456, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.653624534606934, "learning_rate": 5e-05, "loss": 0.8297, "mean_token_accuracy": 0.884836733341217, "num_tokens": 674960031.0, "step": 17772 }, { "epoch": 2.260908281389136, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.002471446990967, "learning_rate": 5e-05, "loss": 0.6935, "mean_token_accuracy": 0.8906358480453491, "num_tokens": 674993354.0, "step": 17773 }, { "epoch": 2.2610354916677267, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.149238586425781, "learning_rate": 5e-05, "loss": 0.8006, "mean_token_accuracy": 0.8858370184898376, "num_tokens": 675029186.0, "step": 17774 }, { "epoch": 2.261162701946317, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.617044925689697, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.890876054763794, "num_tokens": 675065170.0, "step": 17775 }, { "epoch": 2.2612899122249077, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.28776216506958, "learning_rate": 5e-05, "loss": 0.6938, "mean_token_accuracy": 0.893867552280426, "num_tokens": 675102774.0, "step": 17776 }, { "epoch": 2.2614171225034982, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.397054672241211, "learning_rate": 5e-05, "loss": 0.8308, "mean_token_accuracy": 0.8769263029098511, "num_tokens": 675136890.0, "step": 17777 }, { "epoch": 2.2615443327820888, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.431159973144531, "learning_rate": 5e-05, "loss": 0.6434, "mean_token_accuracy": 0.8897645473480225, "num_tokens": 675171959.0, "step": 17778 }, { "epoch": 2.2616715430606793, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.321320533752441, "learning_rate": 5e-05, "loss": 0.866, "mean_token_accuracy": 0.886349618434906, "num_tokens": 675203931.0, "step": 17779 }, { "epoch": 2.26179875333927, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.820285797119141, "learning_rate": 5e-05, "loss": 0.6081, "mean_token_accuracy": 0.8938177824020386, "num_tokens": 675243645.0, "step": 17780 }, { "epoch": 2.2619259636178604, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.32371997833252, "learning_rate": 5e-05, "loss": 0.9861, "mean_token_accuracy": 0.8847365975379944, "num_tokens": 675280318.0, "step": 17781 }, { "epoch": 2.262053173896451, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.375552177429199, "learning_rate": 5e-05, "loss": 0.7756, "mean_token_accuracy": 0.8785023093223572, "num_tokens": 675321442.0, "step": 17782 }, { "epoch": 2.2621803841750414, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 10.546032905578613, "learning_rate": 5e-05, "loss": 1.0616, "mean_token_accuracy": 0.8973126411437988, "num_tokens": 675351039.0, "step": 17783 }, { "epoch": 2.262307594453632, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 8.348421096801758, "learning_rate": 5e-05, "loss": 1.0399, "mean_token_accuracy": 0.8848837614059448, "num_tokens": 675385756.0, "step": 17784 }, { "epoch": 2.2624348047322225, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.807640075683594, "learning_rate": 5e-05, "loss": 0.875, "mean_token_accuracy": 0.8981534838676453, "num_tokens": 675429282.0, "step": 17785 }, { "epoch": 2.262562015010813, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 9.578042984008789, "learning_rate": 5e-05, "loss": 1.0841, "mean_token_accuracy": 0.8925552368164062, "num_tokens": 675466661.0, "step": 17786 }, { "epoch": 2.2626892252894035, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.865906715393066, "learning_rate": 5e-05, "loss": 0.7322, "mean_token_accuracy": 0.8897322416305542, "num_tokens": 675501885.0, "step": 17787 }, { "epoch": 2.262816435567994, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0007781982421875, "grad_norm": 11.159552574157715, "learning_rate": 5e-05, "loss": 1.2172, "mean_token_accuracy": 0.8721970319747925, "num_tokens": 675544423.0, "step": 17788 }, { "epoch": 2.2629436458465846, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.697473049163818, "learning_rate": 5e-05, "loss": 0.9706, "mean_token_accuracy": 0.8792872428894043, "num_tokens": 675585301.0, "step": 17789 }, { "epoch": 2.263070856125175, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 11.346932411193848, "learning_rate": 5e-05, "loss": 1.135, "mean_token_accuracy": 0.8814560174942017, "num_tokens": 675625244.0, "step": 17790 }, { "epoch": 2.263198066403765, "ewc_loss": 0.859375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 9.102095603942871, "learning_rate": 5e-05, "loss": 1.1559, "mean_token_accuracy": 0.8953405618667603, "num_tokens": 675657310.0, "step": 17791 }, { "epoch": 2.263325276682356, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.352148056030273, "learning_rate": 5e-05, "loss": 1.013, "mean_token_accuracy": 0.8836915493011475, "num_tokens": 675690702.0, "step": 17792 }, { "epoch": 2.2634524869609463, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 8.793770790100098, "learning_rate": 5e-05, "loss": 1.0065, "mean_token_accuracy": 0.8889535665512085, "num_tokens": 675725211.0, "step": 17793 }, { "epoch": 2.263579697239537, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.635899543762207, "learning_rate": 5e-05, "loss": 0.9688, "mean_token_accuracy": 0.8940532803535461, "num_tokens": 675762418.0, "step": 17794 }, { "epoch": 2.2637069075181273, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.177170753479004, "learning_rate": 5e-05, "loss": 0.9582, "mean_token_accuracy": 0.8701703548431396, "num_tokens": 675807640.0, "step": 17795 }, { "epoch": 2.263834117796718, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.38431167602539, "learning_rate": 5e-05, "loss": 0.9698, "mean_token_accuracy": 0.8839391469955444, "num_tokens": 675843032.0, "step": 17796 }, { "epoch": 2.2639613280753084, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.596200466156006, "learning_rate": 5e-05, "loss": 0.8306, "mean_token_accuracy": 0.8835653066635132, "num_tokens": 675875150.0, "step": 17797 }, { "epoch": 2.264088538353899, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.624971389770508, "learning_rate": 5e-05, "loss": 1.0318, "mean_token_accuracy": 0.8697958588600159, "num_tokens": 675917635.0, "step": 17798 }, { "epoch": 2.2642157486324894, "ewc_loss": 0.5, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.409336566925049, "learning_rate": 5e-05, "loss": 0.8432, "mean_token_accuracy": 0.8756903409957886, "num_tokens": 675955577.0, "step": 17799 }, { "epoch": 2.26434295891108, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.927386283874512, "learning_rate": 5e-05, "loss": 1.0451, "mean_token_accuracy": 0.8687340021133423, "num_tokens": 675996326.0, "step": 17800 }, { "epoch": 2.2644701691896705, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 7.71195125579834, "learning_rate": 5e-05, "loss": 0.9464, "mean_token_accuracy": 0.8671950101852417, "num_tokens": 676040857.0, "step": 17801 }, { "epoch": 2.264597379468261, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.848512649536133, "learning_rate": 5e-05, "loss": 0.9638, "mean_token_accuracy": 0.8944764733314514, "num_tokens": 676077601.0, "step": 17802 }, { "epoch": 2.2647245897468515, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.130125999450684, "learning_rate": 5e-05, "loss": 0.9495, "mean_token_accuracy": 0.8794285655021667, "num_tokens": 676115558.0, "step": 17803 }, { "epoch": 2.264851800025442, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.347081184387207, "learning_rate": 5e-05, "loss": 0.9885, "mean_token_accuracy": 0.8787667751312256, "num_tokens": 676157828.0, "step": 17804 }, { "epoch": 2.2649790103040326, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.267148971557617, "learning_rate": 5e-05, "loss": 0.9607, "mean_token_accuracy": 0.8717004656791687, "num_tokens": 676199147.0, "step": 17805 }, { "epoch": 2.265106220582623, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.71407413482666, "learning_rate": 5e-05, "loss": 0.8912, "mean_token_accuracy": 0.8920491933822632, "num_tokens": 676241271.0, "step": 17806 }, { "epoch": 2.2652334308612136, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.068896293640137, "learning_rate": 5e-05, "loss": 0.8609, "mean_token_accuracy": 0.883170485496521, "num_tokens": 676276990.0, "step": 17807 }, { "epoch": 2.265360641139804, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.494588851928711, "learning_rate": 5e-05, "loss": 0.8973, "mean_token_accuracy": 0.8827639222145081, "num_tokens": 676313441.0, "step": 17808 }, { "epoch": 2.2654878514183947, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.568393707275391, "learning_rate": 5e-05, "loss": 0.8045, "mean_token_accuracy": 0.8809682726860046, "num_tokens": 676350322.0, "step": 17809 }, { "epoch": 2.2656150616969852, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.838889122009277, "learning_rate": 5e-05, "loss": 0.9096, "mean_token_accuracy": 0.8864939212799072, "num_tokens": 676386113.0, "step": 17810 }, { "epoch": 2.2657422719755758, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.168941974639893, "learning_rate": 5e-05, "loss": 0.7158, "mean_token_accuracy": 0.894904613494873, "num_tokens": 676421995.0, "step": 17811 }, { "epoch": 2.2658694822541663, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.240696907043457, "learning_rate": 5e-05, "loss": 0.9569, "mean_token_accuracy": 0.882952094078064, "num_tokens": 676459761.0, "step": 17812 }, { "epoch": 2.265996692532757, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.3049702644348145, "learning_rate": 5e-05, "loss": 0.8091, "mean_token_accuracy": 0.879337728023529, "num_tokens": 676503870.0, "step": 17813 }, { "epoch": 2.2661239028113473, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.20192813873291, "learning_rate": 5e-05, "loss": 0.9374, "mean_token_accuracy": 0.8934446573257446, "num_tokens": 676544173.0, "step": 17814 }, { "epoch": 2.266251113089938, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.8074212074279785, "learning_rate": 5e-05, "loss": 0.8627, "mean_token_accuracy": 0.8827073574066162, "num_tokens": 676580896.0, "step": 17815 }, { "epoch": 2.266378323368528, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.587508201599121, "learning_rate": 5e-05, "loss": 0.8909, "mean_token_accuracy": 0.8880960941314697, "num_tokens": 676620127.0, "step": 17816 }, { "epoch": 2.266505533647119, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.112764358520508, "learning_rate": 5e-05, "loss": 0.8373, "mean_token_accuracy": 0.8940672874450684, "num_tokens": 676655988.0, "step": 17817 }, { "epoch": 2.266632743925709, "ewc_loss": 0.5, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.884126663208008, "learning_rate": 5e-05, "loss": 0.8153, "mean_token_accuracy": 0.8840179443359375, "num_tokens": 676689125.0, "step": 17818 }, { "epoch": 2.2667599542042995, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.069332122802734, "learning_rate": 5e-05, "loss": 0.8328, "mean_token_accuracy": 0.8808445930480957, "num_tokens": 676729231.0, "step": 17819 }, { "epoch": 2.26688716448289, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.694466590881348, "learning_rate": 5e-05, "loss": 0.7955, "mean_token_accuracy": 0.8784058094024658, "num_tokens": 676770130.0, "step": 17820 }, { "epoch": 2.2670143747614806, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.7044854164123535, "learning_rate": 5e-05, "loss": 0.7475, "mean_token_accuracy": 0.887818455696106, "num_tokens": 676807847.0, "step": 17821 }, { "epoch": 2.267141585040071, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.004958152770996, "learning_rate": 5e-05, "loss": 0.7692, "mean_token_accuracy": 0.8900889158248901, "num_tokens": 676841050.0, "step": 17822 }, { "epoch": 2.2672687953186617, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.234825134277344, "learning_rate": 5e-05, "loss": 0.7144, "mean_token_accuracy": 0.8799121975898743, "num_tokens": 676877114.0, "step": 17823 }, { "epoch": 2.267396005597252, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.359841346740723, "learning_rate": 5e-05, "loss": 0.7561, "mean_token_accuracy": 0.9003852605819702, "num_tokens": 676917913.0, "step": 17824 }, { "epoch": 2.2675232158758427, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.972701549530029, "learning_rate": 5e-05, "loss": 0.653, "mean_token_accuracy": 0.8946983218193054, "num_tokens": 676955974.0, "step": 17825 }, { "epoch": 2.2676504261544332, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.554303169250488, "learning_rate": 5e-05, "loss": 0.8099, "mean_token_accuracy": 0.895085334777832, "num_tokens": 676991285.0, "step": 17826 }, { "epoch": 2.2677776364330238, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.947992324829102, "learning_rate": 5e-05, "loss": 0.7256, "mean_token_accuracy": 0.8771172165870667, "num_tokens": 677025472.0, "step": 17827 }, { "epoch": 2.2679048467116143, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.605816841125488, "learning_rate": 5e-05, "loss": 0.8372, "mean_token_accuracy": 0.889335572719574, "num_tokens": 677065047.0, "step": 17828 }, { "epoch": 2.268032056990205, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.064833164215088, "learning_rate": 5e-05, "loss": 0.7346, "mean_token_accuracy": 0.8772569894790649, "num_tokens": 677103027.0, "step": 17829 }, { "epoch": 2.2681592672687954, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.465641975402832, "learning_rate": 5e-05, "loss": 0.8104, "mean_token_accuracy": 0.8952781558036804, "num_tokens": 677142302.0, "step": 17830 }, { "epoch": 2.268286477547386, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.338244915008545, "learning_rate": 5e-05, "loss": 0.757, "mean_token_accuracy": 0.8774337768554688, "num_tokens": 677186963.0, "step": 17831 }, { "epoch": 2.2684136878259764, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.073505401611328, "learning_rate": 5e-05, "loss": 0.8657, "mean_token_accuracy": 0.8629334568977356, "num_tokens": 677232227.0, "step": 17832 }, { "epoch": 2.268540898104567, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.646332263946533, "learning_rate": 5e-05, "loss": 0.7088, "mean_token_accuracy": 0.8979188203811646, "num_tokens": 677274213.0, "step": 17833 }, { "epoch": 2.2686681083831575, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.637938022613525, "learning_rate": 5e-05, "loss": 0.7322, "mean_token_accuracy": 0.8889245390892029, "num_tokens": 677313870.0, "step": 17834 }, { "epoch": 2.268795318661748, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.865938663482666, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.8864796161651611, "num_tokens": 677350478.0, "step": 17835 }, { "epoch": 2.2689225289403385, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.347909450531006, "learning_rate": 5e-05, "loss": 0.7225, "mean_token_accuracy": 0.8789932131767273, "num_tokens": 677392253.0, "step": 17836 }, { "epoch": 2.269049739218929, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.986732482910156, "learning_rate": 5e-05, "loss": 0.7392, "mean_token_accuracy": 0.8914375305175781, "num_tokens": 677429024.0, "step": 17837 }, { "epoch": 2.2691769494975196, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.189273834228516, "learning_rate": 5e-05, "loss": 0.6593, "mean_token_accuracy": 0.8972752094268799, "num_tokens": 677461705.0, "step": 17838 }, { "epoch": 2.2693041597761097, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.07637882232666, "learning_rate": 5e-05, "loss": 0.761, "mean_token_accuracy": 0.8848043084144592, "num_tokens": 677505586.0, "step": 17839 }, { "epoch": 2.2694313700547006, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.093474864959717, "learning_rate": 5e-05, "loss": 0.6852, "mean_token_accuracy": 0.8820339441299438, "num_tokens": 677544892.0, "step": 17840 }, { "epoch": 2.2695585803332907, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.16519832611084, "learning_rate": 5e-05, "loss": 0.7664, "mean_token_accuracy": 0.8814011812210083, "num_tokens": 677586721.0, "step": 17841 }, { "epoch": 2.2696857906118812, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.988622188568115, "learning_rate": 5e-05, "loss": 0.718, "mean_token_accuracy": 0.8739479780197144, "num_tokens": 677627657.0, "step": 17842 }, { "epoch": 2.2698130008904718, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.274805068969727, "learning_rate": 5e-05, "loss": 0.7256, "mean_token_accuracy": 0.8998054265975952, "num_tokens": 677667513.0, "step": 17843 }, { "epoch": 2.2699402111690623, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.891602039337158, "learning_rate": 5e-05, "loss": 0.6726, "mean_token_accuracy": 0.8848841190338135, "num_tokens": 677702120.0, "step": 17844 }, { "epoch": 2.270067421447653, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.434578895568848, "learning_rate": 5e-05, "loss": 0.8357, "mean_token_accuracy": 0.8709155917167664, "num_tokens": 677739353.0, "step": 17845 }, { "epoch": 2.2701946317262434, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.794636249542236, "learning_rate": 5e-05, "loss": 0.7254, "mean_token_accuracy": 0.8686025738716125, "num_tokens": 677778464.0, "step": 17846 }, { "epoch": 2.270321842004834, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.539752006530762, "learning_rate": 5e-05, "loss": 0.8148, "mean_token_accuracy": 0.8810067176818848, "num_tokens": 677815169.0, "step": 17847 }, { "epoch": 2.2704490522834244, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.800853252410889, "learning_rate": 5e-05, "loss": 0.6468, "mean_token_accuracy": 0.8939974308013916, "num_tokens": 677852824.0, "step": 17848 }, { "epoch": 2.270576262562015, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.541736602783203, "learning_rate": 5e-05, "loss": 0.8418, "mean_token_accuracy": 0.8768688440322876, "num_tokens": 677895274.0, "step": 17849 }, { "epoch": 2.2707034728406055, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.975814342498779, "learning_rate": 5e-05, "loss": 0.6718, "mean_token_accuracy": 0.8921120166778564, "num_tokens": 677929956.0, "step": 17850 }, { "epoch": 2.270830683119196, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.286544799804688, "learning_rate": 5e-05, "loss": 0.8248, "mean_token_accuracy": 0.8776040077209473, "num_tokens": 677965921.0, "step": 17851 }, { "epoch": 2.2709578933977865, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.285693645477295, "learning_rate": 5e-05, "loss": 0.694, "mean_token_accuracy": 0.8929693698883057, "num_tokens": 678005701.0, "step": 17852 }, { "epoch": 2.271085103676377, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.893032550811768, "learning_rate": 5e-05, "loss": 0.7642, "mean_token_accuracy": 0.8839077353477478, "num_tokens": 678040275.0, "step": 17853 }, { "epoch": 2.2712123139549676, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.575863838195801, "learning_rate": 5e-05, "loss": 0.7219, "mean_token_accuracy": 0.885844349861145, "num_tokens": 678071963.0, "step": 17854 }, { "epoch": 2.271339524233558, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.549918174743652, "learning_rate": 5e-05, "loss": 0.7157, "mean_token_accuracy": 0.8804055452346802, "num_tokens": 678103422.0, "step": 17855 }, { "epoch": 2.2714667345121486, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.744287967681885, "learning_rate": 5e-05, "loss": 0.7433, "mean_token_accuracy": 0.8828524351119995, "num_tokens": 678144288.0, "step": 17856 }, { "epoch": 2.271593944790739, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.413912296295166, "learning_rate": 5e-05, "loss": 0.6688, "mean_token_accuracy": 0.8930568099021912, "num_tokens": 678175575.0, "step": 17857 }, { "epoch": 2.2717211550693297, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.714815139770508, "learning_rate": 5e-05, "loss": 0.7597, "mean_token_accuracy": 0.8713836669921875, "num_tokens": 678218089.0, "step": 17858 }, { "epoch": 2.2718483653479202, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.491220951080322, "learning_rate": 5e-05, "loss": 0.7324, "mean_token_accuracy": 0.8741436004638672, "num_tokens": 678258777.0, "step": 17859 }, { "epoch": 2.2719755756265108, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.524269104003906, "learning_rate": 5e-05, "loss": 0.728, "mean_token_accuracy": 0.8786979913711548, "num_tokens": 678300354.0, "step": 17860 }, { "epoch": 2.2721027859051013, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.71312952041626, "learning_rate": 5e-05, "loss": 0.7241, "mean_token_accuracy": 0.8800948858261108, "num_tokens": 678343579.0, "step": 17861 }, { "epoch": 2.272229996183692, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.21762228012085, "learning_rate": 5e-05, "loss": 0.6941, "mean_token_accuracy": 0.8843044638633728, "num_tokens": 678377059.0, "step": 17862 }, { "epoch": 2.2723572064622823, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.055591583251953, "learning_rate": 5e-05, "loss": 0.7665, "mean_token_accuracy": 0.8766975998878479, "num_tokens": 678417608.0, "step": 17863 }, { "epoch": 2.2724844167408724, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.822051048278809, "learning_rate": 5e-05, "loss": 0.6694, "mean_token_accuracy": 0.8836498260498047, "num_tokens": 678455473.0, "step": 17864 }, { "epoch": 2.2726116270194634, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.576611518859863, "learning_rate": 5e-05, "loss": 0.8306, "mean_token_accuracy": 0.8727091550827026, "num_tokens": 678500258.0, "step": 17865 }, { "epoch": 2.2727388372980535, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.297724723815918, "learning_rate": 5e-05, "loss": 0.6537, "mean_token_accuracy": 0.8786141872406006, "num_tokens": 678542219.0, "step": 17866 }, { "epoch": 2.272866047576644, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 9.377779960632324, "learning_rate": 5e-05, "loss": 0.872, "mean_token_accuracy": 0.8818253874778748, "num_tokens": 678579364.0, "step": 17867 }, { "epoch": 2.2729932578552345, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00023937225341796875, "grad_norm": 5.805888652801514, "learning_rate": 5e-05, "loss": 0.6792, "mean_token_accuracy": 0.8709745407104492, "num_tokens": 678616549.0, "step": 17868 }, { "epoch": 2.273120468133825, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.284035682678223, "learning_rate": 5e-05, "loss": 0.9546, "mean_token_accuracy": 0.895350456237793, "num_tokens": 678652324.0, "step": 17869 }, { "epoch": 2.2732476784124156, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 6.431163311004639, "learning_rate": 5e-05, "loss": 0.7614, "mean_token_accuracy": 0.8807149529457092, "num_tokens": 678693574.0, "step": 17870 }, { "epoch": 2.273374888691006, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.237605094909668, "learning_rate": 5e-05, "loss": 1.1163, "mean_token_accuracy": 0.8683563470840454, "num_tokens": 678730097.0, "step": 17871 }, { "epoch": 2.2735020989695967, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 8.292128562927246, "learning_rate": 5e-05, "loss": 1.018, "mean_token_accuracy": 0.8809744119644165, "num_tokens": 678768392.0, "step": 17872 }, { "epoch": 2.273629309248187, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.214088439941406, "learning_rate": 5e-05, "loss": 0.8839, "mean_token_accuracy": 0.8761940002441406, "num_tokens": 678804290.0, "step": 17873 }, { "epoch": 2.2737565195267777, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 9.90621280670166, "learning_rate": 5e-05, "loss": 1.024, "mean_token_accuracy": 0.9015002846717834, "num_tokens": 678839535.0, "step": 17874 }, { "epoch": 2.2738837298053682, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.2843427658081055, "learning_rate": 5e-05, "loss": 0.6936, "mean_token_accuracy": 0.8836504220962524, "num_tokens": 678879179.0, "step": 17875 }, { "epoch": 2.2740109400839588, "ewc_loss": 0.98828125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000919342041015625, "grad_norm": 12.379798889160156, "learning_rate": 5e-05, "loss": 1.2924, "mean_token_accuracy": 0.890056848526001, "num_tokens": 678916804.0, "step": 17876 }, { "epoch": 2.2741381503625493, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 8.805733680725098, "learning_rate": 5e-05, "loss": 1.1138, "mean_token_accuracy": 0.8815288543701172, "num_tokens": 678950386.0, "step": 17877 }, { "epoch": 2.27426536064114, "ewc_loss": 0.8515625, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 12.282164573669434, "learning_rate": 5e-05, "loss": 1.1747, "mean_token_accuracy": 0.8833398818969727, "num_tokens": 678988367.0, "step": 17878 }, { "epoch": 2.2743925709197303, "ewc_loss": 0.99609375, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000926971435546875, "grad_norm": 10.137150764465332, "learning_rate": 5e-05, "loss": 1.2952, "mean_token_accuracy": 0.8920297026634216, "num_tokens": 679022577.0, "step": 17879 }, { "epoch": 2.274519781198321, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.000743865966796875, "grad_norm": 11.675161361694336, "learning_rate": 5e-05, "loss": 1.144, "mean_token_accuracy": 0.8830872774124146, "num_tokens": 679059613.0, "step": 17880 }, { "epoch": 2.2746469914769114, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000698089599609375, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.186938285827637, "learning_rate": 5e-05, "loss": 0.9956, "mean_token_accuracy": 0.8929444551467896, "num_tokens": 679092956.0, "step": 17881 }, { "epoch": 2.274774201755502, "ewc_loss": 0.875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 11.692662239074707, "learning_rate": 5e-05, "loss": 1.2104, "mean_token_accuracy": 0.8834767937660217, "num_tokens": 679131124.0, "step": 17882 }, { "epoch": 2.2749014120340925, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.030600547790527, "learning_rate": 5e-05, "loss": 0.9314, "mean_token_accuracy": 0.8963794708251953, "num_tokens": 679166077.0, "step": 17883 }, { "epoch": 2.275028622312683, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00077056884765625, "grad_norm": 11.684981346130371, "learning_rate": 5e-05, "loss": 1.183, "mean_token_accuracy": 0.880695641040802, "num_tokens": 679204427.0, "step": 17884 }, { "epoch": 2.2751558325912735, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 9.670003890991211, "learning_rate": 5e-05, "loss": 1.0634, "mean_token_accuracy": 0.8854634165763855, "num_tokens": 679236003.0, "step": 17885 }, { "epoch": 2.275283042869864, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 11.097250938415527, "learning_rate": 5e-05, "loss": 1.0825, "mean_token_accuracy": 0.8851711750030518, "num_tokens": 679276347.0, "step": 17886 }, { "epoch": 2.2754102531484546, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.596019744873047, "learning_rate": 5e-05, "loss": 1.0275, "mean_token_accuracy": 0.8859355449676514, "num_tokens": 679318335.0, "step": 17887 }, { "epoch": 2.275537463427045, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.55657958984375, "learning_rate": 5e-05, "loss": 1.06, "mean_token_accuracy": 0.8705146312713623, "num_tokens": 679361253.0, "step": 17888 }, { "epoch": 2.275664673705635, "ewc_loss": 0.625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.11174488067627, "learning_rate": 5e-05, "loss": 0.969, "mean_token_accuracy": 0.8790309429168701, "num_tokens": 679400234.0, "step": 17889 }, { "epoch": 2.275791883984226, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.257234573364258, "learning_rate": 5e-05, "loss": 1.0221, "mean_token_accuracy": 0.8813425898551941, "num_tokens": 679438538.0, "step": 17890 }, { "epoch": 2.2759190942628162, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.632157325744629, "learning_rate": 5e-05, "loss": 0.899, "mean_token_accuracy": 0.880656361579895, "num_tokens": 679475425.0, "step": 17891 }, { "epoch": 2.2760463045414068, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.274907112121582, "learning_rate": 5e-05, "loss": 0.9957, "mean_token_accuracy": 0.8873630166053772, "num_tokens": 679517447.0, "step": 17892 }, { "epoch": 2.2761735148199973, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.411874771118164, "learning_rate": 5e-05, "loss": 0.8465, "mean_token_accuracy": 0.8860782384872437, "num_tokens": 679549933.0, "step": 17893 }, { "epoch": 2.276300725098588, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.436352729797363, "learning_rate": 5e-05, "loss": 1.0234, "mean_token_accuracy": 0.8879045844078064, "num_tokens": 679590498.0, "step": 17894 }, { "epoch": 2.2764279353771784, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.541560173034668, "learning_rate": 5e-05, "loss": 0.8995, "mean_token_accuracy": 0.8787558674812317, "num_tokens": 679634582.0, "step": 17895 }, { "epoch": 2.276555145655769, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.443205833435059, "learning_rate": 5e-05, "loss": 1.0278, "mean_token_accuracy": 0.8879502415657043, "num_tokens": 679672324.0, "step": 17896 }, { "epoch": 2.2766823559343594, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.856162071228027, "learning_rate": 5e-05, "loss": 0.9095, "mean_token_accuracy": 0.8853521347045898, "num_tokens": 679706426.0, "step": 17897 }, { "epoch": 2.27680956621295, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 10.006507873535156, "learning_rate": 5e-05, "loss": 0.9521, "mean_token_accuracy": 0.8976826667785645, "num_tokens": 679744706.0, "step": 17898 }, { "epoch": 2.2769367764915405, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.867148399353027, "learning_rate": 5e-05, "loss": 0.9077, "mean_token_accuracy": 0.8839941024780273, "num_tokens": 679779939.0, "step": 17899 }, { "epoch": 2.277063986770131, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.423481941223145, "learning_rate": 5e-05, "loss": 0.9311, "mean_token_accuracy": 0.8845973014831543, "num_tokens": 679820286.0, "step": 17900 }, { "epoch": 2.2771911970487215, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.496089935302734, "learning_rate": 5e-05, "loss": 0.8545, "mean_token_accuracy": 0.8781672716140747, "num_tokens": 679857985.0, "step": 17901 }, { "epoch": 2.277318407327312, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.160667419433594, "learning_rate": 5e-05, "loss": 0.8822, "mean_token_accuracy": 0.8917962312698364, "num_tokens": 679896379.0, "step": 17902 }, { "epoch": 2.2774456176059026, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.980082988739014, "learning_rate": 5e-05, "loss": 0.8106, "mean_token_accuracy": 0.8750253915786743, "num_tokens": 679933162.0, "step": 17903 }, { "epoch": 2.277572827884493, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.317012786865234, "learning_rate": 5e-05, "loss": 0.9669, "mean_token_accuracy": 0.8674334287643433, "num_tokens": 679975828.0, "step": 17904 }, { "epoch": 2.2777000381630836, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.705783367156982, "learning_rate": 5e-05, "loss": 0.8222, "mean_token_accuracy": 0.8664899468421936, "num_tokens": 680014678.0, "step": 17905 }, { "epoch": 2.277827248441674, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.581552505493164, "learning_rate": 5e-05, "loss": 0.9415, "mean_token_accuracy": 0.884369969367981, "num_tokens": 680050912.0, "step": 17906 }, { "epoch": 2.2779544587202647, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.889671802520752, "learning_rate": 5e-05, "loss": 0.775, "mean_token_accuracy": 0.8891441822052002, "num_tokens": 680082591.0, "step": 17907 }, { "epoch": 2.2780816689988552, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.47848129272461, "learning_rate": 5e-05, "loss": 0.9363, "mean_token_accuracy": 0.8817404508590698, "num_tokens": 680122100.0, "step": 17908 }, { "epoch": 2.2782088792774458, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.242813110351562, "learning_rate": 5e-05, "loss": 0.8541, "mean_token_accuracy": 0.8729026913642883, "num_tokens": 680160926.0, "step": 17909 }, { "epoch": 2.2783360895560363, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.894034385681152, "learning_rate": 5e-05, "loss": 0.8966, "mean_token_accuracy": 0.8794448375701904, "num_tokens": 680202621.0, "step": 17910 }, { "epoch": 2.278463299834627, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.249486923217773, "learning_rate": 5e-05, "loss": 0.811, "mean_token_accuracy": 0.8815922737121582, "num_tokens": 680241129.0, "step": 17911 }, { "epoch": 2.2785905101132173, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.346229553222656, "learning_rate": 5e-05, "loss": 0.8337, "mean_token_accuracy": 0.8835946321487427, "num_tokens": 680279832.0, "step": 17912 }, { "epoch": 2.278717720391808, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.945670127868652, "learning_rate": 5e-05, "loss": 0.7993, "mean_token_accuracy": 0.8741349577903748, "num_tokens": 680313541.0, "step": 17913 }, { "epoch": 2.278844930670398, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.220901489257812, "learning_rate": 5e-05, "loss": 0.82, "mean_token_accuracy": 0.8779535293579102, "num_tokens": 680355756.0, "step": 17914 }, { "epoch": 2.278972140948989, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.559677600860596, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.87898850440979, "num_tokens": 680396276.0, "step": 17915 }, { "epoch": 2.279099351227579, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.321015357971191, "learning_rate": 5e-05, "loss": 0.8169, "mean_token_accuracy": 0.8776715993881226, "num_tokens": 680435196.0, "step": 17916 }, { "epoch": 2.2792265615061695, "ewc_loss": 0.375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.360141754150391, "learning_rate": 5e-05, "loss": 0.6721, "mean_token_accuracy": 0.8938759565353394, "num_tokens": 680473088.0, "step": 17917 }, { "epoch": 2.27935377178476, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.341431617736816, "learning_rate": 5e-05, "loss": 0.8267, "mean_token_accuracy": 0.8745288848876953, "num_tokens": 680506269.0, "step": 17918 }, { "epoch": 2.2794809820633506, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.267642021179199, "learning_rate": 5e-05, "loss": 0.7524, "mean_token_accuracy": 0.8682131767272949, "num_tokens": 680543464.0, "step": 17919 }, { "epoch": 2.279608192341941, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.383949279785156, "learning_rate": 5e-05, "loss": 0.8127, "mean_token_accuracy": 0.8776061534881592, "num_tokens": 680582730.0, "step": 17920 }, { "epoch": 2.2797354026205316, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.200392723083496, "learning_rate": 5e-05, "loss": 0.6875, "mean_token_accuracy": 0.8870160579681396, "num_tokens": 680625773.0, "step": 17921 }, { "epoch": 2.279862612899122, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.445487022399902, "learning_rate": 5e-05, "loss": 0.8042, "mean_token_accuracy": 0.8805829882621765, "num_tokens": 680655890.0, "step": 17922 }, { "epoch": 2.2799898231777127, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.263959884643555, "learning_rate": 5e-05, "loss": 0.6989, "mean_token_accuracy": 0.8827637434005737, "num_tokens": 680689921.0, "step": 17923 }, { "epoch": 2.2801170334563032, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.262091636657715, "learning_rate": 5e-05, "loss": 0.8209, "mean_token_accuracy": 0.8722239136695862, "num_tokens": 680735744.0, "step": 17924 }, { "epoch": 2.2802442437348938, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.478389739990234, "learning_rate": 5e-05, "loss": 0.7278, "mean_token_accuracy": 0.8800500631332397, "num_tokens": 680766925.0, "step": 17925 }, { "epoch": 2.2803714540134843, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.8925251960754395, "learning_rate": 5e-05, "loss": 0.7289, "mean_token_accuracy": 0.8869179487228394, "num_tokens": 680804324.0, "step": 17926 }, { "epoch": 2.280498664292075, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.703441619873047, "learning_rate": 5e-05, "loss": 0.699, "mean_token_accuracy": 0.8879916071891785, "num_tokens": 680842005.0, "step": 17927 }, { "epoch": 2.2806258745706653, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.566287040710449, "learning_rate": 5e-05, "loss": 0.6892, "mean_token_accuracy": 0.888441264629364, "num_tokens": 680873475.0, "step": 17928 }, { "epoch": 2.280753084849256, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.91795539855957, "learning_rate": 5e-05, "loss": 0.7174, "mean_token_accuracy": 0.8888044357299805, "num_tokens": 680911322.0, "step": 17929 }, { "epoch": 2.2808802951278464, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.252040386199951, "learning_rate": 5e-05, "loss": 0.742, "mean_token_accuracy": 0.8610445261001587, "num_tokens": 680949778.0, "step": 17930 }, { "epoch": 2.281007505406437, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.175006866455078, "learning_rate": 5e-05, "loss": 0.761, "mean_token_accuracy": 0.8785102963447571, "num_tokens": 680992795.0, "step": 17931 }, { "epoch": 2.2811347156850275, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.966608047485352, "learning_rate": 5e-05, "loss": 0.712, "mean_token_accuracy": 0.8688896894454956, "num_tokens": 681026430.0, "step": 17932 }, { "epoch": 2.281261925963618, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.454032897949219, "learning_rate": 5e-05, "loss": 0.6794, "mean_token_accuracy": 0.9154959917068481, "num_tokens": 681062998.0, "step": 17933 }, { "epoch": 2.2813891362422085, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.804214954376221, "learning_rate": 5e-05, "loss": 0.6672, "mean_token_accuracy": 0.880448579788208, "num_tokens": 681100956.0, "step": 17934 }, { "epoch": 2.281516346520799, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.579137802124023, "learning_rate": 5e-05, "loss": 0.7837, "mean_token_accuracy": 0.8859920501708984, "num_tokens": 681136406.0, "step": 17935 }, { "epoch": 2.2816435567993896, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.899316310882568, "learning_rate": 5e-05, "loss": 0.7135, "mean_token_accuracy": 0.8716627359390259, "num_tokens": 681174899.0, "step": 17936 }, { "epoch": 2.2817707670779797, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000690460205078125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.390573501586914, "learning_rate": 5e-05, "loss": 0.806, "mean_token_accuracy": 0.8782934546470642, "num_tokens": 681207662.0, "step": 17937 }, { "epoch": 2.2818979773565706, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.250337600708008, "learning_rate": 5e-05, "loss": 0.7233, "mean_token_accuracy": 0.8748140931129456, "num_tokens": 681248819.0, "step": 17938 }, { "epoch": 2.2820251876351607, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.9094696044921875, "learning_rate": 5e-05, "loss": 0.742, "mean_token_accuracy": 0.8818796277046204, "num_tokens": 681286416.0, "step": 17939 }, { "epoch": 2.2821523979137512, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.745595455169678, "learning_rate": 5e-05, "loss": 0.7668, "mean_token_accuracy": 0.8669978380203247, "num_tokens": 681325629.0, "step": 17940 }, { "epoch": 2.2822796081923418, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.27765417098999, "learning_rate": 5e-05, "loss": 0.7325, "mean_token_accuracy": 0.8680698275566101, "num_tokens": 681362310.0, "step": 17941 }, { "epoch": 2.2824068184709323, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.268482208251953, "learning_rate": 5e-05, "loss": 0.7587, "mean_token_accuracy": 0.883030354976654, "num_tokens": 681401123.0, "step": 17942 }, { "epoch": 2.282534028749523, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.811612606048584, "learning_rate": 5e-05, "loss": 0.6654, "mean_token_accuracy": 0.882846474647522, "num_tokens": 681439801.0, "step": 17943 }, { "epoch": 2.2826612390281134, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.662908554077148, "learning_rate": 5e-05, "loss": 0.7513, "mean_token_accuracy": 0.8940249681472778, "num_tokens": 681470032.0, "step": 17944 }, { "epoch": 2.282788449306704, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.601940631866455, "learning_rate": 5e-05, "loss": 0.6662, "mean_token_accuracy": 0.8875274658203125, "num_tokens": 681512421.0, "step": 17945 }, { "epoch": 2.2829156595852944, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.806315422058105, "learning_rate": 5e-05, "loss": 0.8548, "mean_token_accuracy": 0.8794760704040527, "num_tokens": 681550226.0, "step": 17946 }, { "epoch": 2.283042869863885, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.740082263946533, "learning_rate": 5e-05, "loss": 0.6572, "mean_token_accuracy": 0.8900773525238037, "num_tokens": 681587885.0, "step": 17947 }, { "epoch": 2.2831700801424755, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.584403991699219, "learning_rate": 5e-05, "loss": 0.7743, "mean_token_accuracy": 0.8987075090408325, "num_tokens": 681623230.0, "step": 17948 }, { "epoch": 2.283297290421066, "ewc_loss": 0.375, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.188188552856445, "learning_rate": 5e-05, "loss": 0.6934, "mean_token_accuracy": 0.8871554136276245, "num_tokens": 681659419.0, "step": 17949 }, { "epoch": 2.2834245006996565, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.005873680114746, "learning_rate": 5e-05, "loss": 0.7699, "mean_token_accuracy": 0.8825292587280273, "num_tokens": 681699845.0, "step": 17950 }, { "epoch": 2.283551710978247, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.69888162612915, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8755563497543335, "num_tokens": 681737579.0, "step": 17951 }, { "epoch": 2.2836789212568376, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00069427490234375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.500092029571533, "learning_rate": 5e-05, "loss": 0.7388, "mean_token_accuracy": 0.8763808012008667, "num_tokens": 681771174.0, "step": 17952 }, { "epoch": 2.283806131535428, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.89044713973999, "learning_rate": 5e-05, "loss": 0.7521, "mean_token_accuracy": 0.8761206269264221, "num_tokens": 681810641.0, "step": 17953 }, { "epoch": 2.2839333418140186, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.447144031524658, "learning_rate": 5e-05, "loss": 0.7521, "mean_token_accuracy": 0.8738781809806824, "num_tokens": 681849273.0, "step": 17954 }, { "epoch": 2.284060552092609, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.639936923980713, "learning_rate": 5e-05, "loss": 0.7037, "mean_token_accuracy": 0.8879804611206055, "num_tokens": 681893105.0, "step": 17955 }, { "epoch": 2.2841877623711997, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.817351818084717, "learning_rate": 5e-05, "loss": 0.723, "mean_token_accuracy": 0.886275053024292, "num_tokens": 681931364.0, "step": 17956 }, { "epoch": 2.28431497264979, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.059515953063965, "learning_rate": 5e-05, "loss": 0.677, "mean_token_accuracy": 0.8857938051223755, "num_tokens": 681970809.0, "step": 17957 }, { "epoch": 2.2844421829283807, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.519240379333496, "learning_rate": 5e-05, "loss": 0.7836, "mean_token_accuracy": 0.8834536075592041, "num_tokens": 682004581.0, "step": 17958 }, { "epoch": 2.2845693932069713, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.3380537033081055, "learning_rate": 5e-05, "loss": 0.6331, "mean_token_accuracy": 0.8872542381286621, "num_tokens": 682041809.0, "step": 17959 }, { "epoch": 2.284696603485562, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 9.492497444152832, "learning_rate": 5e-05, "loss": 0.8997, "mean_token_accuracy": 0.877533495426178, "num_tokens": 682078058.0, "step": 17960 }, { "epoch": 2.2848238137641523, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00023365020751953125, "grad_norm": 5.805802345275879, "learning_rate": 5e-05, "loss": 0.6148, "mean_token_accuracy": 0.8901426196098328, "num_tokens": 682111743.0, "step": 17961 }, { "epoch": 2.2849510240427424, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.431889533996582, "learning_rate": 5e-05, "loss": 1.0035, "mean_token_accuracy": 0.8778807520866394, "num_tokens": 682147710.0, "step": 17962 }, { "epoch": 2.2850782343213334, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.573263168334961, "learning_rate": 5e-05, "loss": 0.7634, "mean_token_accuracy": 0.8784253597259521, "num_tokens": 682187139.0, "step": 17963 }, { "epoch": 2.2852054445999235, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.343693733215332, "learning_rate": 5e-05, "loss": 1.0332, "mean_token_accuracy": 0.8885531425476074, "num_tokens": 682226345.0, "step": 17964 }, { "epoch": 2.285332654878514, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.446234703063965, "learning_rate": 5e-05, "loss": 0.9801, "mean_token_accuracy": 0.8860211372375488, "num_tokens": 682262479.0, "step": 17965 }, { "epoch": 2.2854598651571045, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.380278587341309, "learning_rate": 5e-05, "loss": 0.8784, "mean_token_accuracy": 0.879706621170044, "num_tokens": 682300062.0, "step": 17966 }, { "epoch": 2.285587075435695, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.73341178894043, "learning_rate": 5e-05, "loss": 1.0217, "mean_token_accuracy": 0.8841620683670044, "num_tokens": 682338075.0, "step": 17967 }, { "epoch": 2.2857142857142856, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.660983085632324, "learning_rate": 5e-05, "loss": 0.7258, "mean_token_accuracy": 0.8809456825256348, "num_tokens": 682373036.0, "step": 17968 }, { "epoch": 2.285841495992876, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 11.191922187805176, "learning_rate": 5e-05, "loss": 1.1636, "mean_token_accuracy": 0.8713229298591614, "num_tokens": 682411131.0, "step": 17969 }, { "epoch": 2.2859687062714666, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 7.7271881103515625, "learning_rate": 5e-05, "loss": 0.9951, "mean_token_accuracy": 0.8630604147911072, "num_tokens": 682449142.0, "step": 17970 }, { "epoch": 2.286095916550057, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 11.100963592529297, "learning_rate": 5e-05, "loss": 1.0615, "mean_token_accuracy": 0.8984420895576477, "num_tokens": 682486369.0, "step": 17971 }, { "epoch": 2.2862231268286477, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 9.109041213989258, "learning_rate": 5e-05, "loss": 1.0959, "mean_token_accuracy": 0.8847890496253967, "num_tokens": 682519756.0, "step": 17972 }, { "epoch": 2.2863503371072382, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.793660163879395, "learning_rate": 5e-05, "loss": 0.9724, "mean_token_accuracy": 0.8785806894302368, "num_tokens": 682559930.0, "step": 17973 }, { "epoch": 2.2864775473858288, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.093511581420898, "learning_rate": 5e-05, "loss": 1.0399, "mean_token_accuracy": 0.8756475448608398, "num_tokens": 682601379.0, "step": 17974 }, { "epoch": 2.2866047576644193, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.759283065795898, "learning_rate": 5e-05, "loss": 0.9151, "mean_token_accuracy": 0.8843804001808167, "num_tokens": 682639070.0, "step": 17975 }, { "epoch": 2.28673196794301, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.762901306152344, "learning_rate": 5e-05, "loss": 0.9063, "mean_token_accuracy": 0.8885200023651123, "num_tokens": 682675684.0, "step": 17976 }, { "epoch": 2.2868591782216003, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.332352638244629, "learning_rate": 5e-05, "loss": 0.888, "mean_token_accuracy": 0.8834571242332458, "num_tokens": 682711026.0, "step": 17977 }, { "epoch": 2.286986388500191, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.167609214782715, "learning_rate": 5e-05, "loss": 0.7864, "mean_token_accuracy": 0.8960584402084351, "num_tokens": 682743074.0, "step": 17978 }, { "epoch": 2.2871135987787814, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.313732147216797, "learning_rate": 5e-05, "loss": 0.8437, "mean_token_accuracy": 0.8823519945144653, "num_tokens": 682778875.0, "step": 17979 }, { "epoch": 2.287240809057372, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.534483432769775, "learning_rate": 5e-05, "loss": 0.7553, "mean_token_accuracy": 0.8855385780334473, "num_tokens": 682817699.0, "step": 17980 }, { "epoch": 2.2873680193359625, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.598770141601562, "learning_rate": 5e-05, "loss": 0.8813, "mean_token_accuracy": 0.8737799525260925, "num_tokens": 682852742.0, "step": 17981 }, { "epoch": 2.287495229614553, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.988574028015137, "learning_rate": 5e-05, "loss": 0.7221, "mean_token_accuracy": 0.877197265625, "num_tokens": 682894911.0, "step": 17982 }, { "epoch": 2.2876224398931435, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.143284797668457, "learning_rate": 5e-05, "loss": 0.9069, "mean_token_accuracy": 0.8788197636604309, "num_tokens": 682932387.0, "step": 17983 }, { "epoch": 2.287749650171734, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.6227617263793945, "learning_rate": 5e-05, "loss": 0.6863, "mean_token_accuracy": 0.8887325525283813, "num_tokens": 682965072.0, "step": 17984 }, { "epoch": 2.2878768604503246, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.783445358276367, "learning_rate": 5e-05, "loss": 0.9497, "mean_token_accuracy": 0.8882937431335449, "num_tokens": 683006008.0, "step": 17985 }, { "epoch": 2.288004070728915, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 6.877522945404053, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8852342367172241, "num_tokens": 683042043.0, "step": 17986 }, { "epoch": 2.288131281007505, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.938627243041992, "learning_rate": 5e-05, "loss": 1.0122, "mean_token_accuracy": 0.8801136612892151, "num_tokens": 683079429.0, "step": 17987 }, { "epoch": 2.288258491286096, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 7.759943962097168, "learning_rate": 5e-05, "loss": 0.9319, "mean_token_accuracy": 0.8762073516845703, "num_tokens": 683115686.0, "step": 17988 }, { "epoch": 2.2883857015646862, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.226183891296387, "learning_rate": 5e-05, "loss": 0.9233, "mean_token_accuracy": 0.8911451101303101, "num_tokens": 683156715.0, "step": 17989 }, { "epoch": 2.2885129118432768, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.238890647888184, "learning_rate": 5e-05, "loss": 0.9418, "mean_token_accuracy": 0.8818427920341492, "num_tokens": 683191758.0, "step": 17990 }, { "epoch": 2.2886401221218673, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.388101577758789, "learning_rate": 5e-05, "loss": 0.8596, "mean_token_accuracy": 0.8813618421554565, "num_tokens": 683225146.0, "step": 17991 }, { "epoch": 2.288767332400458, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.146238327026367, "learning_rate": 5e-05, "loss": 0.8618, "mean_token_accuracy": 0.8924532532691956, "num_tokens": 683259552.0, "step": 17992 }, { "epoch": 2.2888945426790483, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.083662033081055, "learning_rate": 5e-05, "loss": 0.8174, "mean_token_accuracy": 0.8926645517349243, "num_tokens": 683296925.0, "step": 17993 }, { "epoch": 2.289021752957639, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.745737552642822, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.8824623823165894, "num_tokens": 683328836.0, "step": 17994 }, { "epoch": 2.2891489632362294, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.24473762512207, "learning_rate": 5e-05, "loss": 0.824, "mean_token_accuracy": 0.8891794681549072, "num_tokens": 683359306.0, "step": 17995 }, { "epoch": 2.28927617351482, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.270998001098633, "learning_rate": 5e-05, "loss": 0.6868, "mean_token_accuracy": 0.8997066617012024, "num_tokens": 683395743.0, "step": 17996 }, { "epoch": 2.2894033837934105, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.595253944396973, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8874906897544861, "num_tokens": 683434528.0, "step": 17997 }, { "epoch": 2.289530594072001, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.984351634979248, "learning_rate": 5e-05, "loss": 0.7198, "mean_token_accuracy": 0.878705620765686, "num_tokens": 683470540.0, "step": 17998 }, { "epoch": 2.2896578043505915, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.828328132629395, "learning_rate": 5e-05, "loss": 0.8877, "mean_token_accuracy": 0.8801122307777405, "num_tokens": 683502805.0, "step": 17999 }, { "epoch": 2.289785014629182, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.9865851402282715, "learning_rate": 5e-05, "loss": 0.7301, "mean_token_accuracy": 0.8843932151794434, "num_tokens": 683540306.0, "step": 18000 }, { "epoch": 2.2899122249077726, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.888497352600098, "learning_rate": 5e-05, "loss": 0.9044, "mean_token_accuracy": 0.8822562098503113, "num_tokens": 683573089.0, "step": 18001 }, { "epoch": 2.290039435186363, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.138633728027344, "learning_rate": 5e-05, "loss": 0.7767, "mean_token_accuracy": 0.8765329122543335, "num_tokens": 683613540.0, "step": 18002 }, { "epoch": 2.2901666454649536, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.801647186279297, "learning_rate": 5e-05, "loss": 0.9302, "mean_token_accuracy": 0.8771052956581116, "num_tokens": 683655016.0, "step": 18003 }, { "epoch": 2.290293855743544, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.443874835968018, "learning_rate": 5e-05, "loss": 0.7759, "mean_token_accuracy": 0.8891391158103943, "num_tokens": 683692607.0, "step": 18004 }, { "epoch": 2.2904210660221347, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.358942031860352, "learning_rate": 5e-05, "loss": 0.8498, "mean_token_accuracy": 0.8841478824615479, "num_tokens": 683726825.0, "step": 18005 }, { "epoch": 2.290548276300725, "ewc_loss": 0.5, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.7863311767578125, "learning_rate": 5e-05, "loss": 0.8706, "mean_token_accuracy": 0.8670614957809448, "num_tokens": 683763905.0, "step": 18006 }, { "epoch": 2.2906754865793157, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.822439670562744, "learning_rate": 5e-05, "loss": 0.8371, "mean_token_accuracy": 0.874286413192749, "num_tokens": 683806926.0, "step": 18007 }, { "epoch": 2.2908026968579063, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.92771577835083, "learning_rate": 5e-05, "loss": 0.8339, "mean_token_accuracy": 0.8767423033714294, "num_tokens": 683846069.0, "step": 18008 }, { "epoch": 2.290929907136497, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.5623393058776855, "learning_rate": 5e-05, "loss": 0.7611, "mean_token_accuracy": 0.8808431625366211, "num_tokens": 683887530.0, "step": 18009 }, { "epoch": 2.2910571174150873, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.92893648147583, "learning_rate": 5e-05, "loss": 0.788, "mean_token_accuracy": 0.8786255121231079, "num_tokens": 683925721.0, "step": 18010 }, { "epoch": 2.291184327693678, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.485024929046631, "learning_rate": 5e-05, "loss": 0.7293, "mean_token_accuracy": 0.8852978944778442, "num_tokens": 683967867.0, "step": 18011 }, { "epoch": 2.291311537972268, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.918470859527588, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8809121251106262, "num_tokens": 684006360.0, "step": 18012 }, { "epoch": 2.291438748250859, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.402983665466309, "learning_rate": 5e-05, "loss": 0.7016, "mean_token_accuracy": 0.8884677290916443, "num_tokens": 684047559.0, "step": 18013 }, { "epoch": 2.291565958529449, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.934412479400635, "learning_rate": 5e-05, "loss": 0.7739, "mean_token_accuracy": 0.8780887126922607, "num_tokens": 684092028.0, "step": 18014 }, { "epoch": 2.2916931688080395, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.331579208374023, "learning_rate": 5e-05, "loss": 0.7275, "mean_token_accuracy": 0.8732147216796875, "num_tokens": 684131441.0, "step": 18015 }, { "epoch": 2.29182037908663, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.985672473907471, "learning_rate": 5e-05, "loss": 0.7398, "mean_token_accuracy": 0.8863458633422852, "num_tokens": 684163864.0, "step": 18016 }, { "epoch": 2.2919475893652206, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.236720085144043, "learning_rate": 5e-05, "loss": 0.6873, "mean_token_accuracy": 0.8816473484039307, "num_tokens": 684200962.0, "step": 18017 }, { "epoch": 2.292074799643811, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.072181701660156, "learning_rate": 5e-05, "loss": 0.7344, "mean_token_accuracy": 0.8886612057685852, "num_tokens": 684237060.0, "step": 18018 }, { "epoch": 2.2922020099224016, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.105103015899658, "learning_rate": 5e-05, "loss": 0.6763, "mean_token_accuracy": 0.8863232135772705, "num_tokens": 684271299.0, "step": 18019 }, { "epoch": 2.292329220200992, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.223200798034668, "learning_rate": 5e-05, "loss": 0.7879, "mean_token_accuracy": 0.8756314516067505, "num_tokens": 684310008.0, "step": 18020 }, { "epoch": 2.2924564304795827, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.924318790435791, "learning_rate": 5e-05, "loss": 0.6802, "mean_token_accuracy": 0.8811714053153992, "num_tokens": 684348477.0, "step": 18021 }, { "epoch": 2.2925836407581732, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.500676155090332, "learning_rate": 5e-05, "loss": 0.8244, "mean_token_accuracy": 0.8723412156105042, "num_tokens": 684383090.0, "step": 18022 }, { "epoch": 2.2927108510367638, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.654978275299072, "learning_rate": 5e-05, "loss": 0.7198, "mean_token_accuracy": 0.8666964769363403, "num_tokens": 684415142.0, "step": 18023 }, { "epoch": 2.2928380613153543, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.859951972961426, "learning_rate": 5e-05, "loss": 0.852, "mean_token_accuracy": 0.8822816014289856, "num_tokens": 684455878.0, "step": 18024 }, { "epoch": 2.292965271593945, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.4407572746276855, "learning_rate": 5e-05, "loss": 0.6981, "mean_token_accuracy": 0.8772076964378357, "num_tokens": 684501140.0, "step": 18025 }, { "epoch": 2.2930924818725353, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.244872093200684, "learning_rate": 5e-05, "loss": 0.9106, "mean_token_accuracy": 0.8782225847244263, "num_tokens": 684542682.0, "step": 18026 }, { "epoch": 2.293219692151126, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.553518295288086, "learning_rate": 5e-05, "loss": 0.7235, "mean_token_accuracy": 0.8739966750144958, "num_tokens": 684581301.0, "step": 18027 }, { "epoch": 2.2933469024297164, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.175432205200195, "learning_rate": 5e-05, "loss": 0.9099, "mean_token_accuracy": 0.8901795744895935, "num_tokens": 684620124.0, "step": 18028 }, { "epoch": 2.293474112708307, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.2505879402160645, "learning_rate": 5e-05, "loss": 0.8326, "mean_token_accuracy": 0.8763024806976318, "num_tokens": 684657237.0, "step": 18029 }, { "epoch": 2.2936013229868975, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000701904296875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.286704063415527, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8864072561264038, "num_tokens": 684696717.0, "step": 18030 }, { "epoch": 2.293728533265488, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.154568672180176, "learning_rate": 5e-05, "loss": 0.8805, "mean_token_accuracy": 0.8802531361579895, "num_tokens": 684734823.0, "step": 18031 }, { "epoch": 2.2938557435440785, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.103674411773682, "learning_rate": 5e-05, "loss": 0.732, "mean_token_accuracy": 0.8786107301712036, "num_tokens": 684773036.0, "step": 18032 }, { "epoch": 2.293982953822669, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.901107788085938, "learning_rate": 5e-05, "loss": 0.8758, "mean_token_accuracy": 0.8886324167251587, "num_tokens": 684811210.0, "step": 18033 }, { "epoch": 2.2941101641012596, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.480785846710205, "learning_rate": 5e-05, "loss": 0.6869, "mean_token_accuracy": 0.8835078477859497, "num_tokens": 684850656.0, "step": 18034 }, { "epoch": 2.2942373743798496, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.328348159790039, "learning_rate": 5e-05, "loss": 0.9409, "mean_token_accuracy": 0.8793404698371887, "num_tokens": 684890113.0, "step": 18035 }, { "epoch": 2.2943645846584406, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.5960211753845215, "learning_rate": 5e-05, "loss": 0.7003, "mean_token_accuracy": 0.8885276913642883, "num_tokens": 684930418.0, "step": 18036 }, { "epoch": 2.2944917949370307, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.176231384277344, "learning_rate": 5e-05, "loss": 0.9521, "mean_token_accuracy": 0.8814388513565063, "num_tokens": 684967068.0, "step": 18037 }, { "epoch": 2.2946190052156212, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.319903373718262, "learning_rate": 5e-05, "loss": 0.9036, "mean_token_accuracy": 0.8611312508583069, "num_tokens": 685007602.0, "step": 18038 }, { "epoch": 2.2947462154942118, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.384297370910645, "learning_rate": 5e-05, "loss": 0.8733, "mean_token_accuracy": 0.8841020464897156, "num_tokens": 685043332.0, "step": 18039 }, { "epoch": 2.2948734257728023, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.954662322998047, "learning_rate": 5e-05, "loss": 0.8584, "mean_token_accuracy": 0.886956512928009, "num_tokens": 685078327.0, "step": 18040 }, { "epoch": 2.295000636051393, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.61928653717041, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.881722629070282, "num_tokens": 685119065.0, "step": 18041 }, { "epoch": 2.2951278463299833, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.165300369262695, "learning_rate": 5e-05, "loss": 0.9363, "mean_token_accuracy": 0.8581745624542236, "num_tokens": 685154988.0, "step": 18042 }, { "epoch": 2.295255056608574, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.476189613342285, "learning_rate": 5e-05, "loss": 0.7741, "mean_token_accuracy": 0.8774380087852478, "num_tokens": 685194739.0, "step": 18043 }, { "epoch": 2.2953822668871644, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.832941055297852, "learning_rate": 5e-05, "loss": 0.8335, "mean_token_accuracy": 0.8698861598968506, "num_tokens": 685235619.0, "step": 18044 }, { "epoch": 2.295509477165755, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.821537494659424, "learning_rate": 5e-05, "loss": 0.7977, "mean_token_accuracy": 0.8772794008255005, "num_tokens": 685275189.0, "step": 18045 }, { "epoch": 2.2956366874443455, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.319535255432129, "learning_rate": 5e-05, "loss": 0.76, "mean_token_accuracy": 0.8734943866729736, "num_tokens": 685319928.0, "step": 18046 }, { "epoch": 2.295763897722936, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.219326972961426, "learning_rate": 5e-05, "loss": 0.7809, "mean_token_accuracy": 0.8886244297027588, "num_tokens": 685355698.0, "step": 18047 }, { "epoch": 2.2958911080015265, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.933824062347412, "learning_rate": 5e-05, "loss": 0.6829, "mean_token_accuracy": 0.8818098902702332, "num_tokens": 685394873.0, "step": 18048 }, { "epoch": 2.296018318280117, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.53377914428711, "learning_rate": 5e-05, "loss": 0.8044, "mean_token_accuracy": 0.8864588737487793, "num_tokens": 685433896.0, "step": 18049 }, { "epoch": 2.2961455285587076, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.700193881988525, "learning_rate": 5e-05, "loss": 0.6937, "mean_token_accuracy": 0.8789860010147095, "num_tokens": 685476036.0, "step": 18050 }, { "epoch": 2.296272738837298, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.89223861694336, "learning_rate": 5e-05, "loss": 0.8541, "mean_token_accuracy": 0.8827692270278931, "num_tokens": 685514020.0, "step": 18051 }, { "epoch": 2.2963999491158886, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.4986443519592285, "learning_rate": 5e-05, "loss": 0.6672, "mean_token_accuracy": 0.8892608284950256, "num_tokens": 685547644.0, "step": 18052 }, { "epoch": 2.296527159394479, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.256826400756836, "learning_rate": 5e-05, "loss": 0.9446, "mean_token_accuracy": 0.8765898942947388, "num_tokens": 685584147.0, "step": 18053 }, { "epoch": 2.2966543696730697, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.622916221618652, "learning_rate": 5e-05, "loss": 0.7288, "mean_token_accuracy": 0.8824423551559448, "num_tokens": 685623093.0, "step": 18054 }, { "epoch": 2.29678157995166, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.260156631469727, "learning_rate": 5e-05, "loss": 0.9515, "mean_token_accuracy": 0.8886399269104004, "num_tokens": 685657837.0, "step": 18055 }, { "epoch": 2.2969087902302507, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00070953369140625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.2089152336120605, "learning_rate": 5e-05, "loss": 0.8307, "mean_token_accuracy": 0.8808460235595703, "num_tokens": 685691077.0, "step": 18056 }, { "epoch": 2.2970360005088413, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.688335418701172, "learning_rate": 5e-05, "loss": 0.9272, "mean_token_accuracy": 0.8785830736160278, "num_tokens": 685725011.0, "step": 18057 }, { "epoch": 2.297163210787432, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.798118591308594, "learning_rate": 5e-05, "loss": 0.9254, "mean_token_accuracy": 0.8661801815032959, "num_tokens": 685764212.0, "step": 18058 }, { "epoch": 2.2972904210660223, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.958384990692139, "learning_rate": 5e-05, "loss": 0.9045, "mean_token_accuracy": 0.8617008924484253, "num_tokens": 685800780.0, "step": 18059 }, { "epoch": 2.2974176313446124, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.000189781188965, "learning_rate": 5e-05, "loss": 0.8588, "mean_token_accuracy": 0.8805688619613647, "num_tokens": 685837728.0, "step": 18060 }, { "epoch": 2.2975448416232034, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.6649603843688965, "learning_rate": 5e-05, "loss": 0.8427, "mean_token_accuracy": 0.8652327060699463, "num_tokens": 685869349.0, "step": 18061 }, { "epoch": 2.2976720519017935, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.804994106292725, "learning_rate": 5e-05, "loss": 0.7836, "mean_token_accuracy": 0.8886271715164185, "num_tokens": 685908114.0, "step": 18062 }, { "epoch": 2.297799262180384, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.762218952178955, "learning_rate": 5e-05, "loss": 0.8164, "mean_token_accuracy": 0.8715696334838867, "num_tokens": 685944980.0, "step": 18063 }, { "epoch": 2.2979264724589745, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000705718994140625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.527420997619629, "learning_rate": 5e-05, "loss": 0.7522, "mean_token_accuracy": 0.8852859139442444, "num_tokens": 685982729.0, "step": 18064 }, { "epoch": 2.298053682737565, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.838949680328369, "learning_rate": 5e-05, "loss": 0.7233, "mean_token_accuracy": 0.8981101512908936, "num_tokens": 686014090.0, "step": 18065 }, { "epoch": 2.2981808930161556, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.499651908874512, "learning_rate": 5e-05, "loss": 0.7294, "mean_token_accuracy": 0.8869285583496094, "num_tokens": 686047345.0, "step": 18066 }, { "epoch": 2.298308103294746, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.669859409332275, "learning_rate": 5e-05, "loss": 0.7892, "mean_token_accuracy": 0.8714514970779419, "num_tokens": 686088662.0, "step": 18067 }, { "epoch": 2.2984353135733366, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.693884372711182, "learning_rate": 5e-05, "loss": 0.7818, "mean_token_accuracy": 0.8713333606719971, "num_tokens": 686124210.0, "step": 18068 }, { "epoch": 2.298562523851927, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.395379543304443, "learning_rate": 5e-05, "loss": 0.7191, "mean_token_accuracy": 0.8823294639587402, "num_tokens": 686157759.0, "step": 18069 }, { "epoch": 2.2986897341305177, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.9299798011779785, "learning_rate": 5e-05, "loss": 0.7866, "mean_token_accuracy": 0.8748310804367065, "num_tokens": 686198486.0, "step": 18070 }, { "epoch": 2.298816944409108, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.142210006713867, "learning_rate": 5e-05, "loss": 0.7504, "mean_token_accuracy": 0.8618423938751221, "num_tokens": 686235859.0, "step": 18071 }, { "epoch": 2.2989441546876987, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.153800010681152, "learning_rate": 5e-05, "loss": 0.7653, "mean_token_accuracy": 0.8874966502189636, "num_tokens": 686274185.0, "step": 18072 }, { "epoch": 2.2990713649662893, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.908134460449219, "learning_rate": 5e-05, "loss": 0.6433, "mean_token_accuracy": 0.895156979560852, "num_tokens": 686308706.0, "step": 18073 }, { "epoch": 2.29919857524488, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.458441734313965, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.891428530216217, "num_tokens": 686347871.0, "step": 18074 }, { "epoch": 2.2993257855234703, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.581527233123779, "learning_rate": 5e-05, "loss": 0.6963, "mean_token_accuracy": 0.8738368153572083, "num_tokens": 686395284.0, "step": 18075 }, { "epoch": 2.299452995802061, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.96422290802002, "learning_rate": 5e-05, "loss": 0.8854, "mean_token_accuracy": 0.8740672469139099, "num_tokens": 686436566.0, "step": 18076 }, { "epoch": 2.2995802060806514, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.217837810516357, "learning_rate": 5e-05, "loss": 0.6479, "mean_token_accuracy": 0.888627827167511, "num_tokens": 686477332.0, "step": 18077 }, { "epoch": 2.299707416359242, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.617520332336426, "learning_rate": 5e-05, "loss": 0.9082, "mean_token_accuracy": 0.8939491510391235, "num_tokens": 686512000.0, "step": 18078 }, { "epoch": 2.2998346266378324, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.31137752532959, "learning_rate": 5e-05, "loss": 0.7037, "mean_token_accuracy": 0.8843395709991455, "num_tokens": 686544732.0, "step": 18079 }, { "epoch": 2.299961836916423, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.755024909973145, "learning_rate": 5e-05, "loss": 1.0028, "mean_token_accuracy": 0.8827610015869141, "num_tokens": 686583100.0, "step": 18080 }, { "epoch": 2.3000890471950135, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 7.401321887969971, "learning_rate": 5e-05, "loss": 0.8921, "mean_token_accuracy": 0.8800652027130127, "num_tokens": 686620824.0, "step": 18081 }, { "epoch": 2.300216257473604, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.665427207946777, "learning_rate": 5e-05, "loss": 0.9141, "mean_token_accuracy": 0.8871657848358154, "num_tokens": 686660410.0, "step": 18082 }, { "epoch": 2.3003434677521946, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.613138198852539, "learning_rate": 5e-05, "loss": 0.9319, "mean_token_accuracy": 0.8929239511489868, "num_tokens": 686694385.0, "step": 18083 }, { "epoch": 2.300470678030785, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.878294944763184, "learning_rate": 5e-05, "loss": 0.7211, "mean_token_accuracy": 0.8855454921722412, "num_tokens": 686736844.0, "step": 18084 }, { "epoch": 2.300597888309375, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.786008834838867, "learning_rate": 5e-05, "loss": 1.0099, "mean_token_accuracy": 0.8755536079406738, "num_tokens": 686781104.0, "step": 18085 }, { "epoch": 2.300725098587966, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.245981693267822, "learning_rate": 5e-05, "loss": 0.6799, "mean_token_accuracy": 0.895092785358429, "num_tokens": 686815244.0, "step": 18086 }, { "epoch": 2.3008523088665562, "ewc_loss": 0.75, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.514345169067383, "learning_rate": 5e-05, "loss": 1.0877, "mean_token_accuracy": 0.8797178268432617, "num_tokens": 686858133.0, "step": 18087 }, { "epoch": 2.3009795191451468, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 7.519474029541016, "learning_rate": 5e-05, "loss": 0.9775, "mean_token_accuracy": 0.8763289451599121, "num_tokens": 686901250.0, "step": 18088 }, { "epoch": 2.3011067294237373, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.762864112854004, "learning_rate": 5e-05, "loss": 0.9962, "mean_token_accuracy": 0.8887132406234741, "num_tokens": 686936408.0, "step": 18089 }, { "epoch": 2.301233939702328, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 8.80707836151123, "learning_rate": 5e-05, "loss": 1.083, "mean_token_accuracy": 0.8797267079353333, "num_tokens": 686981924.0, "step": 18090 }, { "epoch": 2.3013611499809183, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.05881404876709, "learning_rate": 5e-05, "loss": 0.8492, "mean_token_accuracy": 0.8780806064605713, "num_tokens": 687016052.0, "step": 18091 }, { "epoch": 2.301488360259509, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.359103202819824, "learning_rate": 5e-05, "loss": 1.0099, "mean_token_accuracy": 0.8861938714981079, "num_tokens": 687059107.0, "step": 18092 }, { "epoch": 2.3016155705380994, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.400245189666748, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.8855316638946533, "num_tokens": 687104015.0, "step": 18093 }, { "epoch": 2.30174278081669, "ewc_loss": 0.625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.188973426818848, "learning_rate": 5e-05, "loss": 0.9714, "mean_token_accuracy": 0.8762483596801758, "num_tokens": 687143719.0, "step": 18094 }, { "epoch": 2.3018699910952805, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.869964122772217, "learning_rate": 5e-05, "loss": 0.9195, "mean_token_accuracy": 0.8832489252090454, "num_tokens": 687179732.0, "step": 18095 }, { "epoch": 2.301997201373871, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.104766845703125, "learning_rate": 5e-05, "loss": 0.8452, "mean_token_accuracy": 0.8795799612998962, "num_tokens": 687222309.0, "step": 18096 }, { "epoch": 2.3021244116524615, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.462905883789062, "learning_rate": 5e-05, "loss": 0.9149, "mean_token_accuracy": 0.8876166343688965, "num_tokens": 687259114.0, "step": 18097 }, { "epoch": 2.302251621931052, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.234233856201172, "learning_rate": 5e-05, "loss": 0.7299, "mean_token_accuracy": 0.8867628574371338, "num_tokens": 687293052.0, "step": 18098 }, { "epoch": 2.3023788322096426, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.907225608825684, "learning_rate": 5e-05, "loss": 0.9107, "mean_token_accuracy": 0.8880738019943237, "num_tokens": 687333245.0, "step": 18099 }, { "epoch": 2.302506042488233, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.942300796508789, "learning_rate": 5e-05, "loss": 0.7616, "mean_token_accuracy": 0.8794422745704651, "num_tokens": 687371740.0, "step": 18100 }, { "epoch": 2.3026332527668236, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.215738296508789, "learning_rate": 5e-05, "loss": 0.9905, "mean_token_accuracy": 0.8591508865356445, "num_tokens": 687407298.0, "step": 18101 }, { "epoch": 2.302760463045414, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.053730487823486, "learning_rate": 5e-05, "loss": 0.8061, "mean_token_accuracy": 0.8892172574996948, "num_tokens": 687447425.0, "step": 18102 }, { "epoch": 2.3028876733240047, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.214710235595703, "learning_rate": 5e-05, "loss": 0.9369, "mean_token_accuracy": 0.8835344314575195, "num_tokens": 687488528.0, "step": 18103 }, { "epoch": 2.303014883602595, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.343713760375977, "learning_rate": 5e-05, "loss": 0.8691, "mean_token_accuracy": 0.8792520761489868, "num_tokens": 687521086.0, "step": 18104 }, { "epoch": 2.3031420938811857, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.915151596069336, "learning_rate": 5e-05, "loss": 0.9382, "mean_token_accuracy": 0.8734610080718994, "num_tokens": 687558882.0, "step": 18105 }, { "epoch": 2.3032693041597763, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.449260711669922, "learning_rate": 5e-05, "loss": 0.8304, "mean_token_accuracy": 0.8920168876647949, "num_tokens": 687592058.0, "step": 18106 }, { "epoch": 2.303396514438367, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.774079322814941, "learning_rate": 5e-05, "loss": 0.91, "mean_token_accuracy": 0.8783479332923889, "num_tokens": 687629715.0, "step": 18107 }, { "epoch": 2.3035237247169573, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.188188552856445, "learning_rate": 5e-05, "loss": 0.8195, "mean_token_accuracy": 0.8796260356903076, "num_tokens": 687666201.0, "step": 18108 }, { "epoch": 2.303650934995548, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.02048110961914, "learning_rate": 5e-05, "loss": 0.9166, "mean_token_accuracy": 0.8859707117080688, "num_tokens": 687706897.0, "step": 18109 }, { "epoch": 2.303778145274138, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 6.958892822265625, "learning_rate": 5e-05, "loss": 0.7858, "mean_token_accuracy": 0.8820489048957825, "num_tokens": 687741449.0, "step": 18110 }, { "epoch": 2.303905355552729, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.309059143066406, "learning_rate": 5e-05, "loss": 0.9334, "mean_token_accuracy": 0.8945298194885254, "num_tokens": 687776758.0, "step": 18111 }, { "epoch": 2.304032565831319, "ewc_loss": 0.5, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.179824352264404, "learning_rate": 5e-05, "loss": 0.8747, "mean_token_accuracy": 0.8677749633789062, "num_tokens": 687814636.0, "step": 18112 }, { "epoch": 2.3041597761099095, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.132938385009766, "learning_rate": 5e-05, "loss": 0.9947, "mean_token_accuracy": 0.865476131439209, "num_tokens": 687843685.0, "step": 18113 }, { "epoch": 2.3042869863885, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.725722789764404, "learning_rate": 5e-05, "loss": 0.8582, "mean_token_accuracy": 0.8896495699882507, "num_tokens": 687875040.0, "step": 18114 }, { "epoch": 2.3044141966670906, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.429752349853516, "learning_rate": 5e-05, "loss": 0.9157, "mean_token_accuracy": 0.8772234320640564, "num_tokens": 687912952.0, "step": 18115 }, { "epoch": 2.304541406945681, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.042074203491211, "learning_rate": 5e-05, "loss": 0.8934, "mean_token_accuracy": 0.8817954659461975, "num_tokens": 687951366.0, "step": 18116 }, { "epoch": 2.3046686172242716, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.78931999206543, "learning_rate": 5e-05, "loss": 0.8096, "mean_token_accuracy": 0.883728563785553, "num_tokens": 687993593.0, "step": 18117 }, { "epoch": 2.304795827502862, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.033958435058594, "learning_rate": 5e-05, "loss": 0.8575, "mean_token_accuracy": 0.8795875310897827, "num_tokens": 688033449.0, "step": 18118 }, { "epoch": 2.3049230377814527, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.646417140960693, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.8781283497810364, "num_tokens": 688073239.0, "step": 18119 }, { "epoch": 2.305050248060043, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.744169235229492, "learning_rate": 5e-05, "loss": 0.7982, "mean_token_accuracy": 0.877148449420929, "num_tokens": 688112760.0, "step": 18120 }, { "epoch": 2.3051774583386337, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.8253960609436035, "learning_rate": 5e-05, "loss": 0.8225, "mean_token_accuracy": 0.8714141845703125, "num_tokens": 688150553.0, "step": 18121 }, { "epoch": 2.3053046686172243, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.433450222015381, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8758645057678223, "num_tokens": 688190582.0, "step": 18122 }, { "epoch": 2.305431878895815, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.985373497009277, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.8793994188308716, "num_tokens": 688226512.0, "step": 18123 }, { "epoch": 2.3055590891744053, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.304268836975098, "learning_rate": 5e-05, "loss": 0.7224, "mean_token_accuracy": 0.8812240362167358, "num_tokens": 688264891.0, "step": 18124 }, { "epoch": 2.305686299452996, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.023384094238281, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8788682222366333, "num_tokens": 688301280.0, "step": 18125 }, { "epoch": 2.3058135097315864, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.229959011077881, "learning_rate": 5e-05, "loss": 0.7101, "mean_token_accuracy": 0.8829050660133362, "num_tokens": 688337178.0, "step": 18126 }, { "epoch": 2.305940720010177, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.086892127990723, "learning_rate": 5e-05, "loss": 0.76, "mean_token_accuracy": 0.8896031975746155, "num_tokens": 688375839.0, "step": 18127 }, { "epoch": 2.3060679302887674, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.134025573730469, "learning_rate": 5e-05, "loss": 0.7083, "mean_token_accuracy": 0.8802696466445923, "num_tokens": 688416395.0, "step": 18128 }, { "epoch": 2.306195140567358, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.199382781982422, "learning_rate": 5e-05, "loss": 0.7696, "mean_token_accuracy": 0.8907005190849304, "num_tokens": 688454801.0, "step": 18129 }, { "epoch": 2.3063223508459485, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.022816181182861, "learning_rate": 5e-05, "loss": 0.7115, "mean_token_accuracy": 0.8810861110687256, "num_tokens": 688489265.0, "step": 18130 }, { "epoch": 2.306449561124539, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.358641624450684, "learning_rate": 5e-05, "loss": 0.8165, "mean_token_accuracy": 0.8760951161384583, "num_tokens": 688525392.0, "step": 18131 }, { "epoch": 2.3065767714031296, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.90477180480957, "learning_rate": 5e-05, "loss": 0.7334, "mean_token_accuracy": 0.865087628364563, "num_tokens": 688562389.0, "step": 18132 }, { "epoch": 2.3067039816817196, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.53105640411377, "learning_rate": 5e-05, "loss": 0.8132, "mean_token_accuracy": 0.8822683691978455, "num_tokens": 688597728.0, "step": 18133 }, { "epoch": 2.3068311919603106, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.806647300720215, "learning_rate": 5e-05, "loss": 0.7066, "mean_token_accuracy": 0.8809217810630798, "num_tokens": 688633065.0, "step": 18134 }, { "epoch": 2.3069584022389007, "ewc_loss": 0.5, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.634493827819824, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.8841394186019897, "num_tokens": 688677095.0, "step": 18135 }, { "epoch": 2.3070856125174912, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.905735969543457, "learning_rate": 5e-05, "loss": 0.6864, "mean_token_accuracy": 0.8867083191871643, "num_tokens": 688710421.0, "step": 18136 }, { "epoch": 2.3072128227960818, "ewc_loss": 0.5, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.478105545043945, "learning_rate": 5e-05, "loss": 0.8227, "mean_token_accuracy": 0.8832632303237915, "num_tokens": 688747300.0, "step": 18137 }, { "epoch": 2.3073400330746723, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.228593349456787, "learning_rate": 5e-05, "loss": 0.7236, "mean_token_accuracy": 0.8890427350997925, "num_tokens": 688781867.0, "step": 18138 }, { "epoch": 2.307467243353263, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.019935607910156, "learning_rate": 5e-05, "loss": 0.7832, "mean_token_accuracy": 0.8840450644493103, "num_tokens": 688822777.0, "step": 18139 }, { "epoch": 2.3075944536318533, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.696837902069092, "learning_rate": 5e-05, "loss": 0.7985, "mean_token_accuracy": 0.8746302127838135, "num_tokens": 688862295.0, "step": 18140 }, { "epoch": 2.307721663910444, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.375081539154053, "learning_rate": 5e-05, "loss": 0.6801, "mean_token_accuracy": 0.8960946202278137, "num_tokens": 688904814.0, "step": 18141 }, { "epoch": 2.3078488741890344, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.19158935546875, "learning_rate": 5e-05, "loss": 0.8231, "mean_token_accuracy": 0.873952329158783, "num_tokens": 688940403.0, "step": 18142 }, { "epoch": 2.307976084467625, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.852400302886963, "learning_rate": 5e-05, "loss": 0.7433, "mean_token_accuracy": 0.8663151860237122, "num_tokens": 688980256.0, "step": 18143 }, { "epoch": 2.3081032947462155, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.672074317932129, "learning_rate": 5e-05, "loss": 0.8789, "mean_token_accuracy": 0.8624833822250366, "num_tokens": 689020629.0, "step": 18144 }, { "epoch": 2.308230505024806, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.468655109405518, "learning_rate": 5e-05, "loss": 0.7166, "mean_token_accuracy": 0.8669476509094238, "num_tokens": 689051257.0, "step": 18145 }, { "epoch": 2.3083577153033965, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 9.118617057800293, "learning_rate": 5e-05, "loss": 0.8624, "mean_token_accuracy": 0.8854653239250183, "num_tokens": 689083723.0, "step": 18146 }, { "epoch": 2.308484925581987, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.333061218261719, "learning_rate": 5e-05, "loss": 0.7404, "mean_token_accuracy": 0.8640537261962891, "num_tokens": 689123785.0, "step": 18147 }, { "epoch": 2.3086121358605776, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.390083312988281, "learning_rate": 5e-05, "loss": 0.9444, "mean_token_accuracy": 0.8797808289527893, "num_tokens": 689164053.0, "step": 18148 }, { "epoch": 2.308739346139168, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.755885124206543, "learning_rate": 5e-05, "loss": 0.7359, "mean_token_accuracy": 0.8832643032073975, "num_tokens": 689198221.0, "step": 18149 }, { "epoch": 2.3088665564177586, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.010713577270508, "learning_rate": 5e-05, "loss": 0.9459, "mean_token_accuracy": 0.8781424760818481, "num_tokens": 689232330.0, "step": 18150 }, { "epoch": 2.308993766696349, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.651493072509766, "learning_rate": 5e-05, "loss": 0.8948, "mean_token_accuracy": 0.8743175864219666, "num_tokens": 689277663.0, "step": 18151 }, { "epoch": 2.3091209769749397, "ewc_loss": 0.5, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.903043270111084, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.8880981206893921, "num_tokens": 689308563.0, "step": 18152 }, { "epoch": 2.30924818725353, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.379438400268555, "learning_rate": 5e-05, "loss": 0.8907, "mean_token_accuracy": 0.8807191848754883, "num_tokens": 689346897.0, "step": 18153 }, { "epoch": 2.3093753975321207, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.079033851623535, "learning_rate": 5e-05, "loss": 0.763, "mean_token_accuracy": 0.8710858821868896, "num_tokens": 689380689.0, "step": 18154 }, { "epoch": 2.3095026078107113, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.554222106933594, "learning_rate": 5e-05, "loss": 0.8503, "mean_token_accuracy": 0.8816345930099487, "num_tokens": 689418587.0, "step": 18155 }, { "epoch": 2.309629818089302, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.162792205810547, "learning_rate": 5e-05, "loss": 0.7397, "mean_token_accuracy": 0.8856085538864136, "num_tokens": 689455742.0, "step": 18156 }, { "epoch": 2.3097570283678923, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.006767272949219, "learning_rate": 5e-05, "loss": 0.8062, "mean_token_accuracy": 0.8797667026519775, "num_tokens": 689495618.0, "step": 18157 }, { "epoch": 2.3098842386464824, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.740261554718018, "learning_rate": 5e-05, "loss": 0.8275, "mean_token_accuracy": 0.8633506298065186, "num_tokens": 689529856.0, "step": 18158 }, { "epoch": 2.3100114489250734, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.261415481567383, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.8765778541564941, "num_tokens": 689566785.0, "step": 18159 }, { "epoch": 2.3101386592036635, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.288398742675781, "learning_rate": 5e-05, "loss": 0.832, "mean_token_accuracy": 0.8766795992851257, "num_tokens": 689604332.0, "step": 18160 }, { "epoch": 2.310265869482254, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.817725658416748, "learning_rate": 5e-05, "loss": 0.6693, "mean_token_accuracy": 0.8884840607643127, "num_tokens": 689639159.0, "step": 18161 }, { "epoch": 2.3103930797608445, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.680554389953613, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.8766878843307495, "num_tokens": 689681660.0, "step": 18162 }, { "epoch": 2.310520290039435, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.606541633605957, "learning_rate": 5e-05, "loss": 0.6665, "mean_token_accuracy": 0.8897876739501953, "num_tokens": 689717070.0, "step": 18163 }, { "epoch": 2.3106475003180256, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.93165397644043, "learning_rate": 5e-05, "loss": 0.8782, "mean_token_accuracy": 0.881970226764679, "num_tokens": 689758885.0, "step": 18164 }, { "epoch": 2.310774710596616, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.607515811920166, "learning_rate": 5e-05, "loss": 0.6555, "mean_token_accuracy": 0.8947418928146362, "num_tokens": 689794943.0, "step": 18165 }, { "epoch": 2.3109019208752066, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.043746948242188, "learning_rate": 5e-05, "loss": 0.9052, "mean_token_accuracy": 0.8829918503761292, "num_tokens": 689836943.0, "step": 18166 }, { "epoch": 2.311029131153797, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.82506799697876, "learning_rate": 5e-05, "loss": 0.744, "mean_token_accuracy": 0.8835340142250061, "num_tokens": 689870787.0, "step": 18167 }, { "epoch": 2.3111563414323877, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.917019844055176, "learning_rate": 5e-05, "loss": 0.9362, "mean_token_accuracy": 0.8781529664993286, "num_tokens": 689908302.0, "step": 18168 }, { "epoch": 2.311283551710978, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.161555290222168, "learning_rate": 5e-05, "loss": 0.826, "mean_token_accuracy": 0.8737174272537231, "num_tokens": 689950031.0, "step": 18169 }, { "epoch": 2.3114107619895687, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.621743202209473, "learning_rate": 5e-05, "loss": 0.8927, "mean_token_accuracy": 0.8847629427909851, "num_tokens": 689994674.0, "step": 18170 }, { "epoch": 2.3115379722681593, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.372186660766602, "learning_rate": 5e-05, "loss": 0.8105, "mean_token_accuracy": 0.8862797617912292, "num_tokens": 690038697.0, "step": 18171 }, { "epoch": 2.31166518254675, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.42454719543457, "learning_rate": 5e-05, "loss": 0.9266, "mean_token_accuracy": 0.8752995133399963, "num_tokens": 690077449.0, "step": 18172 }, { "epoch": 2.3117923928253403, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.376287460327148, "learning_rate": 5e-05, "loss": 0.8147, "mean_token_accuracy": 0.8769795894622803, "num_tokens": 690113659.0, "step": 18173 }, { "epoch": 2.311919603103931, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.386212348937988, "learning_rate": 5e-05, "loss": 0.9115, "mean_token_accuracy": 0.8761520385742188, "num_tokens": 690153928.0, "step": 18174 }, { "epoch": 2.3120468133825214, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.300806999206543, "learning_rate": 5e-05, "loss": 0.7849, "mean_token_accuracy": 0.883188009262085, "num_tokens": 690189526.0, "step": 18175 }, { "epoch": 2.312174023661112, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.364520072937012, "learning_rate": 5e-05, "loss": 0.9034, "mean_token_accuracy": 0.8748496770858765, "num_tokens": 690230220.0, "step": 18176 }, { "epoch": 2.3123012339397024, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.309970855712891, "learning_rate": 5e-05, "loss": 0.7991, "mean_token_accuracy": 0.8781600594520569, "num_tokens": 690264998.0, "step": 18177 }, { "epoch": 2.312428444218293, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.225815773010254, "learning_rate": 5e-05, "loss": 0.7922, "mean_token_accuracy": 0.9059780836105347, "num_tokens": 690304323.0, "step": 18178 }, { "epoch": 2.3125556544968835, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.399848461151123, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.8802869915962219, "num_tokens": 690346832.0, "step": 18179 }, { "epoch": 2.312682864775474, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.010823249816895, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.8836753368377686, "num_tokens": 690385829.0, "step": 18180 }, { "epoch": 2.3128100750540646, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.477032661437988, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8840953707695007, "num_tokens": 690429949.0, "step": 18181 }, { "epoch": 2.312937285332655, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.881885051727295, "learning_rate": 5e-05, "loss": 0.7908, "mean_token_accuracy": 0.8844019174575806, "num_tokens": 690473252.0, "step": 18182 }, { "epoch": 2.313064495611245, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.4847612380981445, "learning_rate": 5e-05, "loss": 0.7493, "mean_token_accuracy": 0.8858808279037476, "num_tokens": 690508765.0, "step": 18183 }, { "epoch": 2.313191705889836, "ewc_loss": 0.4375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.7987895011901855, "learning_rate": 5e-05, "loss": 0.7676, "mean_token_accuracy": 0.8810676336288452, "num_tokens": 690546749.0, "step": 18184 }, { "epoch": 2.313318916168426, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.470069885253906, "learning_rate": 5e-05, "loss": 0.74, "mean_token_accuracy": 0.8808803558349609, "num_tokens": 690579808.0, "step": 18185 }, { "epoch": 2.3134461264470167, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.784436225891113, "learning_rate": 5e-05, "loss": 0.7661, "mean_token_accuracy": 0.8801659345626831, "num_tokens": 690611667.0, "step": 18186 }, { "epoch": 2.3135733367256073, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.409583568572998, "learning_rate": 5e-05, "loss": 0.7553, "mean_token_accuracy": 0.8722515106201172, "num_tokens": 690654991.0, "step": 18187 }, { "epoch": 2.313700547004198, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.805319309234619, "learning_rate": 5e-05, "loss": 0.8045, "mean_token_accuracy": 0.8684380054473877, "num_tokens": 690692344.0, "step": 18188 }, { "epoch": 2.3138277572827883, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.349179267883301, "learning_rate": 5e-05, "loss": 0.689, "mean_token_accuracy": 0.8890236020088196, "num_tokens": 690726812.0, "step": 18189 }, { "epoch": 2.313954967561379, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.813648223876953, "learning_rate": 5e-05, "loss": 0.7633, "mean_token_accuracy": 0.8780084848403931, "num_tokens": 690768229.0, "step": 18190 }, { "epoch": 2.3140821778399694, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.305031776428223, "learning_rate": 5e-05, "loss": 0.7029, "mean_token_accuracy": 0.8832038044929504, "num_tokens": 690810614.0, "step": 18191 }, { "epoch": 2.31420938811856, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.874570846557617, "learning_rate": 5e-05, "loss": 0.7925, "mean_token_accuracy": 0.8695387840270996, "num_tokens": 690855422.0, "step": 18192 }, { "epoch": 2.3143365983971504, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.2074127197265625, "learning_rate": 5e-05, "loss": 0.6722, "mean_token_accuracy": 0.8883717656135559, "num_tokens": 690893751.0, "step": 18193 }, { "epoch": 2.314463808675741, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.972726821899414, "learning_rate": 5e-05, "loss": 0.7514, "mean_token_accuracy": 0.8830350041389465, "num_tokens": 690936734.0, "step": 18194 }, { "epoch": 2.3145910189543315, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.055616855621338, "learning_rate": 5e-05, "loss": 0.6832, "mean_token_accuracy": 0.8865114450454712, "num_tokens": 690975005.0, "step": 18195 }, { "epoch": 2.314718229232922, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.173783302307129, "learning_rate": 5e-05, "loss": 0.7549, "mean_token_accuracy": 0.8848543763160706, "num_tokens": 691011607.0, "step": 18196 }, { "epoch": 2.3148454395115126, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.856209754943848, "learning_rate": 5e-05, "loss": 0.6752, "mean_token_accuracy": 0.8859860301017761, "num_tokens": 691047635.0, "step": 18197 }, { "epoch": 2.314972649790103, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.421377182006836, "learning_rate": 5e-05, "loss": 0.8188, "mean_token_accuracy": 0.8758859634399414, "num_tokens": 691095290.0, "step": 18198 }, { "epoch": 2.3150998600686936, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.661056041717529, "learning_rate": 5e-05, "loss": 0.6779, "mean_token_accuracy": 0.8808977007865906, "num_tokens": 691140641.0, "step": 18199 }, { "epoch": 2.315227070347284, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.69690990447998, "learning_rate": 5e-05, "loss": 0.831, "mean_token_accuracy": 0.879734992980957, "num_tokens": 691181744.0, "step": 18200 }, { "epoch": 2.3153542806258747, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.55092716217041, "learning_rate": 5e-05, "loss": 0.6922, "mean_token_accuracy": 0.8822593688964844, "num_tokens": 691221279.0, "step": 18201 }, { "epoch": 2.315481490904465, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.853153228759766, "learning_rate": 5e-05, "loss": 0.8573, "mean_token_accuracy": 0.8839924335479736, "num_tokens": 691258519.0, "step": 18202 }, { "epoch": 2.3156087011830557, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.752252101898193, "learning_rate": 5e-05, "loss": 0.7226, "mean_token_accuracy": 0.8769707083702087, "num_tokens": 691297063.0, "step": 18203 }, { "epoch": 2.3157359114616463, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.53376579284668, "learning_rate": 5e-05, "loss": 0.8587, "mean_token_accuracy": 0.8811862468719482, "num_tokens": 691337146.0, "step": 18204 }, { "epoch": 2.315863121740237, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.374580383300781, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8882806897163391, "num_tokens": 691373475.0, "step": 18205 }, { "epoch": 2.3159903320188273, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.6139235496521, "learning_rate": 5e-05, "loss": 0.7777, "mean_token_accuracy": 0.877991795539856, "num_tokens": 691413286.0, "step": 18206 }, { "epoch": 2.316117542297418, "ewc_loss": 0.5, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.247160911560059, "learning_rate": 5e-05, "loss": 0.7982, "mean_token_accuracy": 0.8945424556732178, "num_tokens": 691444950.0, "step": 18207 }, { "epoch": 2.316244752576008, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.545547962188721, "learning_rate": 5e-05, "loss": 0.6901, "mean_token_accuracy": 0.8798019886016846, "num_tokens": 691486767.0, "step": 18208 }, { "epoch": 2.316371962854599, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.27883243560791, "learning_rate": 5e-05, "loss": 0.905, "mean_token_accuracy": 0.8770666122436523, "num_tokens": 691527149.0, "step": 18209 }, { "epoch": 2.316499173133189, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00024318695068359375, "grad_norm": 5.794928550720215, "learning_rate": 5e-05, "loss": 0.6674, "mean_token_accuracy": 0.8735268115997314, "num_tokens": 691563963.0, "step": 18210 }, { "epoch": 2.3166263834117795, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.429401397705078, "learning_rate": 5e-05, "loss": 1.0317, "mean_token_accuracy": 0.8751978874206543, "num_tokens": 691596423.0, "step": 18211 }, { "epoch": 2.31675359369037, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.282159805297852, "learning_rate": 5e-05, "loss": 0.7856, "mean_token_accuracy": 0.8779594302177429, "num_tokens": 691636928.0, "step": 18212 }, { "epoch": 2.3168808039689606, "ewc_loss": 0.84375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000774383544921875, "grad_norm": 10.866704940795898, "learning_rate": 5e-05, "loss": 1.1803, "mean_token_accuracy": 0.8811643123626709, "num_tokens": 691673427.0, "step": 18213 }, { "epoch": 2.317008014247551, "ewc_loss": 0.78125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 8.495320320129395, "learning_rate": 5e-05, "loss": 1.1251, "mean_token_accuracy": 0.8780403137207031, "num_tokens": 691710427.0, "step": 18214 }, { "epoch": 2.3171352245261416, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.180035591125488, "learning_rate": 5e-05, "loss": 0.9904, "mean_token_accuracy": 0.8714724779129028, "num_tokens": 691746809.0, "step": 18215 }, { "epoch": 2.317262434804732, "ewc_loss": 0.83203125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 9.613635063171387, "learning_rate": 5e-05, "loss": 1.1586, "mean_token_accuracy": 0.8842605948448181, "num_tokens": 691786037.0, "step": 18216 }, { "epoch": 2.3173896450833227, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.170350551605225, "learning_rate": 5e-05, "loss": 0.8024, "mean_token_accuracy": 0.8755394220352173, "num_tokens": 691822270.0, "step": 18217 }, { "epoch": 2.317516855361913, "ewc_loss": 0.87109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 11.186009407043457, "learning_rate": 5e-05, "loss": 1.1547, "mean_token_accuracy": 0.8995401859283447, "num_tokens": 691858538.0, "step": 18218 }, { "epoch": 2.3176440656405037, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 7.838601112365723, "learning_rate": 5e-05, "loss": 0.969, "mean_token_accuracy": 0.8833775520324707, "num_tokens": 691898810.0, "step": 18219 }, { "epoch": 2.3177712759190943, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 11.503673553466797, "learning_rate": 5e-05, "loss": 1.1465, "mean_token_accuracy": 0.8812808394432068, "num_tokens": 691933708.0, "step": 18220 }, { "epoch": 2.317898486197685, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 9.163168907165527, "learning_rate": 5e-05, "loss": 1.1767, "mean_token_accuracy": 0.8927023410797119, "num_tokens": 691968293.0, "step": 18221 }, { "epoch": 2.3180256964762753, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.685249328613281, "learning_rate": 5e-05, "loss": 1.0715, "mean_token_accuracy": 0.872249960899353, "num_tokens": 692004532.0, "step": 18222 }, { "epoch": 2.318152906754866, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 8.72494125366211, "learning_rate": 5e-05, "loss": 1.0329, "mean_token_accuracy": 0.8833354115486145, "num_tokens": 692042323.0, "step": 18223 }, { "epoch": 2.3182801170334564, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.22515869140625, "learning_rate": 5e-05, "loss": 1.0564, "mean_token_accuracy": 0.8795977830886841, "num_tokens": 692080970.0, "step": 18224 }, { "epoch": 2.318407327312047, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.102148056030273, "learning_rate": 5e-05, "loss": 0.9212, "mean_token_accuracy": 0.8857755064964294, "num_tokens": 692115027.0, "step": 18225 }, { "epoch": 2.3185345375906374, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.117622375488281, "learning_rate": 5e-05, "loss": 1.0547, "mean_token_accuracy": 0.879615306854248, "num_tokens": 692146602.0, "step": 18226 }, { "epoch": 2.318661747869228, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.836815357208252, "learning_rate": 5e-05, "loss": 0.9574, "mean_token_accuracy": 0.8657689690589905, "num_tokens": 692189721.0, "step": 18227 }, { "epoch": 2.3187889581478185, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.301498413085938, "learning_rate": 5e-05, "loss": 1.0765, "mean_token_accuracy": 0.8741006851196289, "num_tokens": 692233032.0, "step": 18228 }, { "epoch": 2.318916168426409, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.087657928466797, "learning_rate": 5e-05, "loss": 0.9753, "mean_token_accuracy": 0.8730154037475586, "num_tokens": 692269033.0, "step": 18229 }, { "epoch": 2.3190433787049995, "ewc_loss": 0.71875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.228958129882812, "learning_rate": 5e-05, "loss": 1.0687, "mean_token_accuracy": 0.8775748014450073, "num_tokens": 692307415.0, "step": 18230 }, { "epoch": 2.3191705889835896, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.401999473571777, "learning_rate": 5e-05, "loss": 0.9555, "mean_token_accuracy": 0.8884145021438599, "num_tokens": 692347003.0, "step": 18231 }, { "epoch": 2.3192977992621806, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.82218074798584, "learning_rate": 5e-05, "loss": 1.0259, "mean_token_accuracy": 0.8810977935791016, "num_tokens": 692385141.0, "step": 18232 }, { "epoch": 2.3194250095407707, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.422737121582031, "learning_rate": 5e-05, "loss": 0.9092, "mean_token_accuracy": 0.8943294286727905, "num_tokens": 692421731.0, "step": 18233 }, { "epoch": 2.319552219819361, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.370580673217773, "learning_rate": 5e-05, "loss": 1.0004, "mean_token_accuracy": 0.877619743347168, "num_tokens": 692466400.0, "step": 18234 }, { "epoch": 2.3196794300979517, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.20444107055664, "learning_rate": 5e-05, "loss": 0.8482, "mean_token_accuracy": 0.8965217471122742, "num_tokens": 692501201.0, "step": 18235 }, { "epoch": 2.3198066403765423, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.117429733276367, "learning_rate": 5e-05, "loss": 0.9186, "mean_token_accuracy": 0.8955577611923218, "num_tokens": 692541549.0, "step": 18236 }, { "epoch": 2.319933850655133, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.855963230133057, "learning_rate": 5e-05, "loss": 0.8518, "mean_token_accuracy": 0.8861833810806274, "num_tokens": 692582156.0, "step": 18237 }, { "epoch": 2.3200610609337233, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.200963973999023, "learning_rate": 5e-05, "loss": 0.9947, "mean_token_accuracy": 0.8662123680114746, "num_tokens": 692620522.0, "step": 18238 }, { "epoch": 2.320188271212314, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.633519172668457, "learning_rate": 5e-05, "loss": 0.8459, "mean_token_accuracy": 0.8761246204376221, "num_tokens": 692658372.0, "step": 18239 }, { "epoch": 2.3203154814909044, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.368953704833984, "learning_rate": 5e-05, "loss": 0.9584, "mean_token_accuracy": 0.8852233290672302, "num_tokens": 692687338.0, "step": 18240 }, { "epoch": 2.320442691769495, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.771865367889404, "learning_rate": 5e-05, "loss": 0.8616, "mean_token_accuracy": 0.8795304298400879, "num_tokens": 692724417.0, "step": 18241 }, { "epoch": 2.3205699020480854, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.169437408447266, "learning_rate": 5e-05, "loss": 0.9782, "mean_token_accuracy": 0.880577802658081, "num_tokens": 692767881.0, "step": 18242 }, { "epoch": 2.320697112326676, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.01525592803955, "learning_rate": 5e-05, "loss": 0.8886, "mean_token_accuracy": 0.8774648308753967, "num_tokens": 692803154.0, "step": 18243 }, { "epoch": 2.3208243226052665, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.695066452026367, "learning_rate": 5e-05, "loss": 0.8883, "mean_token_accuracy": 0.8957769274711609, "num_tokens": 692842435.0, "step": 18244 }, { "epoch": 2.320951532883857, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.955966472625732, "learning_rate": 5e-05, "loss": 0.8707, "mean_token_accuracy": 0.8751358985900879, "num_tokens": 692884491.0, "step": 18245 }, { "epoch": 2.3210787431624476, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.441329956054688, "learning_rate": 5e-05, "loss": 0.905, "mean_token_accuracy": 0.8759443759918213, "num_tokens": 692920825.0, "step": 18246 }, { "epoch": 2.321205953441038, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.682023525238037, "learning_rate": 5e-05, "loss": 0.7862, "mean_token_accuracy": 0.8844218254089355, "num_tokens": 692956413.0, "step": 18247 }, { "epoch": 2.3213331637196286, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.459877014160156, "learning_rate": 5e-05, "loss": 0.9187, "mean_token_accuracy": 0.8743699789047241, "num_tokens": 692995860.0, "step": 18248 }, { "epoch": 2.321460373998219, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.4005279541015625, "learning_rate": 5e-05, "loss": 0.7424, "mean_token_accuracy": 0.8967474699020386, "num_tokens": 693031077.0, "step": 18249 }, { "epoch": 2.3215875842768097, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.63353443145752, "learning_rate": 5e-05, "loss": 0.9356, "mean_token_accuracy": 0.8695818185806274, "num_tokens": 693067181.0, "step": 18250 }, { "epoch": 2.3217147945554, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.284317970275879, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8833184242248535, "num_tokens": 693103919.0, "step": 18251 }, { "epoch": 2.3218420048339907, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.687352180480957, "learning_rate": 5e-05, "loss": 0.9418, "mean_token_accuracy": 0.8659438490867615, "num_tokens": 693138471.0, "step": 18252 }, { "epoch": 2.3219692151125813, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.38615083694458, "learning_rate": 5e-05, "loss": 0.8009, "mean_token_accuracy": 0.8747445344924927, "num_tokens": 693174220.0, "step": 18253 }, { "epoch": 2.322096425391172, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.499999046325684, "learning_rate": 5e-05, "loss": 0.8811, "mean_token_accuracy": 0.8902369737625122, "num_tokens": 693217222.0, "step": 18254 }, { "epoch": 2.3222236356697623, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.546996593475342, "learning_rate": 5e-05, "loss": 0.7483, "mean_token_accuracy": 0.8958505988121033, "num_tokens": 693260052.0, "step": 18255 }, { "epoch": 2.3223508459483524, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.20776653289795, "learning_rate": 5e-05, "loss": 0.8744, "mean_token_accuracy": 0.8734670281410217, "num_tokens": 693293775.0, "step": 18256 }, { "epoch": 2.3224780562269434, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.620661735534668, "learning_rate": 5e-05, "loss": 0.7587, "mean_token_accuracy": 0.887494683265686, "num_tokens": 693335876.0, "step": 18257 }, { "epoch": 2.3226052665055335, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.002501487731934, "learning_rate": 5e-05, "loss": 0.8491, "mean_token_accuracy": 0.8720977306365967, "num_tokens": 693376886.0, "step": 18258 }, { "epoch": 2.322732476784124, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.5935444831848145, "learning_rate": 5e-05, "loss": 0.7959, "mean_token_accuracy": 0.8723387122154236, "num_tokens": 693416386.0, "step": 18259 }, { "epoch": 2.3228596870627145, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.911898612976074, "learning_rate": 5e-05, "loss": 0.7656, "mean_token_accuracy": 0.892648458480835, "num_tokens": 693454926.0, "step": 18260 }, { "epoch": 2.322986897341305, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.529139041900635, "learning_rate": 5e-05, "loss": 0.7404, "mean_token_accuracy": 0.8802409172058105, "num_tokens": 693492926.0, "step": 18261 }, { "epoch": 2.3231141076198956, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.880049705505371, "learning_rate": 5e-05, "loss": 0.7843, "mean_token_accuracy": 0.882382869720459, "num_tokens": 693534684.0, "step": 18262 }, { "epoch": 2.323241317898486, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.452881336212158, "learning_rate": 5e-05, "loss": 0.7544, "mean_token_accuracy": 0.8697702288627625, "num_tokens": 693570856.0, "step": 18263 }, { "epoch": 2.3233685281770766, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.885273456573486, "learning_rate": 5e-05, "loss": 0.7963, "mean_token_accuracy": 0.8723832964897156, "num_tokens": 693613286.0, "step": 18264 }, { "epoch": 2.323495738455667, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.3797526359558105, "learning_rate": 5e-05, "loss": 0.6929, "mean_token_accuracy": 0.8879063129425049, "num_tokens": 693648748.0, "step": 18265 }, { "epoch": 2.3236229487342577, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.924325466156006, "learning_rate": 5e-05, "loss": 0.7253, "mean_token_accuracy": 0.8956024646759033, "num_tokens": 693688997.0, "step": 18266 }, { "epoch": 2.323750159012848, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.306746959686279, "learning_rate": 5e-05, "loss": 0.7231, "mean_token_accuracy": 0.8772218227386475, "num_tokens": 693727653.0, "step": 18267 }, { "epoch": 2.3238773692914387, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.961946964263916, "learning_rate": 5e-05, "loss": 0.7734, "mean_token_accuracy": 0.8809922933578491, "num_tokens": 693771146.0, "step": 18268 }, { "epoch": 2.3240045795700293, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.265828609466553, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.8576347827911377, "num_tokens": 693809135.0, "step": 18269 }, { "epoch": 2.32413178984862, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.968720436096191, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.8891305923461914, "num_tokens": 693847127.0, "step": 18270 }, { "epoch": 2.3242590001272103, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.2516021728515625, "learning_rate": 5e-05, "loss": 0.694, "mean_token_accuracy": 0.884234607219696, "num_tokens": 693887186.0, "step": 18271 }, { "epoch": 2.324386210405801, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.953344345092773, "learning_rate": 5e-05, "loss": 0.7331, "mean_token_accuracy": 0.8904997706413269, "num_tokens": 693925943.0, "step": 18272 }, { "epoch": 2.3245134206843914, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.273037910461426, "learning_rate": 5e-05, "loss": 0.6492, "mean_token_accuracy": 0.8963993787765503, "num_tokens": 693958825.0, "step": 18273 }, { "epoch": 2.324640630962982, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.901695728302002, "learning_rate": 5e-05, "loss": 0.7797, "mean_token_accuracy": 0.8696345686912537, "num_tokens": 694003787.0, "step": 18274 }, { "epoch": 2.3247678412415724, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.311797142028809, "learning_rate": 5e-05, "loss": 0.701, "mean_token_accuracy": 0.8814648389816284, "num_tokens": 694047253.0, "step": 18275 }, { "epoch": 2.324895051520163, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.826882839202881, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.877007007598877, "num_tokens": 694085913.0, "step": 18276 }, { "epoch": 2.3250222617987535, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.388024806976318, "learning_rate": 5e-05, "loss": 0.7441, "mean_token_accuracy": 0.8672657012939453, "num_tokens": 694122848.0, "step": 18277 }, { "epoch": 2.325149472077344, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.691998481750488, "learning_rate": 5e-05, "loss": 0.7388, "mean_token_accuracy": 0.8755787014961243, "num_tokens": 694160545.0, "step": 18278 }, { "epoch": 2.3252766823559345, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.501676082611084, "learning_rate": 5e-05, "loss": 0.6993, "mean_token_accuracy": 0.8833515644073486, "num_tokens": 694194272.0, "step": 18279 }, { "epoch": 2.325403892634525, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.543581485748291, "learning_rate": 5e-05, "loss": 0.7002, "mean_token_accuracy": 0.8857230544090271, "num_tokens": 694227164.0, "step": 18280 }, { "epoch": 2.325531102913115, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.634337902069092, "learning_rate": 5e-05, "loss": 0.6729, "mean_token_accuracy": 0.8934589624404907, "num_tokens": 694266589.0, "step": 18281 }, { "epoch": 2.325658313191706, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.392200469970703, "learning_rate": 5e-05, "loss": 0.7085, "mean_token_accuracy": 0.8724673986434937, "num_tokens": 694301558.0, "step": 18282 }, { "epoch": 2.325785523470296, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.786055088043213, "learning_rate": 5e-05, "loss": 0.7547, "mean_token_accuracy": 0.8701894283294678, "num_tokens": 694343332.0, "step": 18283 }, { "epoch": 2.3259127337488867, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.184417724609375, "learning_rate": 5e-05, "loss": 0.652, "mean_token_accuracy": 0.8905452489852905, "num_tokens": 694383602.0, "step": 18284 }, { "epoch": 2.3260399440274773, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 8.032219886779785, "learning_rate": 5e-05, "loss": 0.7698, "mean_token_accuracy": 0.8708685636520386, "num_tokens": 694417664.0, "step": 18285 }, { "epoch": 2.326167154306068, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.881175518035889, "learning_rate": 5e-05, "loss": 0.7104, "mean_token_accuracy": 0.8642657995223999, "num_tokens": 694456176.0, "step": 18286 }, { "epoch": 2.3262943645846583, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.454075813293457, "learning_rate": 5e-05, "loss": 0.806, "mean_token_accuracy": 0.8754526376724243, "num_tokens": 694497114.0, "step": 18287 }, { "epoch": 2.326421574863249, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 6.3990797996521, "learning_rate": 5e-05, "loss": 0.6214, "mean_token_accuracy": 0.888038694858551, "num_tokens": 694540832.0, "step": 18288 }, { "epoch": 2.3265487851418394, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 9.169958114624023, "learning_rate": 5e-05, "loss": 0.8331, "mean_token_accuracy": 0.8866692185401917, "num_tokens": 694578673.0, "step": 18289 }, { "epoch": 2.32667599542043, "ewc_loss": 0.2890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00021839141845703125, "grad_norm": 5.8051438331604, "learning_rate": 5e-05, "loss": 0.6979, "mean_token_accuracy": 0.8640608787536621, "num_tokens": 694624736.0, "step": 18290 }, { "epoch": 2.3268032056990204, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 10.208648681640625, "learning_rate": 5e-05, "loss": 0.9344, "mean_token_accuracy": 0.8965035676956177, "num_tokens": 694659123.0, "step": 18291 }, { "epoch": 2.326930415977611, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 5.855135917663574, "learning_rate": 5e-05, "loss": 0.6629, "mean_token_accuracy": 0.8912751078605652, "num_tokens": 694695310.0, "step": 18292 }, { "epoch": 2.3270576262562015, "ewc_loss": 0.796875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.874755859375, "learning_rate": 5e-05, "loss": 1.149, "mean_token_accuracy": 0.8734106421470642, "num_tokens": 694732470.0, "step": 18293 }, { "epoch": 2.327184836534792, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 7.831173419952393, "learning_rate": 5e-05, "loss": 0.9833, "mean_token_accuracy": 0.8767328262329102, "num_tokens": 694763693.0, "step": 18294 }, { "epoch": 2.3273120468133826, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.541153907775879, "learning_rate": 5e-05, "loss": 0.9535, "mean_token_accuracy": 0.8887726068496704, "num_tokens": 694798127.0, "step": 18295 }, { "epoch": 2.327439257091973, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 9.450701713562012, "learning_rate": 5e-05, "loss": 1.1195, "mean_token_accuracy": 0.8845003843307495, "num_tokens": 694833462.0, "step": 18296 }, { "epoch": 2.3275664673705636, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.012032508850098, "learning_rate": 5e-05, "loss": 0.754, "mean_token_accuracy": 0.8754916191101074, "num_tokens": 694870257.0, "step": 18297 }, { "epoch": 2.327693677649154, "ewc_loss": 0.95703125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000885009765625, "grad_norm": 11.948763847351074, "learning_rate": 5e-05, "loss": 1.3062, "mean_token_accuracy": 0.8744093775749207, "num_tokens": 694910455.0, "step": 18298 }, { "epoch": 2.3278208879277447, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.054180145263672, "learning_rate": 5e-05, "loss": 0.9575, "mean_token_accuracy": 0.8839418292045593, "num_tokens": 694947040.0, "step": 18299 }, { "epoch": 2.327948098206335, "ewc_loss": 0.98828125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000919342041015625, "grad_norm": 13.071316719055176, "learning_rate": 5e-05, "loss": 1.3344, "mean_token_accuracy": 0.8776426911354065, "num_tokens": 694989043.0, "step": 18300 }, { "epoch": 2.3280753084849257, "ewc_loss": 1.0703125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00099945068359375, "grad_norm": 10.747184753417969, "learning_rate": 5e-05, "loss": 1.3758, "mean_token_accuracy": 0.8927342891693115, "num_tokens": 695027094.0, "step": 18301 }, { "epoch": 2.3282025187635162, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 12.117198944091797, "learning_rate": 5e-05, "loss": 1.1559, "mean_token_accuracy": 0.881592333316803, "num_tokens": 695056419.0, "step": 18302 }, { "epoch": 2.3283297290421068, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 9.604825019836426, "learning_rate": 5e-05, "loss": 1.1112, "mean_token_accuracy": 0.876484751701355, "num_tokens": 695101430.0, "step": 18303 }, { "epoch": 2.3284569393206973, "ewc_loss": 0.9765625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00090789794921875, "grad_norm": 12.686945915222168, "learning_rate": 5e-05, "loss": 1.3478, "mean_token_accuracy": 0.8705576658248901, "num_tokens": 695141446.0, "step": 18304 }, { "epoch": 2.328584149599288, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.569744110107422, "learning_rate": 5e-05, "loss": 1.024, "mean_token_accuracy": 0.8805375099182129, "num_tokens": 695179761.0, "step": 18305 }, { "epoch": 2.328711359877878, "ewc_loss": 0.97265625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000904083251953125, "grad_norm": 12.938895225524902, "learning_rate": 5e-05, "loss": 1.2607, "mean_token_accuracy": 0.8991050720214844, "num_tokens": 695217224.0, "step": 18306 }, { "epoch": 2.328838570156469, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 11.07461166381836, "learning_rate": 5e-05, "loss": 1.2059, "mean_token_accuracy": 0.8789913058280945, "num_tokens": 695253481.0, "step": 18307 }, { "epoch": 2.328965780435059, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 11.625956535339355, "learning_rate": 5e-05, "loss": 1.1117, "mean_token_accuracy": 0.8808600902557373, "num_tokens": 695287002.0, "step": 18308 }, { "epoch": 2.3290929907136495, "ewc_loss": 0.75, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.450643539428711, "learning_rate": 5e-05, "loss": 1.0651, "mean_token_accuracy": 0.8863729238510132, "num_tokens": 695328296.0, "step": 18309 }, { "epoch": 2.32922020099224, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 11.200262069702148, "learning_rate": 5e-05, "loss": 1.0814, "mean_token_accuracy": 0.8805763125419617, "num_tokens": 695364291.0, "step": 18310 }, { "epoch": 2.3293474112708306, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.588109970092773, "learning_rate": 5e-05, "loss": 1.0228, "mean_token_accuracy": 0.865254282951355, "num_tokens": 695407263.0, "step": 18311 }, { "epoch": 2.329474621549421, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 11.136892318725586, "learning_rate": 5e-05, "loss": 1.1107, "mean_token_accuracy": 0.864949643611908, "num_tokens": 695444243.0, "step": 18312 }, { "epoch": 2.3296018318280116, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.201131820678711, "learning_rate": 5e-05, "loss": 0.8978, "mean_token_accuracy": 0.8907628059387207, "num_tokens": 695475207.0, "step": 18313 }, { "epoch": 2.329729042106602, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 11.260027885437012, "learning_rate": 5e-05, "loss": 1.0402, "mean_token_accuracy": 0.8881935477256775, "num_tokens": 695515915.0, "step": 18314 }, { "epoch": 2.3298562523851927, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.165992736816406, "learning_rate": 5e-05, "loss": 0.9267, "mean_token_accuracy": 0.8832292556762695, "num_tokens": 695555168.0, "step": 18315 }, { "epoch": 2.329983462663783, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 11.486494064331055, "learning_rate": 5e-05, "loss": 1.0977, "mean_token_accuracy": 0.881322979927063, "num_tokens": 695591829.0, "step": 18316 }, { "epoch": 2.3301106729423737, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.48306941986084, "learning_rate": 5e-05, "loss": 0.9577, "mean_token_accuracy": 0.8770749568939209, "num_tokens": 695628518.0, "step": 18317 }, { "epoch": 2.3302378832209643, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 11.37447738647461, "learning_rate": 5e-05, "loss": 1.0721, "mean_token_accuracy": 0.8839774131774902, "num_tokens": 695661769.0, "step": 18318 }, { "epoch": 2.330365093499555, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.784157752990723, "learning_rate": 5e-05, "loss": 1.0183, "mean_token_accuracy": 0.8642905950546265, "num_tokens": 695699830.0, "step": 18319 }, { "epoch": 2.3304923037781453, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.792533874511719, "learning_rate": 5e-05, "loss": 0.9963, "mean_token_accuracy": 0.8854191303253174, "num_tokens": 695738871.0, "step": 18320 }, { "epoch": 2.330619514056736, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.586256980895996, "learning_rate": 5e-05, "loss": 0.9679, "mean_token_accuracy": 0.8746176958084106, "num_tokens": 695771868.0, "step": 18321 }, { "epoch": 2.3307467243353264, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 10.22592544555664, "learning_rate": 5e-05, "loss": 1.0078, "mean_token_accuracy": 0.8721604347229004, "num_tokens": 695808145.0, "step": 18322 }, { "epoch": 2.330873934613917, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 9.059374809265137, "learning_rate": 5e-05, "loss": 0.8399, "mean_token_accuracy": 0.8989676833152771, "num_tokens": 695837611.0, "step": 18323 }, { "epoch": 2.3310011448925074, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 10.010902404785156, "learning_rate": 5e-05, "loss": 0.933, "mean_token_accuracy": 0.893791139125824, "num_tokens": 695874640.0, "step": 18324 }, { "epoch": 2.331128355171098, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.49616527557373, "learning_rate": 5e-05, "loss": 0.8633, "mean_token_accuracy": 0.8697569370269775, "num_tokens": 695912580.0, "step": 18325 }, { "epoch": 2.3312555654496885, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 10.23912239074707, "learning_rate": 5e-05, "loss": 0.951, "mean_token_accuracy": 0.8903058767318726, "num_tokens": 695952061.0, "step": 18326 }, { "epoch": 2.331382775728279, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.365076065063477, "learning_rate": 5e-05, "loss": 0.7788, "mean_token_accuracy": 0.8914838433265686, "num_tokens": 695986389.0, "step": 18327 }, { "epoch": 2.3315099860068695, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.525771141052246, "learning_rate": 5e-05, "loss": 1.0022, "mean_token_accuracy": 0.8863401412963867, "num_tokens": 696026153.0, "step": 18328 }, { "epoch": 2.3316371962854596, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.811456680297852, "learning_rate": 5e-05, "loss": 0.8975, "mean_token_accuracy": 0.8664138317108154, "num_tokens": 696062693.0, "step": 18329 }, { "epoch": 2.3317644065640506, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 10.17219066619873, "learning_rate": 5e-05, "loss": 0.9536, "mean_token_accuracy": 0.8869990110397339, "num_tokens": 696101933.0, "step": 18330 }, { "epoch": 2.3318916168426407, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 9.048225402832031, "learning_rate": 5e-05, "loss": 0.8932, "mean_token_accuracy": 0.8760198950767517, "num_tokens": 696137607.0, "step": 18331 }, { "epoch": 2.332018827121231, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.410385131835938, "learning_rate": 5e-05, "loss": 0.9272, "mean_token_accuracy": 0.8806886672973633, "num_tokens": 696181277.0, "step": 18332 }, { "epoch": 2.3321460373998217, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.693153381347656, "learning_rate": 5e-05, "loss": 0.839, "mean_token_accuracy": 0.8798049688339233, "num_tokens": 696216592.0, "step": 18333 }, { "epoch": 2.3322732476784123, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.945257186889648, "learning_rate": 5e-05, "loss": 0.8193, "mean_token_accuracy": 0.8972296118736267, "num_tokens": 696250762.0, "step": 18334 }, { "epoch": 2.332400457957003, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.103528022766113, "learning_rate": 5e-05, "loss": 0.7605, "mean_token_accuracy": 0.8797032237052917, "num_tokens": 696289819.0, "step": 18335 }, { "epoch": 2.3325276682355933, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.927277565002441, "learning_rate": 5e-05, "loss": 0.9333, "mean_token_accuracy": 0.8540051579475403, "num_tokens": 696335551.0, "step": 18336 }, { "epoch": 2.332654878514184, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.642743110656738, "learning_rate": 5e-05, "loss": 0.7076, "mean_token_accuracy": 0.8864511847496033, "num_tokens": 696369973.0, "step": 18337 }, { "epoch": 2.3327820887927744, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 9.12023639678955, "learning_rate": 5e-05, "loss": 0.8849, "mean_token_accuracy": 0.875006914138794, "num_tokens": 696403718.0, "step": 18338 }, { "epoch": 2.332909299071365, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.501911640167236, "learning_rate": 5e-05, "loss": 0.7123, "mean_token_accuracy": 0.8764725923538208, "num_tokens": 696440769.0, "step": 18339 }, { "epoch": 2.3330365093499554, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.287534713745117, "learning_rate": 5e-05, "loss": 0.8889, "mean_token_accuracy": 0.8815485835075378, "num_tokens": 696482570.0, "step": 18340 }, { "epoch": 2.333163719628546, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.678058624267578, "learning_rate": 5e-05, "loss": 0.6955, "mean_token_accuracy": 0.8939999341964722, "num_tokens": 696519199.0, "step": 18341 }, { "epoch": 2.3332909299071365, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 9.152687072753906, "learning_rate": 5e-05, "loss": 0.9161, "mean_token_accuracy": 0.8688939213752747, "num_tokens": 696558460.0, "step": 18342 }, { "epoch": 2.333418140185727, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.002473831176758, "learning_rate": 5e-05, "loss": 0.8414, "mean_token_accuracy": 0.856149435043335, "num_tokens": 696600395.0, "step": 18343 }, { "epoch": 2.3335453504643175, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.566170692443848, "learning_rate": 5e-05, "loss": 0.8349, "mean_token_accuracy": 0.8783066272735596, "num_tokens": 696631478.0, "step": 18344 }, { "epoch": 2.333672560742908, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.119425773620605, "learning_rate": 5e-05, "loss": 0.7577, "mean_token_accuracy": 0.8855608701705933, "num_tokens": 696669634.0, "step": 18345 }, { "epoch": 2.3337997710214986, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.91131591796875, "learning_rate": 5e-05, "loss": 0.7738, "mean_token_accuracy": 0.875865638256073, "num_tokens": 696707642.0, "step": 18346 }, { "epoch": 2.333926981300089, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 8.093758583068848, "learning_rate": 5e-05, "loss": 0.7291, "mean_token_accuracy": 0.8887801170349121, "num_tokens": 696747426.0, "step": 18347 }, { "epoch": 2.3340541915786797, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.540406227111816, "learning_rate": 5e-05, "loss": 0.767, "mean_token_accuracy": 0.8659424185752869, "num_tokens": 696784114.0, "step": 18348 }, { "epoch": 2.33418140185727, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 8.154865264892578, "learning_rate": 5e-05, "loss": 0.722, "mean_token_accuracy": 0.8890377283096313, "num_tokens": 696822303.0, "step": 18349 }, { "epoch": 2.3343086121358607, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.250723838806152, "learning_rate": 5e-05, "loss": 0.7067, "mean_token_accuracy": 0.8776076436042786, "num_tokens": 696855460.0, "step": 18350 }, { "epoch": 2.3344358224144512, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.421283721923828, "learning_rate": 5e-05, "loss": 0.6989, "mean_token_accuracy": 0.9046868085861206, "num_tokens": 696891070.0, "step": 18351 }, { "epoch": 2.3345630326930418, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.959871292114258, "learning_rate": 5e-05, "loss": 0.6707, "mean_token_accuracy": 0.8813011646270752, "num_tokens": 696924679.0, "step": 18352 }, { "epoch": 2.3346902429716323, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.751778602600098, "learning_rate": 5e-05, "loss": 0.7861, "mean_token_accuracy": 0.886616051197052, "num_tokens": 696958894.0, "step": 18353 }, { "epoch": 2.3348174532502224, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.91114616394043, "learning_rate": 5e-05, "loss": 0.698, "mean_token_accuracy": 0.8754592537879944, "num_tokens": 696995371.0, "step": 18354 }, { "epoch": 2.3349446635288134, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.802035331726074, "learning_rate": 5e-05, "loss": 0.7902, "mean_token_accuracy": 0.8931760191917419, "num_tokens": 697034284.0, "step": 18355 }, { "epoch": 2.3350718738074034, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.21814489364624, "learning_rate": 5e-05, "loss": 0.6656, "mean_token_accuracy": 0.8927105665206909, "num_tokens": 697069805.0, "step": 18356 }, { "epoch": 2.335199084085994, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.450769424438477, "learning_rate": 5e-05, "loss": 0.7813, "mean_token_accuracy": 0.8847103118896484, "num_tokens": 697105326.0, "step": 18357 }, { "epoch": 2.3353262943645845, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.636889934539795, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.883073627948761, "num_tokens": 697144157.0, "step": 18358 }, { "epoch": 2.335453504643175, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.908881664276123, "learning_rate": 5e-05, "loss": 0.7238, "mean_token_accuracy": 0.8878119587898254, "num_tokens": 697178261.0, "step": 18359 }, { "epoch": 2.3355807149217656, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.8627190589904785, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8856632709503174, "num_tokens": 697224044.0, "step": 18360 }, { "epoch": 2.335707925200356, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.61225700378418, "learning_rate": 5e-05, "loss": 0.7047, "mean_token_accuracy": 0.8855363130569458, "num_tokens": 697262007.0, "step": 18361 }, { "epoch": 2.3358351354789466, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.873414039611816, "learning_rate": 5e-05, "loss": 0.7091, "mean_token_accuracy": 0.8868169188499451, "num_tokens": 697301689.0, "step": 18362 }, { "epoch": 2.335962345757537, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.624885559082031, "learning_rate": 5e-05, "loss": 0.7021, "mean_token_accuracy": 0.8848591446876526, "num_tokens": 697343509.0, "step": 18363 }, { "epoch": 2.3360895560361277, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.629319190979004, "learning_rate": 5e-05, "loss": 0.7169, "mean_token_accuracy": 0.8835509419441223, "num_tokens": 697380620.0, "step": 18364 }, { "epoch": 2.336216766314718, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.889490604400635, "learning_rate": 5e-05, "loss": 0.7384, "mean_token_accuracy": 0.8766748905181885, "num_tokens": 697422811.0, "step": 18365 }, { "epoch": 2.3363439765933087, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.239259719848633, "learning_rate": 5e-05, "loss": 0.7071, "mean_token_accuracy": 0.8766876459121704, "num_tokens": 697460527.0, "step": 18366 }, { "epoch": 2.3364711868718993, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.303458213806152, "learning_rate": 5e-05, "loss": 0.749, "mean_token_accuracy": 0.8855481147766113, "num_tokens": 697495417.0, "step": 18367 }, { "epoch": 2.33659839715049, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.814875602722168, "learning_rate": 5e-05, "loss": 0.6481, "mean_token_accuracy": 0.8866776823997498, "num_tokens": 697534745.0, "step": 18368 }, { "epoch": 2.3367256074290803, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.777628898620605, "learning_rate": 5e-05, "loss": 0.7923, "mean_token_accuracy": 0.8848655819892883, "num_tokens": 697572074.0, "step": 18369 }, { "epoch": 2.336852817707671, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.537804126739502, "learning_rate": 5e-05, "loss": 0.6539, "mean_token_accuracy": 0.8801498413085938, "num_tokens": 697606943.0, "step": 18370 }, { "epoch": 2.3369800279862614, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 9.111729621887207, "learning_rate": 5e-05, "loss": 0.8461, "mean_token_accuracy": 0.8813880681991577, "num_tokens": 697643634.0, "step": 18371 }, { "epoch": 2.337107238264852, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.679690361022949, "learning_rate": 5e-05, "loss": 0.6792, "mean_token_accuracy": 0.8799329996109009, "num_tokens": 697682429.0, "step": 18372 }, { "epoch": 2.3372344485434424, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.900659561157227, "learning_rate": 5e-05, "loss": 0.8444, "mean_token_accuracy": 0.8811737298965454, "num_tokens": 697718039.0, "step": 18373 }, { "epoch": 2.337361658822033, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.379561424255371, "learning_rate": 5e-05, "loss": 0.7115, "mean_token_accuracy": 0.8884257078170776, "num_tokens": 697756179.0, "step": 18374 }, { "epoch": 2.3374888691006235, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.947861194610596, "learning_rate": 5e-05, "loss": 0.7481, "mean_token_accuracy": 0.8848811388015747, "num_tokens": 697795327.0, "step": 18375 }, { "epoch": 2.337616079379214, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.185967445373535, "learning_rate": 5e-05, "loss": 0.777, "mean_token_accuracy": 0.8801376819610596, "num_tokens": 697830423.0, "step": 18376 }, { "epoch": 2.3377432896578045, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.003082752227783, "learning_rate": 5e-05, "loss": 0.7456, "mean_token_accuracy": 0.8620963096618652, "num_tokens": 697870890.0, "step": 18377 }, { "epoch": 2.337870499936395, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.820590019226074, "learning_rate": 5e-05, "loss": 0.8412, "mean_token_accuracy": 0.8719020485877991, "num_tokens": 697910880.0, "step": 18378 }, { "epoch": 2.337997710214985, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.538640022277832, "learning_rate": 5e-05, "loss": 0.6446, "mean_token_accuracy": 0.8871192932128906, "num_tokens": 697945509.0, "step": 18379 }, { "epoch": 2.338124920493576, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 9.203475952148438, "learning_rate": 5e-05, "loss": 0.9358, "mean_token_accuracy": 0.8654265999794006, "num_tokens": 697987374.0, "step": 18380 }, { "epoch": 2.338252130772166, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.533163070678711, "learning_rate": 5e-05, "loss": 0.6769, "mean_token_accuracy": 0.8799940347671509, "num_tokens": 698025130.0, "step": 18381 }, { "epoch": 2.3383793410507567, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.290704727172852, "learning_rate": 5e-05, "loss": 0.9071, "mean_token_accuracy": 0.8820476531982422, "num_tokens": 698057785.0, "step": 18382 }, { "epoch": 2.3385065513293473, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.998660564422607, "learning_rate": 5e-05, "loss": 0.7133, "mean_token_accuracy": 0.8890069723129272, "num_tokens": 698098952.0, "step": 18383 }, { "epoch": 2.338633761607938, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.876233100891113, "learning_rate": 5e-05, "loss": 0.8785, "mean_token_accuracy": 0.8892862796783447, "num_tokens": 698142280.0, "step": 18384 }, { "epoch": 2.3387609718865283, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.633715629577637, "learning_rate": 5e-05, "loss": 0.8002, "mean_token_accuracy": 0.8768843412399292, "num_tokens": 698183083.0, "step": 18385 }, { "epoch": 2.338888182165119, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.169856071472168, "learning_rate": 5e-05, "loss": 0.8586, "mean_token_accuracy": 0.878787636756897, "num_tokens": 698215446.0, "step": 18386 }, { "epoch": 2.3390153924437094, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.915132522583008, "learning_rate": 5e-05, "loss": 0.7982, "mean_token_accuracy": 0.8764398097991943, "num_tokens": 698257353.0, "step": 18387 }, { "epoch": 2.3391426027223, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.882655620574951, "learning_rate": 5e-05, "loss": 0.7954, "mean_token_accuracy": 0.8781411647796631, "num_tokens": 698292753.0, "step": 18388 }, { "epoch": 2.3392698130008904, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.63388204574585, "learning_rate": 5e-05, "loss": 0.736, "mean_token_accuracy": 0.8841654062271118, "num_tokens": 698327282.0, "step": 18389 }, { "epoch": 2.339397023279481, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.194314956665039, "learning_rate": 5e-05, "loss": 0.8259, "mean_token_accuracy": 0.880265474319458, "num_tokens": 698368006.0, "step": 18390 }, { "epoch": 2.3395242335580715, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.11024808883667, "learning_rate": 5e-05, "loss": 0.7771, "mean_token_accuracy": 0.8532779812812805, "num_tokens": 698406898.0, "step": 18391 }, { "epoch": 2.339651443836662, "ewc_loss": 0.5, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.631731033325195, "learning_rate": 5e-05, "loss": 0.8481, "mean_token_accuracy": 0.8770837783813477, "num_tokens": 698448159.0, "step": 18392 }, { "epoch": 2.3397786541152525, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.801762104034424, "learning_rate": 5e-05, "loss": 0.6709, "mean_token_accuracy": 0.8902157545089722, "num_tokens": 698489075.0, "step": 18393 }, { "epoch": 2.339905864393843, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.858797073364258, "learning_rate": 5e-05, "loss": 0.8733, "mean_token_accuracy": 0.8779325485229492, "num_tokens": 698527619.0, "step": 18394 }, { "epoch": 2.3400330746724336, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.894318580627441, "learning_rate": 5e-05, "loss": 0.7306, "mean_token_accuracy": 0.8719650506973267, "num_tokens": 698565250.0, "step": 18395 }, { "epoch": 2.340160284951024, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.722404479980469, "learning_rate": 5e-05, "loss": 0.8585, "mean_token_accuracy": 0.8835366368293762, "num_tokens": 698603207.0, "step": 18396 }, { "epoch": 2.3402874952296147, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.267536163330078, "learning_rate": 5e-05, "loss": 0.7006, "mean_token_accuracy": 0.8960468173027039, "num_tokens": 698636747.0, "step": 18397 }, { "epoch": 2.340414705508205, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.25959587097168, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.8887549638748169, "num_tokens": 698677885.0, "step": 18398 }, { "epoch": 2.3405419157867957, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.645380973815918, "learning_rate": 5e-05, "loss": 0.7671, "mean_token_accuracy": 0.8796889185905457, "num_tokens": 698721924.0, "step": 18399 }, { "epoch": 2.3406691260653862, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.795082092285156, "learning_rate": 5e-05, "loss": 0.8121, "mean_token_accuracy": 0.8700615763664246, "num_tokens": 698754847.0, "step": 18400 }, { "epoch": 2.3407963363439768, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.846597671508789, "learning_rate": 5e-05, "loss": 0.7462, "mean_token_accuracy": 0.890966534614563, "num_tokens": 698798567.0, "step": 18401 }, { "epoch": 2.3409235466225673, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.529648303985596, "learning_rate": 5e-05, "loss": 0.7131, "mean_token_accuracy": 0.8853923082351685, "num_tokens": 698837574.0, "step": 18402 }, { "epoch": 2.341050756901158, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.944929599761963, "learning_rate": 5e-05, "loss": 0.8001, "mean_token_accuracy": 0.8781634569168091, "num_tokens": 698867550.0, "step": 18403 }, { "epoch": 2.341177967179748, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.368436813354492, "learning_rate": 5e-05, "loss": 0.6853, "mean_token_accuracy": 0.8897475600242615, "num_tokens": 698898859.0, "step": 18404 }, { "epoch": 2.341305177458339, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.008780479431152, "learning_rate": 5e-05, "loss": 0.7516, "mean_token_accuracy": 0.8849374055862427, "num_tokens": 698933449.0, "step": 18405 }, { "epoch": 2.341432387736929, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.246729373931885, "learning_rate": 5e-05, "loss": 0.6734, "mean_token_accuracy": 0.8936654329299927, "num_tokens": 698969045.0, "step": 18406 }, { "epoch": 2.3415595980155195, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.121600151062012, "learning_rate": 5e-05, "loss": 0.7979, "mean_token_accuracy": 0.8679391741752625, "num_tokens": 699006201.0, "step": 18407 }, { "epoch": 2.34168680829411, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.114376544952393, "learning_rate": 5e-05, "loss": 0.6754, "mean_token_accuracy": 0.887187123298645, "num_tokens": 699044091.0, "step": 18408 }, { "epoch": 2.3418140185727006, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.239496231079102, "learning_rate": 5e-05, "loss": 0.7702, "mean_token_accuracy": 0.8811842203140259, "num_tokens": 699075519.0, "step": 18409 }, { "epoch": 2.341941228851291, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.982668399810791, "learning_rate": 5e-05, "loss": 0.6754, "mean_token_accuracy": 0.8888605833053589, "num_tokens": 699113129.0, "step": 18410 }, { "epoch": 2.3420684391298816, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.401263236999512, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8820251822471619, "num_tokens": 699147515.0, "step": 18411 }, { "epoch": 2.342195649408472, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.867011547088623, "learning_rate": 5e-05, "loss": 0.7011, "mean_token_accuracy": 0.8776826858520508, "num_tokens": 699190578.0, "step": 18412 }, { "epoch": 2.3423228596870627, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.526674270629883, "learning_rate": 5e-05, "loss": 0.8215, "mean_token_accuracy": 0.876860499382019, "num_tokens": 699230422.0, "step": 18413 }, { "epoch": 2.342450069965653, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.813440799713135, "learning_rate": 5e-05, "loss": 0.6835, "mean_token_accuracy": 0.8828554153442383, "num_tokens": 699268662.0, "step": 18414 }, { "epoch": 2.3425772802442437, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.612936973571777, "learning_rate": 5e-05, "loss": 0.837, "mean_token_accuracy": 0.8765125274658203, "num_tokens": 699312498.0, "step": 18415 }, { "epoch": 2.3427044905228342, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.861852169036865, "learning_rate": 5e-05, "loss": 0.6647, "mean_token_accuracy": 0.8900541067123413, "num_tokens": 699354143.0, "step": 18416 }, { "epoch": 2.3428317008014248, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.529630661010742, "learning_rate": 5e-05, "loss": 0.8075, "mean_token_accuracy": 0.8870258331298828, "num_tokens": 699392099.0, "step": 18417 }, { "epoch": 2.3429589110800153, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.051943778991699, "learning_rate": 5e-05, "loss": 0.6984, "mean_token_accuracy": 0.8844877481460571, "num_tokens": 699428449.0, "step": 18418 }, { "epoch": 2.343086121358606, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.280400276184082, "learning_rate": 5e-05, "loss": 0.8198, "mean_token_accuracy": 0.8788448572158813, "num_tokens": 699469359.0, "step": 18419 }, { "epoch": 2.3432133316371964, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.3285441398620605, "learning_rate": 5e-05, "loss": 0.7087, "mean_token_accuracy": 0.8883522748947144, "num_tokens": 699507481.0, "step": 18420 }, { "epoch": 2.343340541915787, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.94730806350708, "learning_rate": 5e-05, "loss": 0.8022, "mean_token_accuracy": 0.8722129464149475, "num_tokens": 699548335.0, "step": 18421 }, { "epoch": 2.3434677521943774, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.579890727996826, "learning_rate": 5e-05, "loss": 0.7622, "mean_token_accuracy": 0.8733007907867432, "num_tokens": 699587460.0, "step": 18422 }, { "epoch": 2.343594962472968, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.646696090698242, "learning_rate": 5e-05, "loss": 0.768, "mean_token_accuracy": 0.8723238706588745, "num_tokens": 699631128.0, "step": 18423 }, { "epoch": 2.3437221727515585, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.74693489074707, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8720158338546753, "num_tokens": 699672133.0, "step": 18424 }, { "epoch": 2.343849383030149, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.502465724945068, "learning_rate": 5e-05, "loss": 0.7307, "mean_token_accuracy": 0.8761743307113647, "num_tokens": 699705280.0, "step": 18425 }, { "epoch": 2.3439765933087395, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.755770683288574, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.8776068687438965, "num_tokens": 699740959.0, "step": 18426 }, { "epoch": 2.3441038035873296, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.501241683959961, "learning_rate": 5e-05, "loss": 0.7507, "mean_token_accuracy": 0.8665980696678162, "num_tokens": 699775546.0, "step": 18427 }, { "epoch": 2.3442310138659206, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.655663967132568, "learning_rate": 5e-05, "loss": 0.7283, "mean_token_accuracy": 0.8817457556724548, "num_tokens": 699809777.0, "step": 18428 }, { "epoch": 2.3443582241445107, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.589005470275879, "learning_rate": 5e-05, "loss": 0.6798, "mean_token_accuracy": 0.8928285837173462, "num_tokens": 699846140.0, "step": 18429 }, { "epoch": 2.344485434423101, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.51546573638916, "learning_rate": 5e-05, "loss": 0.6926, "mean_token_accuracy": 0.8818047046661377, "num_tokens": 699882769.0, "step": 18430 }, { "epoch": 2.3446126447016917, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.726791858673096, "learning_rate": 5e-05, "loss": 0.7274, "mean_token_accuracy": 0.8772662878036499, "num_tokens": 699923658.0, "step": 18431 }, { "epoch": 2.3447398549802823, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.326498031616211, "learning_rate": 5e-05, "loss": 0.6505, "mean_token_accuracy": 0.8915551900863647, "num_tokens": 699957457.0, "step": 18432 }, { "epoch": 2.344867065258873, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.9386210441589355, "learning_rate": 5e-05, "loss": 0.7078, "mean_token_accuracy": 0.8884192109107971, "num_tokens": 699994852.0, "step": 18433 }, { "epoch": 2.3449942755374633, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 7.048738956451416, "learning_rate": 5e-05, "loss": 0.6526, "mean_token_accuracy": 0.8877326250076294, "num_tokens": 700030983.0, "step": 18434 }, { "epoch": 2.345121485816054, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.316900253295898, "learning_rate": 5e-05, "loss": 0.7644, "mean_token_accuracy": 0.884976863861084, "num_tokens": 700067907.0, "step": 18435 }, { "epoch": 2.3452486960946444, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.609462738037109, "learning_rate": 5e-05, "loss": 0.6558, "mean_token_accuracy": 0.8817126750946045, "num_tokens": 700107989.0, "step": 18436 }, { "epoch": 2.345375906373235, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.956623077392578, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8936814069747925, "num_tokens": 700145870.0, "step": 18437 }, { "epoch": 2.3455031166518254, "ewc_loss": 0.296875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000225067138671875, "grad_norm": 5.998012065887451, "learning_rate": 5e-05, "loss": 0.608, "mean_token_accuracy": 0.8904317617416382, "num_tokens": 700186762.0, "step": 18438 }, { "epoch": 2.345630326930416, "ewc_loss": 0.609375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.986454963684082, "learning_rate": 5e-05, "loss": 0.9244, "mean_token_accuracy": 0.8877004384994507, "num_tokens": 700227307.0, "step": 18439 }, { "epoch": 2.3457575372090065, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 5.709210395812988, "learning_rate": 5e-05, "loss": 0.684, "mean_token_accuracy": 0.8771113753318787, "num_tokens": 700263660.0, "step": 18440 }, { "epoch": 2.345884747487597, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 10.964459419250488, "learning_rate": 5e-05, "loss": 1.0984, "mean_token_accuracy": 0.8945268988609314, "num_tokens": 700294835.0, "step": 18441 }, { "epoch": 2.3460119577661875, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.334428310394287, "learning_rate": 5e-05, "loss": 0.9045, "mean_token_accuracy": 0.8860143423080444, "num_tokens": 700335977.0, "step": 18442 }, { "epoch": 2.346139168044778, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.29602336883545, "learning_rate": 5e-05, "loss": 1.0499, "mean_token_accuracy": 0.8803085088729858, "num_tokens": 700369388.0, "step": 18443 }, { "epoch": 2.3462663783233686, "ewc_loss": 0.85546875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 9.233694076538086, "learning_rate": 5e-05, "loss": 1.1903, "mean_token_accuracy": 0.8794375658035278, "num_tokens": 700409575.0, "step": 18444 }, { "epoch": 2.346393588601959, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.7639851570129395, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.890363335609436, "num_tokens": 700449463.0, "step": 18445 }, { "epoch": 2.3465207988805497, "ewc_loss": 0.9140625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000843048095703125, "grad_norm": 11.22323226928711, "learning_rate": 5e-05, "loss": 1.236, "mean_token_accuracy": 0.8848422169685364, "num_tokens": 700483765.0, "step": 18446 }, { "epoch": 2.34664800915914, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.292693614959717, "learning_rate": 5e-05, "loss": 0.86, "mean_token_accuracy": 0.8846831321716309, "num_tokens": 700521953.0, "step": 18447 }, { "epoch": 2.3467752194377307, "ewc_loss": 1.078125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.001007080078125, "grad_norm": 13.200525283813477, "learning_rate": 5e-05, "loss": 1.4287, "mean_token_accuracy": 0.8728638887405396, "num_tokens": 700559662.0, "step": 18448 }, { "epoch": 2.3469024297163212, "ewc_loss": 1.09375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0010223388671875, "grad_norm": 10.717381477355957, "learning_rate": 5e-05, "loss": 1.4261, "mean_token_accuracy": 0.8831971883773804, "num_tokens": 700598880.0, "step": 18449 }, { "epoch": 2.3470296399949118, "ewc_loss": 0.78125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 11.876188278198242, "learning_rate": 5e-05, "loss": 1.0633, "mean_token_accuracy": 0.8970341682434082, "num_tokens": 700631712.0, "step": 18450 }, { "epoch": 2.3471568502735023, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000766754150390625, "grad_norm": 9.713419914245605, "learning_rate": 5e-05, "loss": 1.1999, "mean_token_accuracy": 0.8717966079711914, "num_tokens": 700665420.0, "step": 18451 }, { "epoch": 2.3472840605520924, "ewc_loss": 0.921875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 12.317780494689941, "learning_rate": 5e-05, "loss": 1.2477, "mean_token_accuracy": 0.8836420774459839, "num_tokens": 700704603.0, "step": 18452 }, { "epoch": 2.3474112708306833, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.870160102844238, "learning_rate": 5e-05, "loss": 0.9448, "mean_token_accuracy": 0.886268138885498, "num_tokens": 700738730.0, "step": 18453 }, { "epoch": 2.3475384811092734, "ewc_loss": 1.0390625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000965118408203125, "grad_norm": 13.094731330871582, "learning_rate": 5e-05, "loss": 1.3269, "mean_token_accuracy": 0.8972152471542358, "num_tokens": 700773848.0, "step": 18454 }, { "epoch": 2.347665691387864, "ewc_loss": 0.83203125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 10.559408187866211, "learning_rate": 5e-05, "loss": 1.1727, "mean_token_accuracy": 0.8808698654174805, "num_tokens": 700810352.0, "step": 18455 }, { "epoch": 2.3477929016664545, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0007781982421875, "grad_norm": 12.228227615356445, "learning_rate": 5e-05, "loss": 1.1964, "mean_token_accuracy": 0.8765606880187988, "num_tokens": 700846432.0, "step": 18456 }, { "epoch": 2.347920111945045, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00077056884765625, "grad_norm": 10.865999221801758, "learning_rate": 5e-05, "loss": 1.1538, "mean_token_accuracy": 0.8872897028923035, "num_tokens": 700883674.0, "step": 18457 }, { "epoch": 2.3480473222236355, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 11.36158561706543, "learning_rate": 5e-05, "loss": 1.1625, "mean_token_accuracy": 0.8611271381378174, "num_tokens": 700925096.0, "step": 18458 }, { "epoch": 2.348174532502226, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.005266189575195, "learning_rate": 5e-05, "loss": 0.9783, "mean_token_accuracy": 0.894197940826416, "num_tokens": 700957301.0, "step": 18459 }, { "epoch": 2.3483017427808166, "ewc_loss": 0.734375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 11.06527042388916, "learning_rate": 5e-05, "loss": 1.0398, "mean_token_accuracy": 0.892424464225769, "num_tokens": 700990636.0, "step": 18460 }, { "epoch": 2.348428953059407, "ewc_loss": 0.625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.402828216552734, "learning_rate": 5e-05, "loss": 0.921, "mean_token_accuracy": 0.8940303921699524, "num_tokens": 701026379.0, "step": 18461 }, { "epoch": 2.3485561633379977, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.940207481384277, "learning_rate": 5e-05, "loss": 1.0221, "mean_token_accuracy": 0.8814163208007812, "num_tokens": 701070377.0, "step": 18462 }, { "epoch": 2.348683373616588, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.019908905029297, "learning_rate": 5e-05, "loss": 0.9001, "mean_token_accuracy": 0.8890925645828247, "num_tokens": 701107691.0, "step": 18463 }, { "epoch": 2.3488105838951787, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 11.168106079101562, "learning_rate": 5e-05, "loss": 1.0492, "mean_token_accuracy": 0.8837059736251831, "num_tokens": 701147775.0, "step": 18464 }, { "epoch": 2.3489377941737692, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.029703140258789, "learning_rate": 5e-05, "loss": 0.8977, "mean_token_accuracy": 0.8914487361907959, "num_tokens": 701187732.0, "step": 18465 }, { "epoch": 2.3490650044523598, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 11.388568878173828, "learning_rate": 5e-05, "loss": 1.0394, "mean_token_accuracy": 0.893254280090332, "num_tokens": 701220916.0, "step": 18466 }, { "epoch": 2.3491922147309503, "ewc_loss": 0.625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.489875793457031, "learning_rate": 5e-05, "loss": 0.9504, "mean_token_accuracy": 0.8838090896606445, "num_tokens": 701258831.0, "step": 18467 }, { "epoch": 2.349319425009541, "ewc_loss": 0.71875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 11.051923751831055, "learning_rate": 5e-05, "loss": 1.0304, "mean_token_accuracy": 0.8895626068115234, "num_tokens": 701299310.0, "step": 18468 }, { "epoch": 2.3494466352881314, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.687943458557129, "learning_rate": 5e-05, "loss": 0.9797, "mean_token_accuracy": 0.8761466145515442, "num_tokens": 701331471.0, "step": 18469 }, { "epoch": 2.349573845566722, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 10.33035659790039, "learning_rate": 5e-05, "loss": 0.9841, "mean_token_accuracy": 0.8848687410354614, "num_tokens": 701371126.0, "step": 18470 }, { "epoch": 2.3497010558453124, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.286328315734863, "learning_rate": 5e-05, "loss": 0.8542, "mean_token_accuracy": 0.9057544469833374, "num_tokens": 701413612.0, "step": 18471 }, { "epoch": 2.349828266123903, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.85750961303711, "learning_rate": 5e-05, "loss": 1.0228, "mean_token_accuracy": 0.8598045110702515, "num_tokens": 701456085.0, "step": 18472 }, { "epoch": 2.3499554764024935, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.665473937988281, "learning_rate": 5e-05, "loss": 0.8427, "mean_token_accuracy": 0.8829118013381958, "num_tokens": 701494409.0, "step": 18473 }, { "epoch": 2.350082686681084, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.829160690307617, "learning_rate": 5e-05, "loss": 0.9463, "mean_token_accuracy": 0.8787282705307007, "num_tokens": 701533610.0, "step": 18474 }, { "epoch": 2.3502098969596745, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.184455871582031, "learning_rate": 5e-05, "loss": 0.8246, "mean_token_accuracy": 0.8692320585250854, "num_tokens": 701571324.0, "step": 18475 }, { "epoch": 2.350337107238265, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.183891296386719, "learning_rate": 5e-05, "loss": 1.0368, "mean_token_accuracy": 0.8704360723495483, "num_tokens": 701609077.0, "step": 18476 }, { "epoch": 2.350464317516855, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.251428604125977, "learning_rate": 5e-05, "loss": 0.7528, "mean_token_accuracy": 0.9014521837234497, "num_tokens": 701647281.0, "step": 18477 }, { "epoch": 2.350591527795446, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.370978355407715, "learning_rate": 5e-05, "loss": 0.9851, "mean_token_accuracy": 0.8926761746406555, "num_tokens": 701680039.0, "step": 18478 }, { "epoch": 2.350718738074036, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.816385269165039, "learning_rate": 5e-05, "loss": 0.8796, "mean_token_accuracy": 0.8791650533676147, "num_tokens": 701721679.0, "step": 18479 }, { "epoch": 2.3508459483526267, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.767518997192383, "learning_rate": 5e-05, "loss": 0.899, "mean_token_accuracy": 0.9001798629760742, "num_tokens": 701758692.0, "step": 18480 }, { "epoch": 2.3509731586312173, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.899930953979492, "learning_rate": 5e-05, "loss": 0.9282, "mean_token_accuracy": 0.866380512714386, "num_tokens": 701793038.0, "step": 18481 }, { "epoch": 2.351100368909808, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.96096420288086, "learning_rate": 5e-05, "loss": 0.8846, "mean_token_accuracy": 0.880893349647522, "num_tokens": 701831172.0, "step": 18482 }, { "epoch": 2.3512275791883983, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.438045501708984, "learning_rate": 5e-05, "loss": 0.7639, "mean_token_accuracy": 0.8945162296295166, "num_tokens": 701865644.0, "step": 18483 }, { "epoch": 2.351354789466989, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.625680923461914, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.8842607736587524, "num_tokens": 701903928.0, "step": 18484 }, { "epoch": 2.3514819997455794, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.890587329864502, "learning_rate": 5e-05, "loss": 0.7654, "mean_token_accuracy": 0.8704730272293091, "num_tokens": 701937440.0, "step": 18485 }, { "epoch": 2.35160921002417, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.606949806213379, "learning_rate": 5e-05, "loss": 0.8307, "mean_token_accuracy": 0.8825098276138306, "num_tokens": 701975936.0, "step": 18486 }, { "epoch": 2.3517364203027604, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.522664546966553, "learning_rate": 5e-05, "loss": 0.7048, "mean_token_accuracy": 0.8824270963668823, "num_tokens": 702009504.0, "step": 18487 }, { "epoch": 2.351863630581351, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.682318687438965, "learning_rate": 5e-05, "loss": 0.8219, "mean_token_accuracy": 0.8802404999732971, "num_tokens": 702046631.0, "step": 18488 }, { "epoch": 2.3519908408599415, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.336007118225098, "learning_rate": 5e-05, "loss": 0.7044, "mean_token_accuracy": 0.8845915198326111, "num_tokens": 702086915.0, "step": 18489 }, { "epoch": 2.352118051138532, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.823406219482422, "learning_rate": 5e-05, "loss": 0.8477, "mean_token_accuracy": 0.8757928609848022, "num_tokens": 702129214.0, "step": 18490 }, { "epoch": 2.3522452614171225, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.279329299926758, "learning_rate": 5e-05, "loss": 0.7149, "mean_token_accuracy": 0.8772765398025513, "num_tokens": 702167450.0, "step": 18491 }, { "epoch": 2.352372471695713, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.932480812072754, "learning_rate": 5e-05, "loss": 0.8514, "mean_token_accuracy": 0.8816825151443481, "num_tokens": 702205967.0, "step": 18492 }, { "epoch": 2.3524996819743036, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.456132888793945, "learning_rate": 5e-05, "loss": 0.7204, "mean_token_accuracy": 0.880294919013977, "num_tokens": 702239570.0, "step": 18493 }, { "epoch": 2.352626892252894, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.679342269897461, "learning_rate": 5e-05, "loss": 0.8457, "mean_token_accuracy": 0.8767969608306885, "num_tokens": 702279276.0, "step": 18494 }, { "epoch": 2.3527541025314846, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.78355073928833, "learning_rate": 5e-05, "loss": 0.7246, "mean_token_accuracy": 0.8885267972946167, "num_tokens": 702319695.0, "step": 18495 }, { "epoch": 2.352881312810075, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.049635887145996, "learning_rate": 5e-05, "loss": 0.7523, "mean_token_accuracy": 0.8876599073410034, "num_tokens": 702361933.0, "step": 18496 }, { "epoch": 2.3530085230886657, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.025915145874023, "learning_rate": 5e-05, "loss": 0.7493, "mean_token_accuracy": 0.8817037343978882, "num_tokens": 702398532.0, "step": 18497 }, { "epoch": 2.3531357333672562, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.486236572265625, "learning_rate": 5e-05, "loss": 0.7098, "mean_token_accuracy": 0.88499516248703, "num_tokens": 702437880.0, "step": 18498 }, { "epoch": 2.3532629436458468, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.290287017822266, "learning_rate": 5e-05, "loss": 0.7784, "mean_token_accuracy": 0.8753139972686768, "num_tokens": 702478908.0, "step": 18499 }, { "epoch": 2.353390153924437, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.044961929321289, "learning_rate": 5e-05, "loss": 0.704, "mean_token_accuracy": 0.8748337030410767, "num_tokens": 702523817.0, "step": 18500 }, { "epoch": 2.353517364203028, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.695149421691895, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.883041262626648, "num_tokens": 702569682.0, "step": 18501 }, { "epoch": 2.353644574481618, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.763710975646973, "learning_rate": 5e-05, "loss": 0.6934, "mean_token_accuracy": 0.8754174709320068, "num_tokens": 702604746.0, "step": 18502 }, { "epoch": 2.353771784760209, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 9.02226734161377, "learning_rate": 5e-05, "loss": 0.8683, "mean_token_accuracy": 0.8738592267036438, "num_tokens": 702637660.0, "step": 18503 }, { "epoch": 2.353898995038799, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.915234088897705, "learning_rate": 5e-05, "loss": 0.6332, "mean_token_accuracy": 0.8990188837051392, "num_tokens": 702676942.0, "step": 18504 }, { "epoch": 2.3540262053173895, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.752361297607422, "learning_rate": 5e-05, "loss": 0.8738, "mean_token_accuracy": 0.870872974395752, "num_tokens": 702718629.0, "step": 18505 }, { "epoch": 2.35415341559598, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.561371326446533, "learning_rate": 5e-05, "loss": 0.7295, "mean_token_accuracy": 0.8834403157234192, "num_tokens": 702757698.0, "step": 18506 }, { "epoch": 2.3542806258745705, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.907180309295654, "learning_rate": 5e-05, "loss": 0.7375, "mean_token_accuracy": 0.8891407251358032, "num_tokens": 702797337.0, "step": 18507 }, { "epoch": 2.354407836153161, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.160768508911133, "learning_rate": 5e-05, "loss": 0.7841, "mean_token_accuracy": 0.8741595149040222, "num_tokens": 702832002.0, "step": 18508 }, { "epoch": 2.3545350464317516, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.239757061004639, "learning_rate": 5e-05, "loss": 0.6886, "mean_token_accuracy": 0.8883371353149414, "num_tokens": 702874472.0, "step": 18509 }, { "epoch": 2.354662256710342, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.480817794799805, "learning_rate": 5e-05, "loss": 0.8177, "mean_token_accuracy": 0.8801703453063965, "num_tokens": 702909867.0, "step": 18510 }, { "epoch": 2.3547894669889327, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.0235772132873535, "learning_rate": 5e-05, "loss": 0.6526, "mean_token_accuracy": 0.890413761138916, "num_tokens": 702945817.0, "step": 18511 }, { "epoch": 2.354916677267523, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.500533103942871, "learning_rate": 5e-05, "loss": 0.7589, "mean_token_accuracy": 0.8972711563110352, "num_tokens": 702986373.0, "step": 18512 }, { "epoch": 2.3550438875461137, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.177095890045166, "learning_rate": 5e-05, "loss": 0.6979, "mean_token_accuracy": 0.8809322714805603, "num_tokens": 703026077.0, "step": 18513 }, { "epoch": 2.3551710978247042, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.182976722717285, "learning_rate": 5e-05, "loss": 0.7744, "mean_token_accuracy": 0.8794109225273132, "num_tokens": 703067042.0, "step": 18514 }, { "epoch": 2.3552983081032948, "ewc_loss": 0.375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.532392978668213, "learning_rate": 5e-05, "loss": 0.6893, "mean_token_accuracy": 0.8893588781356812, "num_tokens": 703105290.0, "step": 18515 }, { "epoch": 2.3554255183818853, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.707186698913574, "learning_rate": 5e-05, "loss": 0.7392, "mean_token_accuracy": 0.8787662982940674, "num_tokens": 703144446.0, "step": 18516 }, { "epoch": 2.355552728660476, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.899390697479248, "learning_rate": 5e-05, "loss": 0.7451, "mean_token_accuracy": 0.8777179718017578, "num_tokens": 703179490.0, "step": 18517 }, { "epoch": 2.3556799389390664, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.373329162597656, "learning_rate": 5e-05, "loss": 0.7076, "mean_token_accuracy": 0.8764989972114563, "num_tokens": 703214218.0, "step": 18518 }, { "epoch": 2.355807149217657, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 8.070730209350586, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.884900689125061, "num_tokens": 703252576.0, "step": 18519 }, { "epoch": 2.3559343594962474, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.293144702911377, "learning_rate": 5e-05, "loss": 0.6301, "mean_token_accuracy": 0.9038358330726624, "num_tokens": 703291129.0, "step": 18520 }, { "epoch": 2.356061569774838, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 8.032172203063965, "learning_rate": 5e-05, "loss": 0.7807, "mean_token_accuracy": 0.8696936368942261, "num_tokens": 703326864.0, "step": 18521 }, { "epoch": 2.3561887800534285, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.378518104553223, "learning_rate": 5e-05, "loss": 0.7009, "mean_token_accuracy": 0.8803844451904297, "num_tokens": 703369719.0, "step": 18522 }, { "epoch": 2.356315990332019, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.8601298332214355, "learning_rate": 5e-05, "loss": 0.7338, "mean_token_accuracy": 0.8822581768035889, "num_tokens": 703415304.0, "step": 18523 }, { "epoch": 2.3564432006106095, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.5457658767700195, "learning_rate": 5e-05, "loss": 0.7193, "mean_token_accuracy": 0.8761200904846191, "num_tokens": 703456273.0, "step": 18524 }, { "epoch": 2.3565704108891996, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.666892051696777, "learning_rate": 5e-05, "loss": 0.7311, "mean_token_accuracy": 0.8766360282897949, "num_tokens": 703491355.0, "step": 18525 }, { "epoch": 2.3566976211677906, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.655327320098877, "learning_rate": 5e-05, "loss": 0.7034, "mean_token_accuracy": 0.883830726146698, "num_tokens": 703532951.0, "step": 18526 }, { "epoch": 2.3568248314463807, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.602559566497803, "learning_rate": 5e-05, "loss": 0.7091, "mean_token_accuracy": 0.8823395371437073, "num_tokens": 703575306.0, "step": 18527 }, { "epoch": 2.356952041724971, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.617301940917969, "learning_rate": 5e-05, "loss": 0.6722, "mean_token_accuracy": 0.8886566758155823, "num_tokens": 703615198.0, "step": 18528 }, { "epoch": 2.3570792520035617, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.677183151245117, "learning_rate": 5e-05, "loss": 0.6683, "mean_token_accuracy": 0.892751932144165, "num_tokens": 703651780.0, "step": 18529 }, { "epoch": 2.3572064622821522, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.482684135437012, "learning_rate": 5e-05, "loss": 0.6994, "mean_token_accuracy": 0.8755580186843872, "num_tokens": 703685273.0, "step": 18530 }, { "epoch": 2.3573336725607428, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.876204490661621, "learning_rate": 5e-05, "loss": 0.6801, "mean_token_accuracy": 0.8922392725944519, "num_tokens": 703722185.0, "step": 18531 }, { "epoch": 2.3574608828393333, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 7.209587097167969, "learning_rate": 5e-05, "loss": 0.658, "mean_token_accuracy": 0.8873218894004822, "num_tokens": 703760229.0, "step": 18532 }, { "epoch": 2.357588093117924, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.199480056762695, "learning_rate": 5e-05, "loss": 0.7194, "mean_token_accuracy": 0.8906559348106384, "num_tokens": 703794286.0, "step": 18533 }, { "epoch": 2.3577153033965144, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.823712348937988, "learning_rate": 5e-05, "loss": 0.6384, "mean_token_accuracy": 0.8860242366790771, "num_tokens": 703827486.0, "step": 18534 }, { "epoch": 2.357842513675105, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.751269340515137, "learning_rate": 5e-05, "loss": 0.8359, "mean_token_accuracy": 0.8681855201721191, "num_tokens": 703861398.0, "step": 18535 }, { "epoch": 2.3579697239536954, "ewc_loss": 0.291015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002193450927734375, "grad_norm": 6.20022439956665, "learning_rate": 5e-05, "loss": 0.6341, "mean_token_accuracy": 0.8786848783493042, "num_tokens": 703901690.0, "step": 18536 }, { "epoch": 2.358096934232286, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.740181922912598, "learning_rate": 5e-05, "loss": 0.8822, "mean_token_accuracy": 0.8857232928276062, "num_tokens": 703943828.0, "step": 18537 }, { "epoch": 2.3582241445108765, "ewc_loss": 0.28125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000209808349609375, "grad_norm": 5.495335578918457, "learning_rate": 5e-05, "loss": 0.6335, "mean_token_accuracy": 0.8749655485153198, "num_tokens": 703984893.0, "step": 18538 }, { "epoch": 2.358351354789467, "ewc_loss": 0.765625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 11.194718360900879, "learning_rate": 5e-05, "loss": 1.1308, "mean_token_accuracy": 0.8703190088272095, "num_tokens": 704023633.0, "step": 18539 }, { "epoch": 2.3584785650680575, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.341461658477783, "learning_rate": 5e-05, "loss": 0.7783, "mean_token_accuracy": 0.8803503513336182, "num_tokens": 704058762.0, "step": 18540 }, { "epoch": 2.358605775346648, "ewc_loss": 0.98828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000919342041015625, "grad_norm": 12.116868019104004, "learning_rate": 5e-05, "loss": 1.322, "mean_token_accuracy": 0.8784416913986206, "num_tokens": 704099100.0, "step": 18541 }, { "epoch": 2.3587329856252386, "ewc_loss": 0.94140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0008697509765625, "grad_norm": 9.440244674682617, "learning_rate": 5e-05, "loss": 1.3319, "mean_token_accuracy": 0.8638670444488525, "num_tokens": 704131900.0, "step": 18542 }, { "epoch": 2.358860195903829, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.33492660522461, "learning_rate": 5e-05, "loss": 0.993, "mean_token_accuracy": 0.8887563943862915, "num_tokens": 704161940.0, "step": 18543 }, { "epoch": 2.3589874061824196, "ewc_loss": 0.95703125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000885009765625, "grad_norm": 10.143363952636719, "learning_rate": 5e-05, "loss": 1.2633, "mean_token_accuracy": 0.8891795873641968, "num_tokens": 704200698.0, "step": 18544 }, { "epoch": 2.35911461646101, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.378090858459473, "learning_rate": 5e-05, "loss": 0.8906, "mean_token_accuracy": 0.8807071447372437, "num_tokens": 704243317.0, "step": 18545 }, { "epoch": 2.3592418267396007, "ewc_loss": 0.93359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00086212158203125, "grad_norm": 11.469654083251953, "learning_rate": 5e-05, "loss": 1.2344, "mean_token_accuracy": 0.8922272324562073, "num_tokens": 704277258.0, "step": 18546 }, { "epoch": 2.3593690370181912, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.481905937194824, "learning_rate": 5e-05, "loss": 0.9792, "mean_token_accuracy": 0.8912626504898071, "num_tokens": 704315932.0, "step": 18547 }, { "epoch": 2.3594962472967818, "ewc_loss": 0.90234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 12.093843460083008, "learning_rate": 5e-05, "loss": 1.277, "mean_token_accuracy": 0.8694207668304443, "num_tokens": 704357355.0, "step": 18548 }, { "epoch": 2.3596234575753723, "ewc_loss": 0.90625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000835418701171875, "grad_norm": 9.73798942565918, "learning_rate": 5e-05, "loss": 1.26, "mean_token_accuracy": 0.8744845390319824, "num_tokens": 704392981.0, "step": 18549 }, { "epoch": 2.3597506678539624, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 11.368817329406738, "learning_rate": 5e-05, "loss": 1.0823, "mean_token_accuracy": 0.8824715614318848, "num_tokens": 704430530.0, "step": 18550 }, { "epoch": 2.3598778781325533, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 9.324392318725586, "learning_rate": 5e-05, "loss": 1.0924, "mean_token_accuracy": 0.890099823474884, "num_tokens": 704467444.0, "step": 18551 }, { "epoch": 2.3600050884111434, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.916818618774414, "learning_rate": 5e-05, "loss": 1.0319, "mean_token_accuracy": 0.8923865556716919, "num_tokens": 704504179.0, "step": 18552 }, { "epoch": 2.360132298689734, "ewc_loss": 0.640625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.56882095336914, "learning_rate": 5e-05, "loss": 0.9594, "mean_token_accuracy": 0.8846811056137085, "num_tokens": 704545346.0, "step": 18553 }, { "epoch": 2.3602595089683245, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.56855297088623, "learning_rate": 5e-05, "loss": 1.0794, "mean_token_accuracy": 0.8685958385467529, "num_tokens": 704583453.0, "step": 18554 }, { "epoch": 2.360386719246915, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.071121215820312, "learning_rate": 5e-05, "loss": 0.9072, "mean_token_accuracy": 0.8795945644378662, "num_tokens": 704617210.0, "step": 18555 }, { "epoch": 2.3605139295255055, "ewc_loss": 0.703125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.627979278564453, "learning_rate": 5e-05, "loss": 1.0164, "mean_token_accuracy": 0.8851444125175476, "num_tokens": 704651002.0, "step": 18556 }, { "epoch": 2.360641139804096, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.027154922485352, "learning_rate": 5e-05, "loss": 0.9264, "mean_token_accuracy": 0.8798438310623169, "num_tokens": 704694718.0, "step": 18557 }, { "epoch": 2.3607683500826866, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.79658031463623, "learning_rate": 5e-05, "loss": 1.0317, "mean_token_accuracy": 0.8914660215377808, "num_tokens": 704730266.0, "step": 18558 }, { "epoch": 2.360895560361277, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.473798751831055, "learning_rate": 5e-05, "loss": 0.9639, "mean_token_accuracy": 0.8861593008041382, "num_tokens": 704766422.0, "step": 18559 }, { "epoch": 2.3610227706398677, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.53372859954834, "learning_rate": 5e-05, "loss": 1.0393, "mean_token_accuracy": 0.8840380907058716, "num_tokens": 704802133.0, "step": 18560 }, { "epoch": 2.361149980918458, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.8302640914917, "learning_rate": 5e-05, "loss": 0.9726, "mean_token_accuracy": 0.8895694613456726, "num_tokens": 704832108.0, "step": 18561 }, { "epoch": 2.3612771911970487, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.851463317871094, "learning_rate": 5e-05, "loss": 1.0036, "mean_token_accuracy": 0.8758924603462219, "num_tokens": 704867056.0, "step": 18562 }, { "epoch": 2.3614044014756392, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.761612892150879, "learning_rate": 5e-05, "loss": 1.0012, "mean_token_accuracy": 0.8699561357498169, "num_tokens": 704905565.0, "step": 18563 }, { "epoch": 2.3615316117542298, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.132286071777344, "learning_rate": 5e-05, "loss": 0.9247, "mean_token_accuracy": 0.8827016949653625, "num_tokens": 704949690.0, "step": 18564 }, { "epoch": 2.3616588220328203, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.428354263305664, "learning_rate": 5e-05, "loss": 0.9525, "mean_token_accuracy": 0.8644989728927612, "num_tokens": 704985123.0, "step": 18565 }, { "epoch": 2.361786032311411, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.745787620544434, "learning_rate": 5e-05, "loss": 0.8743, "mean_token_accuracy": 0.8837103247642517, "num_tokens": 705020317.0, "step": 18566 }, { "epoch": 2.3619132425900013, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.911080360412598, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8888436555862427, "num_tokens": 705059025.0, "step": 18567 }, { "epoch": 2.362040452868592, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.720927238464355, "learning_rate": 5e-05, "loss": 0.8255, "mean_token_accuracy": 0.8969103693962097, "num_tokens": 705094157.0, "step": 18568 }, { "epoch": 2.3621676631471824, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.5244245529174805, "learning_rate": 5e-05, "loss": 0.762, "mean_token_accuracy": 0.8838996291160583, "num_tokens": 705133642.0, "step": 18569 }, { "epoch": 2.362294873425773, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.792747497558594, "learning_rate": 5e-05, "loss": 0.9024, "mean_token_accuracy": 0.8737462759017944, "num_tokens": 705165253.0, "step": 18570 }, { "epoch": 2.3624220837043635, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.3045878410339355, "learning_rate": 5e-05, "loss": 0.7111, "mean_token_accuracy": 0.892257809638977, "num_tokens": 705203847.0, "step": 18571 }, { "epoch": 2.362549293982954, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.041509628295898, "learning_rate": 5e-05, "loss": 0.9318, "mean_token_accuracy": 0.8793581128120422, "num_tokens": 705241984.0, "step": 18572 }, { "epoch": 2.3626765042615445, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.151260852813721, "learning_rate": 5e-05, "loss": 0.7635, "mean_token_accuracy": 0.8728700280189514, "num_tokens": 705278462.0, "step": 18573 }, { "epoch": 2.362803714540135, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.418245315551758, "learning_rate": 5e-05, "loss": 0.9456, "mean_token_accuracy": 0.8884158134460449, "num_tokens": 705313004.0, "step": 18574 }, { "epoch": 2.362930924818725, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.413751125335693, "learning_rate": 5e-05, "loss": 0.7485, "mean_token_accuracy": 0.8924617171287537, "num_tokens": 705350225.0, "step": 18575 }, { "epoch": 2.363058135097316, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.333063125610352, "learning_rate": 5e-05, "loss": 0.9532, "mean_token_accuracy": 0.8890026807785034, "num_tokens": 705380213.0, "step": 18576 }, { "epoch": 2.363185345375906, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.045936584472656, "learning_rate": 5e-05, "loss": 0.8862, "mean_token_accuracy": 0.8739850521087646, "num_tokens": 705419059.0, "step": 18577 }, { "epoch": 2.3633125556544967, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.483854293823242, "learning_rate": 5e-05, "loss": 0.8996, "mean_token_accuracy": 0.8780027627944946, "num_tokens": 705459506.0, "step": 18578 }, { "epoch": 2.3634397659330872, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.303938865661621, "learning_rate": 5e-05, "loss": 0.871, "mean_token_accuracy": 0.8818843960762024, "num_tokens": 705495719.0, "step": 18579 }, { "epoch": 2.3635669762116778, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.808343887329102, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.8707310557365417, "num_tokens": 705536040.0, "step": 18580 }, { "epoch": 2.3636941864902683, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.165238380432129, "learning_rate": 5e-05, "loss": 0.8105, "mean_token_accuracy": 0.883083164691925, "num_tokens": 705571541.0, "step": 18581 }, { "epoch": 2.363821396768859, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.730833530426025, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8866553902626038, "num_tokens": 705610048.0, "step": 18582 }, { "epoch": 2.3639486070474494, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.875971794128418, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.8865900039672852, "num_tokens": 705645098.0, "step": 18583 }, { "epoch": 2.36407581732604, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.867019176483154, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8828954696655273, "num_tokens": 705680962.0, "step": 18584 }, { "epoch": 2.3642030276046304, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.571689605712891, "learning_rate": 5e-05, "loss": 0.7375, "mean_token_accuracy": 0.8784133195877075, "num_tokens": 705716983.0, "step": 18585 }, { "epoch": 2.364330237883221, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.101237297058105, "learning_rate": 5e-05, "loss": 0.7692, "mean_token_accuracy": 0.8837273120880127, "num_tokens": 705749735.0, "step": 18586 }, { "epoch": 2.3644574481618115, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.242384910583496, "learning_rate": 5e-05, "loss": 0.7289, "mean_token_accuracy": 0.8744438290596008, "num_tokens": 705788747.0, "step": 18587 }, { "epoch": 2.364584658440402, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.446348190307617, "learning_rate": 5e-05, "loss": 0.811, "mean_token_accuracy": 0.877598762512207, "num_tokens": 705827475.0, "step": 18588 }, { "epoch": 2.3647118687189925, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.927888870239258, "learning_rate": 5e-05, "loss": 0.6498, "mean_token_accuracy": 0.8979167938232422, "num_tokens": 705864940.0, "step": 18589 }, { "epoch": 2.364839078997583, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.821081161499023, "learning_rate": 5e-05, "loss": 0.7986, "mean_token_accuracy": 0.8974400758743286, "num_tokens": 705898274.0, "step": 18590 }, { "epoch": 2.3649662892761736, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.758983612060547, "learning_rate": 5e-05, "loss": 0.6909, "mean_token_accuracy": 0.8799210786819458, "num_tokens": 705936619.0, "step": 18591 }, { "epoch": 2.365093499554764, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.01768970489502, "learning_rate": 5e-05, "loss": 0.8241, "mean_token_accuracy": 0.905971884727478, "num_tokens": 705975208.0, "step": 18592 }, { "epoch": 2.3652207098333546, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.951488018035889, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8734384775161743, "num_tokens": 706019948.0, "step": 18593 }, { "epoch": 2.365347920111945, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.795976638793945, "learning_rate": 5e-05, "loss": 0.8756, "mean_token_accuracy": 0.884297251701355, "num_tokens": 706056862.0, "step": 18594 }, { "epoch": 2.3654751303905357, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.4033522605896, "learning_rate": 5e-05, "loss": 0.7808, "mean_token_accuracy": 0.8781750202178955, "num_tokens": 706092591.0, "step": 18595 }, { "epoch": 2.3656023406691262, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.284782409667969, "learning_rate": 5e-05, "loss": 0.8618, "mean_token_accuracy": 0.8788958191871643, "num_tokens": 706132572.0, "step": 18596 }, { "epoch": 2.3657295509477168, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.774107456207275, "learning_rate": 5e-05, "loss": 0.7754, "mean_token_accuracy": 0.8834459781646729, "num_tokens": 706168264.0, "step": 18597 }, { "epoch": 2.365856761226307, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.800015449523926, "learning_rate": 5e-05, "loss": 0.7897, "mean_token_accuracy": 0.8818665742874146, "num_tokens": 706206684.0, "step": 18598 }, { "epoch": 2.365983971504898, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.932617664337158, "learning_rate": 5e-05, "loss": 0.7321, "mean_token_accuracy": 0.896373450756073, "num_tokens": 706247547.0, "step": 18599 }, { "epoch": 2.366111181783488, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.550168037414551, "learning_rate": 5e-05, "loss": 0.7083, "mean_token_accuracy": 0.8936781883239746, "num_tokens": 706286286.0, "step": 18600 }, { "epoch": 2.3662383920620784, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.019229888916016, "learning_rate": 5e-05, "loss": 0.7904, "mean_token_accuracy": 0.8767014145851135, "num_tokens": 706329990.0, "step": 18601 }, { "epoch": 2.366365602340669, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.358921527862549, "learning_rate": 5e-05, "loss": 0.7258, "mean_token_accuracy": 0.8778626322746277, "num_tokens": 706370423.0, "step": 18602 }, { "epoch": 2.3664928126192595, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.16193962097168, "learning_rate": 5e-05, "loss": 0.7799, "mean_token_accuracy": 0.8841769695281982, "num_tokens": 706414380.0, "step": 18603 }, { "epoch": 2.36662002289785, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.121483325958252, "learning_rate": 5e-05, "loss": 0.7094, "mean_token_accuracy": 0.8813735246658325, "num_tokens": 706455403.0, "step": 18604 }, { "epoch": 2.3667472331764405, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.459293365478516, "learning_rate": 5e-05, "loss": 0.7469, "mean_token_accuracy": 0.9005577564239502, "num_tokens": 706490114.0, "step": 18605 }, { "epoch": 2.366874443455031, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.795720100402832, "learning_rate": 5e-05, "loss": 0.6428, "mean_token_accuracy": 0.8950555324554443, "num_tokens": 706523926.0, "step": 18606 }, { "epoch": 2.3670016537336216, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.875654220581055, "learning_rate": 5e-05, "loss": 0.7945, "mean_token_accuracy": 0.896704375743866, "num_tokens": 706562626.0, "step": 18607 }, { "epoch": 2.367128864012212, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.532365322113037, "learning_rate": 5e-05, "loss": 0.6361, "mean_token_accuracy": 0.895828366279602, "num_tokens": 706598735.0, "step": 18608 }, { "epoch": 2.3672560742908026, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.222846031188965, "learning_rate": 5e-05, "loss": 0.9356, "mean_token_accuracy": 0.8692804574966431, "num_tokens": 706638399.0, "step": 18609 }, { "epoch": 2.367383284569393, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.653269290924072, "learning_rate": 5e-05, "loss": 0.6988, "mean_token_accuracy": 0.8807231187820435, "num_tokens": 706680274.0, "step": 18610 }, { "epoch": 2.3675104948479837, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.128535270690918, "learning_rate": 5e-05, "loss": 0.9666, "mean_token_accuracy": 0.8668493628501892, "num_tokens": 706713918.0, "step": 18611 }, { "epoch": 2.3676377051265742, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.285283088684082, "learning_rate": 5e-05, "loss": 0.7836, "mean_token_accuracy": 0.8803452253341675, "num_tokens": 706751758.0, "step": 18612 }, { "epoch": 2.3677649154051648, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.385612487792969, "learning_rate": 5e-05, "loss": 0.8463, "mean_token_accuracy": 0.8905198574066162, "num_tokens": 706788243.0, "step": 18613 }, { "epoch": 2.3678921256837553, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.963798999786377, "learning_rate": 5e-05, "loss": 0.8266, "mean_token_accuracy": 0.8907467722892761, "num_tokens": 706821960.0, "step": 18614 }, { "epoch": 2.368019335962346, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.553860664367676, "learning_rate": 5e-05, "loss": 0.761, "mean_token_accuracy": 0.8845810294151306, "num_tokens": 706858768.0, "step": 18615 }, { "epoch": 2.3681465462409363, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.278624534606934, "learning_rate": 5e-05, "loss": 0.7863, "mean_token_accuracy": 0.8966039419174194, "num_tokens": 706898889.0, "step": 18616 }, { "epoch": 2.368273756519527, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.343352317810059, "learning_rate": 5e-05, "loss": 0.7377, "mean_token_accuracy": 0.8804123401641846, "num_tokens": 706934245.0, "step": 18617 }, { "epoch": 2.3684009667981174, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.058454513549805, "learning_rate": 5e-05, "loss": 0.8056, "mean_token_accuracy": 0.8766975402832031, "num_tokens": 706982092.0, "step": 18618 }, { "epoch": 2.368528177076708, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.63918924331665, "learning_rate": 5e-05, "loss": 0.737, "mean_token_accuracy": 0.8868712186813354, "num_tokens": 707014748.0, "step": 18619 }, { "epoch": 2.3686553873552985, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.559981346130371, "learning_rate": 5e-05, "loss": 0.7233, "mean_token_accuracy": 0.8862963318824768, "num_tokens": 707048363.0, "step": 18620 }, { "epoch": 2.368782597633889, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.084195137023926, "learning_rate": 5e-05, "loss": 0.8036, "mean_token_accuracy": 0.8757633566856384, "num_tokens": 707084089.0, "step": 18621 }, { "epoch": 2.3689098079124795, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.050707817077637, "learning_rate": 5e-05, "loss": 0.7272, "mean_token_accuracy": 0.8712538480758667, "num_tokens": 707127334.0, "step": 18622 }, { "epoch": 2.3690370181910696, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.570270538330078, "learning_rate": 5e-05, "loss": 0.8158, "mean_token_accuracy": 0.884816586971283, "num_tokens": 707166485.0, "step": 18623 }, { "epoch": 2.3691642284696606, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.634461402893066, "learning_rate": 5e-05, "loss": 0.6652, "mean_token_accuracy": 0.8870949745178223, "num_tokens": 707202031.0, "step": 18624 }, { "epoch": 2.3692914387482507, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 9.099075317382812, "learning_rate": 5e-05, "loss": 0.8578, "mean_token_accuracy": 0.8846514225006104, "num_tokens": 707240649.0, "step": 18625 }, { "epoch": 2.369418649026841, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.367440223693848, "learning_rate": 5e-05, "loss": 0.681, "mean_token_accuracy": 0.8820924162864685, "num_tokens": 707275996.0, "step": 18626 }, { "epoch": 2.3695458593054317, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.53408145904541, "learning_rate": 5e-05, "loss": 0.9968, "mean_token_accuracy": 0.867588460445404, "num_tokens": 707312776.0, "step": 18627 }, { "epoch": 2.3696730695840222, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.6416707038879395, "learning_rate": 5e-05, "loss": 0.7499, "mean_token_accuracy": 0.8756901025772095, "num_tokens": 707349519.0, "step": 18628 }, { "epoch": 2.3698002798626128, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000713348388671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.33918285369873, "learning_rate": 5e-05, "loss": 0.9665, "mean_token_accuracy": 0.878331184387207, "num_tokens": 707389180.0, "step": 18629 }, { "epoch": 2.3699274901412033, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.529921531677246, "learning_rate": 5e-05, "loss": 0.8475, "mean_token_accuracy": 0.882766604423523, "num_tokens": 707428603.0, "step": 18630 }, { "epoch": 2.370054700419794, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.33898639678955, "learning_rate": 5e-05, "loss": 0.9063, "mean_token_accuracy": 0.8704081773757935, "num_tokens": 707470297.0, "step": 18631 }, { "epoch": 2.3701819106983844, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.294931411743164, "learning_rate": 5e-05, "loss": 0.8976, "mean_token_accuracy": 0.8825809359550476, "num_tokens": 707507353.0, "step": 18632 }, { "epoch": 2.370309120976975, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.384757041931152, "learning_rate": 5e-05, "loss": 0.7583, "mean_token_accuracy": 0.8877427577972412, "num_tokens": 707541795.0, "step": 18633 }, { "epoch": 2.3704363312555654, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.535858154296875, "learning_rate": 5e-05, "loss": 0.866, "mean_token_accuracy": 0.8809102177619934, "num_tokens": 707577333.0, "step": 18634 }, { "epoch": 2.370563541534156, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.250048637390137, "learning_rate": 5e-05, "loss": 0.7863, "mean_token_accuracy": 0.8702946901321411, "num_tokens": 707611306.0, "step": 18635 }, { "epoch": 2.3706907518127465, "ewc_loss": 0.5, "ewc_loss_diag": 0.0007171630859375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.263246536254883, "learning_rate": 5e-05, "loss": 0.8499, "mean_token_accuracy": 0.8748959898948669, "num_tokens": 707647698.0, "step": 18636 }, { "epoch": 2.370817962091337, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.442610740661621, "learning_rate": 5e-05, "loss": 0.7392, "mean_token_accuracy": 0.8866847157478333, "num_tokens": 707683934.0, "step": 18637 }, { "epoch": 2.3709451723699275, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.048379898071289, "learning_rate": 5e-05, "loss": 0.8179, "mean_token_accuracy": 0.8766286373138428, "num_tokens": 707723216.0, "step": 18638 }, { "epoch": 2.371072382648518, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.390112400054932, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8858870267868042, "num_tokens": 707761194.0, "step": 18639 }, { "epoch": 2.3711995929271086, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.195984840393066, "learning_rate": 5e-05, "loss": 0.8552, "mean_token_accuracy": 0.8662292957305908, "num_tokens": 707807327.0, "step": 18640 }, { "epoch": 2.371326803205699, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.07452392578125, "learning_rate": 5e-05, "loss": 0.7149, "mean_token_accuracy": 0.8834161758422852, "num_tokens": 707846394.0, "step": 18641 }, { "epoch": 2.3714540134842896, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.591279983520508, "learning_rate": 5e-05, "loss": 0.8476, "mean_token_accuracy": 0.8778839111328125, "num_tokens": 707889099.0, "step": 18642 }, { "epoch": 2.37158122376288, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.723305702209473, "learning_rate": 5e-05, "loss": 0.6619, "mean_token_accuracy": 0.8906525373458862, "num_tokens": 707924729.0, "step": 18643 }, { "epoch": 2.3717084340414707, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 9.006231307983398, "learning_rate": 5e-05, "loss": 0.947, "mean_token_accuracy": 0.8599143028259277, "num_tokens": 707965096.0, "step": 18644 }, { "epoch": 2.371835644320061, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.583953857421875, "learning_rate": 5e-05, "loss": 0.6906, "mean_token_accuracy": 0.8844974040985107, "num_tokens": 708002277.0, "step": 18645 }, { "epoch": 2.3719628545986517, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.211791038513184, "learning_rate": 5e-05, "loss": 0.95, "mean_token_accuracy": 0.8778012990951538, "num_tokens": 708038978.0, "step": 18646 }, { "epoch": 2.3720900648772423, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.885693550109863, "learning_rate": 5e-05, "loss": 0.7175, "mean_token_accuracy": 0.8963035345077515, "num_tokens": 708076022.0, "step": 18647 }, { "epoch": 2.3722172751558324, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.921666145324707, "learning_rate": 5e-05, "loss": 0.9506, "mean_token_accuracy": 0.8717399835586548, "num_tokens": 708119606.0, "step": 18648 }, { "epoch": 2.3723444854344233, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.536919116973877, "learning_rate": 5e-05, "loss": 0.8182, "mean_token_accuracy": 0.8922704458236694, "num_tokens": 708159570.0, "step": 18649 }, { "epoch": 2.3724716957130134, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.122749328613281, "learning_rate": 5e-05, "loss": 0.8989, "mean_token_accuracy": 0.8689368963241577, "num_tokens": 708201092.0, "step": 18650 }, { "epoch": 2.372598905991604, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.08161449432373, "learning_rate": 5e-05, "loss": 0.8581, "mean_token_accuracy": 0.882089376449585, "num_tokens": 708235808.0, "step": 18651 }, { "epoch": 2.3727261162701945, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.462589740753174, "learning_rate": 5e-05, "loss": 0.726, "mean_token_accuracy": 0.8994677066802979, "num_tokens": 708270863.0, "step": 18652 }, { "epoch": 2.372853326548785, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.238459587097168, "learning_rate": 5e-05, "loss": 0.8906, "mean_token_accuracy": 0.8686497211456299, "num_tokens": 708309653.0, "step": 18653 }, { "epoch": 2.3729805368273755, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.3228440284729, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.8841887712478638, "num_tokens": 708346242.0, "step": 18654 }, { "epoch": 2.373107747105966, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.099021911621094, "learning_rate": 5e-05, "loss": 0.828, "mean_token_accuracy": 0.8757805824279785, "num_tokens": 708383391.0, "step": 18655 }, { "epoch": 2.3732349573845566, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.453545570373535, "learning_rate": 5e-05, "loss": 0.7309, "mean_token_accuracy": 0.8880810737609863, "num_tokens": 708417094.0, "step": 18656 }, { "epoch": 2.373362167663147, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.857735633850098, "learning_rate": 5e-05, "loss": 0.7749, "mean_token_accuracy": 0.8794905543327332, "num_tokens": 708456648.0, "step": 18657 }, { "epoch": 2.3734893779417376, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.636839389801025, "learning_rate": 5e-05, "loss": 0.7773, "mean_token_accuracy": 0.8736732006072998, "num_tokens": 708499125.0, "step": 18658 }, { "epoch": 2.373616588220328, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.623433589935303, "learning_rate": 5e-05, "loss": 0.7299, "mean_token_accuracy": 0.8841632604598999, "num_tokens": 708544844.0, "step": 18659 }, { "epoch": 2.3737437984989187, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.781644821166992, "learning_rate": 5e-05, "loss": 0.76, "mean_token_accuracy": 0.8788076639175415, "num_tokens": 708586852.0, "step": 18660 }, { "epoch": 2.3738710087775092, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.435892105102539, "learning_rate": 5e-05, "loss": 0.7157, "mean_token_accuracy": 0.881283164024353, "num_tokens": 708624108.0, "step": 18661 }, { "epoch": 2.3739982190560998, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.948235511779785, "learning_rate": 5e-05, "loss": 0.7062, "mean_token_accuracy": 0.8979200720787048, "num_tokens": 708658214.0, "step": 18662 }, { "epoch": 2.3741254293346903, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.203639507293701, "learning_rate": 5e-05, "loss": 0.6624, "mean_token_accuracy": 0.8942089080810547, "num_tokens": 708695315.0, "step": 18663 }, { "epoch": 2.374252639613281, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.20991325378418, "learning_rate": 5e-05, "loss": 0.7976, "mean_token_accuracy": 0.8718495965003967, "num_tokens": 708734034.0, "step": 18664 }, { "epoch": 2.3743798498918713, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.899956226348877, "learning_rate": 5e-05, "loss": 0.6874, "mean_token_accuracy": 0.8809739351272583, "num_tokens": 708769758.0, "step": 18665 }, { "epoch": 2.374507060170462, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.611286163330078, "learning_rate": 5e-05, "loss": 0.8282, "mean_token_accuracy": 0.8749638199806213, "num_tokens": 708803639.0, "step": 18666 }, { "epoch": 2.3746342704490524, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.501391887664795, "learning_rate": 5e-05, "loss": 0.6921, "mean_token_accuracy": 0.8767917156219482, "num_tokens": 708837085.0, "step": 18667 }, { "epoch": 2.374761480727643, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 9.205718994140625, "learning_rate": 5e-05, "loss": 0.864, "mean_token_accuracy": 0.8817074298858643, "num_tokens": 708876102.0, "step": 18668 }, { "epoch": 2.3748886910062335, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.142845630645752, "learning_rate": 5e-05, "loss": 0.6516, "mean_token_accuracy": 0.8904352784156799, "num_tokens": 708914260.0, "step": 18669 }, { "epoch": 2.375015901284824, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.844502449035645, "learning_rate": 5e-05, "loss": 1.004, "mean_token_accuracy": 0.8703972697257996, "num_tokens": 708952071.0, "step": 18670 }, { "epoch": 2.3751431115634145, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.4171223640441895, "learning_rate": 5e-05, "loss": 0.7459, "mean_token_accuracy": 0.8750916719436646, "num_tokens": 708989394.0, "step": 18671 }, { "epoch": 2.375270321842005, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.88467788696289, "learning_rate": 5e-05, "loss": 1.1006, "mean_token_accuracy": 0.8588714003562927, "num_tokens": 709033558.0, "step": 18672 }, { "epoch": 2.375397532120595, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.664416790008545, "learning_rate": 5e-05, "loss": 0.9403, "mean_token_accuracy": 0.8762811422348022, "num_tokens": 709071420.0, "step": 18673 }, { "epoch": 2.375524742399186, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.60822582244873, "learning_rate": 5e-05, "loss": 0.899, "mean_token_accuracy": 0.8908195495605469, "num_tokens": 709104810.0, "step": 18674 }, { "epoch": 2.375651952677776, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.835816383361816, "learning_rate": 5e-05, "loss": 1.0006, "mean_token_accuracy": 0.8744628429412842, "num_tokens": 709142789.0, "step": 18675 }, { "epoch": 2.3757791629563667, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.899066925048828, "learning_rate": 5e-05, "loss": 0.7272, "mean_token_accuracy": 0.8896788954734802, "num_tokens": 709177590.0, "step": 18676 }, { "epoch": 2.3759063732349572, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.823294639587402, "learning_rate": 5e-05, "loss": 0.9819, "mean_token_accuracy": 0.8885723352432251, "num_tokens": 709217177.0, "step": 18677 }, { "epoch": 2.3760335835135478, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.604081153869629, "learning_rate": 5e-05, "loss": 0.8484, "mean_token_accuracy": 0.8510924577713013, "num_tokens": 709252566.0, "step": 18678 }, { "epoch": 2.3761607937921383, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.95462417602539, "learning_rate": 5e-05, "loss": 0.9774, "mean_token_accuracy": 0.8932871222496033, "num_tokens": 709290386.0, "step": 18679 }, { "epoch": 2.376288004070729, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 7.658473491668701, "learning_rate": 5e-05, "loss": 0.9417, "mean_token_accuracy": 0.8821724653244019, "num_tokens": 709331110.0, "step": 18680 }, { "epoch": 2.3764152143493193, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.915238380432129, "learning_rate": 5e-05, "loss": 0.9187, "mean_token_accuracy": 0.8941810131072998, "num_tokens": 709371390.0, "step": 18681 }, { "epoch": 2.37654242462791, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.501815795898438, "learning_rate": 5e-05, "loss": 1.0086, "mean_token_accuracy": 0.8780750036239624, "num_tokens": 709413428.0, "step": 18682 }, { "epoch": 2.3766696349065004, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.731869697570801, "learning_rate": 5e-05, "loss": 0.822, "mean_token_accuracy": 0.8824340105056763, "num_tokens": 709450332.0, "step": 18683 }, { "epoch": 2.376796845185091, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.736100196838379, "learning_rate": 5e-05, "loss": 1.0137, "mean_token_accuracy": 0.8584514856338501, "num_tokens": 709490039.0, "step": 18684 }, { "epoch": 2.3769240554636815, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.478208065032959, "learning_rate": 5e-05, "loss": 0.8009, "mean_token_accuracy": 0.887648344039917, "num_tokens": 709531839.0, "step": 18685 }, { "epoch": 2.377051265742272, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.366337776184082, "learning_rate": 5e-05, "loss": 0.8735, "mean_token_accuracy": 0.8869227170944214, "num_tokens": 709568126.0, "step": 18686 }, { "epoch": 2.3771784760208625, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.721553802490234, "learning_rate": 5e-05, "loss": 0.8304, "mean_token_accuracy": 0.8861351609230042, "num_tokens": 709610546.0, "step": 18687 }, { "epoch": 2.377305686299453, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.896600246429443, "learning_rate": 5e-05, "loss": 0.8385, "mean_token_accuracy": 0.8742920160293579, "num_tokens": 709656869.0, "step": 18688 }, { "epoch": 2.3774328965780436, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.835714817047119, "learning_rate": 5e-05, "loss": 0.8037, "mean_token_accuracy": 0.889738917350769, "num_tokens": 709695691.0, "step": 18689 }, { "epoch": 2.377560106856634, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.642897129058838, "learning_rate": 5e-05, "loss": 0.7706, "mean_token_accuracy": 0.8802287578582764, "num_tokens": 709736691.0, "step": 18690 }, { "epoch": 2.3776873171352246, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.903097629547119, "learning_rate": 5e-05, "loss": 0.8274, "mean_token_accuracy": 0.8737361431121826, "num_tokens": 709776277.0, "step": 18691 }, { "epoch": 2.377814527413815, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.43603515625, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8724445104598999, "num_tokens": 709817222.0, "step": 18692 }, { "epoch": 2.3779417376924057, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.104780197143555, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.8777251243591309, "num_tokens": 709859060.0, "step": 18693 }, { "epoch": 2.378068947970996, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.162567615509033, "learning_rate": 5e-05, "loss": 0.7167, "mean_token_accuracy": 0.880431056022644, "num_tokens": 709897396.0, "step": 18694 }, { "epoch": 2.3781961582495867, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.382704734802246, "learning_rate": 5e-05, "loss": 0.8036, "mean_token_accuracy": 0.8846515417098999, "num_tokens": 709939797.0, "step": 18695 }, { "epoch": 2.378323368528177, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.888772010803223, "learning_rate": 5e-05, "loss": 0.7108, "mean_token_accuracy": 0.8778768181800842, "num_tokens": 709977102.0, "step": 18696 }, { "epoch": 2.378450578806768, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.680063247680664, "learning_rate": 5e-05, "loss": 0.8554, "mean_token_accuracy": 0.8775988221168518, "num_tokens": 710020824.0, "step": 18697 }, { "epoch": 2.378577789085358, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.65647029876709, "learning_rate": 5e-05, "loss": 0.6915, "mean_token_accuracy": 0.8807291984558105, "num_tokens": 710052263.0, "step": 18698 }, { "epoch": 2.3787049993639484, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 9.018473625183105, "learning_rate": 5e-05, "loss": 0.8706, "mean_token_accuracy": 0.8826843500137329, "num_tokens": 710095196.0, "step": 18699 }, { "epoch": 2.378832209642539, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.5097222328186035, "learning_rate": 5e-05, "loss": 0.7344, "mean_token_accuracy": 0.8743420839309692, "num_tokens": 710127255.0, "step": 18700 }, { "epoch": 2.3789594199211295, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.331350326538086, "learning_rate": 5e-05, "loss": 0.9386, "mean_token_accuracy": 0.8800867795944214, "num_tokens": 710159129.0, "step": 18701 }, { "epoch": 2.37908663019972, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.65128755569458, "learning_rate": 5e-05, "loss": 0.7706, "mean_token_accuracy": 0.876213788986206, "num_tokens": 710197244.0, "step": 18702 }, { "epoch": 2.3792138404783105, "ewc_loss": 0.625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.342463493347168, "learning_rate": 5e-05, "loss": 1.0129, "mean_token_accuracy": 0.866875171661377, "num_tokens": 710241133.0, "step": 18703 }, { "epoch": 2.379341050756901, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.229079246520996, "learning_rate": 5e-05, "loss": 0.8611, "mean_token_accuracy": 0.8830689787864685, "num_tokens": 710283264.0, "step": 18704 }, { "epoch": 2.3794682610354916, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.766212463378906, "learning_rate": 5e-05, "loss": 0.9525, "mean_token_accuracy": 0.8721586465835571, "num_tokens": 710329203.0, "step": 18705 }, { "epoch": 2.379595471314082, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.814154148101807, "learning_rate": 5e-05, "loss": 0.8806, "mean_token_accuracy": 0.8850849270820618, "num_tokens": 710366772.0, "step": 18706 }, { "epoch": 2.3797226815926726, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.022293090820312, "learning_rate": 5e-05, "loss": 0.8674, "mean_token_accuracy": 0.8780156970024109, "num_tokens": 710407458.0, "step": 18707 }, { "epoch": 2.379849891871263, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.009596824645996, "learning_rate": 5e-05, "loss": 0.8565, "mean_token_accuracy": 0.8828527331352234, "num_tokens": 710449200.0, "step": 18708 }, { "epoch": 2.3799771021498537, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.710457801818848, "learning_rate": 5e-05, "loss": 0.7628, "mean_token_accuracy": 0.8994090557098389, "num_tokens": 710490764.0, "step": 18709 }, { "epoch": 2.3801043124284442, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.811314105987549, "learning_rate": 5e-05, "loss": 0.8201, "mean_token_accuracy": 0.8801501989364624, "num_tokens": 710526282.0, "step": 18710 }, { "epoch": 2.3802315227070348, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.85972261428833, "learning_rate": 5e-05, "loss": 0.7906, "mean_token_accuracy": 0.8883793354034424, "num_tokens": 710566523.0, "step": 18711 }, { "epoch": 2.3803587329856253, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.432407379150391, "learning_rate": 5e-05, "loss": 0.7495, "mean_token_accuracy": 0.8838399648666382, "num_tokens": 710607640.0, "step": 18712 }, { "epoch": 2.380485943264216, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.112504959106445, "learning_rate": 5e-05, "loss": 0.8176, "mean_token_accuracy": 0.8783316016197205, "num_tokens": 710644785.0, "step": 18713 }, { "epoch": 2.3806131535428063, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.150875091552734, "learning_rate": 5e-05, "loss": 0.7187, "mean_token_accuracy": 0.883041262626648, "num_tokens": 710681857.0, "step": 18714 }, { "epoch": 2.380740363821397, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.322509765625, "learning_rate": 5e-05, "loss": 0.8012, "mean_token_accuracy": 0.8884172439575195, "num_tokens": 710718119.0, "step": 18715 }, { "epoch": 2.3808675740999874, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.982390880584717, "learning_rate": 5e-05, "loss": 0.7606, "mean_token_accuracy": 0.8698456287384033, "num_tokens": 710758573.0, "step": 18716 }, { "epoch": 2.380994784378578, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.494847297668457, "learning_rate": 5e-05, "loss": 0.7975, "mean_token_accuracy": 0.8950244188308716, "num_tokens": 710791567.0, "step": 18717 }, { "epoch": 2.3811219946571685, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.8589558601379395, "learning_rate": 5e-05, "loss": 0.6855, "mean_token_accuracy": 0.8888287544250488, "num_tokens": 710833060.0, "step": 18718 }, { "epoch": 2.381249204935759, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.689796447753906, "learning_rate": 5e-05, "loss": 0.8764, "mean_token_accuracy": 0.8776690363883972, "num_tokens": 710872175.0, "step": 18719 }, { "epoch": 2.3813764152143495, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.825616836547852, "learning_rate": 5e-05, "loss": 0.7027, "mean_token_accuracy": 0.8848775625228882, "num_tokens": 710910442.0, "step": 18720 }, { "epoch": 2.3815036254929396, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.722753524780273, "learning_rate": 5e-05, "loss": 0.8761, "mean_token_accuracy": 0.8813198804855347, "num_tokens": 710948656.0, "step": 18721 }, { "epoch": 2.3816308357715306, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.056933403015137, "learning_rate": 5e-05, "loss": 0.7733, "mean_token_accuracy": 0.8791091442108154, "num_tokens": 710989300.0, "step": 18722 }, { "epoch": 2.3817580460501206, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.425193786621094, "learning_rate": 5e-05, "loss": 0.8337, "mean_token_accuracy": 0.8932751417160034, "num_tokens": 711027661.0, "step": 18723 }, { "epoch": 2.381885256328711, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.452652454376221, "learning_rate": 5e-05, "loss": 0.8253, "mean_token_accuracy": 0.8722127676010132, "num_tokens": 711067895.0, "step": 18724 }, { "epoch": 2.3820124666073017, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.858037948608398, "learning_rate": 5e-05, "loss": 0.8125, "mean_token_accuracy": 0.8769303560256958, "num_tokens": 711105567.0, "step": 18725 }, { "epoch": 2.3821396768858922, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.861386775970459, "learning_rate": 5e-05, "loss": 0.7658, "mean_token_accuracy": 0.8941888213157654, "num_tokens": 711145263.0, "step": 18726 }, { "epoch": 2.3822668871644828, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.3756103515625, "learning_rate": 5e-05, "loss": 0.7333, "mean_token_accuracy": 0.8859655857086182, "num_tokens": 711187024.0, "step": 18727 }, { "epoch": 2.3823940974430733, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.162720680236816, "learning_rate": 5e-05, "loss": 0.8093, "mean_token_accuracy": 0.8790152072906494, "num_tokens": 711226261.0, "step": 18728 }, { "epoch": 2.382521307721664, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.040770530700684, "learning_rate": 5e-05, "loss": 0.7202, "mean_token_accuracy": 0.8748461008071899, "num_tokens": 711267005.0, "step": 18729 }, { "epoch": 2.3826485180002543, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.39079761505127, "learning_rate": 5e-05, "loss": 0.8457, "mean_token_accuracy": 0.8747111558914185, "num_tokens": 711307145.0, "step": 18730 }, { "epoch": 2.382775728278845, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.8475341796875, "learning_rate": 5e-05, "loss": 0.6881, "mean_token_accuracy": 0.8842520713806152, "num_tokens": 711351128.0, "step": 18731 }, { "epoch": 2.3829029385574354, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.611249923706055, "learning_rate": 5e-05, "loss": 0.8145, "mean_token_accuracy": 0.8882273435592651, "num_tokens": 711385344.0, "step": 18732 }, { "epoch": 2.383030148836026, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.694321155548096, "learning_rate": 5e-05, "loss": 0.709, "mean_token_accuracy": 0.8764389753341675, "num_tokens": 711421548.0, "step": 18733 }, { "epoch": 2.3831573591146165, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.866656303405762, "learning_rate": 5e-05, "loss": 0.853, "mean_token_accuracy": 0.8840662837028503, "num_tokens": 711456952.0, "step": 18734 }, { "epoch": 2.383284569393207, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.625504493713379, "learning_rate": 5e-05, "loss": 0.6887, "mean_token_accuracy": 0.8851435780525208, "num_tokens": 711486985.0, "step": 18735 }, { "epoch": 2.3834117796717975, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.025091171264648, "learning_rate": 5e-05, "loss": 0.8791, "mean_token_accuracy": 0.8979290723800659, "num_tokens": 711521047.0, "step": 18736 }, { "epoch": 2.383538989950388, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.865437984466553, "learning_rate": 5e-05, "loss": 0.7347, "mean_token_accuracy": 0.8874741196632385, "num_tokens": 711561151.0, "step": 18737 }, { "epoch": 2.3836662002289786, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.753024101257324, "learning_rate": 5e-05, "loss": 0.9597, "mean_token_accuracy": 0.8628749251365662, "num_tokens": 711597975.0, "step": 18738 }, { "epoch": 2.383793410507569, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.385216236114502, "learning_rate": 5e-05, "loss": 0.8218, "mean_token_accuracy": 0.8792420625686646, "num_tokens": 711637923.0, "step": 18739 }, { "epoch": 2.3839206207861596, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.115074157714844, "learning_rate": 5e-05, "loss": 0.873, "mean_token_accuracy": 0.8787351846694946, "num_tokens": 711675408.0, "step": 18740 }, { "epoch": 2.38404783106475, "ewc_loss": 0.5, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.833139419555664, "learning_rate": 5e-05, "loss": 0.8149, "mean_token_accuracy": 0.8880544900894165, "num_tokens": 711712781.0, "step": 18741 }, { "epoch": 2.3841750413433407, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.600286960601807, "learning_rate": 5e-05, "loss": 0.7819, "mean_token_accuracy": 0.8822849988937378, "num_tokens": 711752974.0, "step": 18742 }, { "epoch": 2.384302251621931, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.900137424468994, "learning_rate": 5e-05, "loss": 0.8011, "mean_token_accuracy": 0.8827449679374695, "num_tokens": 711793526.0, "step": 18743 }, { "epoch": 2.3844294619005217, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.5977396965026855, "learning_rate": 5e-05, "loss": 0.717, "mean_token_accuracy": 0.8961517214775085, "num_tokens": 711830409.0, "step": 18744 }, { "epoch": 2.3845566721791123, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.5811662673950195, "learning_rate": 5e-05, "loss": 0.7322, "mean_token_accuracy": 0.8907883167266846, "num_tokens": 711873476.0, "step": 18745 }, { "epoch": 2.3846838824577024, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.945374488830566, "learning_rate": 5e-05, "loss": 0.7765, "mean_token_accuracy": 0.8880751729011536, "num_tokens": 711912507.0, "step": 18746 }, { "epoch": 2.3848110927362933, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.132870197296143, "learning_rate": 5e-05, "loss": 0.6919, "mean_token_accuracy": 0.8906898498535156, "num_tokens": 711946884.0, "step": 18747 }, { "epoch": 2.3849383030148834, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.35405158996582, "learning_rate": 5e-05, "loss": 0.8353, "mean_token_accuracy": 0.8712276220321655, "num_tokens": 711987060.0, "step": 18748 }, { "epoch": 2.385065513293474, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.755588531494141, "learning_rate": 5e-05, "loss": 0.6433, "mean_token_accuracy": 0.8985114097595215, "num_tokens": 712023197.0, "step": 18749 }, { "epoch": 2.3851927235720645, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.78500747680664, "learning_rate": 5e-05, "loss": 0.8157, "mean_token_accuracy": 0.8926717042922974, "num_tokens": 712060337.0, "step": 18750 }, { "epoch": 2.385319933850655, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.502643585205078, "learning_rate": 5e-05, "loss": 0.6704, "mean_token_accuracy": 0.8905165195465088, "num_tokens": 712101601.0, "step": 18751 }, { "epoch": 2.3854471441292455, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.126747131347656, "learning_rate": 5e-05, "loss": 0.8802, "mean_token_accuracy": 0.8903911113739014, "num_tokens": 712136311.0, "step": 18752 }, { "epoch": 2.385574354407836, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.610454082489014, "learning_rate": 5e-05, "loss": 0.7373, "mean_token_accuracy": 0.8738413453102112, "num_tokens": 712174386.0, "step": 18753 }, { "epoch": 2.3857015646864266, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.042709350585938, "learning_rate": 5e-05, "loss": 0.9384, "mean_token_accuracy": 0.8840107917785645, "num_tokens": 712209540.0, "step": 18754 }, { "epoch": 2.385828774965017, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.224986553192139, "learning_rate": 5e-05, "loss": 0.8301, "mean_token_accuracy": 0.8744776248931885, "num_tokens": 712245405.0, "step": 18755 }, { "epoch": 2.3859559852436076, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.301950454711914, "learning_rate": 5e-05, "loss": 0.9309, "mean_token_accuracy": 0.8606119155883789, "num_tokens": 712285013.0, "step": 18756 }, { "epoch": 2.386083195522198, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.980060577392578, "learning_rate": 5e-05, "loss": 0.9049, "mean_token_accuracy": 0.8655841946601868, "num_tokens": 712324850.0, "step": 18757 }, { "epoch": 2.3862104058007887, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.28978967666626, "learning_rate": 5e-05, "loss": 0.7783, "mean_token_accuracy": 0.8766540288925171, "num_tokens": 712364336.0, "step": 18758 }, { "epoch": 2.386337616079379, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.595041275024414, "learning_rate": 5e-05, "loss": 0.925, "mean_token_accuracy": 0.8674389719963074, "num_tokens": 712403773.0, "step": 18759 }, { "epoch": 2.3864648263579697, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.68828821182251, "learning_rate": 5e-05, "loss": 0.7386, "mean_token_accuracy": 0.8689137697219849, "num_tokens": 712440307.0, "step": 18760 }, { "epoch": 2.3865920366365603, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.022096633911133, "learning_rate": 5e-05, "loss": 0.8905, "mean_token_accuracy": 0.8822283744812012, "num_tokens": 712472008.0, "step": 18761 }, { "epoch": 2.386719246915151, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.513393878936768, "learning_rate": 5e-05, "loss": 0.6855, "mean_token_accuracy": 0.8859618902206421, "num_tokens": 712512935.0, "step": 18762 }, { "epoch": 2.3868464571937413, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.287849426269531, "learning_rate": 5e-05, "loss": 0.9344, "mean_token_accuracy": 0.8873372673988342, "num_tokens": 712552783.0, "step": 18763 }, { "epoch": 2.386973667472332, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.778582572937012, "learning_rate": 5e-05, "loss": 0.7082, "mean_token_accuracy": 0.902681827545166, "num_tokens": 712583752.0, "step": 18764 }, { "epoch": 2.3871008777509224, "ewc_loss": 0.625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.18729019165039, "learning_rate": 5e-05, "loss": 1.0029, "mean_token_accuracy": 0.871442973613739, "num_tokens": 712623125.0, "step": 18765 }, { "epoch": 2.387228088029513, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.420928001403809, "learning_rate": 5e-05, "loss": 0.8873, "mean_token_accuracy": 0.8736273050308228, "num_tokens": 712664965.0, "step": 18766 }, { "epoch": 2.3873552983081034, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.545269012451172, "learning_rate": 5e-05, "loss": 0.9334, "mean_token_accuracy": 0.8778729438781738, "num_tokens": 712703571.0, "step": 18767 }, { "epoch": 2.387482508586694, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.913647651672363, "learning_rate": 5e-05, "loss": 0.888, "mean_token_accuracy": 0.8842194080352783, "num_tokens": 712739830.0, "step": 18768 }, { "epoch": 2.3876097188652845, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.886203289031982, "learning_rate": 5e-05, "loss": 0.8782, "mean_token_accuracy": 0.8727124929428101, "num_tokens": 712782367.0, "step": 18769 }, { "epoch": 2.387736929143875, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.916685581207275, "learning_rate": 5e-05, "loss": 0.8406, "mean_token_accuracy": 0.8847669959068298, "num_tokens": 712816937.0, "step": 18770 }, { "epoch": 2.387864139422465, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.885279655456543, "learning_rate": 5e-05, "loss": 0.7983, "mean_token_accuracy": 0.8902444243431091, "num_tokens": 712850198.0, "step": 18771 }, { "epoch": 2.387991349701056, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.41129732131958, "learning_rate": 5e-05, "loss": 0.7726, "mean_token_accuracy": 0.8830163478851318, "num_tokens": 712886201.0, "step": 18772 }, { "epoch": 2.388118559979646, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.366667747497559, "learning_rate": 5e-05, "loss": 0.8878, "mean_token_accuracy": 0.8737174272537231, "num_tokens": 712924309.0, "step": 18773 }, { "epoch": 2.3882457702582367, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.854998588562012, "learning_rate": 5e-05, "loss": 0.6861, "mean_token_accuracy": 0.8957719802856445, "num_tokens": 712964508.0, "step": 18774 }, { "epoch": 2.3883729805368272, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.851630210876465, "learning_rate": 5e-05, "loss": 0.9176, "mean_token_accuracy": 0.8752864003181458, "num_tokens": 713002737.0, "step": 18775 }, { "epoch": 2.3885001908154178, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.644045352935791, "learning_rate": 5e-05, "loss": 0.7135, "mean_token_accuracy": 0.8795729875564575, "num_tokens": 713040859.0, "step": 18776 }, { "epoch": 2.3886274010940083, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.106573104858398, "learning_rate": 5e-05, "loss": 0.9199, "mean_token_accuracy": 0.8841955661773682, "num_tokens": 713073504.0, "step": 18777 }, { "epoch": 2.388754611372599, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 6.898233413696289, "learning_rate": 5e-05, "loss": 0.7639, "mean_token_accuracy": 0.8856772780418396, "num_tokens": 713110285.0, "step": 18778 }, { "epoch": 2.3888818216511893, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.90685749053955, "learning_rate": 5e-05, "loss": 0.9451, "mean_token_accuracy": 0.8764299154281616, "num_tokens": 713146227.0, "step": 18779 }, { "epoch": 2.38900903192978, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.455968856811523, "learning_rate": 5e-05, "loss": 0.8603, "mean_token_accuracy": 0.8791940212249756, "num_tokens": 713187632.0, "step": 18780 }, { "epoch": 2.3891362422083704, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.242412567138672, "learning_rate": 5e-05, "loss": 0.8885, "mean_token_accuracy": 0.8785203099250793, "num_tokens": 713223023.0, "step": 18781 }, { "epoch": 2.389263452486961, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.967744827270508, "learning_rate": 5e-05, "loss": 0.8943, "mean_token_accuracy": 0.8718447685241699, "num_tokens": 713264390.0, "step": 18782 }, { "epoch": 2.3893906627655515, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.490543842315674, "learning_rate": 5e-05, "loss": 0.7674, "mean_token_accuracy": 0.8869615197181702, "num_tokens": 713306504.0, "step": 18783 }, { "epoch": 2.389517873044142, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.2566556930542, "learning_rate": 5e-05, "loss": 0.8477, "mean_token_accuracy": 0.8825955986976624, "num_tokens": 713347556.0, "step": 18784 }, { "epoch": 2.3896450833227325, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.143059253692627, "learning_rate": 5e-05, "loss": 0.7977, "mean_token_accuracy": 0.857516348361969, "num_tokens": 713386645.0, "step": 18785 }, { "epoch": 2.389772293601323, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.289811134338379, "learning_rate": 5e-05, "loss": 0.8226, "mean_token_accuracy": 0.8845947980880737, "num_tokens": 713419754.0, "step": 18786 }, { "epoch": 2.3898995038799136, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.158586025238037, "learning_rate": 5e-05, "loss": 0.731, "mean_token_accuracy": 0.8801845908164978, "num_tokens": 713456374.0, "step": 18787 }, { "epoch": 2.390026714158504, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.161221504211426, "learning_rate": 5e-05, "loss": 0.8085, "mean_token_accuracy": 0.8823609948158264, "num_tokens": 713487961.0, "step": 18788 }, { "epoch": 2.3901539244370946, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.124134540557861, "learning_rate": 5e-05, "loss": 0.7342, "mean_token_accuracy": 0.8802604675292969, "num_tokens": 713525697.0, "step": 18789 }, { "epoch": 2.390281134715685, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.292466163635254, "learning_rate": 5e-05, "loss": 0.83, "mean_token_accuracy": 0.8787047266960144, "num_tokens": 713563902.0, "step": 18790 }, { "epoch": 2.3904083449942757, "ewc_loss": 0.375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.9153008460998535, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8715506792068481, "num_tokens": 713602996.0, "step": 18791 }, { "epoch": 2.390535555272866, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.548819541931152, "learning_rate": 5e-05, "loss": 0.8506, "mean_token_accuracy": 0.8806315064430237, "num_tokens": 713641725.0, "step": 18792 }, { "epoch": 2.3906627655514567, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.783999443054199, "learning_rate": 5e-05, "loss": 0.7587, "mean_token_accuracy": 0.8606196641921997, "num_tokens": 713682973.0, "step": 18793 }, { "epoch": 2.390789975830047, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.70359992980957, "learning_rate": 5e-05, "loss": 0.8529, "mean_token_accuracy": 0.886267900466919, "num_tokens": 713718919.0, "step": 18794 }, { "epoch": 2.390917186108638, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.836294651031494, "learning_rate": 5e-05, "loss": 0.7407, "mean_token_accuracy": 0.8778361082077026, "num_tokens": 713757416.0, "step": 18795 }, { "epoch": 2.391044396387228, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.616973876953125, "learning_rate": 5e-05, "loss": 0.9313, "mean_token_accuracy": 0.8658051490783691, "num_tokens": 713795648.0, "step": 18796 }, { "epoch": 2.3911716066658184, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.099432945251465, "learning_rate": 5e-05, "loss": 0.7252, "mean_token_accuracy": 0.897307276725769, "num_tokens": 713830095.0, "step": 18797 }, { "epoch": 2.391298816944409, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.2974214553833, "learning_rate": 5e-05, "loss": 0.8773, "mean_token_accuracy": 0.8753313422203064, "num_tokens": 713868958.0, "step": 18798 }, { "epoch": 2.3914260272229995, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.431057929992676, "learning_rate": 5e-05, "loss": 0.8423, "mean_token_accuracy": 0.8702885508537292, "num_tokens": 713911533.0, "step": 18799 }, { "epoch": 2.39155323750159, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.880077838897705, "learning_rate": 5e-05, "loss": 0.8356, "mean_token_accuracy": 0.8742270469665527, "num_tokens": 713952585.0, "step": 18800 }, { "epoch": 2.3916804477801805, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.674925804138184, "learning_rate": 5e-05, "loss": 0.7595, "mean_token_accuracy": 0.8964152336120605, "num_tokens": 713990929.0, "step": 18801 }, { "epoch": 2.391807658058771, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.604434967041016, "learning_rate": 5e-05, "loss": 0.7885, "mean_token_accuracy": 0.8703423738479614, "num_tokens": 714027098.0, "step": 18802 }, { "epoch": 2.3919348683373616, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.712671756744385, "learning_rate": 5e-05, "loss": 0.795, "mean_token_accuracy": 0.8740918636322021, "num_tokens": 714062444.0, "step": 18803 }, { "epoch": 2.392062078615952, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.556132793426514, "learning_rate": 5e-05, "loss": 0.7054, "mean_token_accuracy": 0.8961081504821777, "num_tokens": 714099091.0, "step": 18804 }, { "epoch": 2.3921892888945426, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.605010032653809, "learning_rate": 5e-05, "loss": 0.772, "mean_token_accuracy": 0.8743792772293091, "num_tokens": 714135694.0, "step": 18805 }, { "epoch": 2.392316499173133, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000736236572265625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.659115791320801, "learning_rate": 5e-05, "loss": 0.7208, "mean_token_accuracy": 0.89216548204422, "num_tokens": 714170324.0, "step": 18806 }, { "epoch": 2.3924437094517237, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000736236572265625, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.406229496002197, "learning_rate": 5e-05, "loss": 0.7187, "mean_token_accuracy": 0.8826534152030945, "num_tokens": 714205080.0, "step": 18807 }, { "epoch": 2.392570919730314, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.846356391906738, "learning_rate": 5e-05, "loss": 0.7644, "mean_token_accuracy": 0.8771654367446899, "num_tokens": 714245297.0, "step": 18808 }, { "epoch": 2.3926981300089047, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.163865089416504, "learning_rate": 5e-05, "loss": 0.6725, "mean_token_accuracy": 0.8913198113441467, "num_tokens": 714284155.0, "step": 18809 }, { "epoch": 2.3928253402874953, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.108010292053223, "learning_rate": 5e-05, "loss": 0.7637, "mean_token_accuracy": 0.8829836845397949, "num_tokens": 714318688.0, "step": 18810 }, { "epoch": 2.392952550566086, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.870555877685547, "learning_rate": 5e-05, "loss": 0.6421, "mean_token_accuracy": 0.8975034952163696, "num_tokens": 714353730.0, "step": 18811 }, { "epoch": 2.3930797608446763, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.499435424804688, "learning_rate": 5e-05, "loss": 0.8096, "mean_token_accuracy": 0.8797999620437622, "num_tokens": 714390156.0, "step": 18812 }, { "epoch": 2.393206971123267, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.488284111022949, "learning_rate": 5e-05, "loss": 0.7012, "mean_token_accuracy": 0.8687562942504883, "num_tokens": 714429435.0, "step": 18813 }, { "epoch": 2.3933341814018574, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 9.037166595458984, "learning_rate": 5e-05, "loss": 0.823, "mean_token_accuracy": 0.8998252749443054, "num_tokens": 714470034.0, "step": 18814 }, { "epoch": 2.393461391680448, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.170202732086182, "learning_rate": 5e-05, "loss": 0.6763, "mean_token_accuracy": 0.8788928985595703, "num_tokens": 714505989.0, "step": 18815 }, { "epoch": 2.3935886019590384, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.598396301269531, "learning_rate": 5e-05, "loss": 0.9789, "mean_token_accuracy": 0.8733054399490356, "num_tokens": 714541727.0, "step": 18816 }, { "epoch": 2.393715812237629, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.437634468078613, "learning_rate": 5e-05, "loss": 0.7337, "mean_token_accuracy": 0.8823415637016296, "num_tokens": 714581429.0, "step": 18817 }, { "epoch": 2.3938430225162195, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.510782241821289, "learning_rate": 5e-05, "loss": 0.9854, "mean_token_accuracy": 0.8792886734008789, "num_tokens": 714621839.0, "step": 18818 }, { "epoch": 2.3939702327948096, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.577488422393799, "learning_rate": 5e-05, "loss": 0.9202, "mean_token_accuracy": 0.8789331912994385, "num_tokens": 714665459.0, "step": 18819 }, { "epoch": 2.3940974430734006, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.166074752807617, "learning_rate": 5e-05, "loss": 0.8546, "mean_token_accuracy": 0.8888716697692871, "num_tokens": 714699259.0, "step": 18820 }, { "epoch": 2.3942246533519906, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.81983757019043, "learning_rate": 5e-05, "loss": 0.9905, "mean_token_accuracy": 0.8742175102233887, "num_tokens": 714740330.0, "step": 18821 }, { "epoch": 2.394351863630581, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.4423627853393555, "learning_rate": 5e-05, "loss": 0.6898, "mean_token_accuracy": 0.8852773308753967, "num_tokens": 714773330.0, "step": 18822 }, { "epoch": 2.3944790739091717, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.238655090332031, "learning_rate": 5e-05, "loss": 1.0513, "mean_token_accuracy": 0.8729115724563599, "num_tokens": 714812707.0, "step": 18823 }, { "epoch": 2.3946062841877622, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.217531204223633, "learning_rate": 5e-05, "loss": 0.7338, "mean_token_accuracy": 0.8846616744995117, "num_tokens": 714853496.0, "step": 18824 }, { "epoch": 2.3947334944663528, "ewc_loss": 0.88671875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 11.164798736572266, "learning_rate": 5e-05, "loss": 1.2083, "mean_token_accuracy": 0.8850841522216797, "num_tokens": 714895766.0, "step": 18825 }, { "epoch": 2.3948607047449433, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 8.373798370361328, "learning_rate": 5e-05, "loss": 1.0985, "mean_token_accuracy": 0.8831790089607239, "num_tokens": 714936089.0, "step": 18826 }, { "epoch": 2.394987915023534, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.048139572143555, "learning_rate": 5e-05, "loss": 0.9858, "mean_token_accuracy": 0.8956243991851807, "num_tokens": 714980759.0, "step": 18827 }, { "epoch": 2.3951151253021243, "ewc_loss": 0.859375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 9.341304779052734, "learning_rate": 5e-05, "loss": 1.1694, "mean_token_accuracy": 0.8867878317832947, "num_tokens": 715019671.0, "step": 18828 }, { "epoch": 2.395242335580715, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.364836692810059, "learning_rate": 5e-05, "loss": 0.9578, "mean_token_accuracy": 0.8648110628128052, "num_tokens": 715059639.0, "step": 18829 }, { "epoch": 2.3953695458593054, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 9.66479206085205, "learning_rate": 5e-05, "loss": 1.0302, "mean_token_accuracy": 0.8913893699645996, "num_tokens": 715096315.0, "step": 18830 }, { "epoch": 2.395496756137896, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 7.824122428894043, "learning_rate": 5e-05, "loss": 0.9205, "mean_token_accuracy": 0.8828461766242981, "num_tokens": 715130800.0, "step": 18831 }, { "epoch": 2.3956239664164865, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.270386695861816, "learning_rate": 5e-05, "loss": 0.9905, "mean_token_accuracy": 0.8753974437713623, "num_tokens": 715175369.0, "step": 18832 }, { "epoch": 2.395751176695077, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.18294906616211, "learning_rate": 5e-05, "loss": 0.9773, "mean_token_accuracy": 0.877400279045105, "num_tokens": 715216480.0, "step": 18833 }, { "epoch": 2.3958783869736675, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.086160659790039, "learning_rate": 5e-05, "loss": 0.8267, "mean_token_accuracy": 0.8869759440422058, "num_tokens": 715257182.0, "step": 18834 }, { "epoch": 2.396005597252258, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.68754768371582, "learning_rate": 5e-05, "loss": 0.9723, "mean_token_accuracy": 0.876366138458252, "num_tokens": 715288793.0, "step": 18835 }, { "epoch": 2.3961328075308486, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.288183689117432, "learning_rate": 5e-05, "loss": 0.7736, "mean_token_accuracy": 0.8791412115097046, "num_tokens": 715323720.0, "step": 18836 }, { "epoch": 2.396260017809439, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.014108657836914, "learning_rate": 5e-05, "loss": 0.973, "mean_token_accuracy": 0.8771538734436035, "num_tokens": 715356104.0, "step": 18837 }, { "epoch": 2.3963872280880296, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.1217122077941895, "learning_rate": 5e-05, "loss": 0.7807, "mean_token_accuracy": 0.889369010925293, "num_tokens": 715390727.0, "step": 18838 }, { "epoch": 2.39651443836662, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.082352638244629, "learning_rate": 5e-05, "loss": 0.9224, "mean_token_accuracy": 0.8854615688323975, "num_tokens": 715426340.0, "step": 18839 }, { "epoch": 2.3966416486452107, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.275356769561768, "learning_rate": 5e-05, "loss": 0.7749, "mean_token_accuracy": 0.9062516689300537, "num_tokens": 715458195.0, "step": 18840 }, { "epoch": 2.396768858923801, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.808588981628418, "learning_rate": 5e-05, "loss": 0.8755, "mean_token_accuracy": 0.8915929794311523, "num_tokens": 715496896.0, "step": 18841 }, { "epoch": 2.3968960692023917, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.504945755004883, "learning_rate": 5e-05, "loss": 0.9024, "mean_token_accuracy": 0.8679295778274536, "num_tokens": 715537985.0, "step": 18842 }, { "epoch": 2.3970232794809823, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.292671203613281, "learning_rate": 5e-05, "loss": 0.8009, "mean_token_accuracy": 0.8949768543243408, "num_tokens": 715574767.0, "step": 18843 }, { "epoch": 2.3971504897595723, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.679887294769287, "learning_rate": 5e-05, "loss": 0.8494, "mean_token_accuracy": 0.8801088333129883, "num_tokens": 715608649.0, "step": 18844 }, { "epoch": 2.3972777000381633, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.925935745239258, "learning_rate": 5e-05, "loss": 0.7955, "mean_token_accuracy": 0.8821923732757568, "num_tokens": 715650278.0, "step": 18845 }, { "epoch": 2.3974049103167534, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.653166770935059, "learning_rate": 5e-05, "loss": 0.8159, "mean_token_accuracy": 0.8758273124694824, "num_tokens": 715692678.0, "step": 18846 }, { "epoch": 2.397532120595344, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.831321716308594, "learning_rate": 5e-05, "loss": 0.8155, "mean_token_accuracy": 0.8735437393188477, "num_tokens": 715733128.0, "step": 18847 }, { "epoch": 2.3976593308739345, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.526918411254883, "learning_rate": 5e-05, "loss": 0.7785, "mean_token_accuracy": 0.8799874782562256, "num_tokens": 715767272.0, "step": 18848 }, { "epoch": 2.397786541152525, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.866485118865967, "learning_rate": 5e-05, "loss": 0.7233, "mean_token_accuracy": 0.8985622525215149, "num_tokens": 715803497.0, "step": 18849 }, { "epoch": 2.3979137514311155, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.348743438720703, "learning_rate": 5e-05, "loss": 0.7113, "mean_token_accuracy": 0.884746789932251, "num_tokens": 715847069.0, "step": 18850 }, { "epoch": 2.398040961709706, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.079602241516113, "learning_rate": 5e-05, "loss": 0.7759, "mean_token_accuracy": 0.8867142200469971, "num_tokens": 715887639.0, "step": 18851 }, { "epoch": 2.3981681719882966, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.986792087554932, "learning_rate": 5e-05, "loss": 0.6305, "mean_token_accuracy": 0.9028525948524475, "num_tokens": 715919427.0, "step": 18852 }, { "epoch": 2.398295382266887, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.57688045501709, "learning_rate": 5e-05, "loss": 0.874, "mean_token_accuracy": 0.867447018623352, "num_tokens": 715961287.0, "step": 18853 }, { "epoch": 2.3984225925454776, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.4807000160217285, "learning_rate": 5e-05, "loss": 0.6899, "mean_token_accuracy": 0.8792442083358765, "num_tokens": 715995021.0, "step": 18854 }, { "epoch": 2.398549802824068, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.236004829406738, "learning_rate": 5e-05, "loss": 0.8685, "mean_token_accuracy": 0.8943668603897095, "num_tokens": 716039517.0, "step": 18855 }, { "epoch": 2.3986770131026587, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.223296165466309, "learning_rate": 5e-05, "loss": 0.724, "mean_token_accuracy": 0.8677575588226318, "num_tokens": 716077116.0, "step": 18856 }, { "epoch": 2.398804223381249, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.821812629699707, "learning_rate": 5e-05, "loss": 0.9948, "mean_token_accuracy": 0.8733780384063721, "num_tokens": 716112800.0, "step": 18857 }, { "epoch": 2.3989314336598397, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 6.6722025871276855, "learning_rate": 5e-05, "loss": 0.753, "mean_token_accuracy": 0.8924500942230225, "num_tokens": 716148165.0, "step": 18858 }, { "epoch": 2.3990586439384303, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.7874755859375, "learning_rate": 5e-05, "loss": 1.0321, "mean_token_accuracy": 0.8755156993865967, "num_tokens": 716185745.0, "step": 18859 }, { "epoch": 2.399185854217021, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 7.809614658355713, "learning_rate": 5e-05, "loss": 0.9094, "mean_token_accuracy": 0.8927425742149353, "num_tokens": 716225177.0, "step": 18860 }, { "epoch": 2.3993130644956113, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.682178497314453, "learning_rate": 5e-05, "loss": 0.8759, "mean_token_accuracy": 0.8996604084968567, "num_tokens": 716265969.0, "step": 18861 }, { "epoch": 2.399440274774202, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.586670875549316, "learning_rate": 5e-05, "loss": 0.9833, "mean_token_accuracy": 0.8794522285461426, "num_tokens": 716302056.0, "step": 18862 }, { "epoch": 2.3995674850527924, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.3765716552734375, "learning_rate": 5e-05, "loss": 0.8068, "mean_token_accuracy": 0.8705791234970093, "num_tokens": 716339452.0, "step": 18863 }, { "epoch": 2.399694695331383, "ewc_loss": 0.625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.020264625549316, "learning_rate": 5e-05, "loss": 0.9645, "mean_token_accuracy": 0.8814239501953125, "num_tokens": 716379930.0, "step": 18864 }, { "epoch": 2.3998219056099734, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 6.99299955368042, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8815658092498779, "num_tokens": 716416367.0, "step": 18865 }, { "epoch": 2.399949115888564, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.983263969421387, "learning_rate": 5e-05, "loss": 0.9878, "mean_token_accuracy": 0.8655714988708496, "num_tokens": 716460112.0, "step": 18866 }, { "epoch": 2.4000763261671545, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.243897438049316, "learning_rate": 5e-05, "loss": 0.8174, "mean_token_accuracy": 0.8767099380493164, "num_tokens": 716497085.0, "step": 18867 }, { "epoch": 2.400203536445745, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.636457443237305, "learning_rate": 5e-05, "loss": 0.9077, "mean_token_accuracy": 0.8814314603805542, "num_tokens": 716542892.0, "step": 18868 }, { "epoch": 2.400330746724335, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.459496974945068, "learning_rate": 5e-05, "loss": 0.8769, "mean_token_accuracy": 0.8688738346099854, "num_tokens": 716578765.0, "step": 18869 }, { "epoch": 2.400457957002926, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.356640815734863, "learning_rate": 5e-05, "loss": 0.8675, "mean_token_accuracy": 0.8853784203529358, "num_tokens": 716618881.0, "step": 18870 }, { "epoch": 2.400585167281516, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.466992378234863, "learning_rate": 5e-05, "loss": 0.8231, "mean_token_accuracy": 0.8810172080993652, "num_tokens": 716651583.0, "step": 18871 }, { "epoch": 2.4007123775601067, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.236470222473145, "learning_rate": 5e-05, "loss": 0.8803, "mean_token_accuracy": 0.8691606521606445, "num_tokens": 716687118.0, "step": 18872 }, { "epoch": 2.400839587838697, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.38781213760376, "learning_rate": 5e-05, "loss": 0.8178, "mean_token_accuracy": 0.8726819753646851, "num_tokens": 716722543.0, "step": 18873 }, { "epoch": 2.4009667981172877, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.163702964782715, "learning_rate": 5e-05, "loss": 0.8236, "mean_token_accuracy": 0.8825262784957886, "num_tokens": 716757447.0, "step": 18874 }, { "epoch": 2.4010940083958783, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.2869744300842285, "learning_rate": 5e-05, "loss": 0.7686, "mean_token_accuracy": 0.8806470036506653, "num_tokens": 716799736.0, "step": 18875 }, { "epoch": 2.401221218674469, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.188084602355957, "learning_rate": 5e-05, "loss": 0.8389, "mean_token_accuracy": 0.8770997524261475, "num_tokens": 716844850.0, "step": 18876 }, { "epoch": 2.4013484289530593, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.190365314483643, "learning_rate": 5e-05, "loss": 0.7557, "mean_token_accuracy": 0.8780912160873413, "num_tokens": 716878801.0, "step": 18877 }, { "epoch": 2.40147563923165, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.226055145263672, "learning_rate": 5e-05, "loss": 0.804, "mean_token_accuracy": 0.8828624486923218, "num_tokens": 716914271.0, "step": 18878 }, { "epoch": 2.4016028495102404, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.133367538452148, "learning_rate": 5e-05, "loss": 0.7378, "mean_token_accuracy": 0.8801398873329163, "num_tokens": 716954094.0, "step": 18879 }, { "epoch": 2.401730059788831, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.21068286895752, "learning_rate": 5e-05, "loss": 0.8001, "mean_token_accuracy": 0.8840271234512329, "num_tokens": 716995061.0, "step": 18880 }, { "epoch": 2.4018572700674214, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.119481086730957, "learning_rate": 5e-05, "loss": 0.7269, "mean_token_accuracy": 0.8786450028419495, "num_tokens": 717032271.0, "step": 18881 }, { "epoch": 2.401984480346012, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.242986679077148, "learning_rate": 5e-05, "loss": 0.7996, "mean_token_accuracy": 0.8824812173843384, "num_tokens": 717068551.0, "step": 18882 }, { "epoch": 2.4021116906246025, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.079290866851807, "learning_rate": 5e-05, "loss": 0.7191, "mean_token_accuracy": 0.8845429420471191, "num_tokens": 717107460.0, "step": 18883 }, { "epoch": 2.402238900903193, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.244576454162598, "learning_rate": 5e-05, "loss": 0.8193, "mean_token_accuracy": 0.877303957939148, "num_tokens": 717144774.0, "step": 18884 }, { "epoch": 2.4023661111817836, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.128021240234375, "learning_rate": 5e-05, "loss": 0.7252, "mean_token_accuracy": 0.8781278729438782, "num_tokens": 717180808.0, "step": 18885 }, { "epoch": 2.402493321460374, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.126601219177246, "learning_rate": 5e-05, "loss": 0.7828, "mean_token_accuracy": 0.8845149874687195, "num_tokens": 717220788.0, "step": 18886 }, { "epoch": 2.4026205317389646, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.242001056671143, "learning_rate": 5e-05, "loss": 0.7219, "mean_token_accuracy": 0.8875861167907715, "num_tokens": 717260257.0, "step": 18887 }, { "epoch": 2.402747742017555, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.9648118019104, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.8872676491737366, "num_tokens": 717291911.0, "step": 18888 }, { "epoch": 2.4028749522961457, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.387871265411377, "learning_rate": 5e-05, "loss": 0.7435, "mean_token_accuracy": 0.8775992393493652, "num_tokens": 717322356.0, "step": 18889 }, { "epoch": 2.403002162574736, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.767529487609863, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.8811817169189453, "num_tokens": 717363439.0, "step": 18890 }, { "epoch": 2.4031293728533267, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.521029472351074, "learning_rate": 5e-05, "loss": 0.7388, "mean_token_accuracy": 0.8781522512435913, "num_tokens": 717401050.0, "step": 18891 }, { "epoch": 2.403256583131917, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.533353805541992, "learning_rate": 5e-05, "loss": 0.7249, "mean_token_accuracy": 0.8782203197479248, "num_tokens": 717442346.0, "step": 18892 }, { "epoch": 2.403383793410508, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.725827693939209, "learning_rate": 5e-05, "loss": 0.7294, "mean_token_accuracy": 0.8846664428710938, "num_tokens": 717483387.0, "step": 18893 }, { "epoch": 2.403511003689098, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.284987449645996, "learning_rate": 5e-05, "loss": 0.7128, "mean_token_accuracy": 0.8812518119812012, "num_tokens": 717522208.0, "step": 18894 }, { "epoch": 2.4036382139676884, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.9632697105407715, "learning_rate": 5e-05, "loss": 0.7323, "mean_token_accuracy": 0.8887484073638916, "num_tokens": 717563338.0, "step": 18895 }, { "epoch": 2.403765424246279, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.003380298614502, "learning_rate": 5e-05, "loss": 0.695, "mean_token_accuracy": 0.8813679218292236, "num_tokens": 717603351.0, "step": 18896 }, { "epoch": 2.4038926345248695, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.273478507995605, "learning_rate": 5e-05, "loss": 0.7376, "mean_token_accuracy": 0.8927581906318665, "num_tokens": 717639886.0, "step": 18897 }, { "epoch": 2.40401984480346, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.661739826202393, "learning_rate": 5e-05, "loss": 0.7414, "mean_token_accuracy": 0.8587321043014526, "num_tokens": 717672228.0, "step": 18898 }, { "epoch": 2.4041470550820505, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.737107276916504, "learning_rate": 5e-05, "loss": 0.8527, "mean_token_accuracy": 0.8723393082618713, "num_tokens": 717709703.0, "step": 18899 }, { "epoch": 2.404274265360641, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.246723651885986, "learning_rate": 5e-05, "loss": 0.6653, "mean_token_accuracy": 0.8795503377914429, "num_tokens": 717746785.0, "step": 18900 }, { "epoch": 2.4044014756392316, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.403558731079102, "learning_rate": 5e-05, "loss": 0.9268, "mean_token_accuracy": 0.873060941696167, "num_tokens": 717785786.0, "step": 18901 }, { "epoch": 2.404528685917822, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.033534526824951, "learning_rate": 5e-05, "loss": 0.6891, "mean_token_accuracy": 0.8775623440742493, "num_tokens": 717821833.0, "step": 18902 }, { "epoch": 2.4046558961964126, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.92776870727539, "learning_rate": 5e-05, "loss": 0.97, "mean_token_accuracy": 0.8882719278335571, "num_tokens": 717858107.0, "step": 18903 }, { "epoch": 2.404783106475003, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 6.753765106201172, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8945748805999756, "num_tokens": 717893090.0, "step": 18904 }, { "epoch": 2.4049103167535937, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.452757835388184, "learning_rate": 5e-05, "loss": 0.9581, "mean_token_accuracy": 0.8896967768669128, "num_tokens": 717919703.0, "step": 18905 }, { "epoch": 2.405037527032184, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.152956008911133, "learning_rate": 5e-05, "loss": 0.9397, "mean_token_accuracy": 0.8918471336364746, "num_tokens": 717956877.0, "step": 18906 }, { "epoch": 2.4051647373107747, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.645045757293701, "learning_rate": 5e-05, "loss": 0.8449, "mean_token_accuracy": 0.8716253638267517, "num_tokens": 718002812.0, "step": 18907 }, { "epoch": 2.4052919475893653, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.449628829956055, "learning_rate": 5e-05, "loss": 0.9956, "mean_token_accuracy": 0.8821251392364502, "num_tokens": 718041170.0, "step": 18908 }, { "epoch": 2.405419157867956, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.260695457458496, "learning_rate": 5e-05, "loss": 0.7136, "mean_token_accuracy": 0.8758810758590698, "num_tokens": 718084168.0, "step": 18909 }, { "epoch": 2.4055463681465463, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.755460739135742, "learning_rate": 5e-05, "loss": 1.0828, "mean_token_accuracy": 0.8905401229858398, "num_tokens": 718118663.0, "step": 18910 }, { "epoch": 2.405673578425137, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.1630940437316895, "learning_rate": 5e-05, "loss": 0.8916, "mean_token_accuracy": 0.8814883232116699, "num_tokens": 718155490.0, "step": 18911 }, { "epoch": 2.4058007887037274, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.780981063842773, "learning_rate": 5e-05, "loss": 1.123, "mean_token_accuracy": 0.8774923086166382, "num_tokens": 718192301.0, "step": 18912 }, { "epoch": 2.405927998982318, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 8.917561531066895, "learning_rate": 5e-05, "loss": 1.1442, "mean_token_accuracy": 0.8906265497207642, "num_tokens": 718231390.0, "step": 18913 }, { "epoch": 2.4060552092609084, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.256996154785156, "learning_rate": 5e-05, "loss": 0.9131, "mean_token_accuracy": 0.8935935497283936, "num_tokens": 718269248.0, "step": 18914 }, { "epoch": 2.406182419539499, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 9.190474510192871, "learning_rate": 5e-05, "loss": 1.071, "mean_token_accuracy": 0.8840956687927246, "num_tokens": 718305961.0, "step": 18915 }, { "epoch": 2.4063096298180895, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.191771507263184, "learning_rate": 5e-05, "loss": 0.9296, "mean_token_accuracy": 0.8753107786178589, "num_tokens": 718344659.0, "step": 18916 }, { "epoch": 2.4064368400966796, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.892818450927734, "learning_rate": 5e-05, "loss": 1.0013, "mean_token_accuracy": 0.8663918972015381, "num_tokens": 718386624.0, "step": 18917 }, { "epoch": 2.4065640503752705, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.197691917419434, "learning_rate": 5e-05, "loss": 0.8971, "mean_token_accuracy": 0.894442081451416, "num_tokens": 718422154.0, "step": 18918 }, { "epoch": 2.4066912606538606, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.84505033493042, "learning_rate": 5e-05, "loss": 0.8529, "mean_token_accuracy": 0.8666245937347412, "num_tokens": 718460253.0, "step": 18919 }, { "epoch": 2.406818470932451, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.778886795043945, "learning_rate": 5e-05, "loss": 0.9478, "mean_token_accuracy": 0.8836417198181152, "num_tokens": 718495328.0, "step": 18920 }, { "epoch": 2.4069456812110417, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.926025867462158, "learning_rate": 5e-05, "loss": 0.7314, "mean_token_accuracy": 0.884484589099884, "num_tokens": 718531244.0, "step": 18921 }, { "epoch": 2.407072891489632, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.528217315673828, "learning_rate": 5e-05, "loss": 0.9754, "mean_token_accuracy": 0.882128119468689, "num_tokens": 718569766.0, "step": 18922 }, { "epoch": 2.4072001017682227, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.821168899536133, "learning_rate": 5e-05, "loss": 0.7978, "mean_token_accuracy": 0.8803873062133789, "num_tokens": 718612496.0, "step": 18923 }, { "epoch": 2.4073273120468133, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.94105052947998, "learning_rate": 5e-05, "loss": 1.0028, "mean_token_accuracy": 0.8835889101028442, "num_tokens": 718651251.0, "step": 18924 }, { "epoch": 2.407454522325404, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 7.49493932723999, "learning_rate": 5e-05, "loss": 0.9314, "mean_token_accuracy": 0.8781299591064453, "num_tokens": 718683204.0, "step": 18925 }, { "epoch": 2.4075817326039943, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.716020584106445, "learning_rate": 5e-05, "loss": 0.9785, "mean_token_accuracy": 0.890616774559021, "num_tokens": 718717114.0, "step": 18926 }, { "epoch": 2.407708942882585, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.0132474899292, "learning_rate": 5e-05, "loss": 0.9964, "mean_token_accuracy": 0.873395562171936, "num_tokens": 718754756.0, "step": 18927 }, { "epoch": 2.4078361531611754, "ewc_loss": 0.625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.108757972717285, "learning_rate": 5e-05, "loss": 0.9918, "mean_token_accuracy": 0.8738465309143066, "num_tokens": 718792312.0, "step": 18928 }, { "epoch": 2.407963363439766, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.991496562957764, "learning_rate": 5e-05, "loss": 0.9405, "mean_token_accuracy": 0.8773680329322815, "num_tokens": 718834771.0, "step": 18929 }, { "epoch": 2.4080905737183564, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.744494438171387, "learning_rate": 5e-05, "loss": 0.9581, "mean_token_accuracy": 0.8739372491836548, "num_tokens": 718871124.0, "step": 18930 }, { "epoch": 2.408217783996947, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.612517833709717, "learning_rate": 5e-05, "loss": 0.845, "mean_token_accuracy": 0.8800156712532043, "num_tokens": 718912037.0, "step": 18931 }, { "epoch": 2.4083449942755375, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.836912155151367, "learning_rate": 5e-05, "loss": 0.9728, "mean_token_accuracy": 0.870960533618927, "num_tokens": 718953048.0, "step": 18932 }, { "epoch": 2.408472204554128, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.316970348358154, "learning_rate": 5e-05, "loss": 0.8355, "mean_token_accuracy": 0.8737567663192749, "num_tokens": 718991825.0, "step": 18933 }, { "epoch": 2.4085994148327186, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.943968772888184, "learning_rate": 5e-05, "loss": 0.9548, "mean_token_accuracy": 0.8812552690505981, "num_tokens": 719026847.0, "step": 18934 }, { "epoch": 2.408726625111309, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.391051769256592, "learning_rate": 5e-05, "loss": 0.8007, "mean_token_accuracy": 0.887357234954834, "num_tokens": 719061415.0, "step": 18935 }, { "epoch": 2.4088538353898996, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.803630828857422, "learning_rate": 5e-05, "loss": 0.9316, "mean_token_accuracy": 0.8875410556793213, "num_tokens": 719097432.0, "step": 18936 }, { "epoch": 2.40898104566849, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.571408271789551, "learning_rate": 5e-05, "loss": 0.8387, "mean_token_accuracy": 0.8820375800132751, "num_tokens": 719129301.0, "step": 18937 }, { "epoch": 2.4091082559470807, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.551733016967773, "learning_rate": 5e-05, "loss": 0.9217, "mean_token_accuracy": 0.8810032606124878, "num_tokens": 719171772.0, "step": 18938 }, { "epoch": 2.409235466225671, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.622971534729004, "learning_rate": 5e-05, "loss": 0.8488, "mean_token_accuracy": 0.8760226964950562, "num_tokens": 719211751.0, "step": 18939 }, { "epoch": 2.4093626765042617, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.351369857788086, "learning_rate": 5e-05, "loss": 0.846, "mean_token_accuracy": 0.8992178440093994, "num_tokens": 719246812.0, "step": 18940 }, { "epoch": 2.4094898867828523, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.608761787414551, "learning_rate": 5e-05, "loss": 0.8812, "mean_token_accuracy": 0.8627997636795044, "num_tokens": 719290962.0, "step": 18941 }, { "epoch": 2.4096170970614423, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.198578834533691, "learning_rate": 5e-05, "loss": 0.8816, "mean_token_accuracy": 0.8783031105995178, "num_tokens": 719323506.0, "step": 18942 }, { "epoch": 2.4097443073400333, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.508116722106934, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.8729478120803833, "num_tokens": 719361435.0, "step": 18943 }, { "epoch": 2.4098715176186234, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.202402114868164, "learning_rate": 5e-05, "loss": 0.8274, "mean_token_accuracy": 0.8902888298034668, "num_tokens": 719405617.0, "step": 18944 }, { "epoch": 2.409998727897214, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.360288619995117, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8811267018318176, "num_tokens": 719450588.0, "step": 18945 }, { "epoch": 2.4101259381758044, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.267143249511719, "learning_rate": 5e-05, "loss": 0.8233, "mean_token_accuracy": 0.8904277086257935, "num_tokens": 719489028.0, "step": 18946 }, { "epoch": 2.410253148454395, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.297414779663086, "learning_rate": 5e-05, "loss": 0.7778, "mean_token_accuracy": 0.8766587972640991, "num_tokens": 719527921.0, "step": 18947 }, { "epoch": 2.4103803587329855, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.221258163452148, "learning_rate": 5e-05, "loss": 0.8466, "mean_token_accuracy": 0.8817183375358582, "num_tokens": 719572461.0, "step": 18948 }, { "epoch": 2.410507569011576, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.319299697875977, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8793588876724243, "num_tokens": 719613478.0, "step": 18949 }, { "epoch": 2.4106347792901666, "ewc_loss": 0.5, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.16899585723877, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.8912348747253418, "num_tokens": 719643665.0, "step": 18950 }, { "epoch": 2.410761989568757, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.3075337409973145, "learning_rate": 5e-05, "loss": 0.7683, "mean_token_accuracy": 0.8726911544799805, "num_tokens": 719684168.0, "step": 18951 }, { "epoch": 2.4108891998473476, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.118157386779785, "learning_rate": 5e-05, "loss": 0.8219, "mean_token_accuracy": 0.8811631202697754, "num_tokens": 719722484.0, "step": 18952 }, { "epoch": 2.411016410125938, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.361584186553955, "learning_rate": 5e-05, "loss": 0.7355, "mean_token_accuracy": 0.8838942050933838, "num_tokens": 719756659.0, "step": 18953 }, { "epoch": 2.4111436204045287, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.9709038734436035, "learning_rate": 5e-05, "loss": 0.7958, "mean_token_accuracy": 0.8839943408966064, "num_tokens": 719787302.0, "step": 18954 }, { "epoch": 2.411270830683119, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.445406913757324, "learning_rate": 5e-05, "loss": 0.7455, "mean_token_accuracy": 0.8818665742874146, "num_tokens": 719823359.0, "step": 18955 }, { "epoch": 2.4113980409617097, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.817552089691162, "learning_rate": 5e-05, "loss": 0.7723, "mean_token_accuracy": 0.8837798833847046, "num_tokens": 719861263.0, "step": 18956 }, { "epoch": 2.4115252512403003, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.529548645019531, "learning_rate": 5e-05, "loss": 0.732, "mean_token_accuracy": 0.8846707940101624, "num_tokens": 719897987.0, "step": 18957 }, { "epoch": 2.411652461518891, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.682899475097656, "learning_rate": 5e-05, "loss": 0.7333, "mean_token_accuracy": 0.889592170715332, "num_tokens": 719931067.0, "step": 18958 }, { "epoch": 2.4117796717974813, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.568580150604248, "learning_rate": 5e-05, "loss": 0.7108, "mean_token_accuracy": 0.8933292627334595, "num_tokens": 719971174.0, "step": 18959 }, { "epoch": 2.411906882076072, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.601274490356445, "learning_rate": 5e-05, "loss": 0.7272, "mean_token_accuracy": 0.8853086829185486, "num_tokens": 720005781.0, "step": 18960 }, { "epoch": 2.4120340923546624, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.591650485992432, "learning_rate": 5e-05, "loss": 0.7042, "mean_token_accuracy": 0.8872033357620239, "num_tokens": 720040917.0, "step": 18961 }, { "epoch": 2.412161302633253, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.5423431396484375, "learning_rate": 5e-05, "loss": 0.7586, "mean_token_accuracy": 0.8698223233222961, "num_tokens": 720080490.0, "step": 18962 }, { "epoch": 2.4122885129118434, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.60998010635376, "learning_rate": 5e-05, "loss": 0.7275, "mean_token_accuracy": 0.8833169937133789, "num_tokens": 720116461.0, "step": 18963 }, { "epoch": 2.412415723190434, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.485410690307617, "learning_rate": 5e-05, "loss": 0.6565, "mean_token_accuracy": 0.9013427495956421, "num_tokens": 720156756.0, "step": 18964 }, { "epoch": 2.4125429334690245, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.652005195617676, "learning_rate": 5e-05, "loss": 0.7345, "mean_token_accuracy": 0.8809452056884766, "num_tokens": 720195271.0, "step": 18965 }, { "epoch": 2.412670143747615, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.397504806518555, "learning_rate": 5e-05, "loss": 0.6743, "mean_token_accuracy": 0.8914311528205872, "num_tokens": 720232159.0, "step": 18966 }, { "epoch": 2.412797354026205, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.719024181365967, "learning_rate": 5e-05, "loss": 0.716, "mean_token_accuracy": 0.8867667317390442, "num_tokens": 720264591.0, "step": 18967 }, { "epoch": 2.412924564304796, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.299742221832275, "learning_rate": 5e-05, "loss": 0.6856, "mean_token_accuracy": 0.8839748501777649, "num_tokens": 720302957.0, "step": 18968 }, { "epoch": 2.413051774583386, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.816704273223877, "learning_rate": 5e-05, "loss": 0.7597, "mean_token_accuracy": 0.874241054058075, "num_tokens": 720335679.0, "step": 18969 }, { "epoch": 2.4131789848619767, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.172454833984375, "learning_rate": 5e-05, "loss": 0.7025, "mean_token_accuracy": 0.8782894015312195, "num_tokens": 720377261.0, "step": 18970 }, { "epoch": 2.413306195140567, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 8.003604888916016, "learning_rate": 5e-05, "loss": 0.7383, "mean_token_accuracy": 0.8813618421554565, "num_tokens": 720415344.0, "step": 18971 }, { "epoch": 2.4134334054191577, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.941442012786865, "learning_rate": 5e-05, "loss": 0.6842, "mean_token_accuracy": 0.876227617263794, "num_tokens": 720449133.0, "step": 18972 }, { "epoch": 2.4135606156977483, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.310103416442871, "learning_rate": 5e-05, "loss": 0.7984, "mean_token_accuracy": 0.8723062872886658, "num_tokens": 720484250.0, "step": 18973 }, { "epoch": 2.413687825976339, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.580090522766113, "learning_rate": 5e-05, "loss": 0.6579, "mean_token_accuracy": 0.8769658207893372, "num_tokens": 720520466.0, "step": 18974 }, { "epoch": 2.4138150362549293, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.858490943908691, "learning_rate": 5e-05, "loss": 0.8181, "mean_token_accuracy": 0.8850289583206177, "num_tokens": 720557013.0, "step": 18975 }, { "epoch": 2.41394224653352, "ewc_loss": 0.302734375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002307891845703125, "grad_norm": 6.045898914337158, "learning_rate": 5e-05, "loss": 0.637, "mean_token_accuracy": 0.8847978115081787, "num_tokens": 720594957.0, "step": 18976 }, { "epoch": 2.4140694568121104, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.760568618774414, "learning_rate": 5e-05, "loss": 0.9429, "mean_token_accuracy": 0.879789412021637, "num_tokens": 720630531.0, "step": 18977 }, { "epoch": 2.414196667090701, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 5.771495819091797, "learning_rate": 5e-05, "loss": 0.692, "mean_token_accuracy": 0.8765976428985596, "num_tokens": 720673502.0, "step": 18978 }, { "epoch": 2.4143238773692914, "ewc_loss": 0.734375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.54567813873291, "learning_rate": 5e-05, "loss": 1.073, "mean_token_accuracy": 0.8776513934135437, "num_tokens": 720716765.0, "step": 18979 }, { "epoch": 2.414451087647882, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.09091854095459, "learning_rate": 5e-05, "loss": 0.8463, "mean_token_accuracy": 0.8922125101089478, "num_tokens": 720754711.0, "step": 18980 }, { "epoch": 2.4145782979264725, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.86216926574707, "learning_rate": 5e-05, "loss": 1.0182, "mean_token_accuracy": 0.8793503642082214, "num_tokens": 720789067.0, "step": 18981 }, { "epoch": 2.414705508205063, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 8.911006927490234, "learning_rate": 5e-05, "loss": 1.113, "mean_token_accuracy": 0.8734622001647949, "num_tokens": 720821725.0, "step": 18982 }, { "epoch": 2.4148327184836536, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.395519256591797, "learning_rate": 5e-05, "loss": 0.7645, "mean_token_accuracy": 0.8850364685058594, "num_tokens": 720857587.0, "step": 18983 }, { "epoch": 2.414959928762244, "ewc_loss": 0.8515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 10.914819717407227, "learning_rate": 5e-05, "loss": 1.1824, "mean_token_accuracy": 0.8843997120857239, "num_tokens": 720900481.0, "step": 18984 }, { "epoch": 2.4150871390408346, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 6.711435317993164, "learning_rate": 5e-05, "loss": 0.7927, "mean_token_accuracy": 0.8814653158187866, "num_tokens": 720933131.0, "step": 18985 }, { "epoch": 2.415214349319425, "ewc_loss": 1.1015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00102996826171875, "grad_norm": 13.10844612121582, "learning_rate": 5e-05, "loss": 1.4426, "mean_token_accuracy": 0.8787378072738647, "num_tokens": 720972142.0, "step": 18986 }, { "epoch": 2.4153415595980157, "ewc_loss": 1.09375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0010223388671875, "grad_norm": 10.511761665344238, "learning_rate": 5e-05, "loss": 1.4036, "mean_token_accuracy": 0.8895776867866516, "num_tokens": 721014251.0, "step": 18987 }, { "epoch": 2.415468769876606, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 11.604043960571289, "learning_rate": 5e-05, "loss": 1.126, "mean_token_accuracy": 0.8703169822692871, "num_tokens": 721053327.0, "step": 18988 }, { "epoch": 2.4155959801551967, "ewc_loss": 0.89453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 9.82402229309082, "learning_rate": 5e-05, "loss": 1.2254, "mean_token_accuracy": 0.8846591114997864, "num_tokens": 721092047.0, "step": 18989 }, { "epoch": 2.415723190433787, "ewc_loss": 0.86328125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 11.643096923828125, "learning_rate": 5e-05, "loss": 1.1471, "mean_token_accuracy": 0.8973842263221741, "num_tokens": 721125318.0, "step": 18990 }, { "epoch": 2.4158504007123778, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.769987106323242, "learning_rate": 5e-05, "loss": 0.9306, "mean_token_accuracy": 0.89476478099823, "num_tokens": 721162058.0, "step": 18991 }, { "epoch": 2.415977610990968, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000835418701171875, "grad_norm": 11.887617111206055, "learning_rate": 5e-05, "loss": 1.2417, "mean_token_accuracy": 0.8844796419143677, "num_tokens": 721199219.0, "step": 18992 }, { "epoch": 2.4161048212695584, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 9.354851722717285, "learning_rate": 5e-05, "loss": 1.0503, "mean_token_accuracy": 0.8769331574440002, "num_tokens": 721239696.0, "step": 18993 }, { "epoch": 2.416232031548149, "ewc_loss": 0.828125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00075531005859375, "grad_norm": 11.680712699890137, "learning_rate": 5e-05, "loss": 1.1896, "mean_token_accuracy": 0.8721040487289429, "num_tokens": 721277660.0, "step": 18994 }, { "epoch": 2.4163592418267394, "ewc_loss": 0.796875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 10.016702651977539, "learning_rate": 5e-05, "loss": 1.1356, "mean_token_accuracy": 0.8804366588592529, "num_tokens": 721311805.0, "step": 18995 }, { "epoch": 2.41648645210533, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.882511138916016, "learning_rate": 5e-05, "loss": 1.1103, "mean_token_accuracy": 0.8627840876579285, "num_tokens": 721355968.0, "step": 18996 }, { "epoch": 2.4166136623839205, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.549884796142578, "learning_rate": 5e-05, "loss": 1.0421, "mean_token_accuracy": 0.8743327856063843, "num_tokens": 721397930.0, "step": 18997 }, { "epoch": 2.416740872662511, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.430641174316406, "learning_rate": 5e-05, "loss": 1.0047, "mean_token_accuracy": 0.8908109664916992, "num_tokens": 721432408.0, "step": 18998 }, { "epoch": 2.4168680829411016, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.967552185058594, "learning_rate": 5e-05, "loss": 0.8862, "mean_token_accuracy": 0.8997930288314819, "num_tokens": 721468903.0, "step": 18999 }, { "epoch": 2.416995293219692, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.261908531188965, "learning_rate": 5e-05, "loss": 0.9725, "mean_token_accuracy": 0.8966547250747681, "num_tokens": 721509687.0, "step": 19000 }, { "epoch": 2.4171225034982826, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.547106742858887, "learning_rate": 5e-05, "loss": 0.9122, "mean_token_accuracy": 0.8741576075553894, "num_tokens": 721548130.0, "step": 19001 }, { "epoch": 2.417249713776873, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.402741432189941, "learning_rate": 5e-05, "loss": 1.0209, "mean_token_accuracy": 0.8889002203941345, "num_tokens": 721588504.0, "step": 19002 }, { "epoch": 2.4173769240554637, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.482281684875488, "learning_rate": 5e-05, "loss": 0.9049, "mean_token_accuracy": 0.8801044225692749, "num_tokens": 721632621.0, "step": 19003 }, { "epoch": 2.417504134334054, "ewc_loss": 0.71875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.594231605529785, "learning_rate": 5e-05, "loss": 1.0634, "mean_token_accuracy": 0.8795948028564453, "num_tokens": 721670898.0, "step": 19004 }, { "epoch": 2.4176313446126447, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.766183853149414, "learning_rate": 5e-05, "loss": 0.9063, "mean_token_accuracy": 0.8889290690422058, "num_tokens": 721709012.0, "step": 19005 }, { "epoch": 2.4177585548912353, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.46677303314209, "learning_rate": 5e-05, "loss": 1.06, "mean_token_accuracy": 0.8807406425476074, "num_tokens": 721747135.0, "step": 19006 }, { "epoch": 2.417885765169826, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.046463012695312, "learning_rate": 5e-05, "loss": 0.9434, "mean_token_accuracy": 0.8805781006813049, "num_tokens": 721784447.0, "step": 19007 }, { "epoch": 2.4180129754484163, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.909103393554688, "learning_rate": 5e-05, "loss": 1.0184, "mean_token_accuracy": 0.8725922107696533, "num_tokens": 721817853.0, "step": 19008 }, { "epoch": 2.418140185727007, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.905714988708496, "learning_rate": 5e-05, "loss": 0.9012, "mean_token_accuracy": 0.8810437321662903, "num_tokens": 721852537.0, "step": 19009 }, { "epoch": 2.4182673960055974, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.359514236450195, "learning_rate": 5e-05, "loss": 0.9372, "mean_token_accuracy": 0.8843963146209717, "num_tokens": 721892156.0, "step": 19010 }, { "epoch": 2.418394606284188, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.421740531921387, "learning_rate": 5e-05, "loss": 0.8467, "mean_token_accuracy": 0.8826830387115479, "num_tokens": 721929425.0, "step": 19011 }, { "epoch": 2.4185218165627784, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.267684936523438, "learning_rate": 5e-05, "loss": 0.949, "mean_token_accuracy": 0.8741297721862793, "num_tokens": 721967492.0, "step": 19012 }, { "epoch": 2.418649026841369, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.905208587646484, "learning_rate": 5e-05, "loss": 0.8044, "mean_token_accuracy": 0.877253532409668, "num_tokens": 722007754.0, "step": 19013 }, { "epoch": 2.4187762371199595, "ewc_loss": 0.625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.575807571411133, "learning_rate": 5e-05, "loss": 0.9576, "mean_token_accuracy": 0.8831440210342407, "num_tokens": 722043380.0, "step": 19014 }, { "epoch": 2.4189034473985496, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.780883312225342, "learning_rate": 5e-05, "loss": 0.8024, "mean_token_accuracy": 0.87620609998703, "num_tokens": 722086200.0, "step": 19015 }, { "epoch": 2.4190306576771405, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.86447525024414, "learning_rate": 5e-05, "loss": 0.9851, "mean_token_accuracy": 0.8862943053245544, "num_tokens": 722122907.0, "step": 19016 }, { "epoch": 2.4191578679557306, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.215316772460938, "learning_rate": 5e-05, "loss": 0.825, "mean_token_accuracy": 0.8841860294342041, "num_tokens": 722155951.0, "step": 19017 }, { "epoch": 2.419285078234321, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.533428192138672, "learning_rate": 5e-05, "loss": 0.9821, "mean_token_accuracy": 0.8772045373916626, "num_tokens": 722194479.0, "step": 19018 }, { "epoch": 2.4194122885129117, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.55471420288086, "learning_rate": 5e-05, "loss": 0.8619, "mean_token_accuracy": 0.8860706686973572, "num_tokens": 722228470.0, "step": 19019 }, { "epoch": 2.419539498791502, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.754916191101074, "learning_rate": 5e-05, "loss": 0.9211, "mean_token_accuracy": 0.8754513263702393, "num_tokens": 722265960.0, "step": 19020 }, { "epoch": 2.4196667090700927, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.321033477783203, "learning_rate": 5e-05, "loss": 0.8491, "mean_token_accuracy": 0.8777393102645874, "num_tokens": 722301377.0, "step": 19021 }, { "epoch": 2.4197939193486833, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.3267183303833, "learning_rate": 5e-05, "loss": 0.8295, "mean_token_accuracy": 0.8853991031646729, "num_tokens": 722338534.0, "step": 19022 }, { "epoch": 2.419921129627274, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.828711032867432, "learning_rate": 5e-05, "loss": 0.7733, "mean_token_accuracy": 0.8783006072044373, "num_tokens": 722377460.0, "step": 19023 }, { "epoch": 2.4200483399058643, "ewc_loss": 0.5, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.409785270690918, "learning_rate": 5e-05, "loss": 0.8708, "mean_token_accuracy": 0.8716437816619873, "num_tokens": 722415167.0, "step": 19024 }, { "epoch": 2.420175550184455, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.415107250213623, "learning_rate": 5e-05, "loss": 0.7757, "mean_token_accuracy": 0.8687310218811035, "num_tokens": 722456114.0, "step": 19025 }, { "epoch": 2.4203027604630454, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.578169822692871, "learning_rate": 5e-05, "loss": 0.8264, "mean_token_accuracy": 0.8865292072296143, "num_tokens": 722496378.0, "step": 19026 }, { "epoch": 2.420429970741636, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.2621893882751465, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.8725920915603638, "num_tokens": 722534903.0, "step": 19027 }, { "epoch": 2.4205571810202264, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.66946792602539, "learning_rate": 5e-05, "loss": 0.8697, "mean_token_accuracy": 0.8760781288146973, "num_tokens": 722576649.0, "step": 19028 }, { "epoch": 2.420684391298817, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.213849067687988, "learning_rate": 5e-05, "loss": 0.7264, "mean_token_accuracy": 0.8735653162002563, "num_tokens": 722618555.0, "step": 19029 }, { "epoch": 2.4208116015774075, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.76334285736084, "learning_rate": 5e-05, "loss": 0.8105, "mean_token_accuracy": 0.8991129994392395, "num_tokens": 722652045.0, "step": 19030 }, { "epoch": 2.420938811855998, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.341422080993652, "learning_rate": 5e-05, "loss": 0.7341, "mean_token_accuracy": 0.8787795305252075, "num_tokens": 722686987.0, "step": 19031 }, { "epoch": 2.4210660221345885, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.59055233001709, "learning_rate": 5e-05, "loss": 0.8653, "mean_token_accuracy": 0.8760012984275818, "num_tokens": 722724161.0, "step": 19032 }, { "epoch": 2.421193232413179, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.622326850891113, "learning_rate": 5e-05, "loss": 0.7455, "mean_token_accuracy": 0.8855876922607422, "num_tokens": 722766473.0, "step": 19033 }, { "epoch": 2.4213204426917696, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.082164764404297, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8826499581336975, "num_tokens": 722809499.0, "step": 19034 }, { "epoch": 2.42144765297036, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.8910088539123535, "learning_rate": 5e-05, "loss": 0.7519, "mean_token_accuracy": 0.8861925005912781, "num_tokens": 722847295.0, "step": 19035 }, { "epoch": 2.4215748632489507, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.549885272979736, "learning_rate": 5e-05, "loss": 0.7324, "mean_token_accuracy": 0.8822359442710876, "num_tokens": 722884313.0, "step": 19036 }, { "epoch": 2.421702073527541, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.060542106628418, "learning_rate": 5e-05, "loss": 0.7544, "mean_token_accuracy": 0.8848057985305786, "num_tokens": 722924530.0, "step": 19037 }, { "epoch": 2.4218292838061317, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.254817485809326, "learning_rate": 5e-05, "loss": 0.7328, "mean_token_accuracy": 0.8711287975311279, "num_tokens": 722962968.0, "step": 19038 }, { "epoch": 2.4219564940847222, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.247268676757812, "learning_rate": 5e-05, "loss": 0.7404, "mean_token_accuracy": 0.8912694454193115, "num_tokens": 722997801.0, "step": 19039 }, { "epoch": 2.4220837043633123, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.028163433074951, "learning_rate": 5e-05, "loss": 0.7235, "mean_token_accuracy": 0.8708338737487793, "num_tokens": 723037724.0, "step": 19040 }, { "epoch": 2.4222109146419033, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.46902084350586, "learning_rate": 5e-05, "loss": 0.8427, "mean_token_accuracy": 0.8714650869369507, "num_tokens": 723078199.0, "step": 19041 }, { "epoch": 2.4223381249204934, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.896086692810059, "learning_rate": 5e-05, "loss": 0.7081, "mean_token_accuracy": 0.8741657137870789, "num_tokens": 723110894.0, "step": 19042 }, { "epoch": 2.422465335199084, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.560521125793457, "learning_rate": 5e-05, "loss": 0.8475, "mean_token_accuracy": 0.8711009621620178, "num_tokens": 723150688.0, "step": 19043 }, { "epoch": 2.4225925454776744, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.031030654907227, "learning_rate": 5e-05, "loss": 0.6956, "mean_token_accuracy": 0.8810831308364868, "num_tokens": 723189863.0, "step": 19044 }, { "epoch": 2.422719755756265, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.348398208618164, "learning_rate": 5e-05, "loss": 0.8238, "mean_token_accuracy": 0.8719082474708557, "num_tokens": 723232333.0, "step": 19045 }, { "epoch": 2.4228469660348555, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.408908367156982, "learning_rate": 5e-05, "loss": 0.7759, "mean_token_accuracy": 0.8661891222000122, "num_tokens": 723267469.0, "step": 19046 }, { "epoch": 2.422974176313446, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.7983574867248535, "learning_rate": 5e-05, "loss": 0.7086, "mean_token_accuracy": 0.8915676474571228, "num_tokens": 723304367.0, "step": 19047 }, { "epoch": 2.4231013865920366, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.822934627532959, "learning_rate": 5e-05, "loss": 0.7423, "mean_token_accuracy": 0.8783677220344543, "num_tokens": 723343247.0, "step": 19048 }, { "epoch": 2.423228596870627, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.358771324157715, "learning_rate": 5e-05, "loss": 0.6857, "mean_token_accuracy": 0.8909852504730225, "num_tokens": 723378022.0, "step": 19049 }, { "epoch": 2.4233558071492176, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.089942932128906, "learning_rate": 5e-05, "loss": 0.7456, "mean_token_accuracy": 0.8835026621818542, "num_tokens": 723409827.0, "step": 19050 }, { "epoch": 2.423483017427808, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.142784118652344, "learning_rate": 5e-05, "loss": 0.6692, "mean_token_accuracy": 0.8900356292724609, "num_tokens": 723447192.0, "step": 19051 }, { "epoch": 2.4236102277063987, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.173763275146484, "learning_rate": 5e-05, "loss": 0.7423, "mean_token_accuracy": 0.8867785334587097, "num_tokens": 723492123.0, "step": 19052 }, { "epoch": 2.423737437984989, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.145076751708984, "learning_rate": 5e-05, "loss": 0.6464, "mean_token_accuracy": 0.8963479995727539, "num_tokens": 723525680.0, "step": 19053 }, { "epoch": 2.4238646482635797, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.085463523864746, "learning_rate": 5e-05, "loss": 0.7635, "mean_token_accuracy": 0.8793238997459412, "num_tokens": 723565667.0, "step": 19054 }, { "epoch": 2.4239918585421703, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.289370536804199, "learning_rate": 5e-05, "loss": 0.7126, "mean_token_accuracy": 0.8751997351646423, "num_tokens": 723599486.0, "step": 19055 }, { "epoch": 2.424119068820761, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000720977783203125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.849766731262207, "learning_rate": 5e-05, "loss": 0.7085, "mean_token_accuracy": 0.8867788314819336, "num_tokens": 723631992.0, "step": 19056 }, { "epoch": 2.4242462790993513, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.518762111663818, "learning_rate": 5e-05, "loss": 0.6731, "mean_token_accuracy": 0.8905715942382812, "num_tokens": 723667654.0, "step": 19057 }, { "epoch": 2.424373489377942, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.554211139678955, "learning_rate": 5e-05, "loss": 0.6864, "mean_token_accuracy": 0.8848375082015991, "num_tokens": 723705368.0, "step": 19058 }, { "epoch": 2.4245006996565324, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.784224987030029, "learning_rate": 5e-05, "loss": 0.7022, "mean_token_accuracy": 0.884680449962616, "num_tokens": 723744973.0, "step": 19059 }, { "epoch": 2.424627909935123, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.31577205657959, "learning_rate": 5e-05, "loss": 0.7227, "mean_token_accuracy": 0.8709396123886108, "num_tokens": 723778868.0, "step": 19060 }, { "epoch": 2.4247551202137134, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.959571361541748, "learning_rate": 5e-05, "loss": 0.7325, "mean_token_accuracy": 0.8823642134666443, "num_tokens": 723822413.0, "step": 19061 }, { "epoch": 2.424882330492304, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 7.1295552253723145, "learning_rate": 5e-05, "loss": 0.6994, "mean_token_accuracy": 0.8729745149612427, "num_tokens": 723865826.0, "step": 19062 }, { "epoch": 2.4250095407708945, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.138627052307129, "learning_rate": 5e-05, "loss": 0.7328, "mean_token_accuracy": 0.8882930874824524, "num_tokens": 723901839.0, "step": 19063 }, { "epoch": 2.425136751049485, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.961358070373535, "learning_rate": 5e-05, "loss": 0.6587, "mean_token_accuracy": 0.8890354037284851, "num_tokens": 723944883.0, "step": 19064 }, { "epoch": 2.425263961328075, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.334159851074219, "learning_rate": 5e-05, "loss": 0.7564, "mean_token_accuracy": 0.8870799541473389, "num_tokens": 723985084.0, "step": 19065 }, { "epoch": 2.425391171606666, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00025177001953125, "grad_norm": 6.735387802124023, "learning_rate": 5e-05, "loss": 0.6679, "mean_token_accuracy": 0.8810274004936218, "num_tokens": 724027919.0, "step": 19066 }, { "epoch": 2.425518381885256, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.676712989807129, "learning_rate": 5e-05, "loss": 0.7881, "mean_token_accuracy": 0.8853542804718018, "num_tokens": 724064593.0, "step": 19067 }, { "epoch": 2.4256455921638467, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.407686233520508, "learning_rate": 5e-05, "loss": 0.6229, "mean_token_accuracy": 0.890580415725708, "num_tokens": 724095967.0, "step": 19068 }, { "epoch": 2.425772802442437, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 9.198625564575195, "learning_rate": 5e-05, "loss": 0.8189, "mean_token_accuracy": 0.8951420783996582, "num_tokens": 724128420.0, "step": 19069 }, { "epoch": 2.4259000127210277, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.105124473571777, "learning_rate": 5e-05, "loss": 0.681, "mean_token_accuracy": 0.8716413378715515, "num_tokens": 724165073.0, "step": 19070 }, { "epoch": 2.4260272229996183, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.777597427368164, "learning_rate": 5e-05, "loss": 0.9637, "mean_token_accuracy": 0.879798173904419, "num_tokens": 724201377.0, "step": 19071 }, { "epoch": 2.426154433278209, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.360005855560303, "learning_rate": 5e-05, "loss": 0.6633, "mean_token_accuracy": 0.8966067433357239, "num_tokens": 724241783.0, "step": 19072 }, { "epoch": 2.4262816435567993, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.783635139465332, "learning_rate": 5e-05, "loss": 0.9804, "mean_token_accuracy": 0.880373477935791, "num_tokens": 724277903.0, "step": 19073 }, { "epoch": 2.42640885383539, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 7.56990385055542, "learning_rate": 5e-05, "loss": 0.848, "mean_token_accuracy": 0.8887119293212891, "num_tokens": 724318834.0, "step": 19074 }, { "epoch": 2.4265360641139804, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.482259750366211, "learning_rate": 5e-05, "loss": 0.8869, "mean_token_accuracy": 0.8836366534233093, "num_tokens": 724356406.0, "step": 19075 }, { "epoch": 2.426663274392571, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.77770709991455, "learning_rate": 5e-05, "loss": 0.9248, "mean_token_accuracy": 0.8903166055679321, "num_tokens": 724394774.0, "step": 19076 }, { "epoch": 2.4267904846711614, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.774518966674805, "learning_rate": 5e-05, "loss": 0.7391, "mean_token_accuracy": 0.8759822249412537, "num_tokens": 724431825.0, "step": 19077 }, { "epoch": 2.426917694949752, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.816628456115723, "learning_rate": 5e-05, "loss": 0.9833, "mean_token_accuracy": 0.8807518482208252, "num_tokens": 724466402.0, "step": 19078 }, { "epoch": 2.4270449052283425, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.413539409637451, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.8722182512283325, "num_tokens": 724503353.0, "step": 19079 }, { "epoch": 2.427172115506933, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.094294548034668, "learning_rate": 5e-05, "loss": 1.0006, "mean_token_accuracy": 0.8870910406112671, "num_tokens": 724540348.0, "step": 19080 }, { "epoch": 2.4272993257855235, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.544544219970703, "learning_rate": 5e-05, "loss": 0.9288, "mean_token_accuracy": 0.8786980509757996, "num_tokens": 724577999.0, "step": 19081 }, { "epoch": 2.427426536064114, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.09915828704834, "learning_rate": 5e-05, "loss": 0.901, "mean_token_accuracy": 0.8951929807662964, "num_tokens": 724614754.0, "step": 19082 }, { "epoch": 2.4275537463427046, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.504631996154785, "learning_rate": 5e-05, "loss": 0.9953, "mean_token_accuracy": 0.8715633153915405, "num_tokens": 724653232.0, "step": 19083 }, { "epoch": 2.427680956621295, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.832393646240234, "learning_rate": 5e-05, "loss": 0.8534, "mean_token_accuracy": 0.8728963136672974, "num_tokens": 724686461.0, "step": 19084 }, { "epoch": 2.4278081668998857, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.71477222442627, "learning_rate": 5e-05, "loss": 0.9097, "mean_token_accuracy": 0.8914245367050171, "num_tokens": 724727330.0, "step": 19085 }, { "epoch": 2.427935377178476, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.574874401092529, "learning_rate": 5e-05, "loss": 0.8437, "mean_token_accuracy": 0.872272789478302, "num_tokens": 724766509.0, "step": 19086 }, { "epoch": 2.4280625874570667, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.193380355834961, "learning_rate": 5e-05, "loss": 0.8087, "mean_token_accuracy": 0.896843671798706, "num_tokens": 724795128.0, "step": 19087 }, { "epoch": 2.428189797735657, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.921266078948975, "learning_rate": 5e-05, "loss": 0.8519, "mean_token_accuracy": 0.8741279244422913, "num_tokens": 724833941.0, "step": 19088 }, { "epoch": 2.4283170080142478, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.505636692047119, "learning_rate": 5e-05, "loss": 0.7473, "mean_token_accuracy": 0.8873164653778076, "num_tokens": 724870429.0, "step": 19089 }, { "epoch": 2.428444218292838, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.273900985717773, "learning_rate": 5e-05, "loss": 0.8325, "mean_token_accuracy": 0.883202075958252, "num_tokens": 724905574.0, "step": 19090 }, { "epoch": 2.4285714285714284, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.136011123657227, "learning_rate": 5e-05, "loss": 0.7049, "mean_token_accuracy": 0.887018084526062, "num_tokens": 724949845.0, "step": 19091 }, { "epoch": 2.428698638850019, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.485033988952637, "learning_rate": 5e-05, "loss": 0.8876, "mean_token_accuracy": 0.8734500408172607, "num_tokens": 724987586.0, "step": 19092 }, { "epoch": 2.4288258491286094, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.947065830230713, "learning_rate": 5e-05, "loss": 0.6863, "mean_token_accuracy": 0.8869801759719849, "num_tokens": 725026432.0, "step": 19093 }, { "epoch": 2.4289530594072, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.713582038879395, "learning_rate": 5e-05, "loss": 0.9309, "mean_token_accuracy": 0.8632841110229492, "num_tokens": 725070627.0, "step": 19094 }, { "epoch": 2.4290802696857905, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.8145976066589355, "learning_rate": 5e-05, "loss": 0.7214, "mean_token_accuracy": 0.8831502199172974, "num_tokens": 725106005.0, "step": 19095 }, { "epoch": 2.429207479964381, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.884669303894043, "learning_rate": 5e-05, "loss": 0.8711, "mean_token_accuracy": 0.8852275013923645, "num_tokens": 725149667.0, "step": 19096 }, { "epoch": 2.4293346902429716, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.832541465759277, "learning_rate": 5e-05, "loss": 0.7415, "mean_token_accuracy": 0.8853886127471924, "num_tokens": 725185526.0, "step": 19097 }, { "epoch": 2.429461900521562, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.870503425598145, "learning_rate": 5e-05, "loss": 0.8889, "mean_token_accuracy": 0.8861855268478394, "num_tokens": 725221789.0, "step": 19098 }, { "epoch": 2.4295891108001526, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.968233108520508, "learning_rate": 5e-05, "loss": 0.777, "mean_token_accuracy": 0.876604437828064, "num_tokens": 725258275.0, "step": 19099 }, { "epoch": 2.429716321078743, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.804635047912598, "learning_rate": 5e-05, "loss": 0.9421, "mean_token_accuracy": 0.8685689568519592, "num_tokens": 725297753.0, "step": 19100 }, { "epoch": 2.4298435313573337, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.044125556945801, "learning_rate": 5e-05, "loss": 0.745, "mean_token_accuracy": 0.8873295187950134, "num_tokens": 725331978.0, "step": 19101 }, { "epoch": 2.429970741635924, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.823237419128418, "learning_rate": 5e-05, "loss": 0.9117, "mean_token_accuracy": 0.8847129344940186, "num_tokens": 725373876.0, "step": 19102 }, { "epoch": 2.4300979519145147, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.123050689697266, "learning_rate": 5e-05, "loss": 0.759, "mean_token_accuracy": 0.8852943181991577, "num_tokens": 725411280.0, "step": 19103 }, { "epoch": 2.4302251621931052, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.712940216064453, "learning_rate": 5e-05, "loss": 0.9431, "mean_token_accuracy": 0.8700487613677979, "num_tokens": 725452050.0, "step": 19104 }, { "epoch": 2.4303523724716958, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.347164630889893, "learning_rate": 5e-05, "loss": 0.804, "mean_token_accuracy": 0.8883416652679443, "num_tokens": 725489922.0, "step": 19105 }, { "epoch": 2.4304795827502863, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.360600471496582, "learning_rate": 5e-05, "loss": 0.8752, "mean_token_accuracy": 0.8786118626594543, "num_tokens": 725527871.0, "step": 19106 }, { "epoch": 2.430606793028877, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.590786457061768, "learning_rate": 5e-05, "loss": 0.8072, "mean_token_accuracy": 0.8912506699562073, "num_tokens": 725565045.0, "step": 19107 }, { "epoch": 2.4307340033074674, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.977146148681641, "learning_rate": 5e-05, "loss": 0.805, "mean_token_accuracy": 0.8945648670196533, "num_tokens": 725596604.0, "step": 19108 }, { "epoch": 2.430861213586058, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.706311225891113, "learning_rate": 5e-05, "loss": 0.8004, "mean_token_accuracy": 0.8826163411140442, "num_tokens": 725632299.0, "step": 19109 }, { "epoch": 2.4309884238646484, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.7425432205200195, "learning_rate": 5e-05, "loss": 0.7881, "mean_token_accuracy": 0.8816620707511902, "num_tokens": 725673778.0, "step": 19110 }, { "epoch": 2.431115634143239, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.722541809082031, "learning_rate": 5e-05, "loss": 0.8235, "mean_token_accuracy": 0.8703936338424683, "num_tokens": 725715087.0, "step": 19111 }, { "epoch": 2.4312428444218295, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.633686065673828, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8839268684387207, "num_tokens": 725754047.0, "step": 19112 }, { "epoch": 2.4313700547004196, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.693680286407471, "learning_rate": 5e-05, "loss": 0.7987, "mean_token_accuracy": 0.8710088729858398, "num_tokens": 725793168.0, "step": 19113 }, { "epoch": 2.4314972649790105, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.582282543182373, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8840084671974182, "num_tokens": 725834713.0, "step": 19114 }, { "epoch": 2.4316244752576006, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.6843953132629395, "learning_rate": 5e-05, "loss": 0.7415, "mean_token_accuracy": 0.8812909722328186, "num_tokens": 725872759.0, "step": 19115 }, { "epoch": 2.431751685536191, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.528134346008301, "learning_rate": 5e-05, "loss": 0.7181, "mean_token_accuracy": 0.8844830989837646, "num_tokens": 725916192.0, "step": 19116 }, { "epoch": 2.4318788958147817, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.730118751525879, "learning_rate": 5e-05, "loss": 0.6856, "mean_token_accuracy": 0.8967541456222534, "num_tokens": 725949838.0, "step": 19117 }, { "epoch": 2.432006106093372, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.425528526306152, "learning_rate": 5e-05, "loss": 0.7321, "mean_token_accuracy": 0.8763161897659302, "num_tokens": 725986538.0, "step": 19118 }, { "epoch": 2.4321333163719627, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.787735462188721, "learning_rate": 5e-05, "loss": 0.7122, "mean_token_accuracy": 0.887468695640564, "num_tokens": 726029007.0, "step": 19119 }, { "epoch": 2.4322605266505533, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.3292765617370605, "learning_rate": 5e-05, "loss": 0.7022, "mean_token_accuracy": 0.8812784552574158, "num_tokens": 726066745.0, "step": 19120 }, { "epoch": 2.432387736929144, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.895480155944824, "learning_rate": 5e-05, "loss": 0.7502, "mean_token_accuracy": 0.8774199485778809, "num_tokens": 726106768.0, "step": 19121 }, { "epoch": 2.4325149472077343, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.189442157745361, "learning_rate": 5e-05, "loss": 0.6826, "mean_token_accuracy": 0.884536862373352, "num_tokens": 726144879.0, "step": 19122 }, { "epoch": 2.432642157486325, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 8.078396797180176, "learning_rate": 5e-05, "loss": 0.7533, "mean_token_accuracy": 0.8813073635101318, "num_tokens": 726181278.0, "step": 19123 }, { "epoch": 2.4327693677649154, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.937348365783691, "learning_rate": 5e-05, "loss": 0.6888, "mean_token_accuracy": 0.8840893507003784, "num_tokens": 726226090.0, "step": 19124 }, { "epoch": 2.432896578043506, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.41895580291748, "learning_rate": 5e-05, "loss": 0.7975, "mean_token_accuracy": 0.8796690106391907, "num_tokens": 726261567.0, "step": 19125 }, { "epoch": 2.4330237883220964, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.570752143859863, "learning_rate": 5e-05, "loss": 0.6648, "mean_token_accuracy": 0.8813855648040771, "num_tokens": 726300527.0, "step": 19126 }, { "epoch": 2.433150998600687, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.99349308013916, "learning_rate": 5e-05, "loss": 0.8296, "mean_token_accuracy": 0.887742280960083, "num_tokens": 726341240.0, "step": 19127 }, { "epoch": 2.4332782088792775, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.10506534576416, "learning_rate": 5e-05, "loss": 0.6802, "mean_token_accuracy": 0.8735601902008057, "num_tokens": 726375286.0, "step": 19128 }, { "epoch": 2.433405419157868, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.760255813598633, "learning_rate": 5e-05, "loss": 0.9161, "mean_token_accuracy": 0.8956356048583984, "num_tokens": 726414810.0, "step": 19129 }, { "epoch": 2.4335326294364585, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.097147464752197, "learning_rate": 5e-05, "loss": 0.6516, "mean_token_accuracy": 0.8977014422416687, "num_tokens": 726451204.0, "step": 19130 }, { "epoch": 2.433659839715049, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.158061027526855, "learning_rate": 5e-05, "loss": 1.0158, "mean_token_accuracy": 0.8831175565719604, "num_tokens": 726491107.0, "step": 19131 }, { "epoch": 2.4337870499936396, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.366935729980469, "learning_rate": 5e-05, "loss": 0.9112, "mean_token_accuracy": 0.8787997364997864, "num_tokens": 726535297.0, "step": 19132 }, { "epoch": 2.43391426027223, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.078994750976562, "learning_rate": 5e-05, "loss": 0.9782, "mean_token_accuracy": 0.8717350363731384, "num_tokens": 726574841.0, "step": 19133 }, { "epoch": 2.4340414705508207, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 8.859846115112305, "learning_rate": 5e-05, "loss": 1.0382, "mean_token_accuracy": 0.8748566508293152, "num_tokens": 726611101.0, "step": 19134 }, { "epoch": 2.434168680829411, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.858766078948975, "learning_rate": 5e-05, "loss": 0.7299, "mean_token_accuracy": 0.8773406147956848, "num_tokens": 726654323.0, "step": 19135 }, { "epoch": 2.4342958911080017, "ewc_loss": 0.765625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.62466812133789, "learning_rate": 5e-05, "loss": 1.0956, "mean_token_accuracy": 0.8849830627441406, "num_tokens": 726691504.0, "step": 19136 }, { "epoch": 2.4344231013865922, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 6.368103981018066, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.8749593496322632, "num_tokens": 726737257.0, "step": 19137 }, { "epoch": 2.4345503116651823, "ewc_loss": 0.9765625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000904083251953125, "grad_norm": 12.186243057250977, "learning_rate": 5e-05, "loss": 1.286, "mean_token_accuracy": 0.8897403478622437, "num_tokens": 726779917.0, "step": 19138 }, { "epoch": 2.4346775219437733, "ewc_loss": 0.93359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 9.330470085144043, "learning_rate": 5e-05, "loss": 1.2588, "mean_token_accuracy": 0.8844506740570068, "num_tokens": 726823130.0, "step": 19139 }, { "epoch": 2.4348047322223634, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.983670234680176, "learning_rate": 5e-05, "loss": 1.0604, "mean_token_accuracy": 0.8841660022735596, "num_tokens": 726861787.0, "step": 19140 }, { "epoch": 2.434931942500954, "ewc_loss": 0.9140625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 9.666696548461914, "learning_rate": 5e-05, "loss": 1.2916, "mean_token_accuracy": 0.8672239780426025, "num_tokens": 726904413.0, "step": 19141 }, { "epoch": 2.4350591527795444, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.971918106079102, "learning_rate": 5e-05, "loss": 1.0042, "mean_token_accuracy": 0.8916686773300171, "num_tokens": 726946984.0, "step": 19142 }, { "epoch": 2.435186363058135, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.06899642944336, "learning_rate": 5e-05, "loss": 1.0128, "mean_token_accuracy": 0.8815836906433105, "num_tokens": 726982587.0, "step": 19143 }, { "epoch": 2.4353135733367255, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.190417289733887, "learning_rate": 5e-05, "loss": 0.9742, "mean_token_accuracy": 0.8854255676269531, "num_tokens": 727017255.0, "step": 19144 }, { "epoch": 2.435440783615316, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.3466157913208, "learning_rate": 5e-05, "loss": 0.8637, "mean_token_accuracy": 0.8890284299850464, "num_tokens": 727046147.0, "step": 19145 }, { "epoch": 2.4355679938939065, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.168757438659668, "learning_rate": 5e-05, "loss": 0.9751, "mean_token_accuracy": 0.8801215291023254, "num_tokens": 727085982.0, "step": 19146 }, { "epoch": 2.435695204172497, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.536604404449463, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8933366537094116, "num_tokens": 727123991.0, "step": 19147 }, { "epoch": 2.4358224144510876, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.708184242248535, "learning_rate": 5e-05, "loss": 1.0329, "mean_token_accuracy": 0.8659976720809937, "num_tokens": 727162155.0, "step": 19148 }, { "epoch": 2.435949624729678, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.365865707397461, "learning_rate": 5e-05, "loss": 0.8567, "mean_token_accuracy": 0.8722594976425171, "num_tokens": 727201234.0, "step": 19149 }, { "epoch": 2.4360768350082687, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.048240661621094, "learning_rate": 5e-05, "loss": 1.0371, "mean_token_accuracy": 0.8810921311378479, "num_tokens": 727243429.0, "step": 19150 }, { "epoch": 2.436204045286859, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.834428787231445, "learning_rate": 5e-05, "loss": 0.9268, "mean_token_accuracy": 0.8818814158439636, "num_tokens": 727279705.0, "step": 19151 }, { "epoch": 2.4363312555654497, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.903693199157715, "learning_rate": 5e-05, "loss": 1.003, "mean_token_accuracy": 0.8854852914810181, "num_tokens": 727315503.0, "step": 19152 }, { "epoch": 2.4364584658440402, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.21699047088623, "learning_rate": 5e-05, "loss": 0.9772, "mean_token_accuracy": 0.8805667757987976, "num_tokens": 727355272.0, "step": 19153 }, { "epoch": 2.4365856761226308, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.438222885131836, "learning_rate": 5e-05, "loss": 1.027, "mean_token_accuracy": 0.8700706362724304, "num_tokens": 727391058.0, "step": 19154 }, { "epoch": 2.4367128864012213, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.282485961914062, "learning_rate": 5e-05, "loss": 0.9484, "mean_token_accuracy": 0.8753427267074585, "num_tokens": 727433011.0, "step": 19155 }, { "epoch": 2.436840096679812, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.919754028320312, "learning_rate": 5e-05, "loss": 0.9441, "mean_token_accuracy": 0.8884629607200623, "num_tokens": 727462282.0, "step": 19156 }, { "epoch": 2.4369673069584024, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.011458396911621, "learning_rate": 5e-05, "loss": 0.8795, "mean_token_accuracy": 0.8802356719970703, "num_tokens": 727501636.0, "step": 19157 }, { "epoch": 2.437094517236993, "ewc_loss": 0.625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.888215065002441, "learning_rate": 5e-05, "loss": 0.9057, "mean_token_accuracy": 0.8966606855392456, "num_tokens": 727536828.0, "step": 19158 }, { "epoch": 2.4372217275155834, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.53391170501709, "learning_rate": 5e-05, "loss": 0.8864, "mean_token_accuracy": 0.8552890419960022, "num_tokens": 727574858.0, "step": 19159 }, { "epoch": 2.437348937794174, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.312037467956543, "learning_rate": 5e-05, "loss": 0.941, "mean_token_accuracy": 0.8967278003692627, "num_tokens": 727618168.0, "step": 19160 }, { "epoch": 2.4374761480727645, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.448007106781006, "learning_rate": 5e-05, "loss": 0.8622, "mean_token_accuracy": 0.8722946643829346, "num_tokens": 727660205.0, "step": 19161 }, { "epoch": 2.437603358351355, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.4986572265625, "learning_rate": 5e-05, "loss": 1.0016, "mean_token_accuracy": 0.8873375058174133, "num_tokens": 727701980.0, "step": 19162 }, { "epoch": 2.437730568629945, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.907815933227539, "learning_rate": 5e-05, "loss": 0.9154, "mean_token_accuracy": 0.8715078830718994, "num_tokens": 727743766.0, "step": 19163 }, { "epoch": 2.437857778908536, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.022635459899902, "learning_rate": 5e-05, "loss": 0.967, "mean_token_accuracy": 0.8830670118331909, "num_tokens": 727785504.0, "step": 19164 }, { "epoch": 2.437984989187126, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.258829116821289, "learning_rate": 5e-05, "loss": 0.9, "mean_token_accuracy": 0.8885664939880371, "num_tokens": 727831296.0, "step": 19165 }, { "epoch": 2.4381121994657167, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.228811264038086, "learning_rate": 5e-05, "loss": 0.8599, "mean_token_accuracy": 0.888235867023468, "num_tokens": 727867343.0, "step": 19166 }, { "epoch": 2.438239409744307, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.213793754577637, "learning_rate": 5e-05, "loss": 0.8716, "mean_token_accuracy": 0.8813644647598267, "num_tokens": 727903856.0, "step": 19167 }, { "epoch": 2.4383666200228977, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.834499359130859, "learning_rate": 5e-05, "loss": 0.8319, "mean_token_accuracy": 0.8814426064491272, "num_tokens": 727938385.0, "step": 19168 }, { "epoch": 2.4384938303014883, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.853844165802002, "learning_rate": 5e-05, "loss": 0.7665, "mean_token_accuracy": 0.8906793594360352, "num_tokens": 727965566.0, "step": 19169 }, { "epoch": 2.438621040580079, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.052851676940918, "learning_rate": 5e-05, "loss": 0.795, "mean_token_accuracy": 0.8947156667709351, "num_tokens": 728002203.0, "step": 19170 }, { "epoch": 2.4387482508586693, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.316860675811768, "learning_rate": 5e-05, "loss": 0.719, "mean_token_accuracy": 0.8880047798156738, "num_tokens": 728037055.0, "step": 19171 }, { "epoch": 2.43887546113726, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.484739303588867, "learning_rate": 5e-05, "loss": 0.87, "mean_token_accuracy": 0.8753095269203186, "num_tokens": 728077226.0, "step": 19172 }, { "epoch": 2.4390026714158504, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.024351119995117, "learning_rate": 5e-05, "loss": 0.7084, "mean_token_accuracy": 0.8865334987640381, "num_tokens": 728111396.0, "step": 19173 }, { "epoch": 2.439129881694441, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.693916320800781, "learning_rate": 5e-05, "loss": 0.8835, "mean_token_accuracy": 0.8798713088035583, "num_tokens": 728149609.0, "step": 19174 }, { "epoch": 2.4392570919730314, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.011944770812988, "learning_rate": 5e-05, "loss": 0.726, "mean_token_accuracy": 0.8830300569534302, "num_tokens": 728186668.0, "step": 19175 }, { "epoch": 2.439384302251622, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.779877662658691, "learning_rate": 5e-05, "loss": 0.9435, "mean_token_accuracy": 0.867093563079834, "num_tokens": 728226504.0, "step": 19176 }, { "epoch": 2.4395115125302125, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.149339199066162, "learning_rate": 5e-05, "loss": 0.7489, "mean_token_accuracy": 0.8809530735015869, "num_tokens": 728264648.0, "step": 19177 }, { "epoch": 2.439638722808803, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00072479248046875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.67209243774414, "learning_rate": 5e-05, "loss": 0.8909, "mean_token_accuracy": 0.8830705285072327, "num_tokens": 728300292.0, "step": 19178 }, { "epoch": 2.4397659330873935, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.496720314025879, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8778330087661743, "num_tokens": 728336169.0, "step": 19179 }, { "epoch": 2.439893143365984, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.16123104095459, "learning_rate": 5e-05, "loss": 0.8255, "mean_token_accuracy": 0.887141227722168, "num_tokens": 728380160.0, "step": 19180 }, { "epoch": 2.4400203536445746, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.84724235534668, "learning_rate": 5e-05, "loss": 0.7994, "mean_token_accuracy": 0.8820351362228394, "num_tokens": 728417978.0, "step": 19181 }, { "epoch": 2.440147563923165, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.610567092895508, "learning_rate": 5e-05, "loss": 0.7542, "mean_token_accuracy": 0.8846789598464966, "num_tokens": 728455264.0, "step": 19182 }, { "epoch": 2.4402747742017556, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.019810676574707, "learning_rate": 5e-05, "loss": 0.8314, "mean_token_accuracy": 0.8688346743583679, "num_tokens": 728498880.0, "step": 19183 }, { "epoch": 2.440401984480346, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.365848541259766, "learning_rate": 5e-05, "loss": 0.7449, "mean_token_accuracy": 0.8785305023193359, "num_tokens": 728531925.0, "step": 19184 }, { "epoch": 2.4405291947589367, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.057567596435547, "learning_rate": 5e-05, "loss": 0.7822, "mean_token_accuracy": 0.8784905672073364, "num_tokens": 728570573.0, "step": 19185 }, { "epoch": 2.440656405037527, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.236268997192383, "learning_rate": 5e-05, "loss": 0.7134, "mean_token_accuracy": 0.8817472457885742, "num_tokens": 728610877.0, "step": 19186 }, { "epoch": 2.4407836153161178, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.161029815673828, "learning_rate": 5e-05, "loss": 0.7892, "mean_token_accuracy": 0.8806583285331726, "num_tokens": 728650963.0, "step": 19187 }, { "epoch": 2.440910825594708, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.067996501922607, "learning_rate": 5e-05, "loss": 0.6663, "mean_token_accuracy": 0.8923912048339844, "num_tokens": 728688699.0, "step": 19188 }, { "epoch": 2.4410380358732984, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.330118179321289, "learning_rate": 5e-05, "loss": 0.8361, "mean_token_accuracy": 0.8703352212905884, "num_tokens": 728730142.0, "step": 19189 }, { "epoch": 2.441165246151889, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.931384086608887, "learning_rate": 5e-05, "loss": 0.7149, "mean_token_accuracy": 0.8761523962020874, "num_tokens": 728767808.0, "step": 19190 }, { "epoch": 2.4412924564304794, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.481792449951172, "learning_rate": 5e-05, "loss": 0.8467, "mean_token_accuracy": 0.8681776523590088, "num_tokens": 728816171.0, "step": 19191 }, { "epoch": 2.44141966670907, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.87741231918335, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.8681670427322388, "num_tokens": 728858350.0, "step": 19192 }, { "epoch": 2.4415468769876605, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.552660942077637, "learning_rate": 5e-05, "loss": 0.7883, "mean_token_accuracy": 0.8943373560905457, "num_tokens": 728894408.0, "step": 19193 }, { "epoch": 2.441674087266251, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.9481329917907715, "learning_rate": 5e-05, "loss": 0.7614, "mean_token_accuracy": 0.8607020974159241, "num_tokens": 728936563.0, "step": 19194 }, { "epoch": 2.4418012975448415, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.444490432739258, "learning_rate": 5e-05, "loss": 0.8587, "mean_token_accuracy": 0.8739675283432007, "num_tokens": 728978796.0, "step": 19195 }, { "epoch": 2.441928507823432, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.172970294952393, "learning_rate": 5e-05, "loss": 0.7085, "mean_token_accuracy": 0.8865407705307007, "num_tokens": 729018524.0, "step": 19196 }, { "epoch": 2.4420557181020226, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.147436141967773, "learning_rate": 5e-05, "loss": 0.8085, "mean_token_accuracy": 0.8770965337753296, "num_tokens": 729055336.0, "step": 19197 }, { "epoch": 2.442182928380613, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.438582420349121, "learning_rate": 5e-05, "loss": 0.7658, "mean_token_accuracy": 0.8711717128753662, "num_tokens": 729095505.0, "step": 19198 }, { "epoch": 2.4423101386592037, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.826443672180176, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.8796209096908569, "num_tokens": 729138384.0, "step": 19199 }, { "epoch": 2.442437348937794, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.618642807006836, "learning_rate": 5e-05, "loss": 0.795, "mean_token_accuracy": 0.8631985187530518, "num_tokens": 729179942.0, "step": 19200 }, { "epoch": 2.4425645592163847, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.638200283050537, "learning_rate": 5e-05, "loss": 0.7974, "mean_token_accuracy": 0.8631293773651123, "num_tokens": 729223051.0, "step": 19201 }, { "epoch": 2.4426917694949752, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.6381306648254395, "learning_rate": 5e-05, "loss": 0.7316, "mean_token_accuracy": 0.8829441070556641, "num_tokens": 729259368.0, "step": 19202 }, { "epoch": 2.4428189797735658, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.664235591888428, "learning_rate": 5e-05, "loss": 0.7471, "mean_token_accuracy": 0.8764020800590515, "num_tokens": 729295148.0, "step": 19203 }, { "epoch": 2.4429461900521563, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.448065757751465, "learning_rate": 5e-05, "loss": 0.7435, "mean_token_accuracy": 0.8733389377593994, "num_tokens": 729330649.0, "step": 19204 }, { "epoch": 2.443073400330747, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.895526885986328, "learning_rate": 5e-05, "loss": 0.7277, "mean_token_accuracy": 0.8881500363349915, "num_tokens": 729369260.0, "step": 19205 }, { "epoch": 2.4432006106093374, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.131948471069336, "learning_rate": 5e-05, "loss": 0.6937, "mean_token_accuracy": 0.8806921243667603, "num_tokens": 729406029.0, "step": 19206 }, { "epoch": 2.443327820887928, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.294158935546875, "learning_rate": 5e-05, "loss": 0.7873, "mean_token_accuracy": 0.8760062456130981, "num_tokens": 729444102.0, "step": 19207 }, { "epoch": 2.4434550311665184, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.6769819259643555, "learning_rate": 5e-05, "loss": 0.6025, "mean_token_accuracy": 0.9036017656326294, "num_tokens": 729477121.0, "step": 19208 }, { "epoch": 2.443582241445109, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.901766777038574, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8727717995643616, "num_tokens": 729511682.0, "step": 19209 }, { "epoch": 2.4437094517236995, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.190469264984131, "learning_rate": 5e-05, "loss": 0.6555, "mean_token_accuracy": 0.8854078650474548, "num_tokens": 729545675.0, "step": 19210 }, { "epoch": 2.4438366620022896, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.614862442016602, "learning_rate": 5e-05, "loss": 0.9521, "mean_token_accuracy": 0.8807273507118225, "num_tokens": 729585323.0, "step": 19211 }, { "epoch": 2.4439638722808805, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.182117462158203, "learning_rate": 5e-05, "loss": 0.7057, "mean_token_accuracy": 0.8749340772628784, "num_tokens": 729623881.0, "step": 19212 }, { "epoch": 2.4440910825594706, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.835442543029785, "learning_rate": 5e-05, "loss": 0.9488, "mean_token_accuracy": 0.8956225514411926, "num_tokens": 729655486.0, "step": 19213 }, { "epoch": 2.444218292838061, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.3041558265686035, "learning_rate": 5e-05, "loss": 0.8521, "mean_token_accuracy": 0.8812224268913269, "num_tokens": 729693606.0, "step": 19214 }, { "epoch": 2.4443455031166517, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.656512260437012, "learning_rate": 5e-05, "loss": 0.9342, "mean_token_accuracy": 0.8724879622459412, "num_tokens": 729729467.0, "step": 19215 }, { "epoch": 2.444472713395242, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.731840133666992, "learning_rate": 5e-05, "loss": 0.9312, "mean_token_accuracy": 0.8884564638137817, "num_tokens": 729767264.0, "step": 19216 }, { "epoch": 2.4445999236738327, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.613518238067627, "learning_rate": 5e-05, "loss": 0.713, "mean_token_accuracy": 0.8821208477020264, "num_tokens": 729807352.0, "step": 19217 }, { "epoch": 2.4447271339524232, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.476656913757324, "learning_rate": 5e-05, "loss": 1.1044, "mean_token_accuracy": 0.8639066219329834, "num_tokens": 729847489.0, "step": 19218 }, { "epoch": 2.4448543442310138, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.076696872711182, "learning_rate": 5e-05, "loss": 0.6979, "mean_token_accuracy": 0.8857092261314392, "num_tokens": 729889675.0, "step": 19219 }, { "epoch": 2.4449815545096043, "ewc_loss": 0.97265625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0009002685546875, "grad_norm": 12.063779830932617, "learning_rate": 5e-05, "loss": 1.3162, "mean_token_accuracy": 0.8789527416229248, "num_tokens": 729926562.0, "step": 19220 }, { "epoch": 2.445108764788195, "ewc_loss": 0.83984375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000766754150390625, "grad_norm": 9.03243350982666, "learning_rate": 5e-05, "loss": 1.2297, "mean_token_accuracy": 0.8687771558761597, "num_tokens": 729964533.0, "step": 19221 }, { "epoch": 2.4452359750667854, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.934104919433594, "learning_rate": 5e-05, "loss": 1.0468, "mean_token_accuracy": 0.8917629718780518, "num_tokens": 730004206.0, "step": 19222 }, { "epoch": 2.445363185345376, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000823974609375, "grad_norm": 9.731474876403809, "learning_rate": 5e-05, "loss": 1.2538, "mean_token_accuracy": 0.8743193745613098, "num_tokens": 730046779.0, "step": 19223 }, { "epoch": 2.4454903956239664, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.613417625427246, "learning_rate": 5e-05, "loss": 1.0195, "mean_token_accuracy": 0.8774982690811157, "num_tokens": 730085882.0, "step": 19224 }, { "epoch": 2.445617605902557, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.422210693359375, "learning_rate": 5e-05, "loss": 1.0453, "mean_token_accuracy": 0.8775937557220459, "num_tokens": 730126072.0, "step": 19225 }, { "epoch": 2.4457448161811475, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.691008567810059, "learning_rate": 5e-05, "loss": 0.9205, "mean_token_accuracy": 0.8964190483093262, "num_tokens": 730160038.0, "step": 19226 }, { "epoch": 2.445872026459738, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.874789237976074, "learning_rate": 5e-05, "loss": 0.9405, "mean_token_accuracy": 0.8803118467330933, "num_tokens": 730195701.0, "step": 19227 }, { "epoch": 2.4459992367383285, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.526128768920898, "learning_rate": 5e-05, "loss": 0.917, "mean_token_accuracy": 0.8858054280281067, "num_tokens": 730233938.0, "step": 19228 }, { "epoch": 2.446126447016919, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.954668045043945, "learning_rate": 5e-05, "loss": 0.7806, "mean_token_accuracy": 0.8968146443367004, "num_tokens": 730272333.0, "step": 19229 }, { "epoch": 2.4462536572955096, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.781037330627441, "learning_rate": 5e-05, "loss": 0.9361, "mean_token_accuracy": 0.872275173664093, "num_tokens": 730312967.0, "step": 19230 }, { "epoch": 2.4463808675741, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.385944366455078, "learning_rate": 5e-05, "loss": 0.768, "mean_token_accuracy": 0.8814691305160522, "num_tokens": 730345525.0, "step": 19231 }, { "epoch": 2.4465080778526906, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.93519401550293, "learning_rate": 5e-05, "loss": 0.9098, "mean_token_accuracy": 0.8843984603881836, "num_tokens": 730384329.0, "step": 19232 }, { "epoch": 2.446635288131281, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.230561256408691, "learning_rate": 5e-05, "loss": 0.7233, "mean_token_accuracy": 0.8952681422233582, "num_tokens": 730423139.0, "step": 19233 }, { "epoch": 2.4467624984098717, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.964129447937012, "learning_rate": 5e-05, "loss": 0.9158, "mean_token_accuracy": 0.8808058500289917, "num_tokens": 730463989.0, "step": 19234 }, { "epoch": 2.4468897086884622, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.211333751678467, "learning_rate": 5e-05, "loss": 0.7887, "mean_token_accuracy": 0.8747590780258179, "num_tokens": 730500854.0, "step": 19235 }, { "epoch": 2.4470169189670523, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.934870719909668, "learning_rate": 5e-05, "loss": 0.8693, "mean_token_accuracy": 0.8960603475570679, "num_tokens": 730537742.0, "step": 19236 }, { "epoch": 2.4471441292456433, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.261119365692139, "learning_rate": 5e-05, "loss": 0.7359, "mean_token_accuracy": 0.898581862449646, "num_tokens": 730573090.0, "step": 19237 }, { "epoch": 2.4472713395242334, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.827249526977539, "learning_rate": 5e-05, "loss": 0.8638, "mean_token_accuracy": 0.8984214067459106, "num_tokens": 730611865.0, "step": 19238 }, { "epoch": 2.447398549802824, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.3085808753967285, "learning_rate": 5e-05, "loss": 0.8559, "mean_token_accuracy": 0.8616079092025757, "num_tokens": 730646831.0, "step": 19239 }, { "epoch": 2.4475257600814144, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.769123077392578, "learning_rate": 5e-05, "loss": 0.9098, "mean_token_accuracy": 0.8835344314575195, "num_tokens": 730689789.0, "step": 19240 }, { "epoch": 2.447652970360005, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.230434894561768, "learning_rate": 5e-05, "loss": 0.7863, "mean_token_accuracy": 0.8826339244842529, "num_tokens": 730720289.0, "step": 19241 }, { "epoch": 2.4477801806385955, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.927933692932129, "learning_rate": 5e-05, "loss": 0.9076, "mean_token_accuracy": 0.8858456611633301, "num_tokens": 730757499.0, "step": 19242 }, { "epoch": 2.447907390917186, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.133141040802002, "learning_rate": 5e-05, "loss": 0.7666, "mean_token_accuracy": 0.8842111825942993, "num_tokens": 730799897.0, "step": 19243 }, { "epoch": 2.4480346011957765, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.068120002746582, "learning_rate": 5e-05, "loss": 0.9275, "mean_token_accuracy": 0.8898786306381226, "num_tokens": 730837709.0, "step": 19244 }, { "epoch": 2.448161811474367, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.318634986877441, "learning_rate": 5e-05, "loss": 0.7495, "mean_token_accuracy": 0.9000354409217834, "num_tokens": 730874128.0, "step": 19245 }, { "epoch": 2.4482890217529576, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.846506118774414, "learning_rate": 5e-05, "loss": 0.9533, "mean_token_accuracy": 0.8793164491653442, "num_tokens": 730917027.0, "step": 19246 }, { "epoch": 2.448416232031548, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.667537212371826, "learning_rate": 5e-05, "loss": 0.8831, "mean_token_accuracy": 0.8709372282028198, "num_tokens": 730955107.0, "step": 19247 }, { "epoch": 2.4485434423101387, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.351557731628418, "learning_rate": 5e-05, "loss": 0.9294, "mean_token_accuracy": 0.8697043657302856, "num_tokens": 730995807.0, "step": 19248 }, { "epoch": 2.448670652588729, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.8665971755981445, "learning_rate": 5e-05, "loss": 0.8115, "mean_token_accuracy": 0.8937235474586487, "num_tokens": 731034935.0, "step": 19249 }, { "epoch": 2.4487978628673197, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.894166946411133, "learning_rate": 5e-05, "loss": 0.8382, "mean_token_accuracy": 0.8787051439285278, "num_tokens": 731070934.0, "step": 19250 }, { "epoch": 2.4489250731459102, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.870429992675781, "learning_rate": 5e-05, "loss": 0.7989, "mean_token_accuracy": 0.8880400061607361, "num_tokens": 731112625.0, "step": 19251 }, { "epoch": 2.4490522834245008, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.755260944366455, "learning_rate": 5e-05, "loss": 0.8097, "mean_token_accuracy": 0.8756494522094727, "num_tokens": 731151011.0, "step": 19252 }, { "epoch": 2.4491794937030913, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.63787317276001, "learning_rate": 5e-05, "loss": 0.7553, "mean_token_accuracy": 0.8832488059997559, "num_tokens": 731190885.0, "step": 19253 }, { "epoch": 2.449306703981682, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.9859843254089355, "learning_rate": 5e-05, "loss": 0.7742, "mean_token_accuracy": 0.8850008249282837, "num_tokens": 731229056.0, "step": 19254 }, { "epoch": 2.4494339142602723, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.25978422164917, "learning_rate": 5e-05, "loss": 0.7398, "mean_token_accuracy": 0.8811978101730347, "num_tokens": 731268098.0, "step": 19255 }, { "epoch": 2.449561124538863, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.307540893554688, "learning_rate": 5e-05, "loss": 0.8603, "mean_token_accuracy": 0.8678783178329468, "num_tokens": 731307809.0, "step": 19256 }, { "epoch": 2.4496883348174534, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.021146297454834, "learning_rate": 5e-05, "loss": 0.7298, "mean_token_accuracy": 0.8753185868263245, "num_tokens": 731343904.0, "step": 19257 }, { "epoch": 2.449815545096044, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.47122573852539, "learning_rate": 5e-05, "loss": 0.7884, "mean_token_accuracy": 0.896134078502655, "num_tokens": 731383120.0, "step": 19258 }, { "epoch": 2.4499427553746345, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.97210168838501, "learning_rate": 5e-05, "loss": 0.709, "mean_token_accuracy": 0.8830021619796753, "num_tokens": 731423020.0, "step": 19259 }, { "epoch": 2.450069965653225, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.509260177612305, "learning_rate": 5e-05, "loss": 0.8546, "mean_token_accuracy": 0.8783036470413208, "num_tokens": 731463522.0, "step": 19260 }, { "epoch": 2.450197175931815, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.080989837646484, "learning_rate": 5e-05, "loss": 0.7349, "mean_token_accuracy": 0.8726169466972351, "num_tokens": 731502327.0, "step": 19261 }, { "epoch": 2.450324386210406, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.32868766784668, "learning_rate": 5e-05, "loss": 0.7918, "mean_token_accuracy": 0.8909758925437927, "num_tokens": 731545253.0, "step": 19262 }, { "epoch": 2.450451596488996, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.3592848777771, "learning_rate": 5e-05, "loss": 0.72, "mean_token_accuracy": 0.8878101110458374, "num_tokens": 731575388.0, "step": 19263 }, { "epoch": 2.4505788067675867, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.926371097564697, "learning_rate": 5e-05, "loss": 0.7937, "mean_token_accuracy": 0.8783504962921143, "num_tokens": 731618195.0, "step": 19264 }, { "epoch": 2.450706017046177, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.687088489532471, "learning_rate": 5e-05, "loss": 0.7929, "mean_token_accuracy": 0.8734214901924133, "num_tokens": 731658551.0, "step": 19265 }, { "epoch": 2.4508332273247677, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.508086681365967, "learning_rate": 5e-05, "loss": 0.7057, "mean_token_accuracy": 0.8929391503334045, "num_tokens": 731700477.0, "step": 19266 }, { "epoch": 2.4509604376033582, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.962643623352051, "learning_rate": 5e-05, "loss": 0.7655, "mean_token_accuracy": 0.8849932551383972, "num_tokens": 731735841.0, "step": 19267 }, { "epoch": 2.4510876478819488, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.190826892852783, "learning_rate": 5e-05, "loss": 0.727, "mean_token_accuracy": 0.877743124961853, "num_tokens": 731774974.0, "step": 19268 }, { "epoch": 2.4512148581605393, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.178641319274902, "learning_rate": 5e-05, "loss": 0.7657, "mean_token_accuracy": 0.8854296803474426, "num_tokens": 731806010.0, "step": 19269 }, { "epoch": 2.45134206843913, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.957198143005371, "learning_rate": 5e-05, "loss": 0.6421, "mean_token_accuracy": 0.8965677618980408, "num_tokens": 731841773.0, "step": 19270 }, { "epoch": 2.4514692787177204, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.440279006958008, "learning_rate": 5e-05, "loss": 0.7755, "mean_token_accuracy": 0.8880062103271484, "num_tokens": 731877974.0, "step": 19271 }, { "epoch": 2.451596488996311, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.699174880981445, "learning_rate": 5e-05, "loss": 0.6769, "mean_token_accuracy": 0.8836358785629272, "num_tokens": 731919375.0, "step": 19272 }, { "epoch": 2.4517236992749014, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.750358581542969, "learning_rate": 5e-05, "loss": 0.8299, "mean_token_accuracy": 0.8836911916732788, "num_tokens": 731957518.0, "step": 19273 }, { "epoch": 2.451850909553492, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.520834922790527, "learning_rate": 5e-05, "loss": 0.656, "mean_token_accuracy": 0.8903846740722656, "num_tokens": 731996597.0, "step": 19274 }, { "epoch": 2.4519781198320825, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 9.044772148132324, "learning_rate": 5e-05, "loss": 0.8882, "mean_token_accuracy": 0.8801162242889404, "num_tokens": 732035888.0, "step": 19275 }, { "epoch": 2.452105330110673, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.596925258636475, "learning_rate": 5e-05, "loss": 0.7089, "mean_token_accuracy": 0.878139078617096, "num_tokens": 732073026.0, "step": 19276 }, { "epoch": 2.4522325403892635, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.968497276306152, "learning_rate": 5e-05, "loss": 0.9758, "mean_token_accuracy": 0.8576405048370361, "num_tokens": 732110794.0, "step": 19277 }, { "epoch": 2.452359750667854, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.124577522277832, "learning_rate": 5e-05, "loss": 0.7333, "mean_token_accuracy": 0.8878118991851807, "num_tokens": 732141051.0, "step": 19278 }, { "epoch": 2.4524869609464446, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.268318176269531, "learning_rate": 5e-05, "loss": 0.8514, "mean_token_accuracy": 0.8849681615829468, "num_tokens": 732187023.0, "step": 19279 }, { "epoch": 2.452614171225035, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.862546443939209, "learning_rate": 5e-05, "loss": 0.8673, "mean_token_accuracy": 0.8682022094726562, "num_tokens": 732220960.0, "step": 19280 }, { "epoch": 2.4527413815036256, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.335652828216553, "learning_rate": 5e-05, "loss": 0.7474, "mean_token_accuracy": 0.8823451995849609, "num_tokens": 732260656.0, "step": 19281 }, { "epoch": 2.452868591782216, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.4566011428833, "learning_rate": 5e-05, "loss": 0.9149, "mean_token_accuracy": 0.8599900603294373, "num_tokens": 732301771.0, "step": 19282 }, { "epoch": 2.4529958020608067, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.8029561042785645, "learning_rate": 5e-05, "loss": 0.7189, "mean_token_accuracy": 0.8772521615028381, "num_tokens": 732341618.0, "step": 19283 }, { "epoch": 2.453123012339397, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.680893898010254, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.9077317118644714, "num_tokens": 732375797.0, "step": 19284 }, { "epoch": 2.4532502226179878, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.851996421813965, "learning_rate": 5e-05, "loss": 0.6893, "mean_token_accuracy": 0.8892514705657959, "num_tokens": 732414234.0, "step": 19285 }, { "epoch": 2.453377432896578, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.438193321228027, "learning_rate": 5e-05, "loss": 0.8401, "mean_token_accuracy": 0.878737211227417, "num_tokens": 732461063.0, "step": 19286 }, { "epoch": 2.4535046431751684, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.241743087768555, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.8835874199867249, "num_tokens": 732501622.0, "step": 19287 }, { "epoch": 2.453631853453759, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.939943790435791, "learning_rate": 5e-05, "loss": 0.804, "mean_token_accuracy": 0.8745678663253784, "num_tokens": 732546938.0, "step": 19288 }, { "epoch": 2.4537590637323494, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.659107208251953, "learning_rate": 5e-05, "loss": 0.7249, "mean_token_accuracy": 0.896884560585022, "num_tokens": 732581250.0, "step": 19289 }, { "epoch": 2.45388627401094, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.463342666625977, "learning_rate": 5e-05, "loss": 0.7079, "mean_token_accuracy": 0.8904743194580078, "num_tokens": 732621081.0, "step": 19290 }, { "epoch": 2.4540134842895305, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.957403659820557, "learning_rate": 5e-05, "loss": 0.7552, "mean_token_accuracy": 0.8888435959815979, "num_tokens": 732660625.0, "step": 19291 }, { "epoch": 2.454140694568121, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.211006164550781, "learning_rate": 5e-05, "loss": 0.705, "mean_token_accuracy": 0.8826901316642761, "num_tokens": 732700100.0, "step": 19292 }, { "epoch": 2.4542679048467115, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.073311805725098, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8821150064468384, "num_tokens": 732735252.0, "step": 19293 }, { "epoch": 2.454395115125302, "ewc_loss": 0.375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.167498588562012, "learning_rate": 5e-05, "loss": 0.7339, "mean_token_accuracy": 0.8745123147964478, "num_tokens": 732767912.0, "step": 19294 }, { "epoch": 2.4545223254038926, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.034931182861328, "learning_rate": 5e-05, "loss": 0.7897, "mean_token_accuracy": 0.87569659948349, "num_tokens": 732808499.0, "step": 19295 }, { "epoch": 2.454649535682483, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.234014987945557, "learning_rate": 5e-05, "loss": 0.7035, "mean_token_accuracy": 0.8809730410575867, "num_tokens": 732840754.0, "step": 19296 }, { "epoch": 2.4547767459610736, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.909517288208008, "learning_rate": 5e-05, "loss": 0.8101, "mean_token_accuracy": 0.86199551820755, "num_tokens": 732882432.0, "step": 19297 }, { "epoch": 2.454903956239664, "ewc_loss": 0.375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.315941333770752, "learning_rate": 5e-05, "loss": 0.6952, "mean_token_accuracy": 0.8899566531181335, "num_tokens": 732912418.0, "step": 19298 }, { "epoch": 2.4550311665182547, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.834019660949707, "learning_rate": 5e-05, "loss": 0.7274, "mean_token_accuracy": 0.8841071128845215, "num_tokens": 732946050.0, "step": 19299 }, { "epoch": 2.4551583767968452, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.2932047843933105, "learning_rate": 5e-05, "loss": 0.6812, "mean_token_accuracy": 0.8877309560775757, "num_tokens": 732980667.0, "step": 19300 }, { "epoch": 2.4552855870754358, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.917224884033203, "learning_rate": 5e-05, "loss": 0.7448, "mean_token_accuracy": 0.8810081481933594, "num_tokens": 733016169.0, "step": 19301 }, { "epoch": 2.4554127973540263, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.104872703552246, "learning_rate": 5e-05, "loss": 0.7103, "mean_token_accuracy": 0.8746926784515381, "num_tokens": 733055120.0, "step": 19302 }, { "epoch": 2.455540007632617, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.201366424560547, "learning_rate": 5e-05, "loss": 0.8069, "mean_token_accuracy": 0.8688352108001709, "num_tokens": 733094550.0, "step": 19303 }, { "epoch": 2.4556672179112073, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.69852352142334, "learning_rate": 5e-05, "loss": 0.6252, "mean_token_accuracy": 0.8947117328643799, "num_tokens": 733132338.0, "step": 19304 }, { "epoch": 2.455794428189798, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.82497787475586, "learning_rate": 5e-05, "loss": 0.8413, "mean_token_accuracy": 0.8775947093963623, "num_tokens": 733167053.0, "step": 19305 }, { "epoch": 2.4559216384683884, "ewc_loss": 0.30859375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00023555755615234375, "grad_norm": 6.053291320800781, "learning_rate": 5e-05, "loss": 0.6785, "mean_token_accuracy": 0.8706308603286743, "num_tokens": 733208175.0, "step": 19306 }, { "epoch": 2.456048848746979, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.892660140991211, "learning_rate": 5e-05, "loss": 0.9717, "mean_token_accuracy": 0.8704347014427185, "num_tokens": 733248336.0, "step": 19307 }, { "epoch": 2.4561760590255695, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.62468957901001, "learning_rate": 5e-05, "loss": 0.6716, "mean_token_accuracy": 0.8791232109069824, "num_tokens": 733283314.0, "step": 19308 }, { "epoch": 2.4563032693041595, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 11.080806732177734, "learning_rate": 5e-05, "loss": 1.1506, "mean_token_accuracy": 0.8806577920913696, "num_tokens": 733317541.0, "step": 19309 }, { "epoch": 2.4564304795827505, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.156518459320068, "learning_rate": 5e-05, "loss": 0.882, "mean_token_accuracy": 0.8988233804702759, "num_tokens": 733349975.0, "step": 19310 }, { "epoch": 2.4565576898613406, "ewc_loss": 0.828125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00075531005859375, "grad_norm": 10.914565086364746, "learning_rate": 5e-05, "loss": 1.1371, "mean_token_accuracy": 0.89093017578125, "num_tokens": 733388085.0, "step": 19311 }, { "epoch": 2.456684900139931, "ewc_loss": 0.9296875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 9.238433837890625, "learning_rate": 5e-05, "loss": 1.2339, "mean_token_accuracy": 0.8906356692314148, "num_tokens": 733428420.0, "step": 19312 }, { "epoch": 2.4568121104185217, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.751995086669922, "learning_rate": 5e-05, "loss": 0.8738, "mean_token_accuracy": 0.8864562511444092, "num_tokens": 733460271.0, "step": 19313 }, { "epoch": 2.456939320697112, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000835418701171875, "grad_norm": 10.384296417236328, "learning_rate": 5e-05, "loss": 1.2555, "mean_token_accuracy": 0.8801724314689636, "num_tokens": 733498610.0, "step": 19314 }, { "epoch": 2.4570665309757027, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.325916290283203, "learning_rate": 5e-05, "loss": 0.8516, "mean_token_accuracy": 0.8734474778175354, "num_tokens": 733530501.0, "step": 19315 }, { "epoch": 2.4571937412542932, "ewc_loss": 1.0078125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000934600830078125, "grad_norm": 12.148268699645996, "learning_rate": 5e-05, "loss": 1.3256, "mean_token_accuracy": 0.8850972652435303, "num_tokens": 733563208.0, "step": 19316 }, { "epoch": 2.4573209515328838, "ewc_loss": 0.9296875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 9.369600296020508, "learning_rate": 5e-05, "loss": 1.2533, "mean_token_accuracy": 0.8881258368492126, "num_tokens": 733598729.0, "step": 19317 }, { "epoch": 2.4574481618114743, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 11.520222663879395, "learning_rate": 5e-05, "loss": 1.1093, "mean_token_accuracy": 0.8797252774238586, "num_tokens": 733632033.0, "step": 19318 }, { "epoch": 2.457575372090065, "ewc_loss": 0.9296875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 9.628273963928223, "learning_rate": 5e-05, "loss": 1.2962, "mean_token_accuracy": 0.8739106059074402, "num_tokens": 733673088.0, "step": 19319 }, { "epoch": 2.4577025823686554, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 11.027188301086426, "learning_rate": 5e-05, "loss": 1.1297, "mean_token_accuracy": 0.8724682331085205, "num_tokens": 733714096.0, "step": 19320 }, { "epoch": 2.457829792647246, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 8.71804428100586, "learning_rate": 5e-05, "loss": 1.0385, "mean_token_accuracy": 0.8758773803710938, "num_tokens": 733754495.0, "step": 19321 }, { "epoch": 2.4579570029258364, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.871552467346191, "learning_rate": 5e-05, "loss": 1.1061, "mean_token_accuracy": 0.8817230463027954, "num_tokens": 733788904.0, "step": 19322 }, { "epoch": 2.458084213204427, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.074536323547363, "learning_rate": 5e-05, "loss": 0.9142, "mean_token_accuracy": 0.8849655389785767, "num_tokens": 733831041.0, "step": 19323 }, { "epoch": 2.4582114234830175, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 11.403861999511719, "learning_rate": 5e-05, "loss": 1.1799, "mean_token_accuracy": 0.8699127435684204, "num_tokens": 733872209.0, "step": 19324 }, { "epoch": 2.458338633761608, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.512280464172363, "learning_rate": 5e-05, "loss": 1.0099, "mean_token_accuracy": 0.8739688992500305, "num_tokens": 733906884.0, "step": 19325 }, { "epoch": 2.4584658440401985, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00079345703125, "grad_norm": 11.826160430908203, "learning_rate": 5e-05, "loss": 1.1576, "mean_token_accuracy": 0.8971571922302246, "num_tokens": 733949262.0, "step": 19326 }, { "epoch": 2.458593054318789, "ewc_loss": 0.78125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 9.665369987487793, "learning_rate": 5e-05, "loss": 1.093, "mean_token_accuracy": 0.8887974619865417, "num_tokens": 733991111.0, "step": 19327 }, { "epoch": 2.4587202645973796, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 11.110894203186035, "learning_rate": 5e-05, "loss": 1.1262, "mean_token_accuracy": 0.8775923252105713, "num_tokens": 734030968.0, "step": 19328 }, { "epoch": 2.45884747487597, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 9.759430885314941, "learning_rate": 5e-05, "loss": 1.1001, "mean_token_accuracy": 0.8716483116149902, "num_tokens": 734070612.0, "step": 19329 }, { "epoch": 2.4589746851545606, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.247055053710938, "learning_rate": 5e-05, "loss": 1.0229, "mean_token_accuracy": 0.8859474658966064, "num_tokens": 734108682.0, "step": 19330 }, { "epoch": 2.459101895433151, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.293701171875, "learning_rate": 5e-05, "loss": 0.9656, "mean_token_accuracy": 0.8862711191177368, "num_tokens": 734141565.0, "step": 19331 }, { "epoch": 2.4592291057117417, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.56100082397461, "learning_rate": 5e-05, "loss": 0.9705, "mean_token_accuracy": 0.8859099745750427, "num_tokens": 734183048.0, "step": 19332 }, { "epoch": 2.459356315990332, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.608685493469238, "learning_rate": 5e-05, "loss": 0.9078, "mean_token_accuracy": 0.8737107515335083, "num_tokens": 734216141.0, "step": 19333 }, { "epoch": 2.4594835262689223, "ewc_loss": 0.625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.439130783081055, "learning_rate": 5e-05, "loss": 0.9625, "mean_token_accuracy": 0.8804250359535217, "num_tokens": 734256365.0, "step": 19334 }, { "epoch": 2.4596107365475133, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.854802131652832, "learning_rate": 5e-05, "loss": 0.7646, "mean_token_accuracy": 0.893258810043335, "num_tokens": 734297679.0, "step": 19335 }, { "epoch": 2.4597379468261034, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.931848526000977, "learning_rate": 5e-05, "loss": 1.0403, "mean_token_accuracy": 0.870098352432251, "num_tokens": 734333245.0, "step": 19336 }, { "epoch": 2.459865157104694, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.6572465896606445, "learning_rate": 5e-05, "loss": 0.8219, "mean_token_accuracy": 0.8739434480667114, "num_tokens": 734374408.0, "step": 19337 }, { "epoch": 2.4599923673832844, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.503321647644043, "learning_rate": 5e-05, "loss": 1.0611, "mean_token_accuracy": 0.88556969165802, "num_tokens": 734408105.0, "step": 19338 }, { "epoch": 2.460119577661875, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.3985595703125, "learning_rate": 5e-05, "loss": 0.8887, "mean_token_accuracy": 0.8850628137588501, "num_tokens": 734451700.0, "step": 19339 }, { "epoch": 2.4602467879404655, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.239063262939453, "learning_rate": 5e-05, "loss": 1.0877, "mean_token_accuracy": 0.8735244274139404, "num_tokens": 734492343.0, "step": 19340 }, { "epoch": 2.460373998219056, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.042961120605469, "learning_rate": 5e-05, "loss": 0.9634, "mean_token_accuracy": 0.881865382194519, "num_tokens": 734531720.0, "step": 19341 }, { "epoch": 2.4605012084976465, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.174506187438965, "learning_rate": 5e-05, "loss": 0.9223, "mean_token_accuracy": 0.8851718306541443, "num_tokens": 734567809.0, "step": 19342 }, { "epoch": 2.460628418776237, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.821172714233398, "learning_rate": 5e-05, "loss": 0.847, "mean_token_accuracy": 0.8993523716926575, "num_tokens": 734605491.0, "step": 19343 }, { "epoch": 2.4607556290548276, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.41817569732666, "learning_rate": 5e-05, "loss": 0.8356, "mean_token_accuracy": 0.889336109161377, "num_tokens": 734637169.0, "step": 19344 }, { "epoch": 2.460882839333418, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.238372802734375, "learning_rate": 5e-05, "loss": 0.7753, "mean_token_accuracy": 0.8863216638565063, "num_tokens": 734668167.0, "step": 19345 }, { "epoch": 2.4610100496120086, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.335301399230957, "learning_rate": 5e-05, "loss": 0.7875, "mean_token_accuracy": 0.8952053189277649, "num_tokens": 734705641.0, "step": 19346 }, { "epoch": 2.461137259890599, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.56142520904541, "learning_rate": 5e-05, "loss": 0.7516, "mean_token_accuracy": 0.877794623374939, "num_tokens": 734742412.0, "step": 19347 }, { "epoch": 2.4612644701691897, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.625595092773438, "learning_rate": 5e-05, "loss": 0.8117, "mean_token_accuracy": 0.8934386968612671, "num_tokens": 734782930.0, "step": 19348 }, { "epoch": 2.4613916804477802, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.093873023986816, "learning_rate": 5e-05, "loss": 0.6865, "mean_token_accuracy": 0.8909693956375122, "num_tokens": 734825051.0, "step": 19349 }, { "epoch": 2.4615188907263708, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 9.046831130981445, "learning_rate": 5e-05, "loss": 0.9144, "mean_token_accuracy": 0.8731021285057068, "num_tokens": 734863524.0, "step": 19350 }, { "epoch": 2.4616461010049613, "ewc_loss": 0.375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.875414848327637, "learning_rate": 5e-05, "loss": 0.7249, "mean_token_accuracy": 0.8751612901687622, "num_tokens": 734899692.0, "step": 19351 }, { "epoch": 2.461773311283552, "ewc_loss": 0.625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.580700874328613, "learning_rate": 5e-05, "loss": 1.0192, "mean_token_accuracy": 0.8651726245880127, "num_tokens": 734933114.0, "step": 19352 }, { "epoch": 2.4619005215621423, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.283456802368164, "learning_rate": 5e-05, "loss": 0.8212, "mean_token_accuracy": 0.8624976873397827, "num_tokens": 734975896.0, "step": 19353 }, { "epoch": 2.462027731840733, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.531559944152832, "learning_rate": 5e-05, "loss": 1.0103, "mean_token_accuracy": 0.8716515302658081, "num_tokens": 735017969.0, "step": 19354 }, { "epoch": 2.4621549421193234, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.260603904724121, "learning_rate": 5e-05, "loss": 0.8746, "mean_token_accuracy": 0.8787247538566589, "num_tokens": 735057052.0, "step": 19355 }, { "epoch": 2.462282152397914, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.274056434631348, "learning_rate": 5e-05, "loss": 0.8048, "mean_token_accuracy": 0.8953052759170532, "num_tokens": 735094211.0, "step": 19356 }, { "epoch": 2.4624093626765045, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.725995063781738, "learning_rate": 5e-05, "loss": 0.9369, "mean_token_accuracy": 0.8619717955589294, "num_tokens": 735134870.0, "step": 19357 }, { "epoch": 2.462536572955095, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.166975975036621, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.870735228061676, "num_tokens": 735177218.0, "step": 19358 }, { "epoch": 2.462663783233685, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.964159965515137, "learning_rate": 5e-05, "loss": 0.9562, "mean_token_accuracy": 0.858164370059967, "num_tokens": 735208310.0, "step": 19359 }, { "epoch": 2.462790993512276, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.942106246948242, "learning_rate": 5e-05, "loss": 0.6785, "mean_token_accuracy": 0.8893876671791077, "num_tokens": 735243514.0, "step": 19360 }, { "epoch": 2.462918203790866, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.136571884155273, "learning_rate": 5e-05, "loss": 0.92, "mean_token_accuracy": 0.8744362592697144, "num_tokens": 735278765.0, "step": 19361 }, { "epoch": 2.4630454140694567, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.185822010040283, "learning_rate": 5e-05, "loss": 0.7692, "mean_token_accuracy": 0.8657652139663696, "num_tokens": 735317966.0, "step": 19362 }, { "epoch": 2.463172624348047, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.934530258178711, "learning_rate": 5e-05, "loss": 0.8726, "mean_token_accuracy": 0.8870013952255249, "num_tokens": 735355552.0, "step": 19363 }, { "epoch": 2.4632998346266377, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.703075885772705, "learning_rate": 5e-05, "loss": 0.7978, "mean_token_accuracy": 0.8745613098144531, "num_tokens": 735398011.0, "step": 19364 }, { "epoch": 2.4634270449052282, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.210890769958496, "learning_rate": 5e-05, "loss": 0.7775, "mean_token_accuracy": 0.8948272466659546, "num_tokens": 735436743.0, "step": 19365 }, { "epoch": 2.4635542551838188, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.042094230651855, "learning_rate": 5e-05, "loss": 0.7537, "mean_token_accuracy": 0.8943877220153809, "num_tokens": 735467932.0, "step": 19366 }, { "epoch": 2.4636814654624093, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.624794006347656, "learning_rate": 5e-05, "loss": 0.73, "mean_token_accuracy": 0.8904730081558228, "num_tokens": 735510057.0, "step": 19367 }, { "epoch": 2.463808675741, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.155820846557617, "learning_rate": 5e-05, "loss": 0.7949, "mean_token_accuracy": 0.8752155900001526, "num_tokens": 735553700.0, "step": 19368 }, { "epoch": 2.4639358860195903, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.480040073394775, "learning_rate": 5e-05, "loss": 0.7802, "mean_token_accuracy": 0.8644461035728455, "num_tokens": 735598490.0, "step": 19369 }, { "epoch": 2.464063096298181, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.05868148803711, "learning_rate": 5e-05, "loss": 0.7748, "mean_token_accuracy": 0.8773350119590759, "num_tokens": 735637987.0, "step": 19370 }, { "epoch": 2.4641903065767714, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.528989315032959, "learning_rate": 5e-05, "loss": 0.6658, "mean_token_accuracy": 0.9012584686279297, "num_tokens": 735676327.0, "step": 19371 }, { "epoch": 2.464317516855362, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.880612373352051, "learning_rate": 5e-05, "loss": 0.7541, "mean_token_accuracy": 0.8787386417388916, "num_tokens": 735713015.0, "step": 19372 }, { "epoch": 2.4644447271339525, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.613097190856934, "learning_rate": 5e-05, "loss": 0.714, "mean_token_accuracy": 0.8834021687507629, "num_tokens": 735750561.0, "step": 19373 }, { "epoch": 2.464571937412543, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.684013843536377, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8740710020065308, "num_tokens": 735788264.0, "step": 19374 }, { "epoch": 2.4646991476911335, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.74673318862915, "learning_rate": 5e-05, "loss": 0.7509, "mean_token_accuracy": 0.8821889758110046, "num_tokens": 735829384.0, "step": 19375 }, { "epoch": 2.464826357969724, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.494112014770508, "learning_rate": 5e-05, "loss": 0.6699, "mean_token_accuracy": 0.8959381580352783, "num_tokens": 735868323.0, "step": 19376 }, { "epoch": 2.4649535682483146, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.881438255310059, "learning_rate": 5e-05, "loss": 0.7007, "mean_token_accuracy": 0.891141414642334, "num_tokens": 735902107.0, "step": 19377 }, { "epoch": 2.465080778526905, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.373654842376709, "learning_rate": 5e-05, "loss": 0.6795, "mean_token_accuracy": 0.8882144689559937, "num_tokens": 735941501.0, "step": 19378 }, { "epoch": 2.4652079888054956, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.9370927810668945, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.8764277696609497, "num_tokens": 735987721.0, "step": 19379 }, { "epoch": 2.465335199084086, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.3091816902160645, "learning_rate": 5e-05, "loss": 0.6595, "mean_token_accuracy": 0.8946149349212646, "num_tokens": 736022126.0, "step": 19380 }, { "epoch": 2.4654624093626767, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.971200942993164, "learning_rate": 5e-05, "loss": 0.7373, "mean_token_accuracy": 0.8861813545227051, "num_tokens": 736059782.0, "step": 19381 }, { "epoch": 2.4655896196412668, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.218502521514893, "learning_rate": 5e-05, "loss": 0.7179, "mean_token_accuracy": 0.8683900833129883, "num_tokens": 736101514.0, "step": 19382 }, { "epoch": 2.4657168299198577, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.128039360046387, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.8883218765258789, "num_tokens": 736137861.0, "step": 19383 }, { "epoch": 2.465844040198448, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.976717948913574, "learning_rate": 5e-05, "loss": 0.6807, "mean_token_accuracy": 0.8769022226333618, "num_tokens": 736177567.0, "step": 19384 }, { "epoch": 2.4659712504770384, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.49116039276123, "learning_rate": 5e-05, "loss": 0.7852, "mean_token_accuracy": 0.8792952299118042, "num_tokens": 736214893.0, "step": 19385 }, { "epoch": 2.466098460755629, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.546957492828369, "learning_rate": 5e-05, "loss": 0.6963, "mean_token_accuracy": 0.870758056640625, "num_tokens": 736250539.0, "step": 19386 }, { "epoch": 2.4662256710342194, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 9.132975578308105, "learning_rate": 5e-05, "loss": 0.8157, "mean_token_accuracy": 0.8921117186546326, "num_tokens": 736280924.0, "step": 19387 }, { "epoch": 2.46635288131281, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00023365020751953125, "grad_norm": 6.013606071472168, "learning_rate": 5e-05, "loss": 0.6459, "mean_token_accuracy": 0.8824107050895691, "num_tokens": 736322446.0, "step": 19388 }, { "epoch": 2.4664800915914005, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 10.064288139343262, "learning_rate": 5e-05, "loss": 0.9567, "mean_token_accuracy": 0.8861061334609985, "num_tokens": 736361370.0, "step": 19389 }, { "epoch": 2.466607301869991, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 5.992039680480957, "learning_rate": 5e-05, "loss": 0.684, "mean_token_accuracy": 0.8823798894882202, "num_tokens": 736401931.0, "step": 19390 }, { "epoch": 2.4667345121485815, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.681416511535645, "learning_rate": 5e-05, "loss": 1.0781, "mean_token_accuracy": 0.884017288684845, "num_tokens": 736445125.0, "step": 19391 }, { "epoch": 2.466861722427172, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.586848735809326, "learning_rate": 5e-05, "loss": 0.8661, "mean_token_accuracy": 0.9003708362579346, "num_tokens": 736488409.0, "step": 19392 }, { "epoch": 2.4669889327057626, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.766276359558105, "learning_rate": 5e-05, "loss": 1.0108, "mean_token_accuracy": 0.8788360357284546, "num_tokens": 736526979.0, "step": 19393 }, { "epoch": 2.467116142984353, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 9.069571495056152, "learning_rate": 5e-05, "loss": 1.0248, "mean_token_accuracy": 0.8885324001312256, "num_tokens": 736571933.0, "step": 19394 }, { "epoch": 2.4672433532629436, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.600413799285889, "learning_rate": 5e-05, "loss": 0.8395, "mean_token_accuracy": 0.8649109601974487, "num_tokens": 736610423.0, "step": 19395 }, { "epoch": 2.467370563541534, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.276119232177734, "learning_rate": 5e-05, "loss": 1.0844, "mean_token_accuracy": 0.8732907176017761, "num_tokens": 736648432.0, "step": 19396 }, { "epoch": 2.4674977738201247, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.847647666931152, "learning_rate": 5e-05, "loss": 0.8216, "mean_token_accuracy": 0.8625039458274841, "num_tokens": 736688032.0, "step": 19397 }, { "epoch": 2.4676249840987152, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 11.053114891052246, "learning_rate": 5e-05, "loss": 1.1354, "mean_token_accuracy": 0.8816778659820557, "num_tokens": 736725785.0, "step": 19398 }, { "epoch": 2.4677521943773058, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 8.343465805053711, "learning_rate": 5e-05, "loss": 1.0726, "mean_token_accuracy": 0.8607698678970337, "num_tokens": 736760229.0, "step": 19399 }, { "epoch": 2.4678794046558963, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.15678882598877, "learning_rate": 5e-05, "loss": 1.0012, "mean_token_accuracy": 0.8814036846160889, "num_tokens": 736794981.0, "step": 19400 }, { "epoch": 2.468006614934487, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 9.057414054870605, "learning_rate": 5e-05, "loss": 1.0658, "mean_token_accuracy": 0.8810664415359497, "num_tokens": 736835387.0, "step": 19401 }, { "epoch": 2.4681338252130773, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.818208694458008, "learning_rate": 5e-05, "loss": 0.8549, "mean_token_accuracy": 0.8987951278686523, "num_tokens": 736877047.0, "step": 19402 }, { "epoch": 2.468261035491668, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.01222038269043, "learning_rate": 5e-05, "loss": 0.9961, "mean_token_accuracy": 0.874010443687439, "num_tokens": 736916637.0, "step": 19403 }, { "epoch": 2.4683882457702584, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.209419250488281, "learning_rate": 5e-05, "loss": 0.8435, "mean_token_accuracy": 0.8884707689285278, "num_tokens": 736954905.0, "step": 19404 }, { "epoch": 2.468515456048849, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.5347900390625, "learning_rate": 5e-05, "loss": 0.8627, "mean_token_accuracy": 0.888735294342041, "num_tokens": 736992264.0, "step": 19405 }, { "epoch": 2.4686426663274394, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.185839653015137, "learning_rate": 5e-05, "loss": 0.9143, "mean_token_accuracy": 0.8652521371841431, "num_tokens": 737032896.0, "step": 19406 }, { "epoch": 2.4687698766060295, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.8098578453063965, "learning_rate": 5e-05, "loss": 0.7619, "mean_token_accuracy": 0.8875618577003479, "num_tokens": 737073734.0, "step": 19407 }, { "epoch": 2.4688970868846205, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.418462753295898, "learning_rate": 5e-05, "loss": 0.8445, "mean_token_accuracy": 0.8865841031074524, "num_tokens": 737112246.0, "step": 19408 }, { "epoch": 2.4690242971632106, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.25178337097168, "learning_rate": 5e-05, "loss": 0.7464, "mean_token_accuracy": 0.8772904872894287, "num_tokens": 737148378.0, "step": 19409 }, { "epoch": 2.469151507441801, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.750239372253418, "learning_rate": 5e-05, "loss": 0.8958, "mean_token_accuracy": 0.8711169958114624, "num_tokens": 737182815.0, "step": 19410 }, { "epoch": 2.4692787177203916, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.918164253234863, "learning_rate": 5e-05, "loss": 0.7031, "mean_token_accuracy": 0.8872872591018677, "num_tokens": 737223301.0, "step": 19411 }, { "epoch": 2.469405927998982, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 9.108431816101074, "learning_rate": 5e-05, "loss": 0.912, "mean_token_accuracy": 0.8758352994918823, "num_tokens": 737254184.0, "step": 19412 }, { "epoch": 2.4695331382775727, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.803448677062988, "learning_rate": 5e-05, "loss": 0.6736, "mean_token_accuracy": 0.8984241485595703, "num_tokens": 737292540.0, "step": 19413 }, { "epoch": 2.4696603485561632, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.298260688781738, "learning_rate": 5e-05, "loss": 0.9766, "mean_token_accuracy": 0.8670940399169922, "num_tokens": 737329822.0, "step": 19414 }, { "epoch": 2.4697875588347538, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.074909687042236, "learning_rate": 5e-05, "loss": 0.766, "mean_token_accuracy": 0.8888957500457764, "num_tokens": 737365416.0, "step": 19415 }, { "epoch": 2.4699147691133443, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.971470832824707, "learning_rate": 5e-05, "loss": 0.9084, "mean_token_accuracy": 0.8845373392105103, "num_tokens": 737411498.0, "step": 19416 }, { "epoch": 2.470041979391935, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.500674247741699, "learning_rate": 5e-05, "loss": 0.8216, "mean_token_accuracy": 0.8897711038589478, "num_tokens": 737453667.0, "step": 19417 }, { "epoch": 2.4701691896705253, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.431885719299316, "learning_rate": 5e-05, "loss": 0.8301, "mean_token_accuracy": 0.8876262903213501, "num_tokens": 737487095.0, "step": 19418 }, { "epoch": 2.470296399949116, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.7155961990356445, "learning_rate": 5e-05, "loss": 0.8113, "mean_token_accuracy": 0.8900818824768066, "num_tokens": 737521513.0, "step": 19419 }, { "epoch": 2.4704236102277064, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.041860580444336, "learning_rate": 5e-05, "loss": 0.7989, "mean_token_accuracy": 0.8867455720901489, "num_tokens": 737559218.0, "step": 19420 }, { "epoch": 2.470550820506297, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.716273784637451, "learning_rate": 5e-05, "loss": 0.8515, "mean_token_accuracy": 0.8685615062713623, "num_tokens": 737600486.0, "step": 19421 }, { "epoch": 2.4706780307848875, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.903985023498535, "learning_rate": 5e-05, "loss": 0.8126, "mean_token_accuracy": 0.8703959584236145, "num_tokens": 737639035.0, "step": 19422 }, { "epoch": 2.470805241063478, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.6107964515686035, "learning_rate": 5e-05, "loss": 0.7064, "mean_token_accuracy": 0.8988457918167114, "num_tokens": 737676582.0, "step": 19423 }, { "epoch": 2.4709324513420685, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.932522296905518, "learning_rate": 5e-05, "loss": 0.7787, "mean_token_accuracy": 0.8832753896713257, "num_tokens": 737714809.0, "step": 19424 }, { "epoch": 2.471059661620659, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.423259735107422, "learning_rate": 5e-05, "loss": 0.7344, "mean_token_accuracy": 0.8751739263534546, "num_tokens": 737757803.0, "step": 19425 }, { "epoch": 2.4711868718992496, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.094639778137207, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.8796370029449463, "num_tokens": 737796563.0, "step": 19426 }, { "epoch": 2.47131408217784, "ewc_loss": 0.375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.207551956176758, "learning_rate": 5e-05, "loss": 0.7418, "mean_token_accuracy": 0.8743786811828613, "num_tokens": 737829711.0, "step": 19427 }, { "epoch": 2.4714412924564306, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.271346092224121, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8845123052597046, "num_tokens": 737862632.0, "step": 19428 }, { "epoch": 2.471568502735021, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.02241849899292, "learning_rate": 5e-05, "loss": 0.7118, "mean_token_accuracy": 0.8796197175979614, "num_tokens": 737900188.0, "step": 19429 }, { "epoch": 2.4716957130136117, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.502081871032715, "learning_rate": 5e-05, "loss": 0.7975, "mean_token_accuracy": 0.8874058127403259, "num_tokens": 737934667.0, "step": 19430 }, { "epoch": 2.471822923292202, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.82718563079834, "learning_rate": 5e-05, "loss": 0.6665, "mean_token_accuracy": 0.8872835040092468, "num_tokens": 737968386.0, "step": 19431 }, { "epoch": 2.4719501335707923, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.789923667907715, "learning_rate": 5e-05, "loss": 0.8528, "mean_token_accuracy": 0.8799397945404053, "num_tokens": 738011761.0, "step": 19432 }, { "epoch": 2.4720773438493833, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.726355075836182, "learning_rate": 5e-05, "loss": 0.6919, "mean_token_accuracy": 0.8858953714370728, "num_tokens": 738049319.0, "step": 19433 }, { "epoch": 2.4722045541279734, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.903860092163086, "learning_rate": 5e-05, "loss": 0.8826, "mean_token_accuracy": 0.8747618198394775, "num_tokens": 738083725.0, "step": 19434 }, { "epoch": 2.472331764406564, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.94343376159668, "learning_rate": 5e-05, "loss": 0.7413, "mean_token_accuracy": 0.876035213470459, "num_tokens": 738126330.0, "step": 19435 }, { "epoch": 2.4724589746851544, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.628034591674805, "learning_rate": 5e-05, "loss": 0.8822, "mean_token_accuracy": 0.8781888484954834, "num_tokens": 738163564.0, "step": 19436 }, { "epoch": 2.472586184963745, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.412879467010498, "learning_rate": 5e-05, "loss": 0.8064, "mean_token_accuracy": 0.8745508193969727, "num_tokens": 738197014.0, "step": 19437 }, { "epoch": 2.4727133952423355, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.981720447540283, "learning_rate": 5e-05, "loss": 0.7967, "mean_token_accuracy": 0.8841140270233154, "num_tokens": 738232263.0, "step": 19438 }, { "epoch": 2.472840605520926, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.9174485206604, "learning_rate": 5e-05, "loss": 0.7803, "mean_token_accuracy": 0.889040470123291, "num_tokens": 738273839.0, "step": 19439 }, { "epoch": 2.4729678157995165, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.335713863372803, "learning_rate": 5e-05, "loss": 0.6881, "mean_token_accuracy": 0.895824134349823, "num_tokens": 738314062.0, "step": 19440 }, { "epoch": 2.473095026078107, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.337329864501953, "learning_rate": 5e-05, "loss": 0.841, "mean_token_accuracy": 0.8762919902801514, "num_tokens": 738352330.0, "step": 19441 }, { "epoch": 2.4732222363566976, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.91615629196167, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.8634880781173706, "num_tokens": 738388527.0, "step": 19442 }, { "epoch": 2.473349446635288, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.657434463500977, "learning_rate": 5e-05, "loss": 0.8609, "mean_token_accuracy": 0.8740345239639282, "num_tokens": 738422913.0, "step": 19443 }, { "epoch": 2.4734766569138786, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.695815086364746, "learning_rate": 5e-05, "loss": 0.7055, "mean_token_accuracy": 0.8753604888916016, "num_tokens": 738463739.0, "step": 19444 }, { "epoch": 2.473603867192469, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.928821563720703, "learning_rate": 5e-05, "loss": 0.8801, "mean_token_accuracy": 0.8741476535797119, "num_tokens": 738499160.0, "step": 19445 }, { "epoch": 2.4737310774710597, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.609967231750488, "learning_rate": 5e-05, "loss": 0.6718, "mean_token_accuracy": 0.8907071948051453, "num_tokens": 738546634.0, "step": 19446 }, { "epoch": 2.47385828774965, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.101149559020996, "learning_rate": 5e-05, "loss": 0.8954, "mean_token_accuracy": 0.8862388134002686, "num_tokens": 738581021.0, "step": 19447 }, { "epoch": 2.4739854980282407, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.867068290710449, "learning_rate": 5e-05, "loss": 0.7255, "mean_token_accuracy": 0.8824626803398132, "num_tokens": 738622606.0, "step": 19448 }, { "epoch": 2.4741127083068313, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.866076469421387, "learning_rate": 5e-05, "loss": 0.8981, "mean_token_accuracy": 0.8798470497131348, "num_tokens": 738656244.0, "step": 19449 }, { "epoch": 2.474239918585422, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.376348972320557, "learning_rate": 5e-05, "loss": 0.833, "mean_token_accuracy": 0.8722454309463501, "num_tokens": 738694743.0, "step": 19450 }, { "epoch": 2.4743671288640123, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.268068313598633, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8745120167732239, "num_tokens": 738731447.0, "step": 19451 }, { "epoch": 2.474494339142603, "ewc_loss": 0.5, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.770211696624756, "learning_rate": 5e-05, "loss": 0.8116, "mean_token_accuracy": 0.8902348279953003, "num_tokens": 738764631.0, "step": 19452 }, { "epoch": 2.4746215494211934, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.823619365692139, "learning_rate": 5e-05, "loss": 0.7794, "mean_token_accuracy": 0.8895090818405151, "num_tokens": 738807685.0, "step": 19453 }, { "epoch": 2.474748759699784, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.753079414367676, "learning_rate": 5e-05, "loss": 0.8092, "mean_token_accuracy": 0.8772106170654297, "num_tokens": 738851717.0, "step": 19454 }, { "epoch": 2.4748759699783744, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.893191814422607, "learning_rate": 5e-05, "loss": 0.8425, "mean_token_accuracy": 0.8659927845001221, "num_tokens": 738889432.0, "step": 19455 }, { "epoch": 2.475003180256965, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.350153923034668, "learning_rate": 5e-05, "loss": 0.7572, "mean_token_accuracy": 0.8734753131866455, "num_tokens": 738923197.0, "step": 19456 }, { "epoch": 2.475130390535555, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.31016731262207, "learning_rate": 5e-05, "loss": 0.8059, "mean_token_accuracy": 0.8860365152359009, "num_tokens": 738959699.0, "step": 19457 }, { "epoch": 2.475257600814146, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.915592193603516, "learning_rate": 5e-05, "loss": 0.6711, "mean_token_accuracy": 0.8946917653083801, "num_tokens": 738990444.0, "step": 19458 }, { "epoch": 2.475384811092736, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.718960762023926, "learning_rate": 5e-05, "loss": 0.8361, "mean_token_accuracy": 0.8922842144966125, "num_tokens": 739030864.0, "step": 19459 }, { "epoch": 2.4755120213713266, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.715945720672607, "learning_rate": 5e-05, "loss": 0.7154, "mean_token_accuracy": 0.8759756088256836, "num_tokens": 739068407.0, "step": 19460 }, { "epoch": 2.475639231649917, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.901433944702148, "learning_rate": 5e-05, "loss": 0.8881, "mean_token_accuracy": 0.8818570375442505, "num_tokens": 739105560.0, "step": 19461 }, { "epoch": 2.4757664419285077, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.8907294273376465, "learning_rate": 5e-05, "loss": 0.7538, "mean_token_accuracy": 0.8708829879760742, "num_tokens": 739142056.0, "step": 19462 }, { "epoch": 2.4758936522070982, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000728607177734375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.653131484985352, "learning_rate": 5e-05, "loss": 0.9406, "mean_token_accuracy": 0.8619334697723389, "num_tokens": 739185407.0, "step": 19463 }, { "epoch": 2.4760208624856888, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.380407810211182, "learning_rate": 5e-05, "loss": 0.8056, "mean_token_accuracy": 0.8759963512420654, "num_tokens": 739222594.0, "step": 19464 }, { "epoch": 2.4761480727642793, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.983038425445557, "learning_rate": 5e-05, "loss": 0.8128, "mean_token_accuracy": 0.883152186870575, "num_tokens": 739261021.0, "step": 19465 }, { "epoch": 2.47627528304287, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.986347198486328, "learning_rate": 5e-05, "loss": 0.8391, "mean_token_accuracy": 0.8783689737319946, "num_tokens": 739300828.0, "step": 19466 }, { "epoch": 2.4764024933214603, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.221194267272949, "learning_rate": 5e-05, "loss": 0.7086, "mean_token_accuracy": 0.8889882564544678, "num_tokens": 739338724.0, "step": 19467 }, { "epoch": 2.476529703600051, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.42673397064209, "learning_rate": 5e-05, "loss": 0.7922, "mean_token_accuracy": 0.889228343963623, "num_tokens": 739377821.0, "step": 19468 }, { "epoch": 2.4766569138786414, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.828608989715576, "learning_rate": 5e-05, "loss": 0.6644, "mean_token_accuracy": 0.8906759023666382, "num_tokens": 739411438.0, "step": 19469 }, { "epoch": 2.476784124157232, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.696599006652832, "learning_rate": 5e-05, "loss": 0.8758, "mean_token_accuracy": 0.8704644441604614, "num_tokens": 739457836.0, "step": 19470 }, { "epoch": 2.4769113344358225, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.666673183441162, "learning_rate": 5e-05, "loss": 0.7254, "mean_token_accuracy": 0.8706729412078857, "num_tokens": 739496336.0, "step": 19471 }, { "epoch": 2.477038544714413, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.893891334533691, "learning_rate": 5e-05, "loss": 0.8697, "mean_token_accuracy": 0.8801602721214294, "num_tokens": 739539777.0, "step": 19472 }, { "epoch": 2.4771657549930035, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.676692485809326, "learning_rate": 5e-05, "loss": 0.6682, "mean_token_accuracy": 0.8944733738899231, "num_tokens": 739572560.0, "step": 19473 }, { "epoch": 2.477292965271594, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.940422058105469, "learning_rate": 5e-05, "loss": 0.892, "mean_token_accuracy": 0.885138750076294, "num_tokens": 739604260.0, "step": 19474 }, { "epoch": 2.4774201755501846, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.95022439956665, "learning_rate": 5e-05, "loss": 0.7646, "mean_token_accuracy": 0.8789888620376587, "num_tokens": 739645241.0, "step": 19475 }, { "epoch": 2.477547385828775, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.645035743713379, "learning_rate": 5e-05, "loss": 0.889, "mean_token_accuracy": 0.8833449482917786, "num_tokens": 739683516.0, "step": 19476 }, { "epoch": 2.4776745961073656, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.406952857971191, "learning_rate": 5e-05, "loss": 0.8152, "mean_token_accuracy": 0.8794283866882324, "num_tokens": 739720422.0, "step": 19477 }, { "epoch": 2.477801806385956, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.07408332824707, "learning_rate": 5e-05, "loss": 0.8042, "mean_token_accuracy": 0.8930368423461914, "num_tokens": 739758071.0, "step": 19478 }, { "epoch": 2.4779290166645467, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.753141403198242, "learning_rate": 5e-05, "loss": 0.8127, "mean_token_accuracy": 0.8830918073654175, "num_tokens": 739796143.0, "step": 19479 }, { "epoch": 2.4780562269431368, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.724916458129883, "learning_rate": 5e-05, "loss": 0.8233, "mean_token_accuracy": 0.8757437467575073, "num_tokens": 739832038.0, "step": 19480 }, { "epoch": 2.4781834372217277, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.7184038162231445, "learning_rate": 5e-05, "loss": 0.7791, "mean_token_accuracy": 0.883462131023407, "num_tokens": 739873232.0, "step": 19481 }, { "epoch": 2.478310647500318, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.7745256423950195, "learning_rate": 5e-05, "loss": 0.7567, "mean_token_accuracy": 0.8871310949325562, "num_tokens": 739912213.0, "step": 19482 }, { "epoch": 2.4784378577789083, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.4365692138671875, "learning_rate": 5e-05, "loss": 0.7511, "mean_token_accuracy": 0.8793846368789673, "num_tokens": 739949363.0, "step": 19483 }, { "epoch": 2.478565068057499, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.059331893920898, "learning_rate": 5e-05, "loss": 0.7826, "mean_token_accuracy": 0.8860431909561157, "num_tokens": 739979360.0, "step": 19484 }, { "epoch": 2.4786922783360894, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.083725929260254, "learning_rate": 5e-05, "loss": 0.7632, "mean_token_accuracy": 0.8655635118484497, "num_tokens": 740019691.0, "step": 19485 }, { "epoch": 2.47881948861468, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.398199081420898, "learning_rate": 5e-05, "loss": 0.8015, "mean_token_accuracy": 0.8879613876342773, "num_tokens": 740053604.0, "step": 19486 }, { "epoch": 2.4789466988932705, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.798614501953125, "learning_rate": 5e-05, "loss": 0.7275, "mean_token_accuracy": 0.8745864629745483, "num_tokens": 740090220.0, "step": 19487 }, { "epoch": 2.479073909171861, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.684908866882324, "learning_rate": 5e-05, "loss": 0.8083, "mean_token_accuracy": 0.8920797109603882, "num_tokens": 740126361.0, "step": 19488 }, { "epoch": 2.4792011194504515, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.668100833892822, "learning_rate": 5e-05, "loss": 0.7038, "mean_token_accuracy": 0.8808516263961792, "num_tokens": 740164752.0, "step": 19489 }, { "epoch": 2.479328329729042, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.86286449432373, "learning_rate": 5e-05, "loss": 0.8592, "mean_token_accuracy": 0.8863623738288879, "num_tokens": 740199622.0, "step": 19490 }, { "epoch": 2.4794555400076326, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.768165588378906, "learning_rate": 5e-05, "loss": 0.7173, "mean_token_accuracy": 0.8809218406677246, "num_tokens": 740236457.0, "step": 19491 }, { "epoch": 2.479582750286223, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.764238357543945, "learning_rate": 5e-05, "loss": 0.9082, "mean_token_accuracy": 0.8740887641906738, "num_tokens": 740275094.0, "step": 19492 }, { "epoch": 2.4797099605648136, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.171690464019775, "learning_rate": 5e-05, "loss": 0.7858, "mean_token_accuracy": 0.8794004917144775, "num_tokens": 740308564.0, "step": 19493 }, { "epoch": 2.479837170843404, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.229758262634277, "learning_rate": 5e-05, "loss": 0.8478, "mean_token_accuracy": 0.8804384469985962, "num_tokens": 740354994.0, "step": 19494 }, { "epoch": 2.4799643811219947, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.723013877868652, "learning_rate": 5e-05, "loss": 0.8036, "mean_token_accuracy": 0.8924447894096375, "num_tokens": 740394924.0, "step": 19495 }, { "epoch": 2.480091591400585, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000732421875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.509947299957275, "learning_rate": 5e-05, "loss": 0.7378, "mean_token_accuracy": 0.8910855054855347, "num_tokens": 740428362.0, "step": 19496 }, { "epoch": 2.4802188016791757, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.158843994140625, "learning_rate": 5e-05, "loss": 0.8554, "mean_token_accuracy": 0.8709741830825806, "num_tokens": 740473987.0, "step": 19497 }, { "epoch": 2.4803460119577663, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.013955116271973, "learning_rate": 5e-05, "loss": 0.7512, "mean_token_accuracy": 0.8669569492340088, "num_tokens": 740512458.0, "step": 19498 }, { "epoch": 2.480473222236357, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.495347023010254, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.8628389835357666, "num_tokens": 740558744.0, "step": 19499 }, { "epoch": 2.4806004325149473, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.692444801330566, "learning_rate": 5e-05, "loss": 0.7317, "mean_token_accuracy": 0.8712339401245117, "num_tokens": 740599889.0, "step": 19500 }, { "epoch": 2.480727642793538, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.852155685424805, "learning_rate": 5e-05, "loss": 0.8735, "mean_token_accuracy": 0.8817005753517151, "num_tokens": 740636162.0, "step": 19501 }, { "epoch": 2.4808548530721284, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.474668025970459, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.8625656962394714, "num_tokens": 740679595.0, "step": 19502 }, { "epoch": 2.480982063350719, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.253623008728027, "learning_rate": 5e-05, "loss": 0.9178, "mean_token_accuracy": 0.8880932331085205, "num_tokens": 740714830.0, "step": 19503 }, { "epoch": 2.4811092736293094, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.568054676055908, "learning_rate": 5e-05, "loss": 0.7076, "mean_token_accuracy": 0.8866986632347107, "num_tokens": 740749748.0, "step": 19504 }, { "epoch": 2.4812364839078995, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.284339904785156, "learning_rate": 5e-05, "loss": 0.9994, "mean_token_accuracy": 0.8691167831420898, "num_tokens": 740785110.0, "step": 19505 }, { "epoch": 2.4813636941864905, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.243223667144775, "learning_rate": 5e-05, "loss": 0.8508, "mean_token_accuracy": 0.8841661214828491, "num_tokens": 740824432.0, "step": 19506 }, { "epoch": 2.4814909044650806, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.54537582397461, "learning_rate": 5e-05, "loss": 0.9377, "mean_token_accuracy": 0.8787146806716919, "num_tokens": 740861403.0, "step": 19507 }, { "epoch": 2.481618114743671, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.024191856384277, "learning_rate": 5e-05, "loss": 0.9103, "mean_token_accuracy": 0.8778584003448486, "num_tokens": 740899299.0, "step": 19508 }, { "epoch": 2.4817453250222616, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.516849040985107, "learning_rate": 5e-05, "loss": 0.7904, "mean_token_accuracy": 0.8861053586006165, "num_tokens": 740934178.0, "step": 19509 }, { "epoch": 2.481872535300852, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.428154945373535, "learning_rate": 5e-05, "loss": 0.8997, "mean_token_accuracy": 0.8768351674079895, "num_tokens": 740967010.0, "step": 19510 }, { "epoch": 2.4819997455794427, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.166842937469482, "learning_rate": 5e-05, "loss": 0.765, "mean_token_accuracy": 0.8761796951293945, "num_tokens": 741004655.0, "step": 19511 }, { "epoch": 2.4821269558580332, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.208213806152344, "learning_rate": 5e-05, "loss": 0.8148, "mean_token_accuracy": 0.8915101289749146, "num_tokens": 741037365.0, "step": 19512 }, { "epoch": 2.4822541661366238, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.487854480743408, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.8768640160560608, "num_tokens": 741072304.0, "step": 19513 }, { "epoch": 2.4823813764152143, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.635503768920898, "learning_rate": 5e-05, "loss": 0.7671, "mean_token_accuracy": 0.8814001679420471, "num_tokens": 741107187.0, "step": 19514 }, { "epoch": 2.482508586693805, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.940526485443115, "learning_rate": 5e-05, "loss": 0.7991, "mean_token_accuracy": 0.8787137269973755, "num_tokens": 741142152.0, "step": 19515 }, { "epoch": 2.4826357969723953, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.1609392166137695, "learning_rate": 5e-05, "loss": 0.7132, "mean_token_accuracy": 0.8865215182304382, "num_tokens": 741175704.0, "step": 19516 }, { "epoch": 2.482763007250986, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.246933937072754, "learning_rate": 5e-05, "loss": 0.7874, "mean_token_accuracy": 0.8907206058502197, "num_tokens": 741212187.0, "step": 19517 }, { "epoch": 2.4828902175295764, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.889443397521973, "learning_rate": 5e-05, "loss": 0.7046, "mean_token_accuracy": 0.8797450661659241, "num_tokens": 741246830.0, "step": 19518 }, { "epoch": 2.483017427808167, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.502076148986816, "learning_rate": 5e-05, "loss": 0.8363, "mean_token_accuracy": 0.8784816265106201, "num_tokens": 741279672.0, "step": 19519 }, { "epoch": 2.4831446380867574, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.67909574508667, "learning_rate": 5e-05, "loss": 0.6645, "mean_token_accuracy": 0.8903465270996094, "num_tokens": 741318205.0, "step": 19520 }, { "epoch": 2.483271848365348, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.837122917175293, "learning_rate": 5e-05, "loss": 0.8448, "mean_token_accuracy": 0.8869694471359253, "num_tokens": 741356312.0, "step": 19521 }, { "epoch": 2.4833990586439385, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.525155544281006, "learning_rate": 5e-05, "loss": 0.7541, "mean_token_accuracy": 0.870375394821167, "num_tokens": 741393370.0, "step": 19522 }, { "epoch": 2.483526268922529, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.163830757141113, "learning_rate": 5e-05, "loss": 0.9575, "mean_token_accuracy": 0.869734525680542, "num_tokens": 741428533.0, "step": 19523 }, { "epoch": 2.4836534792011196, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.635042190551758, "learning_rate": 5e-05, "loss": 0.7639, "mean_token_accuracy": 0.8718228340148926, "num_tokens": 741464614.0, "step": 19524 }, { "epoch": 2.48378068947971, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.161199569702148, "learning_rate": 5e-05, "loss": 0.9597, "mean_token_accuracy": 0.8770477175712585, "num_tokens": 741499126.0, "step": 19525 }, { "epoch": 2.4839078997583006, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.146822929382324, "learning_rate": 5e-05, "loss": 0.8137, "mean_token_accuracy": 0.8867813348770142, "num_tokens": 741544381.0, "step": 19526 }, { "epoch": 2.484035110036891, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.622817039489746, "learning_rate": 5e-05, "loss": 0.8901, "mean_token_accuracy": 0.8956671953201294, "num_tokens": 741588627.0, "step": 19527 }, { "epoch": 2.4841623203154817, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.732316017150879, "learning_rate": 5e-05, "loss": 0.8682, "mean_token_accuracy": 0.8853491544723511, "num_tokens": 741631165.0, "step": 19528 }, { "epoch": 2.484289530594072, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.885290622711182, "learning_rate": 5e-05, "loss": 0.8411, "mean_token_accuracy": 0.8836357593536377, "num_tokens": 741668667.0, "step": 19529 }, { "epoch": 2.4844167408726623, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.988161563873291, "learning_rate": 5e-05, "loss": 0.8603, "mean_token_accuracy": 0.8814264535903931, "num_tokens": 741703625.0, "step": 19530 }, { "epoch": 2.4845439511512533, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.573157787322998, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8773848414421082, "num_tokens": 741744371.0, "step": 19531 }, { "epoch": 2.4846711614298433, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.807504653930664, "learning_rate": 5e-05, "loss": 0.8402, "mean_token_accuracy": 0.8723156452178955, "num_tokens": 741786733.0, "step": 19532 }, { "epoch": 2.484798371708434, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.729995250701904, "learning_rate": 5e-05, "loss": 0.7933, "mean_token_accuracy": 0.8813990354537964, "num_tokens": 741825693.0, "step": 19533 }, { "epoch": 2.4849255819870244, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.437527179718018, "learning_rate": 5e-05, "loss": 0.8176, "mean_token_accuracy": 0.8631047010421753, "num_tokens": 741866176.0, "step": 19534 }, { "epoch": 2.485052792265615, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.978980541229248, "learning_rate": 5e-05, "loss": 0.8404, "mean_token_accuracy": 0.8713089227676392, "num_tokens": 741902523.0, "step": 19535 }, { "epoch": 2.4851800025442055, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.1901702880859375, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.8712700605392456, "num_tokens": 741939005.0, "step": 19536 }, { "epoch": 2.485307212822796, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.10128116607666, "learning_rate": 5e-05, "loss": 0.8555, "mean_token_accuracy": 0.8655539751052856, "num_tokens": 741975012.0, "step": 19537 }, { "epoch": 2.4854344231013865, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.1255574226379395, "learning_rate": 5e-05, "loss": 0.7268, "mean_token_accuracy": 0.8827272653579712, "num_tokens": 742008351.0, "step": 19538 }, { "epoch": 2.485561633379977, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.091678619384766, "learning_rate": 5e-05, "loss": 0.7927, "mean_token_accuracy": 0.8837469816207886, "num_tokens": 742045477.0, "step": 19539 }, { "epoch": 2.4856888436585676, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.134645938873291, "learning_rate": 5e-05, "loss": 0.7486, "mean_token_accuracy": 0.8685817718505859, "num_tokens": 742080070.0, "step": 19540 }, { "epoch": 2.485816053937158, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.077849388122559, "learning_rate": 5e-05, "loss": 0.7735, "mean_token_accuracy": 0.8843634128570557, "num_tokens": 742116735.0, "step": 19541 }, { "epoch": 2.4859432642157486, "ewc_loss": 0.375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.112205505371094, "learning_rate": 5e-05, "loss": 0.7194, "mean_token_accuracy": 0.8786394000053406, "num_tokens": 742151492.0, "step": 19542 }, { "epoch": 2.486070474494339, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.13977336883545, "learning_rate": 5e-05, "loss": 0.7545, "mean_token_accuracy": 0.8926367163658142, "num_tokens": 742196913.0, "step": 19543 }, { "epoch": 2.4861976847729297, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.007360458374023, "learning_rate": 5e-05, "loss": 0.6639, "mean_token_accuracy": 0.8958929181098938, "num_tokens": 742241894.0, "step": 19544 }, { "epoch": 2.48632489505152, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.324493408203125, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.8873856067657471, "num_tokens": 742276256.0, "step": 19545 }, { "epoch": 2.4864521053301107, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.84398889541626, "learning_rate": 5e-05, "loss": 0.6586, "mean_token_accuracy": 0.8959463834762573, "num_tokens": 742316834.0, "step": 19546 }, { "epoch": 2.4865793156087013, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.525556564331055, "learning_rate": 5e-05, "loss": 0.8234, "mean_token_accuracy": 0.8819741606712341, "num_tokens": 742359132.0, "step": 19547 }, { "epoch": 2.486706525887292, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.73919153213501, "learning_rate": 5e-05, "loss": 0.7423, "mean_token_accuracy": 0.8669799566268921, "num_tokens": 742399746.0, "step": 19548 }, { "epoch": 2.4868337361658823, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.652490615844727, "learning_rate": 5e-05, "loss": 0.8893, "mean_token_accuracy": 0.8689974546432495, "num_tokens": 742435400.0, "step": 19549 }, { "epoch": 2.486960946444473, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.812382221221924, "learning_rate": 5e-05, "loss": 0.6863, "mean_token_accuracy": 0.8887844085693359, "num_tokens": 742472935.0, "step": 19550 }, { "epoch": 2.4870881567230634, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.524828910827637, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8865456581115723, "num_tokens": 742516841.0, "step": 19551 }, { "epoch": 2.487215367001654, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.170280456542969, "learning_rate": 5e-05, "loss": 0.7946, "mean_token_accuracy": 0.8735198974609375, "num_tokens": 742551057.0, "step": 19552 }, { "epoch": 2.4873425772802444, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.970357894897461, "learning_rate": 5e-05, "loss": 0.817, "mean_token_accuracy": 0.879047155380249, "num_tokens": 742584891.0, "step": 19553 }, { "epoch": 2.487469787558835, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.758500099182129, "learning_rate": 5e-05, "loss": 0.7834, "mean_token_accuracy": 0.8873629570007324, "num_tokens": 742622574.0, "step": 19554 }, { "epoch": 2.487596997837425, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.179944038391113, "learning_rate": 5e-05, "loss": 0.7229, "mean_token_accuracy": 0.8796423673629761, "num_tokens": 742664458.0, "step": 19555 }, { "epoch": 2.487724208116016, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.424107551574707, "learning_rate": 5e-05, "loss": 0.814, "mean_token_accuracy": 0.8830529451370239, "num_tokens": 742704006.0, "step": 19556 }, { "epoch": 2.487851418394606, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.504960060119629, "learning_rate": 5e-05, "loss": 0.7114, "mean_token_accuracy": 0.8699272871017456, "num_tokens": 742736885.0, "step": 19557 }, { "epoch": 2.4879786286731966, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 9.130576133728027, "learning_rate": 5e-05, "loss": 0.9063, "mean_token_accuracy": 0.8764574527740479, "num_tokens": 742773496.0, "step": 19558 }, { "epoch": 2.488105838951787, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.0467448234558105, "learning_rate": 5e-05, "loss": 0.668, "mean_token_accuracy": 0.8835408687591553, "num_tokens": 742812444.0, "step": 19559 }, { "epoch": 2.4882330492303777, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.887332916259766, "learning_rate": 5e-05, "loss": 0.9805, "mean_token_accuracy": 0.8816933035850525, "num_tokens": 742849910.0, "step": 19560 }, { "epoch": 2.488360259508968, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.313929557800293, "learning_rate": 5e-05, "loss": 0.7187, "mean_token_accuracy": 0.8896796703338623, "num_tokens": 742891412.0, "step": 19561 }, { "epoch": 2.4884874697875587, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.052656173706055, "learning_rate": 5e-05, "loss": 1.0701, "mean_token_accuracy": 0.876017689704895, "num_tokens": 742929510.0, "step": 19562 }, { "epoch": 2.4886146800661493, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.7804694175720215, "learning_rate": 5e-05, "loss": 1.0013, "mean_token_accuracy": 0.8717904686927795, "num_tokens": 742969767.0, "step": 19563 }, { "epoch": 2.48874189034474, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.710336685180664, "learning_rate": 5e-05, "loss": 0.9122, "mean_token_accuracy": 0.8863632082939148, "num_tokens": 743001416.0, "step": 19564 }, { "epoch": 2.4888691006233303, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 8.995893478393555, "learning_rate": 5e-05, "loss": 1.0487, "mean_token_accuracy": 0.8770058155059814, "num_tokens": 743037273.0, "step": 19565 }, { "epoch": 2.488996310901921, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.860797882080078, "learning_rate": 5e-05, "loss": 0.7798, "mean_token_accuracy": 0.8747109770774841, "num_tokens": 743072774.0, "step": 19566 }, { "epoch": 2.4891235211805114, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.16684627532959, "learning_rate": 5e-05, "loss": 1.0691, "mean_token_accuracy": 0.8704286217689514, "num_tokens": 743115931.0, "step": 19567 }, { "epoch": 2.489250731459102, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 6.790223121643066, "learning_rate": 5e-05, "loss": 0.8091, "mean_token_accuracy": 0.8904458284378052, "num_tokens": 743158500.0, "step": 19568 }, { "epoch": 2.4893779417376924, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.3792142868042, "learning_rate": 5e-05, "loss": 1.0792, "mean_token_accuracy": 0.8842474818229675, "num_tokens": 743195311.0, "step": 19569 }, { "epoch": 2.489505152016283, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 8.206448554992676, "learning_rate": 5e-05, "loss": 1.041, "mean_token_accuracy": 0.8842393755912781, "num_tokens": 743227978.0, "step": 19570 }, { "epoch": 2.4896323622948735, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.105108261108398, "learning_rate": 5e-05, "loss": 0.934, "mean_token_accuracy": 0.8846873641014099, "num_tokens": 743264297.0, "step": 19571 }, { "epoch": 2.489759572573464, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 8.832719802856445, "learning_rate": 5e-05, "loss": 1.0563, "mean_token_accuracy": 0.8747181296348572, "num_tokens": 743305351.0, "step": 19572 }, { "epoch": 2.4898867828520546, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000743865966796875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.927145481109619, "learning_rate": 5e-05, "loss": 0.8836, "mean_token_accuracy": 0.8754657506942749, "num_tokens": 743345105.0, "step": 19573 }, { "epoch": 2.490013993130645, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.868820190429688, "learning_rate": 5e-05, "loss": 0.9439, "mean_token_accuracy": 0.8884689211845398, "num_tokens": 743383889.0, "step": 19574 }, { "epoch": 2.4901412034092356, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.7630133628845215, "learning_rate": 5e-05, "loss": 0.8886, "mean_token_accuracy": 0.8810244798660278, "num_tokens": 743420045.0, "step": 19575 }, { "epoch": 2.490268413687826, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.259793281555176, "learning_rate": 5e-05, "loss": 0.8729, "mean_token_accuracy": 0.880445659160614, "num_tokens": 743455358.0, "step": 19576 }, { "epoch": 2.4903956239664167, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.055356979370117, "learning_rate": 5e-05, "loss": 0.9521, "mean_token_accuracy": 0.8609906435012817, "num_tokens": 743492337.0, "step": 19577 }, { "epoch": 2.4905228342450068, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.496029376983643, "learning_rate": 5e-05, "loss": 0.7583, "mean_token_accuracy": 0.8868272304534912, "num_tokens": 743526023.0, "step": 19578 }, { "epoch": 2.4906500445235977, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.452109336853027, "learning_rate": 5e-05, "loss": 0.8751, "mean_token_accuracy": 0.8832902908325195, "num_tokens": 743560224.0, "step": 19579 }, { "epoch": 2.490777254802188, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.990261077880859, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.8533917665481567, "num_tokens": 743601403.0, "step": 19580 }, { "epoch": 2.4909044650807783, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.913671493530273, "learning_rate": 5e-05, "loss": 0.9123, "mean_token_accuracy": 0.881895899772644, "num_tokens": 743641677.0, "step": 19581 }, { "epoch": 2.491031675359369, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.713785648345947, "learning_rate": 5e-05, "loss": 0.7523, "mean_token_accuracy": 0.8795623779296875, "num_tokens": 743684615.0, "step": 19582 }, { "epoch": 2.4911588856379594, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.330300331115723, "learning_rate": 5e-05, "loss": 0.9009, "mean_token_accuracy": 0.8955176472663879, "num_tokens": 743718960.0, "step": 19583 }, { "epoch": 2.49128609591655, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 6.780473232269287, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8893436193466187, "num_tokens": 743758336.0, "step": 19584 }, { "epoch": 2.4914133061951405, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.43731689453125, "learning_rate": 5e-05, "loss": 1.0026, "mean_token_accuracy": 0.8713523149490356, "num_tokens": 743803243.0, "step": 19585 }, { "epoch": 2.491540516473731, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.191272258758545, "learning_rate": 5e-05, "loss": 0.8548, "mean_token_accuracy": 0.885768473148346, "num_tokens": 743840945.0, "step": 19586 }, { "epoch": 2.4916677267523215, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.152539253234863, "learning_rate": 5e-05, "loss": 0.9648, "mean_token_accuracy": 0.8831602931022644, "num_tokens": 743884338.0, "step": 19587 }, { "epoch": 2.491794937030912, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.482176303863525, "learning_rate": 5e-05, "loss": 0.9044, "mean_token_accuracy": 0.8804048895835876, "num_tokens": 743925499.0, "step": 19588 }, { "epoch": 2.4919221473095026, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.875962257385254, "learning_rate": 5e-05, "loss": 0.9794, "mean_token_accuracy": 0.8669625520706177, "num_tokens": 743965016.0, "step": 19589 }, { "epoch": 2.492049357588093, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.430535316467285, "learning_rate": 5e-05, "loss": 0.875, "mean_token_accuracy": 0.8752672076225281, "num_tokens": 744000523.0, "step": 19590 }, { "epoch": 2.4921765678666836, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.857295036315918, "learning_rate": 5e-05, "loss": 0.9554, "mean_token_accuracy": 0.8764681816101074, "num_tokens": 744043862.0, "step": 19591 }, { "epoch": 2.492303778145274, "ewc_loss": 0.5, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.28810453414917, "learning_rate": 5e-05, "loss": 0.89, "mean_token_accuracy": 0.8641449809074402, "num_tokens": 744085161.0, "step": 19592 }, { "epoch": 2.4924309884238647, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.917949676513672, "learning_rate": 5e-05, "loss": 0.9374, "mean_token_accuracy": 0.8827751874923706, "num_tokens": 744117575.0, "step": 19593 }, { "epoch": 2.492558198702455, "ewc_loss": 0.5, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.335323810577393, "learning_rate": 5e-05, "loss": 0.9164, "mean_token_accuracy": 0.8565047979354858, "num_tokens": 744156312.0, "step": 19594 }, { "epoch": 2.4926854089810457, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.786473274230957, "learning_rate": 5e-05, "loss": 0.932, "mean_token_accuracy": 0.8816426396369934, "num_tokens": 744191177.0, "step": 19595 }, { "epoch": 2.4928126192596363, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.563126564025879, "learning_rate": 5e-05, "loss": 0.87, "mean_token_accuracy": 0.8751319050788879, "num_tokens": 744230407.0, "step": 19596 }, { "epoch": 2.492939829538227, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.355323791503906, "learning_rate": 5e-05, "loss": 0.9123, "mean_token_accuracy": 0.8772214651107788, "num_tokens": 744270265.0, "step": 19597 }, { "epoch": 2.4930670398168173, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.778098106384277, "learning_rate": 5e-05, "loss": 0.8618, "mean_token_accuracy": 0.8807855248451233, "num_tokens": 744310913.0, "step": 19598 }, { "epoch": 2.493194250095408, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.871265888214111, "learning_rate": 5e-05, "loss": 0.8634, "mean_token_accuracy": 0.8736101984977722, "num_tokens": 744342918.0, "step": 19599 }, { "epoch": 2.4933214603739984, "ewc_loss": 0.5, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.826785564422607, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8768734931945801, "num_tokens": 744379000.0, "step": 19600 }, { "epoch": 2.493448670652589, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.690951347351074, "learning_rate": 5e-05, "loss": 0.7954, "mean_token_accuracy": 0.8779767751693726, "num_tokens": 744414641.0, "step": 19601 }, { "epoch": 2.4935758809311794, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.707524299621582, "learning_rate": 5e-05, "loss": 0.8076, "mean_token_accuracy": 0.8748595714569092, "num_tokens": 744444860.0, "step": 19602 }, { "epoch": 2.4937030912097695, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.698050498962402, "learning_rate": 5e-05, "loss": 0.7621, "mean_token_accuracy": 0.8910362124443054, "num_tokens": 744479657.0, "step": 19603 }, { "epoch": 2.4938303014883605, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.565021991729736, "learning_rate": 5e-05, "loss": 0.7642, "mean_token_accuracy": 0.8805367946624756, "num_tokens": 744520048.0, "step": 19604 }, { "epoch": 2.4939575117669506, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.741909503936768, "learning_rate": 5e-05, "loss": 0.7927, "mean_token_accuracy": 0.8732935786247253, "num_tokens": 744555265.0, "step": 19605 }, { "epoch": 2.494084722045541, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.472358226776123, "learning_rate": 5e-05, "loss": 0.7459, "mean_token_accuracy": 0.8754463195800781, "num_tokens": 744591011.0, "step": 19606 }, { "epoch": 2.4942119323241316, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.757924556732178, "learning_rate": 5e-05, "loss": 0.7777, "mean_token_accuracy": 0.874321460723877, "num_tokens": 744632302.0, "step": 19607 }, { "epoch": 2.494339142602722, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.403562545776367, "learning_rate": 5e-05, "loss": 0.6938, "mean_token_accuracy": 0.8895766735076904, "num_tokens": 744669827.0, "step": 19608 }, { "epoch": 2.4944663528813127, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.813037395477295, "learning_rate": 5e-05, "loss": 0.7327, "mean_token_accuracy": 0.8847123384475708, "num_tokens": 744708438.0, "step": 19609 }, { "epoch": 2.494593563159903, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.297935485839844, "learning_rate": 5e-05, "loss": 0.72, "mean_token_accuracy": 0.881538450717926, "num_tokens": 744744218.0, "step": 19610 }, { "epoch": 2.4947207734384937, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.928242206573486, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.8846579790115356, "num_tokens": 744782805.0, "step": 19611 }, { "epoch": 2.4948479837170843, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.12151575088501, "learning_rate": 5e-05, "loss": 0.6391, "mean_token_accuracy": 0.8989332914352417, "num_tokens": 744814011.0, "step": 19612 }, { "epoch": 2.494975193995675, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.140278816223145, "learning_rate": 5e-05, "loss": 0.7786, "mean_token_accuracy": 0.8811400532722473, "num_tokens": 744857385.0, "step": 19613 }, { "epoch": 2.4951024042742653, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.903833389282227, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8622125387191772, "num_tokens": 744898212.0, "step": 19614 }, { "epoch": 2.495229614552856, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.443270683288574, "learning_rate": 5e-05, "loss": 0.8095, "mean_token_accuracy": 0.8786193132400513, "num_tokens": 744930665.0, "step": 19615 }, { "epoch": 2.4953568248314464, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.621739387512207, "learning_rate": 5e-05, "loss": 0.7142, "mean_token_accuracy": 0.8723483681678772, "num_tokens": 744968725.0, "step": 19616 }, { "epoch": 2.495484035110037, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.844745635986328, "learning_rate": 5e-05, "loss": 0.877, "mean_token_accuracy": 0.8712729215621948, "num_tokens": 745008386.0, "step": 19617 }, { "epoch": 2.4956112453886274, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.428396224975586, "learning_rate": 5e-05, "loss": 0.697, "mean_token_accuracy": 0.8757520914077759, "num_tokens": 745047059.0, "step": 19618 }, { "epoch": 2.495738455667218, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.133536338806152, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.8732568025588989, "num_tokens": 745089768.0, "step": 19619 }, { "epoch": 2.4958656659458085, "ewc_loss": 0.375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.633431911468506, "learning_rate": 5e-05, "loss": 0.7263, "mean_token_accuracy": 0.8776116967201233, "num_tokens": 745130848.0, "step": 19620 }, { "epoch": 2.495992876224399, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.927075386047363, "learning_rate": 5e-05, "loss": 0.9256, "mean_token_accuracy": 0.8770822286605835, "num_tokens": 745166715.0, "step": 19621 }, { "epoch": 2.4961200865029896, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.362998962402344, "learning_rate": 5e-05, "loss": 0.838, "mean_token_accuracy": 0.8714367747306824, "num_tokens": 745200718.0, "step": 19622 }, { "epoch": 2.49624729678158, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.925963401794434, "learning_rate": 5e-05, "loss": 0.8509, "mean_token_accuracy": 0.8726652264595032, "num_tokens": 745240736.0, "step": 19623 }, { "epoch": 2.4963745070601706, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.261990547180176, "learning_rate": 5e-05, "loss": 0.8794, "mean_token_accuracy": 0.8796271085739136, "num_tokens": 745279040.0, "step": 19624 }, { "epoch": 2.496501717338761, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.765821933746338, "learning_rate": 5e-05, "loss": 0.7276, "mean_token_accuracy": 0.8756927847862244, "num_tokens": 745316936.0, "step": 19625 }, { "epoch": 2.4966289276173517, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.152050971984863, "learning_rate": 5e-05, "loss": 0.9833, "mean_token_accuracy": 0.860094428062439, "num_tokens": 745355438.0, "step": 19626 }, { "epoch": 2.496756137895942, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.133275985717773, "learning_rate": 5e-05, "loss": 0.6721, "mean_token_accuracy": 0.8849635124206543, "num_tokens": 745394509.0, "step": 19627 }, { "epoch": 2.4968833481745323, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.913588523864746, "learning_rate": 5e-05, "loss": 1.0017, "mean_token_accuracy": 0.8774349689483643, "num_tokens": 745426347.0, "step": 19628 }, { "epoch": 2.4970105584531233, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.543910026550293, "learning_rate": 5e-05, "loss": 0.7429, "mean_token_accuracy": 0.8861533403396606, "num_tokens": 745457523.0, "step": 19629 }, { "epoch": 2.4971377687317133, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.875475883483887, "learning_rate": 5e-05, "loss": 1.0826, "mean_token_accuracy": 0.8649898767471313, "num_tokens": 745495155.0, "step": 19630 }, { "epoch": 2.497264979010304, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.9180521965026855, "learning_rate": 5e-05, "loss": 0.9831, "mean_token_accuracy": 0.8760647177696228, "num_tokens": 745535466.0, "step": 19631 }, { "epoch": 2.4973921892888944, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.55343246459961, "learning_rate": 5e-05, "loss": 0.9347, "mean_token_accuracy": 0.8766891956329346, "num_tokens": 745569184.0, "step": 19632 }, { "epoch": 2.497519399567485, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.808571815490723, "learning_rate": 5e-05, "loss": 1.0068, "mean_token_accuracy": 0.8736973404884338, "num_tokens": 745610735.0, "step": 19633 }, { "epoch": 2.4976466098460754, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.2347259521484375, "learning_rate": 5e-05, "loss": 0.8158, "mean_token_accuracy": 0.8693928718566895, "num_tokens": 745653513.0, "step": 19634 }, { "epoch": 2.497773820124666, "ewc_loss": 0.625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.096880912780762, "learning_rate": 5e-05, "loss": 0.9719, "mean_token_accuracy": 0.8770892024040222, "num_tokens": 745690007.0, "step": 19635 }, { "epoch": 2.4979010304032565, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.294404029846191, "learning_rate": 5e-05, "loss": 0.8577, "mean_token_accuracy": 0.8673905730247498, "num_tokens": 745724886.0, "step": 19636 }, { "epoch": 2.498028240681847, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.313061714172363, "learning_rate": 5e-05, "loss": 0.8558, "mean_token_accuracy": 0.8861409425735474, "num_tokens": 745760782.0, "step": 19637 }, { "epoch": 2.4981554509604376, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.994500160217285, "learning_rate": 5e-05, "loss": 0.8234, "mean_token_accuracy": 0.8928543329238892, "num_tokens": 745794535.0, "step": 19638 }, { "epoch": 2.498282661239028, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.318310737609863, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8807911276817322, "num_tokens": 745831010.0, "step": 19639 }, { "epoch": 2.4984098715176186, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.54446029663086, "learning_rate": 5e-05, "loss": 0.9057, "mean_token_accuracy": 0.8754123449325562, "num_tokens": 745866761.0, "step": 19640 }, { "epoch": 2.498537081796209, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.911923408508301, "learning_rate": 5e-05, "loss": 0.7166, "mean_token_accuracy": 0.8806003928184509, "num_tokens": 745909081.0, "step": 19641 }, { "epoch": 2.4986642920747997, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.762134552001953, "learning_rate": 5e-05, "loss": 0.8548, "mean_token_accuracy": 0.8924583792686462, "num_tokens": 745942907.0, "step": 19642 }, { "epoch": 2.49879150235339, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.890563488006592, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.8771148920059204, "num_tokens": 745981558.0, "step": 19643 }, { "epoch": 2.4989187126319807, "ewc_loss": 0.5625, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.808734893798828, "learning_rate": 5e-05, "loss": 0.9007, "mean_token_accuracy": 0.8808792233467102, "num_tokens": 746015568.0, "step": 19644 }, { "epoch": 2.4990459229105713, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.9731903076171875, "learning_rate": 5e-05, "loss": 0.7491, "mean_token_accuracy": 0.8858317732810974, "num_tokens": 746049503.0, "step": 19645 }, { "epoch": 2.499173133189162, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.783466339111328, "learning_rate": 5e-05, "loss": 0.9127, "mean_token_accuracy": 0.8817174434661865, "num_tokens": 746086979.0, "step": 19646 }, { "epoch": 2.4993003434677523, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.039591312408447, "learning_rate": 5e-05, "loss": 0.7794, "mean_token_accuracy": 0.8815423250198364, "num_tokens": 746125073.0, "step": 19647 }, { "epoch": 2.499427553746343, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.749103546142578, "learning_rate": 5e-05, "loss": 0.902, "mean_token_accuracy": 0.8841172456741333, "num_tokens": 746161621.0, "step": 19648 }, { "epoch": 2.4995547640249334, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.009215354919434, "learning_rate": 5e-05, "loss": 0.7948, "mean_token_accuracy": 0.8767540454864502, "num_tokens": 746203032.0, "step": 19649 }, { "epoch": 2.499681974303524, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.882566452026367, "learning_rate": 5e-05, "loss": 0.9386, "mean_token_accuracy": 0.87749183177948, "num_tokens": 746233939.0, "step": 19650 }, { "epoch": 2.499809184582114, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 6.95274019241333, "learning_rate": 5e-05, "loss": 0.7596, "mean_token_accuracy": 0.8888771533966064, "num_tokens": 746269175.0, "step": 19651 }, { "epoch": 2.499936394860705, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.976447105407715, "learning_rate": 5e-05, "loss": 0.95, "mean_token_accuracy": 0.8751307725906372, "num_tokens": 746310634.0, "step": 19652 }, { "epoch": 2.500063605139295, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.1273322105407715, "learning_rate": 5e-05, "loss": 0.8111, "mean_token_accuracy": 0.8849532604217529, "num_tokens": 746353683.0, "step": 19653 }, { "epoch": 2.500190815417886, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.78038501739502, "learning_rate": 5e-05, "loss": 0.9491, "mean_token_accuracy": 0.8769586086273193, "num_tokens": 746392299.0, "step": 19654 }, { "epoch": 2.500318025696476, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.485325336456299, "learning_rate": 5e-05, "loss": 0.8456, "mean_token_accuracy": 0.8840451240539551, "num_tokens": 746435276.0, "step": 19655 }, { "epoch": 2.5004452359750666, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.307377815246582, "learning_rate": 5e-05, "loss": 0.9316, "mean_token_accuracy": 0.8721935153007507, "num_tokens": 746476175.0, "step": 19656 }, { "epoch": 2.500572446253657, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.714163780212402, "learning_rate": 5e-05, "loss": 0.8311, "mean_token_accuracy": 0.8873080015182495, "num_tokens": 746517084.0, "step": 19657 }, { "epoch": 2.5006996565322477, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.903965950012207, "learning_rate": 5e-05, "loss": 0.8914, "mean_token_accuracy": 0.8658174276351929, "num_tokens": 746556073.0, "step": 19658 }, { "epoch": 2.500826866810838, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.731585502624512, "learning_rate": 5e-05, "loss": 0.7934, "mean_token_accuracy": 0.8930803537368774, "num_tokens": 746590188.0, "step": 19659 }, { "epoch": 2.5009540770894287, "ewc_loss": 0.46875, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.743443965911865, "learning_rate": 5e-05, "loss": 0.9051, "mean_token_accuracy": 0.852962076663971, "num_tokens": 746634667.0, "step": 19660 }, { "epoch": 2.5010812873680193, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.621766090393066, "learning_rate": 5e-05, "loss": 0.772, "mean_token_accuracy": 0.8843396902084351, "num_tokens": 746677398.0, "step": 19661 }, { "epoch": 2.50120849764661, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.774556636810303, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8853321671485901, "num_tokens": 746711607.0, "step": 19662 }, { "epoch": 2.5013357079252003, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.4829840660095215, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.8864489793777466, "num_tokens": 746751853.0, "step": 19663 }, { "epoch": 2.501462918203791, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.775798320770264, "learning_rate": 5e-05, "loss": 0.8031, "mean_token_accuracy": 0.8690226674079895, "num_tokens": 746792134.0, "step": 19664 }, { "epoch": 2.5015901284823814, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.464315891265869, "learning_rate": 5e-05, "loss": 0.77, "mean_token_accuracy": 0.8727003931999207, "num_tokens": 746830087.0, "step": 19665 }, { "epoch": 2.501717338760972, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.704064846038818, "learning_rate": 5e-05, "loss": 0.7443, "mean_token_accuracy": 0.8866559863090515, "num_tokens": 746864738.0, "step": 19666 }, { "epoch": 2.5018445490395624, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.490626811981201, "learning_rate": 5e-05, "loss": 0.782, "mean_token_accuracy": 0.8636844158172607, "num_tokens": 746904021.0, "step": 19667 }, { "epoch": 2.501971759318153, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.640400409698486, "learning_rate": 5e-05, "loss": 0.7197, "mean_token_accuracy": 0.8844226598739624, "num_tokens": 746942710.0, "step": 19668 }, { "epoch": 2.5020989695967435, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.491670608520508, "learning_rate": 5e-05, "loss": 0.715, "mean_token_accuracy": 0.881227970123291, "num_tokens": 746986694.0, "step": 19669 }, { "epoch": 2.502226179875334, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.636037826538086, "learning_rate": 5e-05, "loss": 0.7025, "mean_token_accuracy": 0.8911113739013672, "num_tokens": 747028440.0, "step": 19670 }, { "epoch": 2.5023533901539246, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.448019981384277, "learning_rate": 5e-05, "loss": 0.692, "mean_token_accuracy": 0.8895370960235596, "num_tokens": 747063862.0, "step": 19671 }, { "epoch": 2.502480600432515, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.659714698791504, "learning_rate": 5e-05, "loss": 0.7393, "mean_token_accuracy": 0.8767566084861755, "num_tokens": 747100469.0, "step": 19672 }, { "epoch": 2.5026078107111056, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.368900775909424, "learning_rate": 5e-05, "loss": 0.6738, "mean_token_accuracy": 0.8912280201911926, "num_tokens": 747133074.0, "step": 19673 }, { "epoch": 2.5027350209896957, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.769268035888672, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8671349287033081, "num_tokens": 747167010.0, "step": 19674 }, { "epoch": 2.5028622312682867, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.214871406555176, "learning_rate": 5e-05, "loss": 0.704, "mean_token_accuracy": 0.877747118473053, "num_tokens": 747201397.0, "step": 19675 }, { "epoch": 2.5029894415468767, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.93942928314209, "learning_rate": 5e-05, "loss": 0.7271, "mean_token_accuracy": 0.888158917427063, "num_tokens": 747238430.0, "step": 19676 }, { "epoch": 2.5031166518254677, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.991889953613281, "learning_rate": 5e-05, "loss": 0.6985, "mean_token_accuracy": 0.8795210123062134, "num_tokens": 747277760.0, "step": 19677 }, { "epoch": 2.503243862104058, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.25246810913086, "learning_rate": 5e-05, "loss": 0.7769, "mean_token_accuracy": 0.8781528472900391, "num_tokens": 747317763.0, "step": 19678 }, { "epoch": 2.5033710723826488, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.613004207611084, "learning_rate": 5e-05, "loss": 0.693, "mean_token_accuracy": 0.8720211386680603, "num_tokens": 747362733.0, "step": 19679 }, { "epoch": 2.503498282661239, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.823423385620117, "learning_rate": 5e-05, "loss": 0.8008, "mean_token_accuracy": 0.8889971375465393, "num_tokens": 747400735.0, "step": 19680 }, { "epoch": 2.5036254929398294, "ewc_loss": 0.3125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00023746490478515625, "grad_norm": 6.090879917144775, "learning_rate": 5e-05, "loss": 0.6585, "mean_token_accuracy": 0.8783597350120544, "num_tokens": 747440263.0, "step": 19681 }, { "epoch": 2.50375270321842, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.690165519714355, "learning_rate": 5e-05, "loss": 0.9323, "mean_token_accuracy": 0.8821732997894287, "num_tokens": 747477918.0, "step": 19682 }, { "epoch": 2.5038799134970104, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 5.859318256378174, "learning_rate": 5e-05, "loss": 0.6762, "mean_token_accuracy": 0.8787441253662109, "num_tokens": 747513235.0, "step": 19683 }, { "epoch": 2.504007123775601, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.418140411376953, "learning_rate": 5e-05, "loss": 1.0059, "mean_token_accuracy": 0.8975284695625305, "num_tokens": 747553280.0, "step": 19684 }, { "epoch": 2.5041343340541915, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.093845844268799, "learning_rate": 5e-05, "loss": 0.8891, "mean_token_accuracy": 0.8751528859138489, "num_tokens": 747594389.0, "step": 19685 }, { "epoch": 2.504261544332782, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.728013038635254, "learning_rate": 5e-05, "loss": 0.984, "mean_token_accuracy": 0.8826086521148682, "num_tokens": 747634805.0, "step": 19686 }, { "epoch": 2.5043887546113726, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 8.811522483825684, "learning_rate": 5e-05, "loss": 1.1103, "mean_token_accuracy": 0.8714919090270996, "num_tokens": 747673963.0, "step": 19687 }, { "epoch": 2.504515964889963, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.34760046005249, "learning_rate": 5e-05, "loss": 0.7829, "mean_token_accuracy": 0.8736308813095093, "num_tokens": 747715067.0, "step": 19688 }, { "epoch": 2.5046431751685536, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 10.744991302490234, "learning_rate": 5e-05, "loss": 1.1356, "mean_token_accuracy": 0.8864448666572571, "num_tokens": 747752799.0, "step": 19689 }, { "epoch": 2.504770385447144, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.558058261871338, "learning_rate": 5e-05, "loss": 0.8126, "mean_token_accuracy": 0.8716999292373657, "num_tokens": 747789325.0, "step": 19690 }, { "epoch": 2.5048975957257347, "ewc_loss": 1.078125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.001007080078125, "grad_norm": 12.918554306030273, "learning_rate": 5e-05, "loss": 1.4169, "mean_token_accuracy": 0.8786802887916565, "num_tokens": 747823438.0, "step": 19691 }, { "epoch": 2.505024806004325, "ewc_loss": 1.03125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000957489013671875, "grad_norm": 10.204301834106445, "learning_rate": 5e-05, "loss": 1.3431, "mean_token_accuracy": 0.8888123631477356, "num_tokens": 747862310.0, "step": 19692 }, { "epoch": 2.5051520162829157, "ewc_loss": 0.765625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 11.543058395385742, "learning_rate": 5e-05, "loss": 1.102, "mean_token_accuracy": 0.8825394511222839, "num_tokens": 747903154.0, "step": 19693 }, { "epoch": 2.5052792265615063, "ewc_loss": 0.8828125, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0008087158203125, "grad_norm": 9.801421165466309, "learning_rate": 5e-05, "loss": 1.2334, "mean_token_accuracy": 0.8764585852622986, "num_tokens": 747937849.0, "step": 19694 }, { "epoch": 2.505406436840097, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000743865966796875, "grad_norm": 11.407272338867188, "learning_rate": 5e-05, "loss": 1.1419, "mean_token_accuracy": 0.8819196224212646, "num_tokens": 747972387.0, "step": 19695 }, { "epoch": 2.5055336471186873, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.750544548034668, "learning_rate": 5e-05, "loss": 1.0295, "mean_token_accuracy": 0.8658673167228699, "num_tokens": 748007181.0, "step": 19696 }, { "epoch": 2.505660857397278, "ewc_loss": 0.86328125, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 11.460610389709473, "learning_rate": 5e-05, "loss": 1.1951, "mean_token_accuracy": 0.8837879300117493, "num_tokens": 748044342.0, "step": 19697 }, { "epoch": 2.5057880676758684, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.919750213623047, "learning_rate": 5e-05, "loss": 1.0045, "mean_token_accuracy": 0.874382734298706, "num_tokens": 748074012.0, "step": 19698 }, { "epoch": 2.5059152779544585, "ewc_loss": 0.83203125, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 11.504987716674805, "learning_rate": 5e-05, "loss": 1.1411, "mean_token_accuracy": 0.8897719383239746, "num_tokens": 748111055.0, "step": 19699 }, { "epoch": 2.5060424882330494, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 9.631196975708008, "learning_rate": 5e-05, "loss": 1.1487, "mean_token_accuracy": 0.8629851937294006, "num_tokens": 748151922.0, "step": 19700 }, { "epoch": 2.5061696985116395, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.856000900268555, "learning_rate": 5e-05, "loss": 1.047, "mean_token_accuracy": 0.8906101584434509, "num_tokens": 748186090.0, "step": 19701 }, { "epoch": 2.5062969087902305, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.462484359741211, "learning_rate": 5e-05, "loss": 1.0081, "mean_token_accuracy": 0.886560320854187, "num_tokens": 748223281.0, "step": 19702 }, { "epoch": 2.5064241190688206, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.3013916015625, "learning_rate": 5e-05, "loss": 1.016, "mean_token_accuracy": 0.8820363879203796, "num_tokens": 748255657.0, "step": 19703 }, { "epoch": 2.5065513293474115, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.952924728393555, "learning_rate": 5e-05, "loss": 0.9958, "mean_token_accuracy": 0.8701826333999634, "num_tokens": 748296954.0, "step": 19704 }, { "epoch": 2.5066785396260016, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.037861824035645, "learning_rate": 5e-05, "loss": 1.0179, "mean_token_accuracy": 0.8752985000610352, "num_tokens": 748334753.0, "step": 19705 }, { "epoch": 2.506805749904592, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.45374870300293, "learning_rate": 5e-05, "loss": 0.9033, "mean_token_accuracy": 0.8814986944198608, "num_tokens": 748377880.0, "step": 19706 }, { "epoch": 2.5069329601831827, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.137106895446777, "learning_rate": 5e-05, "loss": 1.0434, "mean_token_accuracy": 0.8747279047966003, "num_tokens": 748418230.0, "step": 19707 }, { "epoch": 2.507060170461773, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.292661666870117, "learning_rate": 5e-05, "loss": 0.8663, "mean_token_accuracy": 0.8873493075370789, "num_tokens": 748459320.0, "step": 19708 }, { "epoch": 2.5071873807403637, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.310776710510254, "learning_rate": 5e-05, "loss": 1.1066, "mean_token_accuracy": 0.8638666868209839, "num_tokens": 748503951.0, "step": 19709 }, { "epoch": 2.5073145910189543, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.49283218383789, "learning_rate": 5e-05, "loss": 0.907, "mean_token_accuracy": 0.8811212778091431, "num_tokens": 748544736.0, "step": 19710 }, { "epoch": 2.507441801297545, "ewc_loss": 0.71875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.259468078613281, "learning_rate": 5e-05, "loss": 1.0598, "mean_token_accuracy": 0.8811993598937988, "num_tokens": 748583174.0, "step": 19711 }, { "epoch": 2.5075690115761353, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.784534454345703, "learning_rate": 5e-05, "loss": 0.924, "mean_token_accuracy": 0.8824985027313232, "num_tokens": 748624580.0, "step": 19712 }, { "epoch": 2.507696221854726, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.776083946228027, "learning_rate": 5e-05, "loss": 1.0382, "mean_token_accuracy": 0.8675168752670288, "num_tokens": 748660343.0, "step": 19713 }, { "epoch": 2.5078234321333164, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.746109962463379, "learning_rate": 5e-05, "loss": 0.951, "mean_token_accuracy": 0.8692562580108643, "num_tokens": 748695006.0, "step": 19714 }, { "epoch": 2.507950642411907, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.211539268493652, "learning_rate": 5e-05, "loss": 0.9362, "mean_token_accuracy": 0.886985182762146, "num_tokens": 748732142.0, "step": 19715 }, { "epoch": 2.5080778526904974, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.32729721069336, "learning_rate": 5e-05, "loss": 0.8834, "mean_token_accuracy": 0.87156081199646, "num_tokens": 748771263.0, "step": 19716 }, { "epoch": 2.508205062969088, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.073917388916016, "learning_rate": 5e-05, "loss": 0.9277, "mean_token_accuracy": 0.8797600269317627, "num_tokens": 748812796.0, "step": 19717 }, { "epoch": 2.5083322732476785, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.8017449378967285, "learning_rate": 5e-05, "loss": 0.7881, "mean_token_accuracy": 0.880033016204834, "num_tokens": 748852701.0, "step": 19718 }, { "epoch": 2.508459483526269, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.332855224609375, "learning_rate": 5e-05, "loss": 0.9947, "mean_token_accuracy": 0.8707749843597412, "num_tokens": 748888799.0, "step": 19719 }, { "epoch": 2.5085866938048595, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.633040428161621, "learning_rate": 5e-05, "loss": 0.7812, "mean_token_accuracy": 0.8812944889068604, "num_tokens": 748927056.0, "step": 19720 }, { "epoch": 2.50871390408345, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.572272300720215, "learning_rate": 5e-05, "loss": 1.0236, "mean_token_accuracy": 0.8677993416786194, "num_tokens": 748961829.0, "step": 19721 }, { "epoch": 2.5088411143620406, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.927989959716797, "learning_rate": 5e-05, "loss": 0.7974, "mean_token_accuracy": 0.8909157514572144, "num_tokens": 748995281.0, "step": 19722 }, { "epoch": 2.508968324640631, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.356139183044434, "learning_rate": 5e-05, "loss": 0.9475, "mean_token_accuracy": 0.8877053260803223, "num_tokens": 749027574.0, "step": 19723 }, { "epoch": 2.509095534919221, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.260051727294922, "learning_rate": 5e-05, "loss": 0.8459, "mean_token_accuracy": 0.8841700553894043, "num_tokens": 749065092.0, "step": 19724 }, { "epoch": 2.509222745197812, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.71609115600586, "learning_rate": 5e-05, "loss": 0.9409, "mean_token_accuracy": 0.8725318312644958, "num_tokens": 749104389.0, "step": 19725 }, { "epoch": 2.5093499554764023, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.180416107177734, "learning_rate": 5e-05, "loss": 0.7866, "mean_token_accuracy": 0.8953258395195007, "num_tokens": 749138433.0, "step": 19726 }, { "epoch": 2.5094771657549932, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.24813461303711, "learning_rate": 5e-05, "loss": 0.8428, "mean_token_accuracy": 0.8868110179901123, "num_tokens": 749178095.0, "step": 19727 }, { "epoch": 2.5096043760335833, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.842278480529785, "learning_rate": 5e-05, "loss": 0.7354, "mean_token_accuracy": 0.8972386121749878, "num_tokens": 749215547.0, "step": 19728 }, { "epoch": 2.5097315863121743, "ewc_loss": 0.5, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.219886779785156, "learning_rate": 5e-05, "loss": 0.8706, "mean_token_accuracy": 0.8680906295776367, "num_tokens": 749254935.0, "step": 19729 }, { "epoch": 2.5098587965907644, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.470792770385742, "learning_rate": 5e-05, "loss": 0.7331, "mean_token_accuracy": 0.8913850784301758, "num_tokens": 749290425.0, "step": 19730 }, { "epoch": 2.509986006869355, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.341894149780273, "learning_rate": 5e-05, "loss": 0.8033, "mean_token_accuracy": 0.8848174810409546, "num_tokens": 749319390.0, "step": 19731 }, { "epoch": 2.5101132171479454, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.295565128326416, "learning_rate": 5e-05, "loss": 0.7609, "mean_token_accuracy": 0.8691176176071167, "num_tokens": 749358517.0, "step": 19732 }, { "epoch": 2.510240427426536, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.371220588684082, "learning_rate": 5e-05, "loss": 0.8261, "mean_token_accuracy": 0.8795911073684692, "num_tokens": 749402219.0, "step": 19733 }, { "epoch": 2.5103676377051265, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.238896369934082, "learning_rate": 5e-05, "loss": 0.7948, "mean_token_accuracy": 0.8543120622634888, "num_tokens": 749437514.0, "step": 19734 }, { "epoch": 2.510494847983717, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.391830444335938, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.8832955360412598, "num_tokens": 749477148.0, "step": 19735 }, { "epoch": 2.5106220582623076, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.195586204528809, "learning_rate": 5e-05, "loss": 0.7114, "mean_token_accuracy": 0.8854478001594543, "num_tokens": 749511174.0, "step": 19736 }, { "epoch": 2.510749268540898, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.438824653625488, "learning_rate": 5e-05, "loss": 0.8179, "mean_token_accuracy": 0.8879585862159729, "num_tokens": 749547117.0, "step": 19737 }, { "epoch": 2.5108764788194886, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.273831367492676, "learning_rate": 5e-05, "loss": 0.6917, "mean_token_accuracy": 0.8899378776550293, "num_tokens": 749583755.0, "step": 19738 }, { "epoch": 2.511003689098079, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.224220275878906, "learning_rate": 5e-05, "loss": 0.7949, "mean_token_accuracy": 0.8843650817871094, "num_tokens": 749621742.0, "step": 19739 }, { "epoch": 2.5111308993766697, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.540737628936768, "learning_rate": 5e-05, "loss": 0.7018, "mean_token_accuracy": 0.8922019600868225, "num_tokens": 749659178.0, "step": 19740 }, { "epoch": 2.51125810965526, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.792174816131592, "learning_rate": 5e-05, "loss": 0.7798, "mean_token_accuracy": 0.8736996650695801, "num_tokens": 749698828.0, "step": 19741 }, { "epoch": 2.5113853199338507, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.80095100402832, "learning_rate": 5e-05, "loss": 0.7896, "mean_token_accuracy": 0.8678072690963745, "num_tokens": 749732609.0, "step": 19742 }, { "epoch": 2.5115125302124413, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.419951915740967, "learning_rate": 5e-05, "loss": 0.6732, "mean_token_accuracy": 0.8993113040924072, "num_tokens": 749768240.0, "step": 19743 }, { "epoch": 2.511639740491032, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.028024673461914, "learning_rate": 5e-05, "loss": 0.7284, "mean_token_accuracy": 0.8866823315620422, "num_tokens": 749809435.0, "step": 19744 }, { "epoch": 2.5117669507696223, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.111400127410889, "learning_rate": 5e-05, "loss": 0.7346, "mean_token_accuracy": 0.870104193687439, "num_tokens": 749852790.0, "step": 19745 }, { "epoch": 2.511894161048213, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.321002006530762, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.8763765096664429, "num_tokens": 749895111.0, "step": 19746 }, { "epoch": 2.5120213713268034, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.828527927398682, "learning_rate": 5e-05, "loss": 0.6605, "mean_token_accuracy": 0.8868882656097412, "num_tokens": 749936142.0, "step": 19747 }, { "epoch": 2.512148581605394, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.606063842773438, "learning_rate": 5e-05, "loss": 0.7965, "mean_token_accuracy": 0.8883233070373535, "num_tokens": 749972106.0, "step": 19748 }, { "epoch": 2.512275791883984, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.729016304016113, "learning_rate": 5e-05, "loss": 0.7216, "mean_token_accuracy": 0.8691655397415161, "num_tokens": 750014090.0, "step": 19749 }, { "epoch": 2.512403002162575, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.669668197631836, "learning_rate": 5e-05, "loss": 0.8181, "mean_token_accuracy": 0.8862234354019165, "num_tokens": 750052542.0, "step": 19750 }, { "epoch": 2.512530212441165, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.983081340789795, "learning_rate": 5e-05, "loss": 0.7479, "mean_token_accuracy": 0.866940438747406, "num_tokens": 750091846.0, "step": 19751 }, { "epoch": 2.512657422719756, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.231781959533691, "learning_rate": 5e-05, "loss": 0.8459, "mean_token_accuracy": 0.8654762506484985, "num_tokens": 750135721.0, "step": 19752 }, { "epoch": 2.512784632998346, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.620968341827393, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8679820895195007, "num_tokens": 750177786.0, "step": 19753 }, { "epoch": 2.5129118432769366, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.361168384552002, "learning_rate": 5e-05, "loss": 0.738, "mean_token_accuracy": 0.8767085075378418, "num_tokens": 750209577.0, "step": 19754 }, { "epoch": 2.513039053555527, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.293481826782227, "learning_rate": 5e-05, "loss": 0.7477, "mean_token_accuracy": 0.8956702351570129, "num_tokens": 750243168.0, "step": 19755 }, { "epoch": 2.5131662638341177, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00074005126953125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.801098346710205, "learning_rate": 5e-05, "loss": 0.699, "mean_token_accuracy": 0.8779025673866272, "num_tokens": 750288304.0, "step": 19756 }, { "epoch": 2.513293474112708, "ewc_loss": 0.484375, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.706223487854004, "learning_rate": 5e-05, "loss": 0.8387, "mean_token_accuracy": 0.8760803937911987, "num_tokens": 750329662.0, "step": 19757 }, { "epoch": 2.5134206843912987, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.602171897888184, "learning_rate": 5e-05, "loss": 0.7037, "mean_token_accuracy": 0.8729614019393921, "num_tokens": 750364581.0, "step": 19758 }, { "epoch": 2.5135478946698893, "ewc_loss": 0.515625, "ewc_loss_diag": 0.0007476806640625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.853994369506836, "learning_rate": 5e-05, "loss": 0.8552, "mean_token_accuracy": 0.8845201730728149, "num_tokens": 750403129.0, "step": 19759 }, { "epoch": 2.51367510494848, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.775299549102783, "learning_rate": 5e-05, "loss": 0.6876, "mean_token_accuracy": 0.8808659911155701, "num_tokens": 750439738.0, "step": 19760 }, { "epoch": 2.5138023152270703, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.597267150878906, "learning_rate": 5e-05, "loss": 0.8849, "mean_token_accuracy": 0.869972825050354, "num_tokens": 750479302.0, "step": 19761 }, { "epoch": 2.513929525505661, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.232158184051514, "learning_rate": 5e-05, "loss": 0.7333, "mean_token_accuracy": 0.8813849687576294, "num_tokens": 750517658.0, "step": 19762 }, { "epoch": 2.5140567357842514, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.01548957824707, "learning_rate": 5e-05, "loss": 0.7602, "mean_token_accuracy": 0.8919124007225037, "num_tokens": 750558192.0, "step": 19763 }, { "epoch": 2.514183946062842, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.701727390289307, "learning_rate": 5e-05, "loss": 0.755, "mean_token_accuracy": 0.8799048662185669, "num_tokens": 750595209.0, "step": 19764 }, { "epoch": 2.5143111563414324, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.563337326049805, "learning_rate": 5e-05, "loss": 0.7667, "mean_token_accuracy": 0.8745153546333313, "num_tokens": 750634560.0, "step": 19765 }, { "epoch": 2.514438366620023, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.8133931159973145, "learning_rate": 5e-05, "loss": 0.7448, "mean_token_accuracy": 0.8825763463973999, "num_tokens": 750675223.0, "step": 19766 }, { "epoch": 2.5145655768986135, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.579792499542236, "learning_rate": 5e-05, "loss": 0.7454, "mean_token_accuracy": 0.8793520331382751, "num_tokens": 750717302.0, "step": 19767 }, { "epoch": 2.514692787177204, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.481062889099121, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8783193826675415, "num_tokens": 750754615.0, "step": 19768 }, { "epoch": 2.5148199974557945, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.030177116394043, "learning_rate": 5e-05, "loss": 0.7291, "mean_token_accuracy": 0.8950304388999939, "num_tokens": 750788745.0, "step": 19769 }, { "epoch": 2.514947207734385, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.855264663696289, "learning_rate": 5e-05, "loss": 0.7338, "mean_token_accuracy": 0.8657923340797424, "num_tokens": 750829727.0, "step": 19770 }, { "epoch": 2.5150744180129756, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.766712188720703, "learning_rate": 5e-05, "loss": 0.8736, "mean_token_accuracy": 0.8681676387786865, "num_tokens": 750873095.0, "step": 19771 }, { "epoch": 2.5152016282915657, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.00024318695068359375, "grad_norm": 6.20700740814209, "learning_rate": 5e-05, "loss": 0.687, "mean_token_accuracy": 0.868229329586029, "num_tokens": 750909513.0, "step": 19772 }, { "epoch": 2.5153288385701567, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.656696319580078, "learning_rate": 5e-05, "loss": 0.9564, "mean_token_accuracy": 0.8778765201568604, "num_tokens": 750947234.0, "step": 19773 }, { "epoch": 2.5154560488487467, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 5.988117218017578, "learning_rate": 5e-05, "loss": 0.6442, "mean_token_accuracy": 0.8890339136123657, "num_tokens": 750984926.0, "step": 19774 }, { "epoch": 2.5155832591273377, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.271001815795898, "learning_rate": 5e-05, "loss": 1.0265, "mean_token_accuracy": 0.8809729814529419, "num_tokens": 751023700.0, "step": 19775 }, { "epoch": 2.515710469405928, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.1068572998046875, "learning_rate": 5e-05, "loss": 0.833, "mean_token_accuracy": 0.8815526962280273, "num_tokens": 751065444.0, "step": 19776 }, { "epoch": 2.5158376796845188, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.56345272064209, "learning_rate": 5e-05, "loss": 1.0435, "mean_token_accuracy": 0.8654383420944214, "num_tokens": 751101627.0, "step": 19777 }, { "epoch": 2.515964889963109, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.617589950561523, "learning_rate": 5e-05, "loss": 0.9701, "mean_token_accuracy": 0.8862624168395996, "num_tokens": 751138006.0, "step": 19778 }, { "epoch": 2.5160921002416994, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.524919033050537, "learning_rate": 5e-05, "loss": 0.7604, "mean_token_accuracy": 0.8900924921035767, "num_tokens": 751179187.0, "step": 19779 }, { "epoch": 2.51621931052029, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.837843894958496, "learning_rate": 5e-05, "loss": 0.9974, "mean_token_accuracy": 0.8843592405319214, "num_tokens": 751217086.0, "step": 19780 }, { "epoch": 2.5163465207988804, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.515087604522705, "learning_rate": 5e-05, "loss": 0.7521, "mean_token_accuracy": 0.8775806427001953, "num_tokens": 751259330.0, "step": 19781 }, { "epoch": 2.516473731077471, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 10.73004150390625, "learning_rate": 5e-05, "loss": 1.108, "mean_token_accuracy": 0.879664957523346, "num_tokens": 751300873.0, "step": 19782 }, { "epoch": 2.5166009413560615, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.705924034118652, "learning_rate": 5e-05, "loss": 0.8985, "mean_token_accuracy": 0.88698410987854, "num_tokens": 751330783.0, "step": 19783 }, { "epoch": 2.516728151634652, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.119620323181152, "learning_rate": 5e-05, "loss": 1.0279, "mean_token_accuracy": 0.8833763599395752, "num_tokens": 751370032.0, "step": 19784 }, { "epoch": 2.5168553619132426, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 8.81864070892334, "learning_rate": 5e-05, "loss": 1.0334, "mean_token_accuracy": 0.8847528696060181, "num_tokens": 751412561.0, "step": 19785 }, { "epoch": 2.516982572191833, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.65745735168457, "learning_rate": 5e-05, "loss": 0.9064, "mean_token_accuracy": 0.8781626224517822, "num_tokens": 751442168.0, "step": 19786 }, { "epoch": 2.5171097824704236, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.02878189086914, "learning_rate": 5e-05, "loss": 1.0507, "mean_token_accuracy": 0.8581621646881104, "num_tokens": 751488656.0, "step": 19787 }, { "epoch": 2.517236992749014, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.865545272827148, "learning_rate": 5e-05, "loss": 0.854, "mean_token_accuracy": 0.8782392740249634, "num_tokens": 751533305.0, "step": 19788 }, { "epoch": 2.5173642030276047, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.759650230407715, "learning_rate": 5e-05, "loss": 0.9295, "mean_token_accuracy": 0.8799514174461365, "num_tokens": 751570693.0, "step": 19789 }, { "epoch": 2.517491413306195, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.489701271057129, "learning_rate": 5e-05, "loss": 0.843, "mean_token_accuracy": 0.8938000202178955, "num_tokens": 751604994.0, "step": 19790 }, { "epoch": 2.5176186235847857, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.482546806335449, "learning_rate": 5e-05, "loss": 0.7687, "mean_token_accuracy": 0.8804789185523987, "num_tokens": 751642552.0, "step": 19791 }, { "epoch": 2.5177458338633762, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.962569236755371, "learning_rate": 5e-05, "loss": 0.9054, "mean_token_accuracy": 0.8888617753982544, "num_tokens": 751677253.0, "step": 19792 }, { "epoch": 2.5178730441419668, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.835801124572754, "learning_rate": 5e-05, "loss": 0.7135, "mean_token_accuracy": 0.884823739528656, "num_tokens": 751717737.0, "step": 19793 }, { "epoch": 2.5180002544205573, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.573144912719727, "learning_rate": 5e-05, "loss": 0.921, "mean_token_accuracy": 0.9009016752243042, "num_tokens": 751750453.0, "step": 19794 }, { "epoch": 2.518127464699148, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 6.778690338134766, "learning_rate": 5e-05, "loss": 0.7663, "mean_token_accuracy": 0.8780168294906616, "num_tokens": 751792179.0, "step": 19795 }, { "epoch": 2.5182546749777384, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.064085960388184, "learning_rate": 5e-05, "loss": 1.0304, "mean_token_accuracy": 0.8818314075469971, "num_tokens": 751833367.0, "step": 19796 }, { "epoch": 2.5183818852563284, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 7.392643451690674, "learning_rate": 5e-05, "loss": 0.8891, "mean_token_accuracy": 0.8835130929946899, "num_tokens": 751874042.0, "step": 19797 }, { "epoch": 2.5185090955349194, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.049777030944824, "learning_rate": 5e-05, "loss": 1.105, "mean_token_accuracy": 0.8582624197006226, "num_tokens": 751914990.0, "step": 19798 }, { "epoch": 2.5186363058135095, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.077191352844238, "learning_rate": 5e-05, "loss": 1.0245, "mean_token_accuracy": 0.8665835857391357, "num_tokens": 751946266.0, "step": 19799 }, { "epoch": 2.5187635160921005, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.465058326721191, "learning_rate": 5e-05, "loss": 0.9671, "mean_token_accuracy": 0.8907021880149841, "num_tokens": 751990359.0, "step": 19800 }, { "epoch": 2.5188907263706906, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.235335350036621, "learning_rate": 5e-05, "loss": 0.9725, "mean_token_accuracy": 0.8776414394378662, "num_tokens": 752026955.0, "step": 19801 }, { "epoch": 2.5190179366492815, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.880450248718262, "learning_rate": 5e-05, "loss": 0.9825, "mean_token_accuracy": 0.8704003095626831, "num_tokens": 752065187.0, "step": 19802 }, { "epoch": 2.5191451469278716, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 7.962052345275879, "learning_rate": 5e-05, "loss": 0.8565, "mean_token_accuracy": 0.8934369683265686, "num_tokens": 752100855.0, "step": 19803 }, { "epoch": 2.519272357206462, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.743348121643066, "learning_rate": 5e-05, "loss": 0.932, "mean_token_accuracy": 0.8807807564735413, "num_tokens": 752136672.0, "step": 19804 }, { "epoch": 2.5193995674850527, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.611862659454346, "learning_rate": 5e-05, "loss": 0.8205, "mean_token_accuracy": 0.8882480263710022, "num_tokens": 752167529.0, "step": 19805 }, { "epoch": 2.519526777763643, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.743675231933594, "learning_rate": 5e-05, "loss": 0.9701, "mean_token_accuracy": 0.8676490783691406, "num_tokens": 752201865.0, "step": 19806 }, { "epoch": 2.5196539880422337, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.48765230178833, "learning_rate": 5e-05, "loss": 0.8178, "mean_token_accuracy": 0.8842601776123047, "num_tokens": 752239315.0, "step": 19807 }, { "epoch": 2.5197811983208243, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.631125450134277, "learning_rate": 5e-05, "loss": 0.8971, "mean_token_accuracy": 0.8872554302215576, "num_tokens": 752278505.0, "step": 19808 }, { "epoch": 2.519908408599415, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.5028862953186035, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.8886823058128357, "num_tokens": 752319346.0, "step": 19809 }, { "epoch": 2.5200356188780053, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.509851455688477, "learning_rate": 5e-05, "loss": 0.9401, "mean_token_accuracy": 0.870236873626709, "num_tokens": 752364210.0, "step": 19810 }, { "epoch": 2.520162829156596, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.4668378829956055, "learning_rate": 5e-05, "loss": 0.766, "mean_token_accuracy": 0.8916016221046448, "num_tokens": 752400679.0, "step": 19811 }, { "epoch": 2.5202900394351864, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.467775344848633, "learning_rate": 5e-05, "loss": 0.9055, "mean_token_accuracy": 0.8814187049865723, "num_tokens": 752436003.0, "step": 19812 }, { "epoch": 2.520417249713777, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.427114009857178, "learning_rate": 5e-05, "loss": 0.821, "mean_token_accuracy": 0.8714855313301086, "num_tokens": 752478690.0, "step": 19813 }, { "epoch": 2.5205444599923674, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.426387786865234, "learning_rate": 5e-05, "loss": 0.9477, "mean_token_accuracy": 0.8646692037582397, "num_tokens": 752513326.0, "step": 19814 }, { "epoch": 2.520671670270958, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.418828964233398, "learning_rate": 5e-05, "loss": 0.8169, "mean_token_accuracy": 0.8717677593231201, "num_tokens": 752556553.0, "step": 19815 }, { "epoch": 2.5207988805495485, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.32834529876709, "learning_rate": 5e-05, "loss": 0.841, "mean_token_accuracy": 0.8898416757583618, "num_tokens": 752599219.0, "step": 19816 }, { "epoch": 2.520926090828139, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.452089309692383, "learning_rate": 5e-05, "loss": 0.7723, "mean_token_accuracy": 0.8832481503486633, "num_tokens": 752639308.0, "step": 19817 }, { "epoch": 2.5210533011067295, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.198456764221191, "learning_rate": 5e-05, "loss": 0.904, "mean_token_accuracy": 0.8625779747962952, "num_tokens": 752679384.0, "step": 19818 }, { "epoch": 2.52118051138532, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.4833292961120605, "learning_rate": 5e-05, "loss": 0.7835, "mean_token_accuracy": 0.8745760917663574, "num_tokens": 752722395.0, "step": 19819 }, { "epoch": 2.5213077216639106, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.07298469543457, "learning_rate": 5e-05, "loss": 0.8619, "mean_token_accuracy": 0.8777315020561218, "num_tokens": 752754491.0, "step": 19820 }, { "epoch": 2.521434931942501, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.515833377838135, "learning_rate": 5e-05, "loss": 0.7317, "mean_token_accuracy": 0.8894541263580322, "num_tokens": 752792595.0, "step": 19821 }, { "epoch": 2.521562142221091, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.952669620513916, "learning_rate": 5e-05, "loss": 0.8112, "mean_token_accuracy": 0.8788182735443115, "num_tokens": 752830689.0, "step": 19822 }, { "epoch": 2.521689352499682, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.537757873535156, "learning_rate": 5e-05, "loss": 0.7103, "mean_token_accuracy": 0.8937989473342896, "num_tokens": 752862848.0, "step": 19823 }, { "epoch": 2.5218165627782723, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.873504161834717, "learning_rate": 5e-05, "loss": 0.7761, "mean_token_accuracy": 0.8816693425178528, "num_tokens": 752900566.0, "step": 19824 }, { "epoch": 2.5219437730568632, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.509570121765137, "learning_rate": 5e-05, "loss": 0.7206, "mean_token_accuracy": 0.8888555765151978, "num_tokens": 752939201.0, "step": 19825 }, { "epoch": 2.5220709833354533, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.853853225708008, "learning_rate": 5e-05, "loss": 0.749, "mean_token_accuracy": 0.8872511386871338, "num_tokens": 752973820.0, "step": 19826 }, { "epoch": 2.522198193614044, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.464103698730469, "learning_rate": 5e-05, "loss": 0.7943, "mean_token_accuracy": 0.8620928525924683, "num_tokens": 753018563.0, "step": 19827 }, { "epoch": 2.5223254038926344, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.833348751068115, "learning_rate": 5e-05, "loss": 0.7828, "mean_token_accuracy": 0.8705494403839111, "num_tokens": 753060097.0, "step": 19828 }, { "epoch": 2.522452614171225, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.452097415924072, "learning_rate": 5e-05, "loss": 0.7355, "mean_token_accuracy": 0.8770800232887268, "num_tokens": 753099200.0, "step": 19829 }, { "epoch": 2.5225798244498154, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.82585334777832, "learning_rate": 5e-05, "loss": 0.7305, "mean_token_accuracy": 0.884541392326355, "num_tokens": 753132421.0, "step": 19830 }, { "epoch": 2.522707034728406, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.414912223815918, "learning_rate": 5e-05, "loss": 0.7399, "mean_token_accuracy": 0.868611216545105, "num_tokens": 753165380.0, "step": 19831 }, { "epoch": 2.5228342450069965, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.83871603012085, "learning_rate": 5e-05, "loss": 0.7287, "mean_token_accuracy": 0.8798657059669495, "num_tokens": 753203501.0, "step": 19832 }, { "epoch": 2.522961455285587, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.357824325561523, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.8696520924568176, "num_tokens": 753247674.0, "step": 19833 }, { "epoch": 2.5230886655641775, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.871652126312256, "learning_rate": 5e-05, "loss": 0.7317, "mean_token_accuracy": 0.8821451663970947, "num_tokens": 753285967.0, "step": 19834 }, { "epoch": 2.523215875842768, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.303089141845703, "learning_rate": 5e-05, "loss": 0.7208, "mean_token_accuracy": 0.8754241466522217, "num_tokens": 753319080.0, "step": 19835 }, { "epoch": 2.5233430861213586, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.945868015289307, "learning_rate": 5e-05, "loss": 0.7742, "mean_token_accuracy": 0.8698055744171143, "num_tokens": 753357924.0, "step": 19836 }, { "epoch": 2.523470296399949, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.201024532318115, "learning_rate": 5e-05, "loss": 0.6916, "mean_token_accuracy": 0.8849472403526306, "num_tokens": 753397922.0, "step": 19837 }, { "epoch": 2.5235975066785397, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.05613899230957, "learning_rate": 5e-05, "loss": 0.8253, "mean_token_accuracy": 0.858850359916687, "num_tokens": 753431268.0, "step": 19838 }, { "epoch": 2.52372471695713, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.060784339904785, "learning_rate": 5e-05, "loss": 0.7174, "mean_token_accuracy": 0.873808741569519, "num_tokens": 753465169.0, "step": 19839 }, { "epoch": 2.5238519272357207, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.24832820892334, "learning_rate": 5e-05, "loss": 0.7749, "mean_token_accuracy": 0.8813308477401733, "num_tokens": 753501017.0, "step": 19840 }, { "epoch": 2.5239791375143112, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.874829292297363, "learning_rate": 5e-05, "loss": 0.6677, "mean_token_accuracy": 0.883643627166748, "num_tokens": 753541039.0, "step": 19841 }, { "epoch": 2.5241063477929018, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.487468719482422, "learning_rate": 5e-05, "loss": 0.8786, "mean_token_accuracy": 0.8513083457946777, "num_tokens": 753585793.0, "step": 19842 }, { "epoch": 2.5242335580714923, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.6761250495910645, "learning_rate": 5e-05, "loss": 0.6408, "mean_token_accuracy": 0.8924847841262817, "num_tokens": 753623752.0, "step": 19843 }, { "epoch": 2.524360768350083, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.757322311401367, "learning_rate": 5e-05, "loss": 0.8436, "mean_token_accuracy": 0.8733387589454651, "num_tokens": 753662939.0, "step": 19844 }, { "epoch": 2.5244879786286734, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.5860748291015625, "learning_rate": 5e-05, "loss": 0.6836, "mean_token_accuracy": 0.8833016753196716, "num_tokens": 753699200.0, "step": 19845 }, { "epoch": 2.524615188907264, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.841146469116211, "learning_rate": 5e-05, "loss": 0.8837, "mean_token_accuracy": 0.869192361831665, "num_tokens": 753743867.0, "step": 19846 }, { "epoch": 2.524742399185854, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.833884239196777, "learning_rate": 5e-05, "loss": 0.7449, "mean_token_accuracy": 0.8681976795196533, "num_tokens": 753781494.0, "step": 19847 }, { "epoch": 2.524869609464445, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.43394947052002, "learning_rate": 5e-05, "loss": 0.8356, "mean_token_accuracy": 0.8805959224700928, "num_tokens": 753823518.0, "step": 19848 }, { "epoch": 2.524996819743035, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.485864639282227, "learning_rate": 5e-05, "loss": 0.7953, "mean_token_accuracy": 0.8754122853279114, "num_tokens": 753855058.0, "step": 19849 }, { "epoch": 2.525124030021626, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.459219455718994, "learning_rate": 5e-05, "loss": 0.73, "mean_token_accuracy": 0.8838268518447876, "num_tokens": 753896721.0, "step": 19850 }, { "epoch": 2.525251240300216, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.42924690246582, "learning_rate": 5e-05, "loss": 0.7893, "mean_token_accuracy": 0.8917206525802612, "num_tokens": 753933225.0, "step": 19851 }, { "epoch": 2.5253784505788066, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.384623050689697, "learning_rate": 5e-05, "loss": 0.6862, "mean_token_accuracy": 0.8726171851158142, "num_tokens": 753970373.0, "step": 19852 }, { "epoch": 2.525505660857397, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.594587326049805, "learning_rate": 5e-05, "loss": 0.8911, "mean_token_accuracy": 0.8885369300842285, "num_tokens": 754007544.0, "step": 19853 }, { "epoch": 2.5256328711359877, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00023555755615234375, "grad_norm": 5.645582675933838, "learning_rate": 5e-05, "loss": 0.6465, "mean_token_accuracy": 0.8837324976921082, "num_tokens": 754040858.0, "step": 19854 }, { "epoch": 2.525760081414578, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 10.93213939666748, "learning_rate": 5e-05, "loss": 1.1618, "mean_token_accuracy": 0.8617172241210938, "num_tokens": 754083877.0, "step": 19855 }, { "epoch": 2.5258872916931687, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 6.62929105758667, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8906828165054321, "num_tokens": 754122792.0, "step": 19856 }, { "epoch": 2.5260145019717593, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 11.305442810058594, "learning_rate": 5e-05, "loss": 1.2352, "mean_token_accuracy": 0.8810968399047852, "num_tokens": 754157986.0, "step": 19857 }, { "epoch": 2.52614171225035, "ewc_loss": 0.890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 9.14056396484375, "learning_rate": 5e-05, "loss": 1.2375, "mean_token_accuracy": 0.8782821297645569, "num_tokens": 754198321.0, "step": 19858 }, { "epoch": 2.5262689225289403, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.34819507598877, "learning_rate": 5e-05, "loss": 1.0145, "mean_token_accuracy": 0.864982008934021, "num_tokens": 754235463.0, "step": 19859 }, { "epoch": 2.526396132807531, "ewc_loss": 0.859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 10.004011154174805, "learning_rate": 5e-05, "loss": 1.2021, "mean_token_accuracy": 0.8799388408660889, "num_tokens": 754274319.0, "step": 19860 }, { "epoch": 2.5265233430861214, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.597761154174805, "learning_rate": 5e-05, "loss": 0.8479, "mean_token_accuracy": 0.882804274559021, "num_tokens": 754315206.0, "step": 19861 }, { "epoch": 2.526650553364712, "ewc_loss": 0.8828125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 11.368685722351074, "learning_rate": 5e-05, "loss": 1.2497, "mean_token_accuracy": 0.8751967549324036, "num_tokens": 754353300.0, "step": 19862 }, { "epoch": 2.5267777636433024, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 8.413531303405762, "learning_rate": 5e-05, "loss": 1.0757, "mean_token_accuracy": 0.8694396018981934, "num_tokens": 754392692.0, "step": 19863 }, { "epoch": 2.526904973921893, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 11.348746299743652, "learning_rate": 5e-05, "loss": 1.1402, "mean_token_accuracy": 0.8786576986312866, "num_tokens": 754428688.0, "step": 19864 }, { "epoch": 2.5270321842004835, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 9.082777976989746, "learning_rate": 5e-05, "loss": 1.1315, "mean_token_accuracy": 0.8811172246932983, "num_tokens": 754467647.0, "step": 19865 }, { "epoch": 2.527159394479074, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.70702075958252, "learning_rate": 5e-05, "loss": 1.0804, "mean_token_accuracy": 0.8679462671279907, "num_tokens": 754501511.0, "step": 19866 }, { "epoch": 2.5272866047576645, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 8.639939308166504, "learning_rate": 5e-05, "loss": 0.9863, "mean_token_accuracy": 0.8922346830368042, "num_tokens": 754543000.0, "step": 19867 }, { "epoch": 2.527413815036255, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.214812278747559, "learning_rate": 5e-05, "loss": 1.0413, "mean_token_accuracy": 0.8819311261177063, "num_tokens": 754583485.0, "step": 19868 }, { "epoch": 2.5275410253148456, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.163191795349121, "learning_rate": 5e-05, "loss": 0.9448, "mean_token_accuracy": 0.8831413984298706, "num_tokens": 754620472.0, "step": 19869 }, { "epoch": 2.5276682355934357, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.942157745361328, "learning_rate": 5e-05, "loss": 1.0217, "mean_token_accuracy": 0.8837552070617676, "num_tokens": 754659906.0, "step": 19870 }, { "epoch": 2.5277954458720266, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.92800235748291, "learning_rate": 5e-05, "loss": 0.9089, "mean_token_accuracy": 0.8824026584625244, "num_tokens": 754696128.0, "step": 19871 }, { "epoch": 2.5279226561506167, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.79639720916748, "learning_rate": 5e-05, "loss": 1.0006, "mean_token_accuracy": 0.8796378970146179, "num_tokens": 754729553.0, "step": 19872 }, { "epoch": 2.5280498664292077, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.867792129516602, "learning_rate": 5e-05, "loss": 0.9543, "mean_token_accuracy": 0.8660733103752136, "num_tokens": 754773689.0, "step": 19873 }, { "epoch": 2.528177076707798, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.637255668640137, "learning_rate": 5e-05, "loss": 1.0089, "mean_token_accuracy": 0.8753852248191833, "num_tokens": 754809326.0, "step": 19874 }, { "epoch": 2.5283042869863888, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.828272819519043, "learning_rate": 5e-05, "loss": 0.8904, "mean_token_accuracy": 0.8826119899749756, "num_tokens": 754850627.0, "step": 19875 }, { "epoch": 2.528431497264979, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.517179489135742, "learning_rate": 5e-05, "loss": 0.9651, "mean_token_accuracy": 0.8855522871017456, "num_tokens": 754887810.0, "step": 19876 }, { "epoch": 2.5285587075435694, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.791302680969238, "learning_rate": 5e-05, "loss": 0.8969, "mean_token_accuracy": 0.877567708492279, "num_tokens": 754923778.0, "step": 19877 }, { "epoch": 2.52868591782216, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.37705135345459, "learning_rate": 5e-05, "loss": 0.9801, "mean_token_accuracy": 0.882266640663147, "num_tokens": 754963486.0, "step": 19878 }, { "epoch": 2.5288131281007504, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.829516410827637, "learning_rate": 5e-05, "loss": 0.8419, "mean_token_accuracy": 0.8920999765396118, "num_tokens": 755000302.0, "step": 19879 }, { "epoch": 2.528940338379341, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.135052680969238, "learning_rate": 5e-05, "loss": 0.9248, "mean_token_accuracy": 0.8914914727210999, "num_tokens": 755040319.0, "step": 19880 }, { "epoch": 2.5290675486579315, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.880107879638672, "learning_rate": 5e-05, "loss": 0.8653, "mean_token_accuracy": 0.8844475746154785, "num_tokens": 755073410.0, "step": 19881 }, { "epoch": 2.529194758936522, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.815597534179688, "learning_rate": 5e-05, "loss": 0.9234, "mean_token_accuracy": 0.884574830532074, "num_tokens": 755112279.0, "step": 19882 }, { "epoch": 2.5293219692151125, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.859043121337891, "learning_rate": 5e-05, "loss": 0.9112, "mean_token_accuracy": 0.8627654314041138, "num_tokens": 755146904.0, "step": 19883 }, { "epoch": 2.529449179493703, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.524675369262695, "learning_rate": 5e-05, "loss": 0.9235, "mean_token_accuracy": 0.8702126741409302, "num_tokens": 755179881.0, "step": 19884 }, { "epoch": 2.5295763897722936, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.703187942504883, "learning_rate": 5e-05, "loss": 0.7936, "mean_token_accuracy": 0.8910186290740967, "num_tokens": 755214504.0, "step": 19885 }, { "epoch": 2.529703600050884, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.407758712768555, "learning_rate": 5e-05, "loss": 0.8901, "mean_token_accuracy": 0.8794666528701782, "num_tokens": 755254229.0, "step": 19886 }, { "epoch": 2.5298308103294747, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.50365686416626, "learning_rate": 5e-05, "loss": 0.8022, "mean_token_accuracy": 0.8749234676361084, "num_tokens": 755296150.0, "step": 19887 }, { "epoch": 2.529958020608065, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.459757804870605, "learning_rate": 5e-05, "loss": 0.8663, "mean_token_accuracy": 0.888144850730896, "num_tokens": 755336508.0, "step": 19888 }, { "epoch": 2.5300852308866557, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.3386125564575195, "learning_rate": 5e-05, "loss": 0.7924, "mean_token_accuracy": 0.8700110912322998, "num_tokens": 755376863.0, "step": 19889 }, { "epoch": 2.5302124411652462, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.554939270019531, "learning_rate": 5e-05, "loss": 0.8731, "mean_token_accuracy": 0.8863509893417358, "num_tokens": 755414209.0, "step": 19890 }, { "epoch": 2.5303396514438368, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.284029006958008, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8657261729240417, "num_tokens": 755454281.0, "step": 19891 }, { "epoch": 2.5304668617224273, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.524717330932617, "learning_rate": 5e-05, "loss": 0.8411, "mean_token_accuracy": 0.8956294655799866, "num_tokens": 755492441.0, "step": 19892 }, { "epoch": 2.530594072001018, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.340795040130615, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8787046074867249, "num_tokens": 755526264.0, "step": 19893 }, { "epoch": 2.5307212822796084, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.442066192626953, "learning_rate": 5e-05, "loss": 0.8809, "mean_token_accuracy": 0.8876221179962158, "num_tokens": 755560819.0, "step": 19894 }, { "epoch": 2.5308484925581984, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.40852689743042, "learning_rate": 5e-05, "loss": 0.776, "mean_token_accuracy": 0.8782014846801758, "num_tokens": 755592103.0, "step": 19895 }, { "epoch": 2.5309757028367894, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.290562629699707, "learning_rate": 5e-05, "loss": 0.8633, "mean_token_accuracy": 0.8891913294792175, "num_tokens": 755631864.0, "step": 19896 }, { "epoch": 2.5311029131153795, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.524413108825684, "learning_rate": 5e-05, "loss": 0.7915, "mean_token_accuracy": 0.8736141324043274, "num_tokens": 755672555.0, "step": 19897 }, { "epoch": 2.5312301233939705, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.050359725952148, "learning_rate": 5e-05, "loss": 0.8062, "mean_token_accuracy": 0.889912486076355, "num_tokens": 755705606.0, "step": 19898 }, { "epoch": 2.5313573336725606, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.600840091705322, "learning_rate": 5e-05, "loss": 0.7982, "mean_token_accuracy": 0.8765575885772705, "num_tokens": 755738844.0, "step": 19899 }, { "epoch": 2.5314845439511515, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.883935928344727, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8899132609367371, "num_tokens": 755774164.0, "step": 19900 }, { "epoch": 2.5316117542297416, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.594634056091309, "learning_rate": 5e-05, "loss": 0.7931, "mean_token_accuracy": 0.8706643581390381, "num_tokens": 755816848.0, "step": 19901 }, { "epoch": 2.531738964508332, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.857190132141113, "learning_rate": 5e-05, "loss": 0.7874, "mean_token_accuracy": 0.8766158819198608, "num_tokens": 755854281.0, "step": 19902 }, { "epoch": 2.5318661747869227, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.502796649932861, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.8653542399406433, "num_tokens": 755893769.0, "step": 19903 }, { "epoch": 2.531993385065513, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.901348114013672, "learning_rate": 5e-05, "loss": 0.7872, "mean_token_accuracy": 0.8771740198135376, "num_tokens": 755931690.0, "step": 19904 }, { "epoch": 2.5321205953441037, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.393327713012695, "learning_rate": 5e-05, "loss": 0.7301, "mean_token_accuracy": 0.8809671401977539, "num_tokens": 755967177.0, "step": 19905 }, { "epoch": 2.5322478056226942, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.980418682098389, "learning_rate": 5e-05, "loss": 0.7464, "mean_token_accuracy": 0.8899275064468384, "num_tokens": 755998440.0, "step": 19906 }, { "epoch": 2.5323750159012848, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.273471832275391, "learning_rate": 5e-05, "loss": 0.7303, "mean_token_accuracy": 0.876049280166626, "num_tokens": 756038049.0, "step": 19907 }, { "epoch": 2.5325022261798753, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.04813289642334, "learning_rate": 5e-05, "loss": 0.7555, "mean_token_accuracy": 0.8890595436096191, "num_tokens": 756073185.0, "step": 19908 }, { "epoch": 2.532629436458466, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.227170467376709, "learning_rate": 5e-05, "loss": 0.7377, "mean_token_accuracy": 0.8723156452178955, "num_tokens": 756113439.0, "step": 19909 }, { "epoch": 2.5327566467370564, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.085930824279785, "learning_rate": 5e-05, "loss": 0.7919, "mean_token_accuracy": 0.8830598592758179, "num_tokens": 756155516.0, "step": 19910 }, { "epoch": 2.532883857015647, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.205477714538574, "learning_rate": 5e-05, "loss": 0.7101, "mean_token_accuracy": 0.8811832070350647, "num_tokens": 756196685.0, "step": 19911 }, { "epoch": 2.5330110672942374, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.079608917236328, "learning_rate": 5e-05, "loss": 0.7838, "mean_token_accuracy": 0.8778040409088135, "num_tokens": 756230150.0, "step": 19912 }, { "epoch": 2.533138277572828, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.222841262817383, "learning_rate": 5e-05, "loss": 0.6882, "mean_token_accuracy": 0.8905696272850037, "num_tokens": 756270806.0, "step": 19913 }, { "epoch": 2.5332654878514185, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.019530296325684, "learning_rate": 5e-05, "loss": 0.729, "mean_token_accuracy": 0.8901034593582153, "num_tokens": 756300387.0, "step": 19914 }, { "epoch": 2.533392698130009, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.293669700622559, "learning_rate": 5e-05, "loss": 0.7086, "mean_token_accuracy": 0.884628415107727, "num_tokens": 756335436.0, "step": 19915 }, { "epoch": 2.5335199084085995, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.912795543670654, "learning_rate": 5e-05, "loss": 0.7948, "mean_token_accuracy": 0.8683869242668152, "num_tokens": 756371942.0, "step": 19916 }, { "epoch": 2.53364711868719, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.398528099060059, "learning_rate": 5e-05, "loss": 0.7318, "mean_token_accuracy": 0.8803988695144653, "num_tokens": 756416102.0, "step": 19917 }, { "epoch": 2.5337743289657806, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.734217643737793, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8682039976119995, "num_tokens": 756457066.0, "step": 19918 }, { "epoch": 2.533901539244371, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.562887668609619, "learning_rate": 5e-05, "loss": 0.7301, "mean_token_accuracy": 0.8824514746665955, "num_tokens": 756493977.0, "step": 19919 }, { "epoch": 2.534028749522961, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.524238586425781, "learning_rate": 5e-05, "loss": 0.7717, "mean_token_accuracy": 0.8618754148483276, "num_tokens": 756535926.0, "step": 19920 }, { "epoch": 2.534155959801552, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.745096206665039, "learning_rate": 5e-05, "loss": 0.7061, "mean_token_accuracy": 0.8939646482467651, "num_tokens": 756578406.0, "step": 19921 }, { "epoch": 2.5342831700801423, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.298799991607666, "learning_rate": 5e-05, "loss": 0.7458, "mean_token_accuracy": 0.864856481552124, "num_tokens": 756615104.0, "step": 19922 }, { "epoch": 2.5344103803587332, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.969202518463135, "learning_rate": 5e-05, "loss": 0.7702, "mean_token_accuracy": 0.872684895992279, "num_tokens": 756652059.0, "step": 19923 }, { "epoch": 2.5345375906373233, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 7.020369052886963, "learning_rate": 5e-05, "loss": 0.6343, "mean_token_accuracy": 0.8968982100486755, "num_tokens": 756688640.0, "step": 19924 }, { "epoch": 2.534664800915914, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.331968307495117, "learning_rate": 5e-05, "loss": 0.7451, "mean_token_accuracy": 0.8914883732795715, "num_tokens": 756721596.0, "step": 19925 }, { "epoch": 2.5347920111945044, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.598160743713379, "learning_rate": 5e-05, "loss": 0.6913, "mean_token_accuracy": 0.8685768246650696, "num_tokens": 756757663.0, "step": 19926 }, { "epoch": 2.534919221473095, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.909046173095703, "learning_rate": 5e-05, "loss": 0.8299, "mean_token_accuracy": 0.8819352388381958, "num_tokens": 756802021.0, "step": 19927 }, { "epoch": 2.5350464317516854, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002288818359375, "grad_norm": 6.0515642166137695, "learning_rate": 5e-05, "loss": 0.666, "mean_token_accuracy": 0.8713721632957458, "num_tokens": 756842039.0, "step": 19928 }, { "epoch": 2.535173642030276, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.830572128295898, "learning_rate": 5e-05, "loss": 0.9446, "mean_token_accuracy": 0.88116055727005, "num_tokens": 756875788.0, "step": 19929 }, { "epoch": 2.5353008523088665, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.734142303466797, "learning_rate": 5e-05, "loss": 0.6508, "mean_token_accuracy": 0.8890640139579773, "num_tokens": 756913292.0, "step": 19930 }, { "epoch": 2.535428062587457, "ewc_loss": 0.75, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.666916847229004, "learning_rate": 5e-05, "loss": 1.0542, "mean_token_accuracy": 0.8963363170623779, "num_tokens": 756953005.0, "step": 19931 }, { "epoch": 2.5355552728660475, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.034809589385986, "learning_rate": 5e-05, "loss": 0.882, "mean_token_accuracy": 0.8808134198188782, "num_tokens": 756987656.0, "step": 19932 }, { "epoch": 2.535682483144638, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.09660530090332, "learning_rate": 5e-05, "loss": 1.0274, "mean_token_accuracy": 0.887505054473877, "num_tokens": 757021040.0, "step": 19933 }, { "epoch": 2.5358096934232286, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 8.914767265319824, "learning_rate": 5e-05, "loss": 1.089, "mean_token_accuracy": 0.893191933631897, "num_tokens": 757053039.0, "step": 19934 }, { "epoch": 2.535936903701819, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.657831192016602, "learning_rate": 5e-05, "loss": 0.812, "mean_token_accuracy": 0.8754621744155884, "num_tokens": 757090057.0, "step": 19935 }, { "epoch": 2.5360641139804097, "ewc_loss": 0.85546875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0007781982421875, "grad_norm": 10.836602210998535, "learning_rate": 5e-05, "loss": 1.1912, "mean_token_accuracy": 0.8826872110366821, "num_tokens": 757129537.0, "step": 19936 }, { "epoch": 2.536191324259, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.758102893829346, "learning_rate": 5e-05, "loss": 0.7787, "mean_token_accuracy": 0.8899365663528442, "num_tokens": 757164892.0, "step": 19937 }, { "epoch": 2.5363185345375907, "ewc_loss": 1.109375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00102996826171875, "grad_norm": 13.035016059875488, "learning_rate": 5e-05, "loss": 1.4692, "mean_token_accuracy": 0.8713049292564392, "num_tokens": 757204717.0, "step": 19938 }, { "epoch": 2.5364457448161812, "ewc_loss": 1.0625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00098419189453125, "grad_norm": 10.32598876953125, "learning_rate": 5e-05, "loss": 1.4081, "mean_token_accuracy": 0.8780307769775391, "num_tokens": 757239725.0, "step": 19939 }, { "epoch": 2.5365729550947718, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 11.71232795715332, "learning_rate": 5e-05, "loss": 1.1676, "mean_token_accuracy": 0.8625611066818237, "num_tokens": 757280356.0, "step": 19940 }, { "epoch": 2.5367001653733623, "ewc_loss": 0.890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000812530517578125, "grad_norm": 9.792756080627441, "learning_rate": 5e-05, "loss": 1.2506, "mean_token_accuracy": 0.8739844560623169, "num_tokens": 757316498.0, "step": 19941 }, { "epoch": 2.536827375651953, "ewc_loss": 0.86328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 11.727832794189453, "learning_rate": 5e-05, "loss": 1.2176, "mean_token_accuracy": 0.8764345049858093, "num_tokens": 757351675.0, "step": 19942 }, { "epoch": 2.5369545859305433, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.796213150024414, "learning_rate": 5e-05, "loss": 0.9924, "mean_token_accuracy": 0.8761413097381592, "num_tokens": 757393310.0, "step": 19943 }, { "epoch": 2.537081796209134, "ewc_loss": 0.9296875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000850677490234375, "grad_norm": 11.99254322052002, "learning_rate": 5e-05, "loss": 1.2931, "mean_token_accuracy": 0.8707528710365295, "num_tokens": 757429800.0, "step": 19944 }, { "epoch": 2.537209006487724, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.388104438781738, "learning_rate": 5e-05, "loss": 0.9986, "mean_token_accuracy": 0.8935093879699707, "num_tokens": 757463456.0, "step": 19945 }, { "epoch": 2.537336216766315, "ewc_loss": 0.84375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000766754150390625, "grad_norm": 11.8237943649292, "learning_rate": 5e-05, "loss": 1.2134, "mean_token_accuracy": 0.8678362965583801, "num_tokens": 757496977.0, "step": 19946 }, { "epoch": 2.537463427044905, "ewc_loss": 0.78125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 10.073698043823242, "learning_rate": 5e-05, "loss": 1.0963, "mean_token_accuracy": 0.8858035802841187, "num_tokens": 757536003.0, "step": 19947 }, { "epoch": 2.537590637323496, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 11.057989120483398, "learning_rate": 5e-05, "loss": 1.1069, "mean_token_accuracy": 0.8732681274414062, "num_tokens": 757575017.0, "step": 19948 }, { "epoch": 2.537717847602086, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.663595199584961, "learning_rate": 5e-05, "loss": 1.0256, "mean_token_accuracy": 0.8789844512939453, "num_tokens": 757617098.0, "step": 19949 }, { "epoch": 2.5378450578806766, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.590224266052246, "learning_rate": 5e-05, "loss": 1.0739, "mean_token_accuracy": 0.8675961494445801, "num_tokens": 757652343.0, "step": 19950 }, { "epoch": 2.537972268159267, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.119057655334473, "learning_rate": 5e-05, "loss": 0.9557, "mean_token_accuracy": 0.8800411224365234, "num_tokens": 757691021.0, "step": 19951 }, { "epoch": 2.5380994784378577, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.362114906311035, "learning_rate": 5e-05, "loss": 1.0177, "mean_token_accuracy": 0.88428795337677, "num_tokens": 757727805.0, "step": 19952 }, { "epoch": 2.538226688716448, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.692095756530762, "learning_rate": 5e-05, "loss": 0.9354, "mean_token_accuracy": 0.8747134208679199, "num_tokens": 757767979.0, "step": 19953 }, { "epoch": 2.5383538989950387, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.417632102966309, "learning_rate": 5e-05, "loss": 1.0219, "mean_token_accuracy": 0.8813694715499878, "num_tokens": 757807488.0, "step": 19954 }, { "epoch": 2.5384811092736292, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.560490608215332, "learning_rate": 5e-05, "loss": 0.9223, "mean_token_accuracy": 0.870564341545105, "num_tokens": 757838298.0, "step": 19955 }, { "epoch": 2.5386083195522198, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.548650741577148, "learning_rate": 5e-05, "loss": 1.0655, "mean_token_accuracy": 0.8749890327453613, "num_tokens": 757875016.0, "step": 19956 }, { "epoch": 2.5387355298308103, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.695717811584473, "learning_rate": 5e-05, "loss": 0.902, "mean_token_accuracy": 0.8845695853233337, "num_tokens": 757921306.0, "step": 19957 }, { "epoch": 2.538862740109401, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.550604820251465, "learning_rate": 5e-05, "loss": 1.063, "mean_token_accuracy": 0.8836542963981628, "num_tokens": 757956555.0, "step": 19958 }, { "epoch": 2.5389899503879914, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.004145622253418, "learning_rate": 5e-05, "loss": 0.9803, "mean_token_accuracy": 0.872434139251709, "num_tokens": 757990433.0, "step": 19959 }, { "epoch": 2.539117160666582, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 10.074769020080566, "learning_rate": 5e-05, "loss": 0.9786, "mean_token_accuracy": 0.891589343547821, "num_tokens": 758026606.0, "step": 19960 }, { "epoch": 2.5392443709451724, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.9808988571167, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8953133821487427, "num_tokens": 758060995.0, "step": 19961 }, { "epoch": 2.539371581223763, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.468670845031738, "learning_rate": 5e-05, "loss": 0.9389, "mean_token_accuracy": 0.8861371874809265, "num_tokens": 758097479.0, "step": 19962 }, { "epoch": 2.5394987915023535, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.56503677368164, "learning_rate": 5e-05, "loss": 0.8784, "mean_token_accuracy": 0.8801696300506592, "num_tokens": 758134541.0, "step": 19963 }, { "epoch": 2.539626001780944, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.247879028320312, "learning_rate": 5e-05, "loss": 0.8997, "mean_token_accuracy": 0.8905892372131348, "num_tokens": 758173885.0, "step": 19964 }, { "epoch": 2.5397532120595345, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.0014009475708, "learning_rate": 5e-05, "loss": 0.8293, "mean_token_accuracy": 0.8716464638710022, "num_tokens": 758219394.0, "step": 19965 }, { "epoch": 2.539880422338125, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.47995662689209, "learning_rate": 5e-05, "loss": 0.9584, "mean_token_accuracy": 0.8762545585632324, "num_tokens": 758261146.0, "step": 19966 }, { "epoch": 2.5400076326167156, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.760679721832275, "learning_rate": 5e-05, "loss": 0.7244, "mean_token_accuracy": 0.8999298810958862, "num_tokens": 758297935.0, "step": 19967 }, { "epoch": 2.5401348428953057, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.782896041870117, "learning_rate": 5e-05, "loss": 1.0044, "mean_token_accuracy": 0.871955156326294, "num_tokens": 758328579.0, "step": 19968 }, { "epoch": 2.5402620531738966, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.052631378173828, "learning_rate": 5e-05, "loss": 0.8355, "mean_token_accuracy": 0.876259446144104, "num_tokens": 758361731.0, "step": 19969 }, { "epoch": 2.5403892634524867, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.619542121887207, "learning_rate": 5e-05, "loss": 0.9749, "mean_token_accuracy": 0.8850487470626831, "num_tokens": 758394651.0, "step": 19970 }, { "epoch": 2.5405164737310777, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.446577072143555, "learning_rate": 5e-05, "loss": 0.8664, "mean_token_accuracy": 0.882079005241394, "num_tokens": 758436455.0, "step": 19971 }, { "epoch": 2.540643684009668, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.906014442443848, "learning_rate": 5e-05, "loss": 0.8988, "mean_token_accuracy": 0.8845171332359314, "num_tokens": 758471880.0, "step": 19972 }, { "epoch": 2.5407708942882588, "ewc_loss": 0.5, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.369853973388672, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.8820805549621582, "num_tokens": 758514354.0, "step": 19973 }, { "epoch": 2.540898104566849, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.35844612121582, "learning_rate": 5e-05, "loss": 0.8269, "mean_token_accuracy": 0.887229323387146, "num_tokens": 758554263.0, "step": 19974 }, { "epoch": 2.5410253148454394, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.961557388305664, "learning_rate": 5e-05, "loss": 0.785, "mean_token_accuracy": 0.8795613050460815, "num_tokens": 758588907.0, "step": 19975 }, { "epoch": 2.54115252512403, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.323205947875977, "learning_rate": 5e-05, "loss": 0.836, "mean_token_accuracy": 0.8758574724197388, "num_tokens": 758618573.0, "step": 19976 }, { "epoch": 2.5412797354026204, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.553271770477295, "learning_rate": 5e-05, "loss": 0.7367, "mean_token_accuracy": 0.8812755346298218, "num_tokens": 758654315.0, "step": 19977 }, { "epoch": 2.541406945681211, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.408360481262207, "learning_rate": 5e-05, "loss": 0.8375, "mean_token_accuracy": 0.8748852014541626, "num_tokens": 758690670.0, "step": 19978 }, { "epoch": 2.5415341559598015, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.374000072479248, "learning_rate": 5e-05, "loss": 0.7164, "mean_token_accuracy": 0.8816848993301392, "num_tokens": 758725686.0, "step": 19979 }, { "epoch": 2.541661366238392, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.445921897888184, "learning_rate": 5e-05, "loss": 0.8, "mean_token_accuracy": 0.8861212730407715, "num_tokens": 758757002.0, "step": 19980 }, { "epoch": 2.5417885765169825, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.281678676605225, "learning_rate": 5e-05, "loss": 0.697, "mean_token_accuracy": 0.8857864141464233, "num_tokens": 758793695.0, "step": 19981 }, { "epoch": 2.541915786795573, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.505287170410156, "learning_rate": 5e-05, "loss": 0.8276, "mean_token_accuracy": 0.8779244422912598, "num_tokens": 758837278.0, "step": 19982 }, { "epoch": 2.5420429970741636, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.22838830947876, "learning_rate": 5e-05, "loss": 0.711, "mean_token_accuracy": 0.8774062395095825, "num_tokens": 758871252.0, "step": 19983 }, { "epoch": 2.542170207352754, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.523877143859863, "learning_rate": 5e-05, "loss": 0.8248, "mean_token_accuracy": 0.8793717622756958, "num_tokens": 758912250.0, "step": 19984 }, { "epoch": 2.5422974176313446, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.348498344421387, "learning_rate": 5e-05, "loss": 0.7231, "mean_token_accuracy": 0.8814760446548462, "num_tokens": 758951773.0, "step": 19985 }, { "epoch": 2.542424627909935, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.309318542480469, "learning_rate": 5e-05, "loss": 0.7883, "mean_token_accuracy": 0.8886330723762512, "num_tokens": 758989054.0, "step": 19986 }, { "epoch": 2.5425518381885257, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.584024429321289, "learning_rate": 5e-05, "loss": 0.7093, "mean_token_accuracy": 0.8865553140640259, "num_tokens": 759026917.0, "step": 19987 }, { "epoch": 2.5426790484671162, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.878676891326904, "learning_rate": 5e-05, "loss": 0.7623, "mean_token_accuracy": 0.8804352879524231, "num_tokens": 759065752.0, "step": 19988 }, { "epoch": 2.5428062587457068, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.837595462799072, "learning_rate": 5e-05, "loss": 0.7156, "mean_token_accuracy": 0.888648509979248, "num_tokens": 759100650.0, "step": 19989 }, { "epoch": 2.5429334690242973, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.49984073638916, "learning_rate": 5e-05, "loss": 0.7569, "mean_token_accuracy": 0.8698686361312866, "num_tokens": 759138274.0, "step": 19990 }, { "epoch": 2.543060679302888, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.080761909484863, "learning_rate": 5e-05, "loss": 0.7919, "mean_token_accuracy": 0.8697623014450073, "num_tokens": 759176060.0, "step": 19991 }, { "epoch": 2.5431878895814783, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.152828693389893, "learning_rate": 5e-05, "loss": 0.674, "mean_token_accuracy": 0.8914863467216492, "num_tokens": 759208323.0, "step": 19992 }, { "epoch": 2.5433150998600684, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.413928031921387, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8943352699279785, "num_tokens": 759239180.0, "step": 19993 }, { "epoch": 2.5434423101386594, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.847774982452393, "learning_rate": 5e-05, "loss": 0.671, "mean_token_accuracy": 0.8840503692626953, "num_tokens": 759278489.0, "step": 19994 }, { "epoch": 2.5435695204172495, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.730262756347656, "learning_rate": 5e-05, "loss": 0.8253, "mean_token_accuracy": 0.8781759738922119, "num_tokens": 759315254.0, "step": 19995 }, { "epoch": 2.5436967306958405, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.756710529327393, "learning_rate": 5e-05, "loss": 0.6835, "mean_token_accuracy": 0.8828080892562866, "num_tokens": 759345969.0, "step": 19996 }, { "epoch": 2.5438239409744305, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.745607376098633, "learning_rate": 5e-05, "loss": 0.8419, "mean_token_accuracy": 0.8770519495010376, "num_tokens": 759386530.0, "step": 19997 }, { "epoch": 2.5439511512530215, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.067923545837402, "learning_rate": 5e-05, "loss": 0.7168, "mean_token_accuracy": 0.879392147064209, "num_tokens": 759428755.0, "step": 19998 }, { "epoch": 2.5440783615316116, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.279860496520996, "learning_rate": 5e-05, "loss": 0.794, "mean_token_accuracy": 0.880092442035675, "num_tokens": 759471335.0, "step": 19999 }, { "epoch": 2.544205571810202, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.6386566162109375, "learning_rate": 5e-05, "loss": 0.7367, "mean_token_accuracy": 0.8812757730484009, "num_tokens": 759512541.0, "step": 20000 }, { "epoch": 2.5443327820887927, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.534892559051514, "learning_rate": 5e-05, "loss": 0.7468, "mean_token_accuracy": 0.8767371773719788, "num_tokens": 759551015.0, "step": 20001 }, { "epoch": 2.544459992367383, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.23195743560791, "learning_rate": 5e-05, "loss": 0.803, "mean_token_accuracy": 0.8709965944290161, "num_tokens": 759588698.0, "step": 20002 }, { "epoch": 2.5445872026459737, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.9389824867248535, "learning_rate": 5e-05, "loss": 0.737, "mean_token_accuracy": 0.8652268648147583, "num_tokens": 759630180.0, "step": 20003 }, { "epoch": 2.5447144129245642, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.698204040527344, "learning_rate": 5e-05, "loss": 0.874, "mean_token_accuracy": 0.8594325184822083, "num_tokens": 759664416.0, "step": 20004 }, { "epoch": 2.5448416232031548, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.666509628295898, "learning_rate": 5e-05, "loss": 0.7129, "mean_token_accuracy": 0.8704015016555786, "num_tokens": 759708868.0, "step": 20005 }, { "epoch": 2.5449688334817453, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.876593589782715, "learning_rate": 5e-05, "loss": 0.8348, "mean_token_accuracy": 0.8824272155761719, "num_tokens": 759742431.0, "step": 20006 }, { "epoch": 2.545096043760336, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.819943428039551, "learning_rate": 5e-05, "loss": 0.6877, "mean_token_accuracy": 0.883232593536377, "num_tokens": 759776504.0, "step": 20007 }, { "epoch": 2.5452232540389264, "ewc_loss": 0.5, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.594135284423828, "learning_rate": 5e-05, "loss": 0.8018, "mean_token_accuracy": 0.8935965299606323, "num_tokens": 759808246.0, "step": 20008 }, { "epoch": 2.545350464317517, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.316365718841553, "learning_rate": 5e-05, "loss": 0.6843, "mean_token_accuracy": 0.8931726813316345, "num_tokens": 759846654.0, "step": 20009 }, { "epoch": 2.5454776745961074, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.964011192321777, "learning_rate": 5e-05, "loss": 0.7948, "mean_token_accuracy": 0.8777269124984741, "num_tokens": 759882594.0, "step": 20010 }, { "epoch": 2.545604884874698, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.809742450714111, "learning_rate": 5e-05, "loss": 0.7395, "mean_token_accuracy": 0.8856588006019592, "num_tokens": 759923557.0, "step": 20011 }, { "epoch": 2.5457320951532885, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.493716239929199, "learning_rate": 5e-05, "loss": 0.7588, "mean_token_accuracy": 0.8731297254562378, "num_tokens": 759961152.0, "step": 20012 }, { "epoch": 2.545859305431879, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.974084377288818, "learning_rate": 5e-05, "loss": 0.7565, "mean_token_accuracy": 0.8808069229125977, "num_tokens": 759995505.0, "step": 20013 }, { "epoch": 2.5459865157104695, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.489274978637695, "learning_rate": 5e-05, "loss": 0.727, "mean_token_accuracy": 0.8787091970443726, "num_tokens": 760033910.0, "step": 20014 }, { "epoch": 2.54611372598906, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.694672107696533, "learning_rate": 5e-05, "loss": 0.6853, "mean_token_accuracy": 0.8938846588134766, "num_tokens": 760067605.0, "step": 20015 }, { "epoch": 2.5462409362676506, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.87399435043335, "learning_rate": 5e-05, "loss": 0.7887, "mean_token_accuracy": 0.8649936318397522, "num_tokens": 760101151.0, "step": 20016 }, { "epoch": 2.546368146546241, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.121702194213867, "learning_rate": 5e-05, "loss": 0.7279, "mean_token_accuracy": 0.871492862701416, "num_tokens": 760140253.0, "step": 20017 }, { "epoch": 2.546495356824831, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.505339622497559, "learning_rate": 5e-05, "loss": 0.797, "mean_token_accuracy": 0.8808408975601196, "num_tokens": 760175988.0, "step": 20018 }, { "epoch": 2.546622567103422, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000244140625, "grad_norm": 6.495944976806641, "learning_rate": 5e-05, "loss": 0.6714, "mean_token_accuracy": 0.8771482706069946, "num_tokens": 760211734.0, "step": 20019 }, { "epoch": 2.5467497773820122, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 9.294954299926758, "learning_rate": 5e-05, "loss": 0.8693, "mean_token_accuracy": 0.8747562170028687, "num_tokens": 760248887.0, "step": 20020 }, { "epoch": 2.546876987660603, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00023651123046875, "grad_norm": 6.012817859649658, "learning_rate": 5e-05, "loss": 0.6599, "mean_token_accuracy": 0.8785179853439331, "num_tokens": 760284836.0, "step": 20021 }, { "epoch": 2.5470041979391933, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 10.067694664001465, "learning_rate": 5e-05, "loss": 0.9823, "mean_token_accuracy": 0.8786613941192627, "num_tokens": 760327328.0, "step": 20022 }, { "epoch": 2.547131408217784, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.382939338684082, "learning_rate": 5e-05, "loss": 0.7033, "mean_token_accuracy": 0.8802705407142639, "num_tokens": 760362876.0, "step": 20023 }, { "epoch": 2.5472586184963744, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000762939453125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.080625534057617, "learning_rate": 5e-05, "loss": 1.0083, "mean_token_accuracy": 0.8847241401672363, "num_tokens": 760399652.0, "step": 20024 }, { "epoch": 2.547385828774965, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.806341648101807, "learning_rate": 5e-05, "loss": 0.9071, "mean_token_accuracy": 0.8865885734558105, "num_tokens": 760436346.0, "step": 20025 }, { "epoch": 2.5475130390535554, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.743762016296387, "learning_rate": 5e-05, "loss": 0.9385, "mean_token_accuracy": 0.8801486492156982, "num_tokens": 760468540.0, "step": 20026 }, { "epoch": 2.547640249332146, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.86208438873291, "learning_rate": 5e-05, "loss": 0.9721, "mean_token_accuracy": 0.8801174759864807, "num_tokens": 760504088.0, "step": 20027 }, { "epoch": 2.5477674596107365, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.202120780944824, "learning_rate": 5e-05, "loss": 0.7848, "mean_token_accuracy": 0.8772835731506348, "num_tokens": 760542543.0, "step": 20028 }, { "epoch": 2.547894669889327, "ewc_loss": 0.625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.415063858032227, "learning_rate": 5e-05, "loss": 0.9911, "mean_token_accuracy": 0.8700417280197144, "num_tokens": 760582472.0, "step": 20029 }, { "epoch": 2.5480218801679175, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.998294830322266, "learning_rate": 5e-05, "loss": 0.7882, "mean_token_accuracy": 0.873735249042511, "num_tokens": 760618410.0, "step": 20030 }, { "epoch": 2.548149090446508, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.058065414428711, "learning_rate": 5e-05, "loss": 0.9179, "mean_token_accuracy": 0.8859311938285828, "num_tokens": 760656223.0, "step": 20031 }, { "epoch": 2.5482763007250986, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.53404426574707, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.8950124979019165, "num_tokens": 760700898.0, "step": 20032 }, { "epoch": 2.548403511003689, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.364968299865723, "learning_rate": 5e-05, "loss": 0.8829, "mean_token_accuracy": 0.8800168037414551, "num_tokens": 760739608.0, "step": 20033 }, { "epoch": 2.5485307212822796, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.77009391784668, "learning_rate": 5e-05, "loss": 0.8684, "mean_token_accuracy": 0.8697887063026428, "num_tokens": 760777808.0, "step": 20034 }, { "epoch": 2.54865793156087, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.100945472717285, "learning_rate": 5e-05, "loss": 0.8208, "mean_token_accuracy": 0.8834012746810913, "num_tokens": 760817497.0, "step": 20035 }, { "epoch": 2.5487851418394607, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.568976402282715, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.8776267170906067, "num_tokens": 760855699.0, "step": 20036 }, { "epoch": 2.5489123521180512, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.218160629272461, "learning_rate": 5e-05, "loss": 0.8802, "mean_token_accuracy": 0.8646605014801025, "num_tokens": 760890796.0, "step": 20037 }, { "epoch": 2.5490395623966418, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.3180975914001465, "learning_rate": 5e-05, "loss": 0.7435, "mean_token_accuracy": 0.8825693726539612, "num_tokens": 760929056.0, "step": 20038 }, { "epoch": 2.5491667726752323, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.281339645385742, "learning_rate": 5e-05, "loss": 0.792, "mean_token_accuracy": 0.8909107446670532, "num_tokens": 760963941.0, "step": 20039 }, { "epoch": 2.549293982953823, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.23362922668457, "learning_rate": 5e-05, "loss": 0.7309, "mean_token_accuracy": 0.8818482160568237, "num_tokens": 761009703.0, "step": 20040 }, { "epoch": 2.5494211932324133, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.256754875183105, "learning_rate": 5e-05, "loss": 0.8097, "mean_token_accuracy": 0.882710337638855, "num_tokens": 761051408.0, "step": 20041 }, { "epoch": 2.549548403511004, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.228176116943359, "learning_rate": 5e-05, "loss": 0.75, "mean_token_accuracy": 0.8712341785430908, "num_tokens": 761093764.0, "step": 20042 }, { "epoch": 2.549675613789594, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.211573600769043, "learning_rate": 5e-05, "loss": 0.8026, "mean_token_accuracy": 0.8813315629959106, "num_tokens": 761129975.0, "step": 20043 }, { "epoch": 2.549802824068185, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.223408222198486, "learning_rate": 5e-05, "loss": 0.7476, "mean_token_accuracy": 0.873153567314148, "num_tokens": 761169845.0, "step": 20044 }, { "epoch": 2.549930034346775, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.190679550170898, "learning_rate": 5e-05, "loss": 0.8263, "mean_token_accuracy": 0.8763707876205444, "num_tokens": 761207667.0, "step": 20045 }, { "epoch": 2.550057244625366, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.203000068664551, "learning_rate": 5e-05, "loss": 0.7222, "mean_token_accuracy": 0.8821051120758057, "num_tokens": 761249649.0, "step": 20046 }, { "epoch": 2.550184454903956, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.215987205505371, "learning_rate": 5e-05, "loss": 0.8984, "mean_token_accuracy": 0.8475961685180664, "num_tokens": 761289879.0, "step": 20047 }, { "epoch": 2.5503116651825466, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.142343997955322, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8723375797271729, "num_tokens": 761331398.0, "step": 20048 }, { "epoch": 2.550438875461137, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.267685890197754, "learning_rate": 5e-05, "loss": 0.7962, "mean_token_accuracy": 0.881000816822052, "num_tokens": 761363928.0, "step": 20049 }, { "epoch": 2.5505660857397277, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.083037376403809, "learning_rate": 5e-05, "loss": 0.7735, "mean_token_accuracy": 0.8634124994277954, "num_tokens": 761405737.0, "step": 20050 }, { "epoch": 2.550693296018318, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.300665855407715, "learning_rate": 5e-05, "loss": 0.7779, "mean_token_accuracy": 0.8869178295135498, "num_tokens": 761435934.0, "step": 20051 }, { "epoch": 2.5508205062969087, "ewc_loss": 0.375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.088225841522217, "learning_rate": 5e-05, "loss": 0.6971, "mean_token_accuracy": 0.8844518065452576, "num_tokens": 761466700.0, "step": 20052 }, { "epoch": 2.5509477165754992, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.27204704284668, "learning_rate": 5e-05, "loss": 0.7988, "mean_token_accuracy": 0.8808191418647766, "num_tokens": 761505682.0, "step": 20053 }, { "epoch": 2.5510749268540898, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.146391868591309, "learning_rate": 5e-05, "loss": 0.7245, "mean_token_accuracy": 0.8774810433387756, "num_tokens": 761544071.0, "step": 20054 }, { "epoch": 2.5512021371326803, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.16938304901123, "learning_rate": 5e-05, "loss": 0.7762, "mean_token_accuracy": 0.8862760066986084, "num_tokens": 761586595.0, "step": 20055 }, { "epoch": 2.551329347411271, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.255603790283203, "learning_rate": 5e-05, "loss": 0.7194, "mean_token_accuracy": 0.8824590444564819, "num_tokens": 761623924.0, "step": 20056 }, { "epoch": 2.5514565576898613, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.989790439605713, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8759346008300781, "num_tokens": 761660952.0, "step": 20057 }, { "epoch": 2.551583767968452, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.437544345855713, "learning_rate": 5e-05, "loss": 0.7015, "mean_token_accuracy": 0.8907058238983154, "num_tokens": 761701451.0, "step": 20058 }, { "epoch": 2.5517109782470424, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.768043518066406, "learning_rate": 5e-05, "loss": 0.7502, "mean_token_accuracy": 0.8803392648696899, "num_tokens": 761733638.0, "step": 20059 }, { "epoch": 2.551838188525633, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.608437538146973, "learning_rate": 5e-05, "loss": 0.7252, "mean_token_accuracy": 0.8818998336791992, "num_tokens": 761763978.0, "step": 20060 }, { "epoch": 2.5519653988042235, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.530778884887695, "learning_rate": 5e-05, "loss": 0.7049, "mean_token_accuracy": 0.8847145438194275, "num_tokens": 761802653.0, "step": 20061 }, { "epoch": 2.552092609082814, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.839927673339844, "learning_rate": 5e-05, "loss": 0.7441, "mean_token_accuracy": 0.8823361992835999, "num_tokens": 761840671.0, "step": 20062 }, { "epoch": 2.5522198193614045, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.2380499839782715, "learning_rate": 5e-05, "loss": 0.7293, "mean_token_accuracy": 0.8735884428024292, "num_tokens": 761883823.0, "step": 20063 }, { "epoch": 2.552347029639995, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.171452522277832, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8804826140403748, "num_tokens": 761921597.0, "step": 20064 }, { "epoch": 2.5524742399185856, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.825489521026611, "learning_rate": 5e-05, "loss": 0.6523, "mean_token_accuracy": 0.8877946138381958, "num_tokens": 761963400.0, "step": 20065 }, { "epoch": 2.5526014501971757, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.69826889038086, "learning_rate": 5e-05, "loss": 0.8227, "mean_token_accuracy": 0.8802536129951477, "num_tokens": 762004442.0, "step": 20066 }, { "epoch": 2.5527286604757666, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00023937225341796875, "grad_norm": 6.300539970397949, "learning_rate": 5e-05, "loss": 0.6958, "mean_token_accuracy": 0.866642951965332, "num_tokens": 762040679.0, "step": 20067 }, { "epoch": 2.5528558707543567, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.500093460083008, "learning_rate": 5e-05, "loss": 0.9088, "mean_token_accuracy": 0.8779956698417664, "num_tokens": 762085249.0, "step": 20068 }, { "epoch": 2.5529830810329477, "ewc_loss": 0.322265625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 5.80128812789917, "learning_rate": 5e-05, "loss": 0.6202, "mean_token_accuracy": 0.8941335678100586, "num_tokens": 762125551.0, "step": 20069 }, { "epoch": 2.5531102913115378, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.495238304138184, "learning_rate": 5e-05, "loss": 1.0218, "mean_token_accuracy": 0.8844534158706665, "num_tokens": 762162503.0, "step": 20070 }, { "epoch": 2.5532375015901287, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.463991641998291, "learning_rate": 5e-05, "loss": 0.7811, "mean_token_accuracy": 0.8829362392425537, "num_tokens": 762197302.0, "step": 20071 }, { "epoch": 2.553364711868719, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 10.633488655090332, "learning_rate": 5e-05, "loss": 1.1167, "mean_token_accuracy": 0.8865396976470947, "num_tokens": 762235069.0, "step": 20072 }, { "epoch": 2.5534919221473094, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 8.504443168640137, "learning_rate": 5e-05, "loss": 1.0442, "mean_token_accuracy": 0.8912067413330078, "num_tokens": 762272014.0, "step": 20073 }, { "epoch": 2.5536191324259, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.662542343139648, "learning_rate": 5e-05, "loss": 0.8615, "mean_token_accuracy": 0.8923407793045044, "num_tokens": 762312840.0, "step": 20074 }, { "epoch": 2.5537463427044904, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 9.974565505981445, "learning_rate": 5e-05, "loss": 1.1622, "mean_token_accuracy": 0.8782956600189209, "num_tokens": 762347997.0, "step": 20075 }, { "epoch": 2.553873552983081, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.639835357666016, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.8749421834945679, "num_tokens": 762390706.0, "step": 20076 }, { "epoch": 2.5540007632616715, "ewc_loss": 0.98828125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 12.268043518066406, "learning_rate": 5e-05, "loss": 1.337, "mean_token_accuracy": 0.8779404163360596, "num_tokens": 762425894.0, "step": 20077 }, { "epoch": 2.554127973540262, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 8.762199401855469, "learning_rate": 5e-05, "loss": 1.16, "mean_token_accuracy": 0.8661103248596191, "num_tokens": 762461962.0, "step": 20078 }, { "epoch": 2.5542551838188525, "ewc_loss": 0.875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 12.238482475280762, "learning_rate": 5e-05, "loss": 1.2174, "mean_token_accuracy": 0.8762951493263245, "num_tokens": 762500128.0, "step": 20079 }, { "epoch": 2.554382394097443, "ewc_loss": 0.9921875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000919342041015625, "grad_norm": 10.06203556060791, "learning_rate": 5e-05, "loss": 1.2827, "mean_token_accuracy": 0.8970800638198853, "num_tokens": 762536256.0, "step": 20080 }, { "epoch": 2.5545096043760336, "ewc_loss": 0.8046875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 11.592645645141602, "learning_rate": 5e-05, "loss": 1.2118, "mean_token_accuracy": 0.860271155834198, "num_tokens": 762569797.0, "step": 20081 }, { "epoch": 2.554636814654624, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 9.150289535522461, "learning_rate": 5e-05, "loss": 1.0859, "mean_token_accuracy": 0.8781136274337769, "num_tokens": 762606849.0, "step": 20082 }, { "epoch": 2.5547640249332146, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 11.558575630187988, "learning_rate": 5e-05, "loss": 1.2424, "mean_token_accuracy": 0.8710610866546631, "num_tokens": 762649526.0, "step": 20083 }, { "epoch": 2.554891235211805, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.801847457885742, "learning_rate": 5e-05, "loss": 0.9709, "mean_token_accuracy": 0.8815918564796448, "num_tokens": 762686240.0, "step": 20084 }, { "epoch": 2.5550184454903957, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 11.749594688415527, "learning_rate": 5e-05, "loss": 1.2329, "mean_token_accuracy": 0.8702762126922607, "num_tokens": 762727908.0, "step": 20085 }, { "epoch": 2.5551456557689862, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 9.466365814208984, "learning_rate": 5e-05, "loss": 1.0321, "mean_token_accuracy": 0.8946707248687744, "num_tokens": 762765966.0, "step": 20086 }, { "epoch": 2.5552728660475768, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 11.386173248291016, "learning_rate": 5e-05, "loss": 1.1339, "mean_token_accuracy": 0.8792812824249268, "num_tokens": 762806031.0, "step": 20087 }, { "epoch": 2.5554000763261673, "ewc_loss": 0.75, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 9.747739791870117, "learning_rate": 5e-05, "loss": 1.0592, "mean_token_accuracy": 0.8901932239532471, "num_tokens": 762847531.0, "step": 20088 }, { "epoch": 2.555527286604758, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.721561431884766, "learning_rate": 5e-05, "loss": 1.0403, "mean_token_accuracy": 0.8838821053504944, "num_tokens": 762878409.0, "step": 20089 }, { "epoch": 2.5556544968833483, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.37966251373291, "learning_rate": 5e-05, "loss": 0.9939, "mean_token_accuracy": 0.8803996443748474, "num_tokens": 762918964.0, "step": 20090 }, { "epoch": 2.5557817071619384, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.232623100280762, "learning_rate": 5e-05, "loss": 0.9988, "mean_token_accuracy": 0.887985348701477, "num_tokens": 762958962.0, "step": 20091 }, { "epoch": 2.5559089174405294, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.84152603149414, "learning_rate": 5e-05, "loss": 0.9774, "mean_token_accuracy": 0.8683620691299438, "num_tokens": 762996062.0, "step": 20092 }, { "epoch": 2.5560361277191195, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 10.07493782043457, "learning_rate": 5e-05, "loss": 1.0071, "mean_token_accuracy": 0.8795214891433716, "num_tokens": 763033387.0, "step": 20093 }, { "epoch": 2.5561633379977104, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.429323196411133, "learning_rate": 5e-05, "loss": 0.8964, "mean_token_accuracy": 0.8774574995040894, "num_tokens": 763067826.0, "step": 20094 }, { "epoch": 2.5562905482763005, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.197141647338867, "learning_rate": 5e-05, "loss": 0.9912, "mean_token_accuracy": 0.8883137702941895, "num_tokens": 763100442.0, "step": 20095 }, { "epoch": 2.5564177585548915, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.319498062133789, "learning_rate": 5e-05, "loss": 0.8368, "mean_token_accuracy": 0.892274022102356, "num_tokens": 763137659.0, "step": 20096 }, { "epoch": 2.5565449688334816, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.418503761291504, "learning_rate": 5e-05, "loss": 1.0414, "mean_token_accuracy": 0.8892909288406372, "num_tokens": 763180000.0, "step": 20097 }, { "epoch": 2.556672179112072, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.623180389404297, "learning_rate": 5e-05, "loss": 0.9569, "mean_token_accuracy": 0.8681365847587585, "num_tokens": 763218863.0, "step": 20098 }, { "epoch": 2.5567993893906626, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.238129615783691, "learning_rate": 5e-05, "loss": 1.0042, "mean_token_accuracy": 0.8904154300689697, "num_tokens": 763252973.0, "step": 20099 }, { "epoch": 2.556926599669253, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.920594215393066, "learning_rate": 5e-05, "loss": 0.9968, "mean_token_accuracy": 0.8649174571037292, "num_tokens": 763288781.0, "step": 20100 }, { "epoch": 2.5570538099478437, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.59308910369873, "learning_rate": 5e-05, "loss": 0.9779, "mean_token_accuracy": 0.8763211965560913, "num_tokens": 763321839.0, "step": 20101 }, { "epoch": 2.5571810202264342, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.723283767700195, "learning_rate": 5e-05, "loss": 0.9009, "mean_token_accuracy": 0.8764767050743103, "num_tokens": 763357607.0, "step": 20102 }, { "epoch": 2.5573082305050248, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.08875560760498, "learning_rate": 5e-05, "loss": 0.9422, "mean_token_accuracy": 0.8756515979766846, "num_tokens": 763399489.0, "step": 20103 }, { "epoch": 2.5574354407836153, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.207596778869629, "learning_rate": 5e-05, "loss": 0.8965, "mean_token_accuracy": 0.8651248812675476, "num_tokens": 763441256.0, "step": 20104 }, { "epoch": 2.557562651062206, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.060535430908203, "learning_rate": 5e-05, "loss": 0.9293, "mean_token_accuracy": 0.8709638118743896, "num_tokens": 763475019.0, "step": 20105 }, { "epoch": 2.5576898613407963, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.7229437828063965, "learning_rate": 5e-05, "loss": 0.7692, "mean_token_accuracy": 0.881937563419342, "num_tokens": 763510456.0, "step": 20106 }, { "epoch": 2.557817071619387, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.355972290039062, "learning_rate": 5e-05, "loss": 0.9568, "mean_token_accuracy": 0.8750114440917969, "num_tokens": 763546181.0, "step": 20107 }, { "epoch": 2.5579442818979774, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.591446876525879, "learning_rate": 5e-05, "loss": 0.7617, "mean_token_accuracy": 0.8833823204040527, "num_tokens": 763578763.0, "step": 20108 }, { "epoch": 2.558071492176568, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.572829246520996, "learning_rate": 5e-05, "loss": 1.0159, "mean_token_accuracy": 0.8685933351516724, "num_tokens": 763615615.0, "step": 20109 }, { "epoch": 2.5581987024551585, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.937517166137695, "learning_rate": 5e-05, "loss": 0.7883, "mean_token_accuracy": 0.8913168907165527, "num_tokens": 763655546.0, "step": 20110 }, { "epoch": 2.558325912733749, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.254801750183105, "learning_rate": 5e-05, "loss": 0.962, "mean_token_accuracy": 0.8762680292129517, "num_tokens": 763695400.0, "step": 20111 }, { "epoch": 2.5584531230123395, "ewc_loss": 0.5, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.267545700073242, "learning_rate": 5e-05, "loss": 0.8871, "mean_token_accuracy": 0.8639736771583557, "num_tokens": 763725968.0, "step": 20112 }, { "epoch": 2.55858033329093, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.578143119812012, "learning_rate": 5e-05, "loss": 0.874, "mean_token_accuracy": 0.8838663697242737, "num_tokens": 763763708.0, "step": 20113 }, { "epoch": 2.5587075435695206, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.16669750213623, "learning_rate": 5e-05, "loss": 0.8504, "mean_token_accuracy": 0.8686394095420837, "num_tokens": 763800836.0, "step": 20114 }, { "epoch": 2.558834753848111, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.171638488769531, "learning_rate": 5e-05, "loss": 0.8592, "mean_token_accuracy": 0.8721164464950562, "num_tokens": 763842119.0, "step": 20115 }, { "epoch": 2.558961964126701, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.858999729156494, "learning_rate": 5e-05, "loss": 0.7746, "mean_token_accuracy": 0.8787767887115479, "num_tokens": 763876797.0, "step": 20116 }, { "epoch": 2.559089174405292, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.132122993469238, "learning_rate": 5e-05, "loss": 0.8194, "mean_token_accuracy": 0.8762598633766174, "num_tokens": 763913131.0, "step": 20117 }, { "epoch": 2.5592163846838822, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.5724945068359375, "learning_rate": 5e-05, "loss": 0.7336, "mean_token_accuracy": 0.8808935880661011, "num_tokens": 763952430.0, "step": 20118 }, { "epoch": 2.559343594962473, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.192075729370117, "learning_rate": 5e-05, "loss": 0.8116, "mean_token_accuracy": 0.8815730214118958, "num_tokens": 763988978.0, "step": 20119 }, { "epoch": 2.5594708052410633, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.42484712600708, "learning_rate": 5e-05, "loss": 0.7496, "mean_token_accuracy": 0.8771711587905884, "num_tokens": 764027473.0, "step": 20120 }, { "epoch": 2.559598015519654, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.180296897888184, "learning_rate": 5e-05, "loss": 0.7781, "mean_token_accuracy": 0.8837368488311768, "num_tokens": 764063681.0, "step": 20121 }, { "epoch": 2.5597252257982444, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.357094764709473, "learning_rate": 5e-05, "loss": 0.708, "mean_token_accuracy": 0.8848710060119629, "num_tokens": 764102463.0, "step": 20122 }, { "epoch": 2.559852436076835, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.258866310119629, "learning_rate": 5e-05, "loss": 0.7693, "mean_token_accuracy": 0.890406608581543, "num_tokens": 764141912.0, "step": 20123 }, { "epoch": 2.5599796463554254, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.224636077880859, "learning_rate": 5e-05, "loss": 0.6727, "mean_token_accuracy": 0.891234815120697, "num_tokens": 764176448.0, "step": 20124 }, { "epoch": 2.560106856634016, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.37921142578125, "learning_rate": 5e-05, "loss": 0.8473, "mean_token_accuracy": 0.8664162158966064, "num_tokens": 764217287.0, "step": 20125 }, { "epoch": 2.5602340669126065, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.160738468170166, "learning_rate": 5e-05, "loss": 0.7119, "mean_token_accuracy": 0.8783040046691895, "num_tokens": 764255593.0, "step": 20126 }, { "epoch": 2.560361277191197, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.398715019226074, "learning_rate": 5e-05, "loss": 0.8192, "mean_token_accuracy": 0.8769526481628418, "num_tokens": 764294624.0, "step": 20127 }, { "epoch": 2.5604884874697875, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.260459899902344, "learning_rate": 5e-05, "loss": 0.6771, "mean_token_accuracy": 0.8903148770332336, "num_tokens": 764328495.0, "step": 20128 }, { "epoch": 2.560615697748378, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.170490264892578, "learning_rate": 5e-05, "loss": 0.758, "mean_token_accuracy": 0.888077437877655, "num_tokens": 764359986.0, "step": 20129 }, { "epoch": 2.5607429080269686, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.505986213684082, "learning_rate": 5e-05, "loss": 0.7525, "mean_token_accuracy": 0.8713175058364868, "num_tokens": 764394039.0, "step": 20130 }, { "epoch": 2.560870118305559, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.842987537384033, "learning_rate": 5e-05, "loss": 0.7846, "mean_token_accuracy": 0.8730511665344238, "num_tokens": 764436106.0, "step": 20131 }, { "epoch": 2.5609973285841496, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.812102794647217, "learning_rate": 5e-05, "loss": 0.7265, "mean_token_accuracy": 0.8828567266464233, "num_tokens": 764470004.0, "step": 20132 }, { "epoch": 2.56112453886274, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.367885589599609, "learning_rate": 5e-05, "loss": 0.6829, "mean_token_accuracy": 0.8887183666229248, "num_tokens": 764507679.0, "step": 20133 }, { "epoch": 2.5612517491413307, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.229615211486816, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8924694061279297, "num_tokens": 764548818.0, "step": 20134 }, { "epoch": 2.561378959419921, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.8949360847473145, "learning_rate": 5e-05, "loss": 0.7163, "mean_token_accuracy": 0.8757867217063904, "num_tokens": 764590289.0, "step": 20135 }, { "epoch": 2.5615061696985117, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00075531005859375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.723652839660645, "learning_rate": 5e-05, "loss": 0.7885, "mean_token_accuracy": 0.8870176672935486, "num_tokens": 764624210.0, "step": 20136 }, { "epoch": 2.5616333799771023, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.521510601043701, "learning_rate": 5e-05, "loss": 0.6847, "mean_token_accuracy": 0.87720787525177, "num_tokens": 764657071.0, "step": 20137 }, { "epoch": 2.561760590255693, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 9.212555885314941, "learning_rate": 5e-05, "loss": 0.8671, "mean_token_accuracy": 0.8798196315765381, "num_tokens": 764691439.0, "step": 20138 }, { "epoch": 2.5618878005342833, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.410894393920898, "learning_rate": 5e-05, "loss": 0.6353, "mean_token_accuracy": 0.8951574563980103, "num_tokens": 764731551.0, "step": 20139 }, { "epoch": 2.562015010812874, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.371268272399902, "learning_rate": 5e-05, "loss": 0.8926, "mean_token_accuracy": 0.8940591216087341, "num_tokens": 764763233.0, "step": 20140 }, { "epoch": 2.562142221091464, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.030079364776611, "learning_rate": 5e-05, "loss": 0.6651, "mean_token_accuracy": 0.8977212905883789, "num_tokens": 764796937.0, "step": 20141 }, { "epoch": 2.562269431370055, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.589319229125977, "learning_rate": 5e-05, "loss": 0.8487, "mean_token_accuracy": 0.8849487900733948, "num_tokens": 764831808.0, "step": 20142 }, { "epoch": 2.562396641648645, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.046844482421875, "learning_rate": 5e-05, "loss": 0.8209, "mean_token_accuracy": 0.8795875310897827, "num_tokens": 764863288.0, "step": 20143 }, { "epoch": 2.562523851927236, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.303061008453369, "learning_rate": 5e-05, "loss": 0.7018, "mean_token_accuracy": 0.8920676708221436, "num_tokens": 764905417.0, "step": 20144 }, { "epoch": 2.562651062205826, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.893071174621582, "learning_rate": 5e-05, "loss": 0.8585, "mean_token_accuracy": 0.8817116618156433, "num_tokens": 764942058.0, "step": 20145 }, { "epoch": 2.5627782724844166, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.64355993270874, "learning_rate": 5e-05, "loss": 0.7048, "mean_token_accuracy": 0.8796056509017944, "num_tokens": 764983652.0, "step": 20146 }, { "epoch": 2.562905482763007, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.213738441467285, "learning_rate": 5e-05, "loss": 0.8932, "mean_token_accuracy": 0.884874701499939, "num_tokens": 765023966.0, "step": 20147 }, { "epoch": 2.5630326930415976, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.900978088378906, "learning_rate": 5e-05, "loss": 0.7165, "mean_token_accuracy": 0.8760740756988525, "num_tokens": 765055847.0, "step": 20148 }, { "epoch": 2.563159903320188, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.759812355041504, "learning_rate": 5e-05, "loss": 0.9125, "mean_token_accuracy": 0.8706137537956238, "num_tokens": 765094387.0, "step": 20149 }, { "epoch": 2.5632871135987787, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.588706016540527, "learning_rate": 5e-05, "loss": 0.8081, "mean_token_accuracy": 0.8708993196487427, "num_tokens": 765128193.0, "step": 20150 }, { "epoch": 2.5634143238773692, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.9500603675842285, "learning_rate": 5e-05, "loss": 0.8175, "mean_token_accuracy": 0.8773396611213684, "num_tokens": 765166550.0, "step": 20151 }, { "epoch": 2.5635415341559598, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.023921966552734, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8788257241249084, "num_tokens": 765204228.0, "step": 20152 }, { "epoch": 2.5636687444345503, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.5625457763671875, "learning_rate": 5e-05, "loss": 0.7443, "mean_token_accuracy": 0.8844727873802185, "num_tokens": 765247148.0, "step": 20153 }, { "epoch": 2.563795954713141, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.962801456451416, "learning_rate": 5e-05, "loss": 0.723, "mean_token_accuracy": 0.8934739232063293, "num_tokens": 765287902.0, "step": 20154 }, { "epoch": 2.5639231649917313, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.731256484985352, "learning_rate": 5e-05, "loss": 0.7566, "mean_token_accuracy": 0.8852167725563049, "num_tokens": 765331118.0, "step": 20155 }, { "epoch": 2.564050375270322, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.462762355804443, "learning_rate": 5e-05, "loss": 0.7055, "mean_token_accuracy": 0.8875974416732788, "num_tokens": 765369942.0, "step": 20156 }, { "epoch": 2.5641775855489124, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.235994338989258, "learning_rate": 5e-05, "loss": 0.8163, "mean_token_accuracy": 0.8786333799362183, "num_tokens": 765407664.0, "step": 20157 }, { "epoch": 2.564304795827503, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.868337631225586, "learning_rate": 5e-05, "loss": 0.7098, "mean_token_accuracy": 0.8798146843910217, "num_tokens": 765453303.0, "step": 20158 }, { "epoch": 2.5644320061060935, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.845039367675781, "learning_rate": 5e-05, "loss": 0.8456, "mean_token_accuracy": 0.8821612000465393, "num_tokens": 765492558.0, "step": 20159 }, { "epoch": 2.564559216384684, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.439568042755127, "learning_rate": 5e-05, "loss": 0.7451, "mean_token_accuracy": 0.8597744703292847, "num_tokens": 765529964.0, "step": 20160 }, { "epoch": 2.5646864266632745, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.389225959777832, "learning_rate": 5e-05, "loss": 0.8849, "mean_token_accuracy": 0.8934107422828674, "num_tokens": 765570158.0, "step": 20161 }, { "epoch": 2.564813636941865, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.360213756561279, "learning_rate": 5e-05, "loss": 0.7461, "mean_token_accuracy": 0.8608571290969849, "num_tokens": 765609304.0, "step": 20162 }, { "epoch": 2.5649408472204556, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.64285659790039, "learning_rate": 5e-05, "loss": 0.9706, "mean_token_accuracy": 0.8787689208984375, "num_tokens": 765651124.0, "step": 20163 }, { "epoch": 2.5650680574990457, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.917339324951172, "learning_rate": 5e-05, "loss": 0.724, "mean_token_accuracy": 0.8946239948272705, "num_tokens": 765685544.0, "step": 20164 }, { "epoch": 2.5651952677776366, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.290006637573242, "learning_rate": 5e-05, "loss": 0.9418, "mean_token_accuracy": 0.8912239670753479, "num_tokens": 765726120.0, "step": 20165 }, { "epoch": 2.5653224780562267, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.697126388549805, "learning_rate": 5e-05, "loss": 0.9039, "mean_token_accuracy": 0.8676048517227173, "num_tokens": 765764145.0, "step": 20166 }, { "epoch": 2.5654496883348177, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.494547843933105, "learning_rate": 5e-05, "loss": 0.8607, "mean_token_accuracy": 0.8918851613998413, "num_tokens": 765795530.0, "step": 20167 }, { "epoch": 2.5655768986134078, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.0340576171875, "learning_rate": 5e-05, "loss": 0.8864, "mean_token_accuracy": 0.8754754066467285, "num_tokens": 765831793.0, "step": 20168 }, { "epoch": 2.5657041088919987, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.039341926574707, "learning_rate": 5e-05, "loss": 0.8691, "mean_token_accuracy": 0.8722737431526184, "num_tokens": 765876724.0, "step": 20169 }, { "epoch": 2.565831319170589, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.779752731323242, "learning_rate": 5e-05, "loss": 0.8112, "mean_token_accuracy": 0.8932768106460571, "num_tokens": 765917604.0, "step": 20170 }, { "epoch": 2.5659585294491793, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.209574699401855, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8816627860069275, "num_tokens": 765953717.0, "step": 20171 }, { "epoch": 2.56608573972777, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.354140758514404, "learning_rate": 5e-05, "loss": 0.7737, "mean_token_accuracy": 0.8772046566009521, "num_tokens": 765989967.0, "step": 20172 }, { "epoch": 2.5662129500063604, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.366695404052734, "learning_rate": 5e-05, "loss": 0.8499, "mean_token_accuracy": 0.8840212821960449, "num_tokens": 766029276.0, "step": 20173 }, { "epoch": 2.566340160284951, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.31853723526001, "learning_rate": 5e-05, "loss": 0.7335, "mean_token_accuracy": 0.8870797753334045, "num_tokens": 766064819.0, "step": 20174 }, { "epoch": 2.5664673705635415, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.127342224121094, "learning_rate": 5e-05, "loss": 0.825, "mean_token_accuracy": 0.8751485347747803, "num_tokens": 766103621.0, "step": 20175 }, { "epoch": 2.566594580842132, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.54695987701416, "learning_rate": 5e-05, "loss": 0.7371, "mean_token_accuracy": 0.886466383934021, "num_tokens": 766138945.0, "step": 20176 }, { "epoch": 2.5667217911207225, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.817233562469482, "learning_rate": 5e-05, "loss": 0.7884, "mean_token_accuracy": 0.876322865486145, "num_tokens": 766180771.0, "step": 20177 }, { "epoch": 2.566849001399313, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.683770179748535, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.8718698620796204, "num_tokens": 766216907.0, "step": 20178 }, { "epoch": 2.5669762116779036, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.730170249938965, "learning_rate": 5e-05, "loss": 0.769, "mean_token_accuracy": 0.8778977394104004, "num_tokens": 766256302.0, "step": 20179 }, { "epoch": 2.567103421956494, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.596205234527588, "learning_rate": 5e-05, "loss": 0.7411, "mean_token_accuracy": 0.8831148743629456, "num_tokens": 766291255.0, "step": 20180 }, { "epoch": 2.5672306322350846, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.836422443389893, "learning_rate": 5e-05, "loss": 0.7369, "mean_token_accuracy": 0.884475588798523, "num_tokens": 766323895.0, "step": 20181 }, { "epoch": 2.567357842513675, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.389485836029053, "learning_rate": 5e-05, "loss": 0.6998, "mean_token_accuracy": 0.8870358467102051, "num_tokens": 766364518.0, "step": 20182 }, { "epoch": 2.5674850527922657, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.001420021057129, "learning_rate": 5e-05, "loss": 0.8051, "mean_token_accuracy": 0.8675872087478638, "num_tokens": 766402256.0, "step": 20183 }, { "epoch": 2.567612263070856, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.240569591522217, "learning_rate": 5e-05, "loss": 0.7291, "mean_token_accuracy": 0.8810545206069946, "num_tokens": 766435186.0, "step": 20184 }, { "epoch": 2.5677394733494467, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.152975082397461, "learning_rate": 5e-05, "loss": 0.8108, "mean_token_accuracy": 0.8761427402496338, "num_tokens": 766476950.0, "step": 20185 }, { "epoch": 2.5678666836280373, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.09269905090332, "learning_rate": 5e-05, "loss": 0.6673, "mean_token_accuracy": 0.8941003084182739, "num_tokens": 766517908.0, "step": 20186 }, { "epoch": 2.567993893906628, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.309557914733887, "learning_rate": 5e-05, "loss": 0.8273, "mean_token_accuracy": 0.8696276545524597, "num_tokens": 766556251.0, "step": 20187 }, { "epoch": 2.5681211041852183, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.9196343421936035, "learning_rate": 5e-05, "loss": 0.73, "mean_token_accuracy": 0.8687679767608643, "num_tokens": 766593865.0, "step": 20188 }, { "epoch": 2.5682483144638084, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.578688621520996, "learning_rate": 5e-05, "loss": 0.8261, "mean_token_accuracy": 0.8758119344711304, "num_tokens": 766627594.0, "step": 20189 }, { "epoch": 2.5683755247423994, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.687533378601074, "learning_rate": 5e-05, "loss": 0.6679, "mean_token_accuracy": 0.8855797648429871, "num_tokens": 766669620.0, "step": 20190 }, { "epoch": 2.5685027350209895, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.903616905212402, "learning_rate": 5e-05, "loss": 0.843, "mean_token_accuracy": 0.8863929510116577, "num_tokens": 766711553.0, "step": 20191 }, { "epoch": 2.5686299452995804, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.594513893127441, "learning_rate": 5e-05, "loss": 0.6942, "mean_token_accuracy": 0.8793846368789673, "num_tokens": 766751717.0, "step": 20192 }, { "epoch": 2.5687571555781705, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 9.07220458984375, "learning_rate": 5e-05, "loss": 0.8867, "mean_token_accuracy": 0.8826799392700195, "num_tokens": 766782207.0, "step": 20193 }, { "epoch": 2.5688843658567615, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.84739351272583, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.8741751313209534, "num_tokens": 766822050.0, "step": 20194 }, { "epoch": 2.5690115761353516, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.685173034667969, "learning_rate": 5e-05, "loss": 0.9096, "mean_token_accuracy": 0.8711268901824951, "num_tokens": 766867713.0, "step": 20195 }, { "epoch": 2.569138786413942, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.578200340270996, "learning_rate": 5e-05, "loss": 0.8001, "mean_token_accuracy": 0.8846552968025208, "num_tokens": 766908191.0, "step": 20196 }, { "epoch": 2.5692659966925326, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.657509803771973, "learning_rate": 5e-05, "loss": 0.8196, "mean_token_accuracy": 0.869901180267334, "num_tokens": 766945774.0, "step": 20197 }, { "epoch": 2.569393206971123, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.444426536560059, "learning_rate": 5e-05, "loss": 0.8196, "mean_token_accuracy": 0.8929632902145386, "num_tokens": 766986255.0, "step": 20198 }, { "epoch": 2.5695204172497137, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.654089450836182, "learning_rate": 5e-05, "loss": 0.6963, "mean_token_accuracy": 0.8781490325927734, "num_tokens": 767020930.0, "step": 20199 }, { "epoch": 2.5696476275283042, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.312503814697266, "learning_rate": 5e-05, "loss": 0.9391, "mean_token_accuracy": 0.874595046043396, "num_tokens": 767059863.0, "step": 20200 }, { "epoch": 2.5697748378068948, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.091773509979248, "learning_rate": 5e-05, "loss": 0.7556, "mean_token_accuracy": 0.856532871723175, "num_tokens": 767100566.0, "step": 20201 }, { "epoch": 2.5699020480854853, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 10.118833541870117, "learning_rate": 5e-05, "loss": 1.0175, "mean_token_accuracy": 0.8732419610023499, "num_tokens": 767137892.0, "step": 20202 }, { "epoch": 2.570029258364076, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 6.544264793395996, "learning_rate": 5e-05, "loss": 0.8401, "mean_token_accuracy": 0.8613981008529663, "num_tokens": 767175340.0, "step": 20203 }, { "epoch": 2.5701564686426663, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.118426322937012, "learning_rate": 5e-05, "loss": 1.1168, "mean_token_accuracy": 0.8641072511672974, "num_tokens": 767209601.0, "step": 20204 }, { "epoch": 2.570283678921257, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.129322052001953, "learning_rate": 5e-05, "loss": 0.9384, "mean_token_accuracy": 0.8958241939544678, "num_tokens": 767247500.0, "step": 20205 }, { "epoch": 2.5704108891998474, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.580554008483887, "learning_rate": 5e-05, "loss": 0.9766, "mean_token_accuracy": 0.8666681051254272, "num_tokens": 767285251.0, "step": 20206 }, { "epoch": 2.570538099478438, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.200575828552246, "learning_rate": 5e-05, "loss": 1.0078, "mean_token_accuracy": 0.886967658996582, "num_tokens": 767321493.0, "step": 20207 }, { "epoch": 2.5706653097570284, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.985881328582764, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8863519430160522, "num_tokens": 767351375.0, "step": 20208 }, { "epoch": 2.570792520035619, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.890420913696289, "learning_rate": 5e-05, "loss": 1.0152, "mean_token_accuracy": 0.8794318437576294, "num_tokens": 767392990.0, "step": 20209 }, { "epoch": 2.5709197303142095, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.181879997253418, "learning_rate": 5e-05, "loss": 0.8392, "mean_token_accuracy": 0.8851269483566284, "num_tokens": 767431584.0, "step": 20210 }, { "epoch": 2.5710469405928, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.2396240234375, "learning_rate": 5e-05, "loss": 0.9823, "mean_token_accuracy": 0.8752076625823975, "num_tokens": 767468451.0, "step": 20211 }, { "epoch": 2.5711741508713906, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.22374439239502, "learning_rate": 5e-05, "loss": 0.9879, "mean_token_accuracy": 0.8735156059265137, "num_tokens": 767508133.0, "step": 20212 }, { "epoch": 2.571301361149981, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.787803649902344, "learning_rate": 5e-05, "loss": 0.8042, "mean_token_accuracy": 0.8851655721664429, "num_tokens": 767539427.0, "step": 20213 }, { "epoch": 2.571428571428571, "ewc_loss": 0.625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.922446250915527, "learning_rate": 5e-05, "loss": 0.9711, "mean_token_accuracy": 0.8790051937103271, "num_tokens": 767574842.0, "step": 20214 }, { "epoch": 2.571555781707162, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.15103816986084, "learning_rate": 5e-05, "loss": 0.7885, "mean_token_accuracy": 0.8763260841369629, "num_tokens": 767609378.0, "step": 20215 }, { "epoch": 2.5716829919857522, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.023198127746582, "learning_rate": 5e-05, "loss": 0.951, "mean_token_accuracy": 0.8718465566635132, "num_tokens": 767642283.0, "step": 20216 }, { "epoch": 2.571810202264343, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.354259967803955, "learning_rate": 5e-05, "loss": 0.8107, "mean_token_accuracy": 0.8864672183990479, "num_tokens": 767680019.0, "step": 20217 }, { "epoch": 2.5719374125429333, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.578852653503418, "learning_rate": 5e-05, "loss": 0.9143, "mean_token_accuracy": 0.8740007877349854, "num_tokens": 767715570.0, "step": 20218 }, { "epoch": 2.572064622821524, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.622853755950928, "learning_rate": 5e-05, "loss": 0.8534, "mean_token_accuracy": 0.8791848421096802, "num_tokens": 767757257.0, "step": 20219 }, { "epoch": 2.5721918331001143, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.1813325881958, "learning_rate": 5e-05, "loss": 0.7722, "mean_token_accuracy": 0.9000335931777954, "num_tokens": 767791684.0, "step": 20220 }, { "epoch": 2.572319043378705, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.625354290008545, "learning_rate": 5e-05, "loss": 0.8262, "mean_token_accuracy": 0.8789108991622925, "num_tokens": 767831687.0, "step": 20221 }, { "epoch": 2.5724462536572954, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.063488006591797, "learning_rate": 5e-05, "loss": 0.8034, "mean_token_accuracy": 0.8859118223190308, "num_tokens": 767871349.0, "step": 20222 }, { "epoch": 2.572573463935886, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.48313045501709, "learning_rate": 5e-05, "loss": 0.7589, "mean_token_accuracy": 0.8864073157310486, "num_tokens": 767908301.0, "step": 20223 }, { "epoch": 2.5727006742144765, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.073464393615723, "learning_rate": 5e-05, "loss": 0.7966, "mean_token_accuracy": 0.8822222352027893, "num_tokens": 767944809.0, "step": 20224 }, { "epoch": 2.572827884493067, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.34905481338501, "learning_rate": 5e-05, "loss": 0.7664, "mean_token_accuracy": 0.8740827441215515, "num_tokens": 767984629.0, "step": 20225 }, { "epoch": 2.5729550947716575, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.118803977966309, "learning_rate": 5e-05, "loss": 0.8232, "mean_token_accuracy": 0.8685534000396729, "num_tokens": 768017697.0, "step": 20226 }, { "epoch": 2.573082305050248, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.210556983947754, "learning_rate": 5e-05, "loss": 0.7394, "mean_token_accuracy": 0.8758325576782227, "num_tokens": 768060259.0, "step": 20227 }, { "epoch": 2.5732095153288386, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.248398780822754, "learning_rate": 5e-05, "loss": 0.8178, "mean_token_accuracy": 0.8709158897399902, "num_tokens": 768096248.0, "step": 20228 }, { "epoch": 2.573336725607429, "ewc_loss": 0.375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.041147708892822, "learning_rate": 5e-05, "loss": 0.715, "mean_token_accuracy": 0.8831174373626709, "num_tokens": 768131730.0, "step": 20229 }, { "epoch": 2.5734639358860196, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.473444938659668, "learning_rate": 5e-05, "loss": 0.795, "mean_token_accuracy": 0.8887237906455994, "num_tokens": 768167551.0, "step": 20230 }, { "epoch": 2.57359114616461, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.779937267303467, "learning_rate": 5e-05, "loss": 0.6964, "mean_token_accuracy": 0.8805665969848633, "num_tokens": 768198212.0, "step": 20231 }, { "epoch": 2.5737183564432007, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.821717262268066, "learning_rate": 5e-05, "loss": 0.8565, "mean_token_accuracy": 0.8790246248245239, "num_tokens": 768233052.0, "step": 20232 }, { "epoch": 2.573845566721791, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.525463581085205, "learning_rate": 5e-05, "loss": 0.6605, "mean_token_accuracy": 0.8925889730453491, "num_tokens": 768272961.0, "step": 20233 }, { "epoch": 2.5739727770003817, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.209437370300293, "learning_rate": 5e-05, "loss": 0.8973, "mean_token_accuracy": 0.8868109583854675, "num_tokens": 768308748.0, "step": 20234 }, { "epoch": 2.5740999872789723, "ewc_loss": 0.375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.508170127868652, "learning_rate": 5e-05, "loss": 0.7437, "mean_token_accuracy": 0.8719265460968018, "num_tokens": 768346592.0, "step": 20235 }, { "epoch": 2.574227197557563, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.349863052368164, "learning_rate": 5e-05, "loss": 0.9505, "mean_token_accuracy": 0.8824657201766968, "num_tokens": 768394367.0, "step": 20236 }, { "epoch": 2.5743544078361533, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 6.953237056732178, "learning_rate": 5e-05, "loss": 0.834, "mean_token_accuracy": 0.8745861053466797, "num_tokens": 768437060.0, "step": 20237 }, { "epoch": 2.574481618114744, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.932317733764648, "learning_rate": 5e-05, "loss": 0.9415, "mean_token_accuracy": 0.8797141313552856, "num_tokens": 768470373.0, "step": 20238 }, { "epoch": 2.574608828393334, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.657089710235596, "learning_rate": 5e-05, "loss": 0.8641, "mean_token_accuracy": 0.8881914019584656, "num_tokens": 768507567.0, "step": 20239 }, { "epoch": 2.574736038671925, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.022482872009277, "learning_rate": 5e-05, "loss": 0.8398, "mean_token_accuracy": 0.8876348733901978, "num_tokens": 768547217.0, "step": 20240 }, { "epoch": 2.574863248950515, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.226241111755371, "learning_rate": 5e-05, "loss": 0.8643, "mean_token_accuracy": 0.8879197835922241, "num_tokens": 768586720.0, "step": 20241 }, { "epoch": 2.574990459229106, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.286167621612549, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.8831877708435059, "num_tokens": 768627890.0, "step": 20242 }, { "epoch": 2.575117669507696, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.521185874938965, "learning_rate": 5e-05, "loss": 0.9015, "mean_token_accuracy": 0.8749047517776489, "num_tokens": 768663393.0, "step": 20243 }, { "epoch": 2.5752448797862866, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.001652717590332, "learning_rate": 5e-05, "loss": 0.7067, "mean_token_accuracy": 0.889690637588501, "num_tokens": 768706950.0, "step": 20244 }, { "epoch": 2.575372090064877, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.632034301757812, "learning_rate": 5e-05, "loss": 0.9139, "mean_token_accuracy": 0.8714019060134888, "num_tokens": 768747391.0, "step": 20245 }, { "epoch": 2.5754993003434676, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.957574844360352, "learning_rate": 5e-05, "loss": 0.6765, "mean_token_accuracy": 0.8951932191848755, "num_tokens": 768781170.0, "step": 20246 }, { "epoch": 2.575626510622058, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.72937297821045, "learning_rate": 5e-05, "loss": 0.9064, "mean_token_accuracy": 0.8753383159637451, "num_tokens": 768820505.0, "step": 20247 }, { "epoch": 2.5757537209006487, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.932681560516357, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8791781067848206, "num_tokens": 768859647.0, "step": 20248 }, { "epoch": 2.575880931179239, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.824150085449219, "learning_rate": 5e-05, "loss": 0.8753, "mean_token_accuracy": 0.891839861869812, "num_tokens": 768894883.0, "step": 20249 }, { "epoch": 2.5760081414578297, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.009376525878906, "learning_rate": 5e-05, "loss": 0.7551, "mean_token_accuracy": 0.8900680541992188, "num_tokens": 768934648.0, "step": 20250 }, { "epoch": 2.5761353517364203, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.804298400878906, "learning_rate": 5e-05, "loss": 0.8925, "mean_token_accuracy": 0.8839385509490967, "num_tokens": 768969568.0, "step": 20251 }, { "epoch": 2.576262562015011, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.207250595092773, "learning_rate": 5e-05, "loss": 0.8543, "mean_token_accuracy": 0.8693723678588867, "num_tokens": 769009423.0, "step": 20252 }, { "epoch": 2.5763897722936013, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.576391220092773, "learning_rate": 5e-05, "loss": 0.8975, "mean_token_accuracy": 0.8826260566711426, "num_tokens": 769043812.0, "step": 20253 }, { "epoch": 2.576516982572192, "ewc_loss": 0.5, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.443613529205322, "learning_rate": 5e-05, "loss": 0.8255, "mean_token_accuracy": 0.8836881518363953, "num_tokens": 769079959.0, "step": 20254 }, { "epoch": 2.5766441928507824, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.273432731628418, "learning_rate": 5e-05, "loss": 0.8344, "mean_token_accuracy": 0.8908595442771912, "num_tokens": 769119375.0, "step": 20255 }, { "epoch": 2.576771403129373, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.526480197906494, "learning_rate": 5e-05, "loss": 0.8107, "mean_token_accuracy": 0.886999785900116, "num_tokens": 769157085.0, "step": 20256 }, { "epoch": 2.5768986134079634, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.136065483093262, "learning_rate": 5e-05, "loss": 0.8463, "mean_token_accuracy": 0.8740358352661133, "num_tokens": 769190883.0, "step": 20257 }, { "epoch": 2.577025823686554, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.406437873840332, "learning_rate": 5e-05, "loss": 0.7784, "mean_token_accuracy": 0.8828272223472595, "num_tokens": 769228370.0, "step": 20258 }, { "epoch": 2.5771530339651445, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.24630069732666, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.9064710736274719, "num_tokens": 769265551.0, "step": 20259 }, { "epoch": 2.577280244243735, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.182585716247559, "learning_rate": 5e-05, "loss": 0.749, "mean_token_accuracy": 0.8831583857536316, "num_tokens": 769299482.0, "step": 20260 }, { "epoch": 2.5774074545223256, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.431090354919434, "learning_rate": 5e-05, "loss": 0.8561, "mean_token_accuracy": 0.8802307844161987, "num_tokens": 769339165.0, "step": 20261 }, { "epoch": 2.5775346648009156, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.051616191864014, "learning_rate": 5e-05, "loss": 0.7236, "mean_token_accuracy": 0.8842282295227051, "num_tokens": 769374066.0, "step": 20262 }, { "epoch": 2.5776618750795066, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.512353897094727, "learning_rate": 5e-05, "loss": 0.8655, "mean_token_accuracy": 0.8801795244216919, "num_tokens": 769415536.0, "step": 20263 }, { "epoch": 2.5777890853580967, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.081214427947998, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8786593079566956, "num_tokens": 769454016.0, "step": 20264 }, { "epoch": 2.5779162956366877, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.417116165161133, "learning_rate": 5e-05, "loss": 0.8586, "mean_token_accuracy": 0.8824798464775085, "num_tokens": 769493150.0, "step": 20265 }, { "epoch": 2.5780435059152778, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.285153865814209, "learning_rate": 5e-05, "loss": 0.7586, "mean_token_accuracy": 0.8781068921089172, "num_tokens": 769524247.0, "step": 20266 }, { "epoch": 2.5781707161938687, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.122292518615723, "learning_rate": 5e-05, "loss": 0.7927, "mean_token_accuracy": 0.891054630279541, "num_tokens": 769566476.0, "step": 20267 }, { "epoch": 2.578297926472459, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.539071559906006, "learning_rate": 5e-05, "loss": 0.7924, "mean_token_accuracy": 0.8759952187538147, "num_tokens": 769603474.0, "step": 20268 }, { "epoch": 2.5784251367510493, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.781593322753906, "learning_rate": 5e-05, "loss": 0.7903, "mean_token_accuracy": 0.8763943910598755, "num_tokens": 769640446.0, "step": 20269 }, { "epoch": 2.57855234702964, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.743186950683594, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8865538239479065, "num_tokens": 769679809.0, "step": 20270 }, { "epoch": 2.5786795573082304, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.508087158203125, "learning_rate": 5e-05, "loss": 0.7049, "mean_token_accuracy": 0.8972352743148804, "num_tokens": 769717221.0, "step": 20271 }, { "epoch": 2.578806767586821, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.875272750854492, "learning_rate": 5e-05, "loss": 0.7269, "mean_token_accuracy": 0.8955997228622437, "num_tokens": 769749350.0, "step": 20272 }, { "epoch": 2.5789339778654115, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.358353137969971, "learning_rate": 5e-05, "loss": 0.7525, "mean_token_accuracy": 0.8724244236946106, "num_tokens": 769786604.0, "step": 20273 }, { "epoch": 2.579061188144002, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.9272260665893555, "learning_rate": 5e-05, "loss": 0.7241, "mean_token_accuracy": 0.8933819532394409, "num_tokens": 769827709.0, "step": 20274 }, { "epoch": 2.5791883984225925, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.293246269226074, "learning_rate": 5e-05, "loss": 0.7263, "mean_token_accuracy": 0.8793434500694275, "num_tokens": 769866713.0, "step": 20275 }, { "epoch": 2.579315608701183, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.953781604766846, "learning_rate": 5e-05, "loss": 0.7782, "mean_token_accuracy": 0.8805099129676819, "num_tokens": 769907365.0, "step": 20276 }, { "epoch": 2.5794428189797736, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.229593276977539, "learning_rate": 5e-05, "loss": 0.7154, "mean_token_accuracy": 0.8797792196273804, "num_tokens": 769945573.0, "step": 20277 }, { "epoch": 2.579570029258364, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.030889511108398, "learning_rate": 5e-05, "loss": 0.7753, "mean_token_accuracy": 0.8769835233688354, "num_tokens": 769980936.0, "step": 20278 }, { "epoch": 2.5796972395369546, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.105566501617432, "learning_rate": 5e-05, "loss": 0.7501, "mean_token_accuracy": 0.8667446970939636, "num_tokens": 770022244.0, "step": 20279 }, { "epoch": 2.579824449815545, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.191295623779297, "learning_rate": 5e-05, "loss": 0.7528, "mean_token_accuracy": 0.8875670433044434, "num_tokens": 770059078.0, "step": 20280 }, { "epoch": 2.5799516600941357, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.924991130828857, "learning_rate": 5e-05, "loss": 0.7367, "mean_token_accuracy": 0.8662351369857788, "num_tokens": 770100816.0, "step": 20281 }, { "epoch": 2.580078870372726, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.437751770019531, "learning_rate": 5e-05, "loss": 0.7816, "mean_token_accuracy": 0.8849880695343018, "num_tokens": 770138912.0, "step": 20282 }, { "epoch": 2.5802060806513167, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.692960262298584, "learning_rate": 5e-05, "loss": 0.6714, "mean_token_accuracy": 0.8859642148017883, "num_tokens": 770178915.0, "step": 20283 }, { "epoch": 2.5803332909299073, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.774688720703125, "learning_rate": 5e-05, "loss": 0.8189, "mean_token_accuracy": 0.8888169527053833, "num_tokens": 770217707.0, "step": 20284 }, { "epoch": 2.580460501208498, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.466437339782715, "learning_rate": 5e-05, "loss": 0.6795, "mean_token_accuracy": 0.8802534341812134, "num_tokens": 770251659.0, "step": 20285 }, { "epoch": 2.5805877114870883, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.140702247619629, "learning_rate": 5e-05, "loss": 0.9229, "mean_token_accuracy": 0.874901294708252, "num_tokens": 770291612.0, "step": 20286 }, { "epoch": 2.5807149217656784, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000759124755859375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.472819805145264, "learning_rate": 5e-05, "loss": 0.7277, "mean_token_accuracy": 0.8713547587394714, "num_tokens": 770333822.0, "step": 20287 }, { "epoch": 2.5808421320442694, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.173523902893066, "learning_rate": 5e-05, "loss": 0.9587, "mean_token_accuracy": 0.876847505569458, "num_tokens": 770369850.0, "step": 20288 }, { "epoch": 2.5809693423228595, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.039361476898193, "learning_rate": 5e-05, "loss": 0.8345, "mean_token_accuracy": 0.8694713115692139, "num_tokens": 770408825.0, "step": 20289 }, { "epoch": 2.5810965526014504, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.446272850036621, "learning_rate": 5e-05, "loss": 0.8563, "mean_token_accuracy": 0.8928235173225403, "num_tokens": 770451320.0, "step": 20290 }, { "epoch": 2.5812237628800405, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.933155059814453, "learning_rate": 5e-05, "loss": 0.8857, "mean_token_accuracy": 0.8738148212432861, "num_tokens": 770488099.0, "step": 20291 }, { "epoch": 2.5813509731586315, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.334909439086914, "learning_rate": 5e-05, "loss": 0.7841, "mean_token_accuracy": 0.8710767030715942, "num_tokens": 770533689.0, "step": 20292 }, { "epoch": 2.5814781834372216, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.657012939453125, "learning_rate": 5e-05, "loss": 0.8999, "mean_token_accuracy": 0.8832367658615112, "num_tokens": 770570998.0, "step": 20293 }, { "epoch": 2.581605393715812, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.574616432189941, "learning_rate": 5e-05, "loss": 0.6629, "mean_token_accuracy": 0.8918765187263489, "num_tokens": 770610248.0, "step": 20294 }, { "epoch": 2.5817326039944026, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.209228515625, "learning_rate": 5e-05, "loss": 0.9758, "mean_token_accuracy": 0.8662142753601074, "num_tokens": 770650230.0, "step": 20295 }, { "epoch": 2.581859814272993, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.3976359367370605, "learning_rate": 5e-05, "loss": 0.6898, "mean_token_accuracy": 0.8873048424720764, "num_tokens": 770694740.0, "step": 20296 }, { "epoch": 2.5819870245515837, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.396927833557129, "learning_rate": 5e-05, "loss": 0.9467, "mean_token_accuracy": 0.8880447745323181, "num_tokens": 770734004.0, "step": 20297 }, { "epoch": 2.582114234830174, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 6.936852931976318, "learning_rate": 5e-05, "loss": 0.8251, "mean_token_accuracy": 0.8774909973144531, "num_tokens": 770770421.0, "step": 20298 }, { "epoch": 2.5822414451087647, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.845843315124512, "learning_rate": 5e-05, "loss": 0.9579, "mean_token_accuracy": 0.8744962215423584, "num_tokens": 770810790.0, "step": 20299 }, { "epoch": 2.5823686553873553, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 7.803511619567871, "learning_rate": 5e-05, "loss": 0.9146, "mean_token_accuracy": 0.8764215707778931, "num_tokens": 770848614.0, "step": 20300 }, { "epoch": 2.582495865665946, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.767510414123535, "learning_rate": 5e-05, "loss": 0.784, "mean_token_accuracy": 0.8998013138771057, "num_tokens": 770889888.0, "step": 20301 }, { "epoch": 2.5826230759445363, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.40983772277832, "learning_rate": 5e-05, "loss": 0.98, "mean_token_accuracy": 0.8575311899185181, "num_tokens": 770925918.0, "step": 20302 }, { "epoch": 2.582750286223127, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.166414260864258, "learning_rate": 5e-05, "loss": 0.7106, "mean_token_accuracy": 0.8988548517227173, "num_tokens": 770963130.0, "step": 20303 }, { "epoch": 2.5828774965017174, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.406381607055664, "learning_rate": 5e-05, "loss": 0.901, "mean_token_accuracy": 0.8712427616119385, "num_tokens": 770998308.0, "step": 20304 }, { "epoch": 2.583004706780308, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.368896484375, "learning_rate": 5e-05, "loss": 0.7987, "mean_token_accuracy": 0.872524082660675, "num_tokens": 771032532.0, "step": 20305 }, { "epoch": 2.5831319170588984, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.859764099121094, "learning_rate": 5e-05, "loss": 0.7934, "mean_token_accuracy": 0.8856974840164185, "num_tokens": 771070859.0, "step": 20306 }, { "epoch": 2.583259127337489, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.801435470581055, "learning_rate": 5e-05, "loss": 0.8224, "mean_token_accuracy": 0.8727551102638245, "num_tokens": 771111216.0, "step": 20307 }, { "epoch": 2.5833863376160795, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.3753533363342285, "learning_rate": 5e-05, "loss": 0.7553, "mean_token_accuracy": 0.8809860944747925, "num_tokens": 771146703.0, "step": 20308 }, { "epoch": 2.58351354789467, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.059418678283691, "learning_rate": 5e-05, "loss": 0.799, "mean_token_accuracy": 0.8819657564163208, "num_tokens": 771178778.0, "step": 20309 }, { "epoch": 2.5836407581732606, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.1515960693359375, "learning_rate": 5e-05, "loss": 0.7501, "mean_token_accuracy": 0.8710520267486572, "num_tokens": 771217272.0, "step": 20310 }, { "epoch": 2.583767968451851, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.224546432495117, "learning_rate": 5e-05, "loss": 0.8254, "mean_token_accuracy": 0.8736217617988586, "num_tokens": 771252085.0, "step": 20311 }, { "epoch": 2.583895178730441, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.975793838500977, "learning_rate": 5e-05, "loss": 0.6954, "mean_token_accuracy": 0.882610559463501, "num_tokens": 771285367.0, "step": 20312 }, { "epoch": 2.584022389009032, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.44714069366455, "learning_rate": 5e-05, "loss": 0.8048, "mean_token_accuracy": 0.887618899345398, "num_tokens": 771324388.0, "step": 20313 }, { "epoch": 2.5841495992876222, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.743429183959961, "learning_rate": 5e-05, "loss": 0.6774, "mean_token_accuracy": 0.8868381977081299, "num_tokens": 771361307.0, "step": 20314 }, { "epoch": 2.584276809566213, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.790614128112793, "learning_rate": 5e-05, "loss": 0.8935, "mean_token_accuracy": 0.8698726892471313, "num_tokens": 771400383.0, "step": 20315 }, { "epoch": 2.5844040198448033, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.503142356872559, "learning_rate": 5e-05, "loss": 0.6754, "mean_token_accuracy": 0.8891782760620117, "num_tokens": 771435065.0, "step": 20316 }, { "epoch": 2.584531230123394, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.188658714294434, "learning_rate": 5e-05, "loss": 0.975, "mean_token_accuracy": 0.8636777997016907, "num_tokens": 771480573.0, "step": 20317 }, { "epoch": 2.5846584404019843, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.454780101776123, "learning_rate": 5e-05, "loss": 0.71, "mean_token_accuracy": 0.8859903812408447, "num_tokens": 771517172.0, "step": 20318 }, { "epoch": 2.584785650680575, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.435848236083984, "learning_rate": 5e-05, "loss": 1.0038, "mean_token_accuracy": 0.8695907592773438, "num_tokens": 771554347.0, "step": 20319 }, { "epoch": 2.5849128609591654, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 6.884665489196777, "learning_rate": 5e-05, "loss": 0.7894, "mean_token_accuracy": 0.887020468711853, "num_tokens": 771591209.0, "step": 20320 }, { "epoch": 2.585040071237756, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.136890411376953, "learning_rate": 5e-05, "loss": 0.947, "mean_token_accuracy": 0.8904467821121216, "num_tokens": 771626604.0, "step": 20321 }, { "epoch": 2.5851672815163464, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.628556251525879, "learning_rate": 5e-05, "loss": 0.9176, "mean_token_accuracy": 0.8773988485336304, "num_tokens": 771662525.0, "step": 20322 }, { "epoch": 2.585294491794937, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.278017044067383, "learning_rate": 5e-05, "loss": 0.8857, "mean_token_accuracy": 0.8906532526016235, "num_tokens": 771701616.0, "step": 20323 }, { "epoch": 2.5854217020735275, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.113243103027344, "learning_rate": 5e-05, "loss": 0.948, "mean_token_accuracy": 0.8709566593170166, "num_tokens": 771736864.0, "step": 20324 }, { "epoch": 2.585548912352118, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.579530715942383, "learning_rate": 5e-05, "loss": 0.799, "mean_token_accuracy": 0.8875832557678223, "num_tokens": 771774900.0, "step": 20325 }, { "epoch": 2.5856761226307086, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.09847354888916, "learning_rate": 5e-05, "loss": 0.8772, "mean_token_accuracy": 0.8757297396659851, "num_tokens": 771812078.0, "step": 20326 }, { "epoch": 2.585803332909299, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.616704940795898, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.8786811232566833, "num_tokens": 771852247.0, "step": 20327 }, { "epoch": 2.5859305431878896, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.600106239318848, "learning_rate": 5e-05, "loss": 0.808, "mean_token_accuracy": 0.8763362765312195, "num_tokens": 771891264.0, "step": 20328 }, { "epoch": 2.58605775346648, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.032536506652832, "learning_rate": 5e-05, "loss": 0.8639, "mean_token_accuracy": 0.8683246374130249, "num_tokens": 771935356.0, "step": 20329 }, { "epoch": 2.5861849637450707, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.147216320037842, "learning_rate": 5e-05, "loss": 0.7962, "mean_token_accuracy": 0.8578174710273743, "num_tokens": 771967284.0, "step": 20330 }, { "epoch": 2.586312174023661, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.3018798828125, "learning_rate": 5e-05, "loss": 0.8218, "mean_token_accuracy": 0.8844146132469177, "num_tokens": 772013465.0, "step": 20331 }, { "epoch": 2.5864393843022517, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.98289155960083, "learning_rate": 5e-05, "loss": 0.7315, "mean_token_accuracy": 0.8757726550102234, "num_tokens": 772046978.0, "step": 20332 }, { "epoch": 2.5865665945808423, "ewc_loss": 0.5, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.381731033325195, "learning_rate": 5e-05, "loss": 0.8291, "mean_token_accuracy": 0.886620819568634, "num_tokens": 772081630.0, "step": 20333 }, { "epoch": 2.586693804859433, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.961135387420654, "learning_rate": 5e-05, "loss": 0.764, "mean_token_accuracy": 0.868990957736969, "num_tokens": 772122197.0, "step": 20334 }, { "epoch": 2.5868210151380233, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.420000076293945, "learning_rate": 5e-05, "loss": 0.8328, "mean_token_accuracy": 0.881055474281311, "num_tokens": 772162744.0, "step": 20335 }, { "epoch": 2.586948225416614, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.9578986167907715, "learning_rate": 5e-05, "loss": 0.7412, "mean_token_accuracy": 0.8744127750396729, "num_tokens": 772207520.0, "step": 20336 }, { "epoch": 2.587075435695204, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.472528457641602, "learning_rate": 5e-05, "loss": 0.8522, "mean_token_accuracy": 0.8803373575210571, "num_tokens": 772244869.0, "step": 20337 }, { "epoch": 2.587202645973795, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.9881205558776855, "learning_rate": 5e-05, "loss": 0.7474, "mean_token_accuracy": 0.8798532485961914, "num_tokens": 772279132.0, "step": 20338 }, { "epoch": 2.587329856252385, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.430716514587402, "learning_rate": 5e-05, "loss": 0.8209, "mean_token_accuracy": 0.8909547328948975, "num_tokens": 772323668.0, "step": 20339 }, { "epoch": 2.587457066530976, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.1294026374816895, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8783352375030518, "num_tokens": 772367834.0, "step": 20340 }, { "epoch": 2.587584276809566, "ewc_loss": 0.5, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.223179817199707, "learning_rate": 5e-05, "loss": 0.7895, "mean_token_accuracy": 0.8964223265647888, "num_tokens": 772398777.0, "step": 20341 }, { "epoch": 2.5877114870881566, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.349191665649414, "learning_rate": 5e-05, "loss": 0.7717, "mean_token_accuracy": 0.8786898851394653, "num_tokens": 772432330.0, "step": 20342 }, { "epoch": 2.587838697366747, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.915853023529053, "learning_rate": 5e-05, "loss": 0.8411, "mean_token_accuracy": 0.8734325766563416, "num_tokens": 772472412.0, "step": 20343 }, { "epoch": 2.5879659076453376, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.589252471923828, "learning_rate": 5e-05, "loss": 0.771, "mean_token_accuracy": 0.8811100721359253, "num_tokens": 772509896.0, "step": 20344 }, { "epoch": 2.588093117923928, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.5729265213012695, "learning_rate": 5e-05, "loss": 0.7394, "mean_token_accuracy": 0.8874891400337219, "num_tokens": 772549443.0, "step": 20345 }, { "epoch": 2.5882203282025187, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.808828353881836, "learning_rate": 5e-05, "loss": 0.8223, "mean_token_accuracy": 0.8688980340957642, "num_tokens": 772592086.0, "step": 20346 }, { "epoch": 2.588347538481109, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.294389247894287, "learning_rate": 5e-05, "loss": 0.7618, "mean_token_accuracy": 0.8684201240539551, "num_tokens": 772632805.0, "step": 20347 }, { "epoch": 2.5884747487596997, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.036447525024414, "learning_rate": 5e-05, "loss": 0.8177, "mean_token_accuracy": 0.8685420155525208, "num_tokens": 772673806.0, "step": 20348 }, { "epoch": 2.5886019590382903, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.99907922744751, "learning_rate": 5e-05, "loss": 0.642, "mean_token_accuracy": 0.9022777080535889, "num_tokens": 772709434.0, "step": 20349 }, { "epoch": 2.588729169316881, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.331178665161133, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8816449642181396, "num_tokens": 772748702.0, "step": 20350 }, { "epoch": 2.5888563795954713, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.704278469085693, "learning_rate": 5e-05, "loss": 0.739, "mean_token_accuracy": 0.8624038696289062, "num_tokens": 772792235.0, "step": 20351 }, { "epoch": 2.588983589874062, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.744891166687012, "learning_rate": 5e-05, "loss": 0.8511, "mean_token_accuracy": 0.8725266456604004, "num_tokens": 772827253.0, "step": 20352 }, { "epoch": 2.5891108001526524, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.35806131362915, "learning_rate": 5e-05, "loss": 0.6487, "mean_token_accuracy": 0.8921831250190735, "num_tokens": 772871951.0, "step": 20353 }, { "epoch": 2.589238010431243, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.294756889343262, "learning_rate": 5e-05, "loss": 0.9005, "mean_token_accuracy": 0.8884962201118469, "num_tokens": 772908585.0, "step": 20354 }, { "epoch": 2.5893652207098334, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.274398326873779, "learning_rate": 5e-05, "loss": 0.7248, "mean_token_accuracy": 0.8724477291107178, "num_tokens": 772946883.0, "step": 20355 }, { "epoch": 2.589492430988424, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.560541152954102, "learning_rate": 5e-05, "loss": 0.9692, "mean_token_accuracy": 0.8846591114997864, "num_tokens": 772985143.0, "step": 20356 }, { "epoch": 2.5896196412670145, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 6.968757152557373, "learning_rate": 5e-05, "loss": 0.8149, "mean_token_accuracy": 0.8831307888031006, "num_tokens": 773027004.0, "step": 20357 }, { "epoch": 2.589746851545605, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.882328987121582, "learning_rate": 5e-05, "loss": 0.9852, "mean_token_accuracy": 0.8675624132156372, "num_tokens": 773067807.0, "step": 20358 }, { "epoch": 2.5898740618241956, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.074652671813965, "learning_rate": 5e-05, "loss": 0.9325, "mean_token_accuracy": 0.8859274387359619, "num_tokens": 773103160.0, "step": 20359 }, { "epoch": 2.5900012721027856, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.38129186630249, "learning_rate": 5e-05, "loss": 0.7748, "mean_token_accuracy": 0.8868778944015503, "num_tokens": 773143036.0, "step": 20360 }, { "epoch": 2.5901284823813766, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.053727149963379, "learning_rate": 5e-05, "loss": 0.9492, "mean_token_accuracy": 0.8853774070739746, "num_tokens": 773180930.0, "step": 20361 }, { "epoch": 2.5902556926599667, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.434122085571289, "learning_rate": 5e-05, "loss": 0.7779, "mean_token_accuracy": 0.8617185354232788, "num_tokens": 773220056.0, "step": 20362 }, { "epoch": 2.5903829029385577, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.64206600189209, "learning_rate": 5e-05, "loss": 0.9357, "mean_token_accuracy": 0.8972818851470947, "num_tokens": 773255890.0, "step": 20363 }, { "epoch": 2.5905101132171477, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 6.746841907501221, "learning_rate": 5e-05, "loss": 0.7855, "mean_token_accuracy": 0.8843415975570679, "num_tokens": 773298727.0, "step": 20364 }, { "epoch": 2.5906373234957387, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.2635498046875, "learning_rate": 5e-05, "loss": 1.031, "mean_token_accuracy": 0.8625240921974182, "num_tokens": 773338137.0, "step": 20365 }, { "epoch": 2.590764533774329, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.7372236251831055, "learning_rate": 5e-05, "loss": 0.9127, "mean_token_accuracy": 0.8857448697090149, "num_tokens": 773381532.0, "step": 20366 }, { "epoch": 2.5908917440529193, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.099822998046875, "learning_rate": 5e-05, "loss": 0.9036, "mean_token_accuracy": 0.8678588271141052, "num_tokens": 773416354.0, "step": 20367 }, { "epoch": 2.59101895433151, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.39631175994873, "learning_rate": 5e-05, "loss": 0.9498, "mean_token_accuracy": 0.8801742792129517, "num_tokens": 773456130.0, "step": 20368 }, { "epoch": 2.5911461646101004, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.335385322570801, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8863525390625, "num_tokens": 773493812.0, "step": 20369 }, { "epoch": 2.591273374888691, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.421640396118164, "learning_rate": 5e-05, "loss": 0.8885, "mean_token_accuracy": 0.8823996782302856, "num_tokens": 773533000.0, "step": 20370 }, { "epoch": 2.5914005851672814, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.421359539031982, "learning_rate": 5e-05, "loss": 0.8018, "mean_token_accuracy": 0.8766596913337708, "num_tokens": 773572355.0, "step": 20371 }, { "epoch": 2.591527795445872, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.9078497886657715, "learning_rate": 5e-05, "loss": 0.7952, "mean_token_accuracy": 0.8909270763397217, "num_tokens": 773609154.0, "step": 20372 }, { "epoch": 2.5916550057244625, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.791224479675293, "learning_rate": 5e-05, "loss": 0.7936, "mean_token_accuracy": 0.8876574039459229, "num_tokens": 773654430.0, "step": 20373 }, { "epoch": 2.591782216003053, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.4394097328186035, "learning_rate": 5e-05, "loss": 0.7443, "mean_token_accuracy": 0.8847341537475586, "num_tokens": 773693710.0, "step": 20374 }, { "epoch": 2.5919094262816436, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.996572494506836, "learning_rate": 5e-05, "loss": 0.872, "mean_token_accuracy": 0.8564923405647278, "num_tokens": 773728858.0, "step": 20375 }, { "epoch": 2.592036636560234, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.2340407371521, "learning_rate": 5e-05, "loss": 0.747, "mean_token_accuracy": 0.8731145858764648, "num_tokens": 773761080.0, "step": 20376 }, { "epoch": 2.5921638468388246, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.124360084533691, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8924934267997742, "num_tokens": 773799084.0, "step": 20377 }, { "epoch": 2.592291057117415, "ewc_loss": 0.375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.01313591003418, "learning_rate": 5e-05, "loss": 0.6934, "mean_token_accuracy": 0.88370281457901, "num_tokens": 773840378.0, "step": 20378 }, { "epoch": 2.5924182673960057, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.42454719543457, "learning_rate": 5e-05, "loss": 0.8184, "mean_token_accuracy": 0.8842368721961975, "num_tokens": 773878627.0, "step": 20379 }, { "epoch": 2.592545477674596, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.703959941864014, "learning_rate": 5e-05, "loss": 0.7299, "mean_token_accuracy": 0.8707352876663208, "num_tokens": 773921492.0, "step": 20380 }, { "epoch": 2.5926726879531867, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.861406326293945, "learning_rate": 5e-05, "loss": 0.9136, "mean_token_accuracy": 0.8663222789764404, "num_tokens": 773959603.0, "step": 20381 }, { "epoch": 2.5927998982317773, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.428178787231445, "learning_rate": 5e-05, "loss": 0.7302, "mean_token_accuracy": 0.8698359727859497, "num_tokens": 773996550.0, "step": 20382 }, { "epoch": 2.592927108510368, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.290396690368652, "learning_rate": 5e-05, "loss": 0.9409, "mean_token_accuracy": 0.8820914030075073, "num_tokens": 774033815.0, "step": 20383 }, { "epoch": 2.5930543187889583, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.443141937255859, "learning_rate": 5e-05, "loss": 0.7296, "mean_token_accuracy": 0.8798419237136841, "num_tokens": 774068930.0, "step": 20384 }, { "epoch": 2.5931815290675484, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.530978202819824, "learning_rate": 5e-05, "loss": 0.9862, "mean_token_accuracy": 0.8788915872573853, "num_tokens": 774112014.0, "step": 20385 }, { "epoch": 2.5933087393461394, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 6.972806930541992, "learning_rate": 5e-05, "loss": 0.8479, "mean_token_accuracy": 0.8815183639526367, "num_tokens": 774149676.0, "step": 20386 }, { "epoch": 2.5934359496247295, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.220510482788086, "learning_rate": 5e-05, "loss": 0.9303, "mean_token_accuracy": 0.8997989296913147, "num_tokens": 774190160.0, "step": 20387 }, { "epoch": 2.5935631599033204, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.733532905578613, "learning_rate": 5e-05, "loss": 0.9415, "mean_token_accuracy": 0.8773834705352783, "num_tokens": 774235133.0, "step": 20388 }, { "epoch": 2.5936903701819105, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.312267303466797, "learning_rate": 5e-05, "loss": 0.8966, "mean_token_accuracy": 0.8850433230400085, "num_tokens": 774277962.0, "step": 20389 }, { "epoch": 2.5938175804605015, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.177868843078613, "learning_rate": 5e-05, "loss": 0.9122, "mean_token_accuracy": 0.8842167258262634, "num_tokens": 774316946.0, "step": 20390 }, { "epoch": 2.5939447907390916, "ewc_loss": 0.5, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.626688480377197, "learning_rate": 5e-05, "loss": 0.8281, "mean_token_accuracy": 0.8836931586265564, "num_tokens": 774356376.0, "step": 20391 }, { "epoch": 2.594072001017682, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.075484275817871, "learning_rate": 5e-05, "loss": 0.8337, "mean_token_accuracy": 0.8900319933891296, "num_tokens": 774392543.0, "step": 20392 }, { "epoch": 2.5941992112962726, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.673367023468018, "learning_rate": 5e-05, "loss": 0.8296, "mean_token_accuracy": 0.8732953667640686, "num_tokens": 774432114.0, "step": 20393 }, { "epoch": 2.594326421574863, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.555795192718506, "learning_rate": 5e-05, "loss": 0.8144, "mean_token_accuracy": 0.8703304529190063, "num_tokens": 774468758.0, "step": 20394 }, { "epoch": 2.5944536318534537, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.044154167175293, "learning_rate": 5e-05, "loss": 0.8647, "mean_token_accuracy": 0.8684017658233643, "num_tokens": 774503136.0, "step": 20395 }, { "epoch": 2.594580842132044, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.165238857269287, "learning_rate": 5e-05, "loss": 0.7085, "mean_token_accuracy": 0.8876725435256958, "num_tokens": 774545162.0, "step": 20396 }, { "epoch": 2.5947080524106347, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.246552467346191, "learning_rate": 5e-05, "loss": 0.8278, "mean_token_accuracy": 0.8831942677497864, "num_tokens": 774581831.0, "step": 20397 }, { "epoch": 2.5948352626892253, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.00551700592041, "learning_rate": 5e-05, "loss": 0.7038, "mean_token_accuracy": 0.8852737545967102, "num_tokens": 774619258.0, "step": 20398 }, { "epoch": 2.594962472967816, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.360522270202637, "learning_rate": 5e-05, "loss": 0.8678, "mean_token_accuracy": 0.8672679662704468, "num_tokens": 774650244.0, "step": 20399 }, { "epoch": 2.5950896832464063, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.937816143035889, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.867178201675415, "num_tokens": 774690060.0, "step": 20400 }, { "epoch": 2.595216893524997, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.461309432983398, "learning_rate": 5e-05, "loss": 0.862, "mean_token_accuracy": 0.8777304291725159, "num_tokens": 774733925.0, "step": 20401 }, { "epoch": 2.5953441038035874, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.891162872314453, "learning_rate": 5e-05, "loss": 0.7006, "mean_token_accuracy": 0.8846155405044556, "num_tokens": 774769344.0, "step": 20402 }, { "epoch": 2.595471314082178, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.527116775512695, "learning_rate": 5e-05, "loss": 0.819, "mean_token_accuracy": 0.8945566415786743, "num_tokens": 774809548.0, "step": 20403 }, { "epoch": 2.5955985243607684, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.95142936706543, "learning_rate": 5e-05, "loss": 0.7558, "mean_token_accuracy": 0.8762615323066711, "num_tokens": 774848042.0, "step": 20404 }, { "epoch": 2.595725734639359, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.480341911315918, "learning_rate": 5e-05, "loss": 0.8893, "mean_token_accuracy": 0.8738405108451843, "num_tokens": 774884226.0, "step": 20405 }, { "epoch": 2.5958529449179495, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.126865386962891, "learning_rate": 5e-05, "loss": 0.7485, "mean_token_accuracy": 0.877137303352356, "num_tokens": 774913697.0, "step": 20406 }, { "epoch": 2.59598015519654, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.246782302856445, "learning_rate": 5e-05, "loss": 0.8065, "mean_token_accuracy": 0.8950057029724121, "num_tokens": 774948369.0, "step": 20407 }, { "epoch": 2.5961073654751305, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.372952461242676, "learning_rate": 5e-05, "loss": 0.7794, "mean_token_accuracy": 0.8838203549385071, "num_tokens": 774992260.0, "step": 20408 }, { "epoch": 2.596234575753721, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.905468940734863, "learning_rate": 5e-05, "loss": 0.913, "mean_token_accuracy": 0.8391715884208679, "num_tokens": 775023613.0, "step": 20409 }, { "epoch": 2.596361786032311, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.597890377044678, "learning_rate": 5e-05, "loss": 0.8083, "mean_token_accuracy": 0.8756978511810303, "num_tokens": 775063874.0, "step": 20410 }, { "epoch": 2.596488996310902, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.572983741760254, "learning_rate": 5e-05, "loss": 0.7824, "mean_token_accuracy": 0.8739820718765259, "num_tokens": 775104447.0, "step": 20411 }, { "epoch": 2.596616206589492, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.794839382171631, "learning_rate": 5e-05, "loss": 0.7606, "mean_token_accuracy": 0.8878257274627686, "num_tokens": 775143472.0, "step": 20412 }, { "epoch": 2.596743416868083, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.30863094329834, "learning_rate": 5e-05, "loss": 0.7196, "mean_token_accuracy": 0.8865251541137695, "num_tokens": 775176274.0, "step": 20413 }, { "epoch": 2.5968706271466733, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.954189300537109, "learning_rate": 5e-05, "loss": 0.8065, "mean_token_accuracy": 0.8749808073043823, "num_tokens": 775216485.0, "step": 20414 }, { "epoch": 2.596997837425264, "ewc_loss": 0.375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.116217136383057, "learning_rate": 5e-05, "loss": 0.6812, "mean_token_accuracy": 0.8915080428123474, "num_tokens": 775256828.0, "step": 20415 }, { "epoch": 2.5971250477038543, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.148848533630371, "learning_rate": 5e-05, "loss": 0.813, "mean_token_accuracy": 0.8712680339813232, "num_tokens": 775296577.0, "step": 20416 }, { "epoch": 2.597252257982445, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.873548984527588, "learning_rate": 5e-05, "loss": 0.7088, "mean_token_accuracy": 0.8787290453910828, "num_tokens": 775329567.0, "step": 20417 }, { "epoch": 2.5973794682610354, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.462782859802246, "learning_rate": 5e-05, "loss": 0.7997, "mean_token_accuracy": 0.8841594457626343, "num_tokens": 775363729.0, "step": 20418 }, { "epoch": 2.597506678539626, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.5751566886901855, "learning_rate": 5e-05, "loss": 0.6843, "mean_token_accuracy": 0.8794764280319214, "num_tokens": 775397547.0, "step": 20419 }, { "epoch": 2.5976338888182164, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.877774238586426, "learning_rate": 5e-05, "loss": 0.8912, "mean_token_accuracy": 0.8709379434585571, "num_tokens": 775441033.0, "step": 20420 }, { "epoch": 2.597761099096807, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.346236705780029, "learning_rate": 5e-05, "loss": 0.6785, "mean_token_accuracy": 0.8847458362579346, "num_tokens": 775477516.0, "step": 20421 }, { "epoch": 2.5978883093753975, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.27563762664795, "learning_rate": 5e-05, "loss": 0.9253, "mean_token_accuracy": 0.885522723197937, "num_tokens": 775515573.0, "step": 20422 }, { "epoch": 2.598015519653988, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.497676372528076, "learning_rate": 5e-05, "loss": 0.7288, "mean_token_accuracy": 0.8814201354980469, "num_tokens": 775552970.0, "step": 20423 }, { "epoch": 2.5981427299325786, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.214766502380371, "learning_rate": 5e-05, "loss": 0.967, "mean_token_accuracy": 0.8774081468582153, "num_tokens": 775590180.0, "step": 20424 }, { "epoch": 2.598269940211169, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.256806373596191, "learning_rate": 5e-05, "loss": 0.8501, "mean_token_accuracy": 0.8834038972854614, "num_tokens": 775627346.0, "step": 20425 }, { "epoch": 2.5983971504897596, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.30900764465332, "learning_rate": 5e-05, "loss": 0.8938, "mean_token_accuracy": 0.8803219795227051, "num_tokens": 775667173.0, "step": 20426 }, { "epoch": 2.59852436076835, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.161499977111816, "learning_rate": 5e-05, "loss": 0.9396, "mean_token_accuracy": 0.8706258535385132, "num_tokens": 775702710.0, "step": 20427 }, { "epoch": 2.5986515710469407, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.0804009437561035, "learning_rate": 5e-05, "loss": 0.7555, "mean_token_accuracy": 0.8805112838745117, "num_tokens": 775743934.0, "step": 20428 }, { "epoch": 2.598778781325531, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.889280319213867, "learning_rate": 5e-05, "loss": 0.9141, "mean_token_accuracy": 0.8831654787063599, "num_tokens": 775783184.0, "step": 20429 }, { "epoch": 2.5989059916041217, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.544314384460449, "learning_rate": 5e-05, "loss": 0.7286, "mean_token_accuracy": 0.8716180324554443, "num_tokens": 775820019.0, "step": 20430 }, { "epoch": 2.5990332018827123, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.114039421081543, "learning_rate": 5e-05, "loss": 0.9391, "mean_token_accuracy": 0.8797322511672974, "num_tokens": 775858740.0, "step": 20431 }, { "epoch": 2.599160412161303, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.800434589385986, "learning_rate": 5e-05, "loss": 0.7685, "mean_token_accuracy": 0.8790127635002136, "num_tokens": 775900615.0, "step": 20432 }, { "epoch": 2.5992876224398933, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.709025382995605, "learning_rate": 5e-05, "loss": 0.8873, "mean_token_accuracy": 0.8891478776931763, "num_tokens": 775938953.0, "step": 20433 }, { "epoch": 2.599414832718484, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.458552360534668, "learning_rate": 5e-05, "loss": 0.8601, "mean_token_accuracy": 0.8772332668304443, "num_tokens": 775973203.0, "step": 20434 }, { "epoch": 2.599542042997074, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000766754150390625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.924931049346924, "learning_rate": 5e-05, "loss": 0.8841, "mean_token_accuracy": 0.8676760196685791, "num_tokens": 776014984.0, "step": 20435 }, { "epoch": 2.599669253275665, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.915663242340088, "learning_rate": 5e-05, "loss": 0.9, "mean_token_accuracy": 0.8641719818115234, "num_tokens": 776059883.0, "step": 20436 }, { "epoch": 2.599796463554255, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.433689117431641, "learning_rate": 5e-05, "loss": 0.8195, "mean_token_accuracy": 0.868466854095459, "num_tokens": 776101025.0, "step": 20437 }, { "epoch": 2.599923673832846, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.974681377410889, "learning_rate": 5e-05, "loss": 0.8016, "mean_token_accuracy": 0.8843477964401245, "num_tokens": 776134356.0, "step": 20438 }, { "epoch": 2.600050884111436, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.439620494842529, "learning_rate": 5e-05, "loss": 0.7458, "mean_token_accuracy": 0.8839194774627686, "num_tokens": 776176158.0, "step": 20439 }, { "epoch": 2.6001780943900266, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.714757442474365, "learning_rate": 5e-05, "loss": 0.7774, "mean_token_accuracy": 0.8808907270431519, "num_tokens": 776209586.0, "step": 20440 }, { "epoch": 2.600305304668617, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.639890193939209, "learning_rate": 5e-05, "loss": 0.7884, "mean_token_accuracy": 0.8725863695144653, "num_tokens": 776240707.0, "step": 20441 }, { "epoch": 2.6004325149472076, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.424994945526123, "learning_rate": 5e-05, "loss": 0.7476, "mean_token_accuracy": 0.8788918256759644, "num_tokens": 776280448.0, "step": 20442 }, { "epoch": 2.600559725225798, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.864058017730713, "learning_rate": 5e-05, "loss": 0.7495, "mean_token_accuracy": 0.8916236162185669, "num_tokens": 776320112.0, "step": 20443 }, { "epoch": 2.6006869355043887, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.1627278327941895, "learning_rate": 5e-05, "loss": 0.7298, "mean_token_accuracy": 0.8780431151390076, "num_tokens": 776362720.0, "step": 20444 }, { "epoch": 2.600814145782979, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.091194152832031, "learning_rate": 5e-05, "loss": 0.7963, "mean_token_accuracy": 0.8777917623519897, "num_tokens": 776403006.0, "step": 20445 }, { "epoch": 2.6009413560615697, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.896451950073242, "learning_rate": 5e-05, "loss": 0.7088, "mean_token_accuracy": 0.8757350444793701, "num_tokens": 776436039.0, "step": 20446 }, { "epoch": 2.6010685663401603, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.426639556884766, "learning_rate": 5e-05, "loss": 0.8129, "mean_token_accuracy": 0.8784213662147522, "num_tokens": 776476337.0, "step": 20447 }, { "epoch": 2.601195776618751, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.538690090179443, "learning_rate": 5e-05, "loss": 0.6562, "mean_token_accuracy": 0.8893502950668335, "num_tokens": 776512552.0, "step": 20448 }, { "epoch": 2.6013229868973413, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.975041389465332, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8768550157546997, "num_tokens": 776549727.0, "step": 20449 }, { "epoch": 2.601450197175932, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.160067558288574, "learning_rate": 5e-05, "loss": 0.7062, "mean_token_accuracy": 0.8756221532821655, "num_tokens": 776586701.0, "step": 20450 }, { "epoch": 2.6015774074545224, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.662225723266602, "learning_rate": 5e-05, "loss": 0.9955, "mean_token_accuracy": 0.8753588795661926, "num_tokens": 776627386.0, "step": 20451 }, { "epoch": 2.601704617733113, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.226710319519043, "learning_rate": 5e-05, "loss": 0.7053, "mean_token_accuracy": 0.8859947323799133, "num_tokens": 776668389.0, "step": 20452 }, { "epoch": 2.6018318280117034, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.036334991455078, "learning_rate": 5e-05, "loss": 1.0507, "mean_token_accuracy": 0.8768742084503174, "num_tokens": 776708106.0, "step": 20453 }, { "epoch": 2.601959038290294, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 7.321074485778809, "learning_rate": 5e-05, "loss": 0.9247, "mean_token_accuracy": 0.8801366090774536, "num_tokens": 776744577.0, "step": 20454 }, { "epoch": 2.6020862485688845, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.27193832397461, "learning_rate": 5e-05, "loss": 1.028, "mean_token_accuracy": 0.868671178817749, "num_tokens": 776784350.0, "step": 20455 }, { "epoch": 2.602213458847475, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 8.491156578063965, "learning_rate": 5e-05, "loss": 1.0135, "mean_token_accuracy": 0.8866458535194397, "num_tokens": 776823728.0, "step": 20456 }, { "epoch": 2.6023406691260655, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.572289943695068, "learning_rate": 5e-05, "loss": 0.8459, "mean_token_accuracy": 0.8788532614707947, "num_tokens": 776860097.0, "step": 20457 }, { "epoch": 2.6024678794046556, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.323638916015625, "learning_rate": 5e-05, "loss": 0.9862, "mean_token_accuracy": 0.8845810294151306, "num_tokens": 776895876.0, "step": 20458 }, { "epoch": 2.6025950896832466, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.692621231079102, "learning_rate": 5e-05, "loss": 0.7869, "mean_token_accuracy": 0.8725692629814148, "num_tokens": 776930003.0, "step": 20459 }, { "epoch": 2.6027222999618367, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.65079116821289, "learning_rate": 5e-05, "loss": 0.9452, "mean_token_accuracy": 0.8976122140884399, "num_tokens": 776969699.0, "step": 20460 }, { "epoch": 2.6028495102404277, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.195531368255615, "learning_rate": 5e-05, "loss": 0.8821, "mean_token_accuracy": 0.88245689868927, "num_tokens": 777006517.0, "step": 20461 }, { "epoch": 2.6029767205190177, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.02975845336914, "learning_rate": 5e-05, "loss": 1.027, "mean_token_accuracy": 0.8626669645309448, "num_tokens": 777044160.0, "step": 20462 }, { "epoch": 2.6031039307976087, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.01646900177002, "learning_rate": 5e-05, "loss": 0.9693, "mean_token_accuracy": 0.8811393976211548, "num_tokens": 777079890.0, "step": 20463 }, { "epoch": 2.603231141076199, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.926496505737305, "learning_rate": 5e-05, "loss": 0.8234, "mean_token_accuracy": 0.891411304473877, "num_tokens": 777120928.0, "step": 20464 }, { "epoch": 2.6033583513547893, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.380833625793457, "learning_rate": 5e-05, "loss": 0.9902, "mean_token_accuracy": 0.8679838180541992, "num_tokens": 777166363.0, "step": 20465 }, { "epoch": 2.60348556163338, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.343813896179199, "learning_rate": 5e-05, "loss": 0.8037, "mean_token_accuracy": 0.8803437948226929, "num_tokens": 777201146.0, "step": 20466 }, { "epoch": 2.6036127719119704, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.411500930786133, "learning_rate": 5e-05, "loss": 0.9634, "mean_token_accuracy": 0.8598924875259399, "num_tokens": 777239884.0, "step": 20467 }, { "epoch": 2.603739982190561, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.270231246948242, "learning_rate": 5e-05, "loss": 0.852, "mean_token_accuracy": 0.8659303784370422, "num_tokens": 777279898.0, "step": 20468 }, { "epoch": 2.6038671924691514, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.271739959716797, "learning_rate": 5e-05, "loss": 0.8933, "mean_token_accuracy": 0.8752073049545288, "num_tokens": 777320614.0, "step": 20469 }, { "epoch": 2.603994402747742, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.286919593811035, "learning_rate": 5e-05, "loss": 0.8086, "mean_token_accuracy": 0.8766846060752869, "num_tokens": 777364782.0, "step": 20470 }, { "epoch": 2.6041216130263325, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.200942993164062, "learning_rate": 5e-05, "loss": 0.8601, "mean_token_accuracy": 0.8726223707199097, "num_tokens": 777398276.0, "step": 20471 }, { "epoch": 2.604248823304923, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.225517749786377, "learning_rate": 5e-05, "loss": 0.776, "mean_token_accuracy": 0.8840475082397461, "num_tokens": 777434510.0, "step": 20472 }, { "epoch": 2.6043760335835135, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.205835342407227, "learning_rate": 5e-05, "loss": 0.8222, "mean_token_accuracy": 0.8813932538032532, "num_tokens": 777469266.0, "step": 20473 }, { "epoch": 2.604503243862104, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.149461269378662, "learning_rate": 5e-05, "loss": 0.7168, "mean_token_accuracy": 0.8923362493515015, "num_tokens": 777500770.0, "step": 20474 }, { "epoch": 2.6046304541406946, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.200248718261719, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.8784171342849731, "num_tokens": 777537977.0, "step": 20475 }, { "epoch": 2.604757664419285, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.095870018005371, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8812848925590515, "num_tokens": 777574065.0, "step": 20476 }, { "epoch": 2.6048848746978757, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.216742515563965, "learning_rate": 5e-05, "loss": 0.8044, "mean_token_accuracy": 0.887668251991272, "num_tokens": 777613366.0, "step": 20477 }, { "epoch": 2.605012084976466, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.010103225708008, "learning_rate": 5e-05, "loss": 0.7405, "mean_token_accuracy": 0.8774254322052002, "num_tokens": 777650954.0, "step": 20478 }, { "epoch": 2.6051392952550567, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.321100234985352, "learning_rate": 5e-05, "loss": 0.8648, "mean_token_accuracy": 0.8676071166992188, "num_tokens": 777694370.0, "step": 20479 }, { "epoch": 2.6052665055336472, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.894089698791504, "learning_rate": 5e-05, "loss": 0.7396, "mean_token_accuracy": 0.8760719299316406, "num_tokens": 777734039.0, "step": 20480 }, { "epoch": 2.6053937158122378, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.416421890258789, "learning_rate": 5e-05, "loss": 0.8293, "mean_token_accuracy": 0.8811236619949341, "num_tokens": 777776237.0, "step": 20481 }, { "epoch": 2.6055209260908283, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.864205837249756, "learning_rate": 5e-05, "loss": 0.7159, "mean_token_accuracy": 0.8836170434951782, "num_tokens": 777813767.0, "step": 20482 }, { "epoch": 2.6056481363694184, "ewc_loss": 0.5, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.449835777282715, "learning_rate": 5e-05, "loss": 0.8204, "mean_token_accuracy": 0.8816186785697937, "num_tokens": 777848642.0, "step": 20483 }, { "epoch": 2.6057753466480094, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.896480083465576, "learning_rate": 5e-05, "loss": 0.7095, "mean_token_accuracy": 0.8827091455459595, "num_tokens": 777885748.0, "step": 20484 }, { "epoch": 2.6059025569265994, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.423135757446289, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.8685216307640076, "num_tokens": 777922366.0, "step": 20485 }, { "epoch": 2.6060297672051904, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.997256278991699, "learning_rate": 5e-05, "loss": 0.7565, "mean_token_accuracy": 0.8787007927894592, "num_tokens": 777965130.0, "step": 20486 }, { "epoch": 2.6061569774837805, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.253746032714844, "learning_rate": 5e-05, "loss": 0.7853, "mean_token_accuracy": 0.8918190598487854, "num_tokens": 777999425.0, "step": 20487 }, { "epoch": 2.6062841877623715, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.221487998962402, "learning_rate": 5e-05, "loss": 0.7966, "mean_token_accuracy": 0.8682364225387573, "num_tokens": 778042378.0, "step": 20488 }, { "epoch": 2.6064113980409616, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.917967796325684, "learning_rate": 5e-05, "loss": 0.7967, "mean_token_accuracy": 0.8791643381118774, "num_tokens": 778083858.0, "step": 20489 }, { "epoch": 2.606538608319552, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.496596336364746, "learning_rate": 5e-05, "loss": 0.7574, "mean_token_accuracy": 0.8850080370903015, "num_tokens": 778122103.0, "step": 20490 }, { "epoch": 2.6066658185981426, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.565710544586182, "learning_rate": 5e-05, "loss": 0.7579, "mean_token_accuracy": 0.8799845576286316, "num_tokens": 778158207.0, "step": 20491 }, { "epoch": 2.606793028876733, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.738544940948486, "learning_rate": 5e-05, "loss": 0.7588, "mean_token_accuracy": 0.8864575624465942, "num_tokens": 778191542.0, "step": 20492 }, { "epoch": 2.6069202391553237, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.288657188415527, "learning_rate": 5e-05, "loss": 0.7308, "mean_token_accuracy": 0.8812883496284485, "num_tokens": 778234929.0, "step": 20493 }, { "epoch": 2.607047449433914, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.913646697998047, "learning_rate": 5e-05, "loss": 0.7158, "mean_token_accuracy": 0.8995003700256348, "num_tokens": 778273225.0, "step": 20494 }, { "epoch": 2.6071746597125047, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.110484600067139, "learning_rate": 5e-05, "loss": 0.7177, "mean_token_accuracy": 0.875971794128418, "num_tokens": 778310998.0, "step": 20495 }, { "epoch": 2.6073018699910953, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.074913024902344, "learning_rate": 5e-05, "loss": 0.8096, "mean_token_accuracy": 0.8715550303459167, "num_tokens": 778348379.0, "step": 20496 }, { "epoch": 2.607429080269686, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.917660713195801, "learning_rate": 5e-05, "loss": 0.7202, "mean_token_accuracy": 0.8743516802787781, "num_tokens": 778390742.0, "step": 20497 }, { "epoch": 2.6075562905482763, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.302950859069824, "learning_rate": 5e-05, "loss": 0.7794, "mean_token_accuracy": 0.8854053616523743, "num_tokens": 778429239.0, "step": 20498 }, { "epoch": 2.607683500826867, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.6782989501953125, "learning_rate": 5e-05, "loss": 0.6614, "mean_token_accuracy": 0.889839768409729, "num_tokens": 778467606.0, "step": 20499 }, { "epoch": 2.6078107111054574, "ewc_loss": 0.5, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.658287048339844, "learning_rate": 5e-05, "loss": 0.8625, "mean_token_accuracy": 0.8710681200027466, "num_tokens": 778500790.0, "step": 20500 }, { "epoch": 2.607937921384048, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.408419609069824, "learning_rate": 5e-05, "loss": 0.679, "mean_token_accuracy": 0.8817446231842041, "num_tokens": 778536412.0, "step": 20501 }, { "epoch": 2.6080651316626384, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 9.077678680419922, "learning_rate": 5e-05, "loss": 0.9202, "mean_token_accuracy": 0.8727754354476929, "num_tokens": 778579812.0, "step": 20502 }, { "epoch": 2.608192341941229, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.334023952484131, "learning_rate": 5e-05, "loss": 0.6812, "mean_token_accuracy": 0.8864579200744629, "num_tokens": 778626135.0, "step": 20503 }, { "epoch": 2.6083195522198195, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.30061149597168, "learning_rate": 5e-05, "loss": 0.9487, "mean_token_accuracy": 0.884856104850769, "num_tokens": 778666163.0, "step": 20504 }, { "epoch": 2.60844676249841, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 6.805417060852051, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8814996480941772, "num_tokens": 778703168.0, "step": 20505 }, { "epoch": 2.6085739727770005, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.810548782348633, "learning_rate": 5e-05, "loss": 0.9768, "mean_token_accuracy": 0.8706259727478027, "num_tokens": 778741696.0, "step": 20506 }, { "epoch": 2.608701183055591, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 7.712071418762207, "learning_rate": 5e-05, "loss": 0.9034, "mean_token_accuracy": 0.8782349824905396, "num_tokens": 778777450.0, "step": 20507 }, { "epoch": 2.608828393334181, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.588058948516846, "learning_rate": 5e-05, "loss": 0.8215, "mean_token_accuracy": 0.8798501491546631, "num_tokens": 778811448.0, "step": 20508 }, { "epoch": 2.608955603612772, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.562201499938965, "learning_rate": 5e-05, "loss": 0.932, "mean_token_accuracy": 0.8792693614959717, "num_tokens": 778849333.0, "step": 20509 }, { "epoch": 2.609082813891362, "ewc_loss": 0.375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.620314121246338, "learning_rate": 5e-05, "loss": 0.6997, "mean_token_accuracy": 0.8864724636077881, "num_tokens": 778885264.0, "step": 20510 }, { "epoch": 2.609210024169953, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.148929595947266, "learning_rate": 5e-05, "loss": 0.9491, "mean_token_accuracy": 0.8793156743049622, "num_tokens": 778918160.0, "step": 20511 }, { "epoch": 2.6093372344485433, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.476266384124756, "learning_rate": 5e-05, "loss": 0.7231, "mean_token_accuracy": 0.8842382431030273, "num_tokens": 778955938.0, "step": 20512 }, { "epoch": 2.609464444727134, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.170248985290527, "learning_rate": 5e-05, "loss": 0.9805, "mean_token_accuracy": 0.8746375441551208, "num_tokens": 778994412.0, "step": 20513 }, { "epoch": 2.6095916550057243, "ewc_loss": 0.5, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.0874834060668945, "learning_rate": 5e-05, "loss": 0.8778, "mean_token_accuracy": 0.8691343665122986, "num_tokens": 779030818.0, "step": 20514 }, { "epoch": 2.609718865284315, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.484811782836914, "learning_rate": 5e-05, "loss": 0.8965, "mean_token_accuracy": 0.8882121443748474, "num_tokens": 779070217.0, "step": 20515 }, { "epoch": 2.6098460755629054, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 7.79067325592041, "learning_rate": 5e-05, "loss": 0.8562, "mean_token_accuracy": 0.8912643194198608, "num_tokens": 779105668.0, "step": 20516 }, { "epoch": 2.609973285841496, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.611987590789795, "learning_rate": 5e-05, "loss": 0.7645, "mean_token_accuracy": 0.8995136618614197, "num_tokens": 779140005.0, "step": 20517 }, { "epoch": 2.6101004961200864, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.173319816589355, "learning_rate": 5e-05, "loss": 0.8739, "mean_token_accuracy": 0.882659912109375, "num_tokens": 779176269.0, "step": 20518 }, { "epoch": 2.610227706398677, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.245124340057373, "learning_rate": 5e-05, "loss": 0.7345, "mean_token_accuracy": 0.8915862441062927, "num_tokens": 779209914.0, "step": 20519 }, { "epoch": 2.6103549166772675, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.103296279907227, "learning_rate": 5e-05, "loss": 0.8228, "mean_token_accuracy": 0.8910518288612366, "num_tokens": 779244994.0, "step": 20520 }, { "epoch": 2.610482126955858, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.360776901245117, "learning_rate": 5e-05, "loss": 0.7588, "mean_token_accuracy": 0.8854497671127319, "num_tokens": 779283700.0, "step": 20521 }, { "epoch": 2.6106093372344485, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.766523361206055, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8904193043708801, "num_tokens": 779324827.0, "step": 20522 }, { "epoch": 2.610736547513039, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.605471611022949, "learning_rate": 5e-05, "loss": 0.8221, "mean_token_accuracy": 0.8697748184204102, "num_tokens": 779363593.0, "step": 20523 }, { "epoch": 2.6108637577916296, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.466614723205566, "learning_rate": 5e-05, "loss": 0.7445, "mean_token_accuracy": 0.883076548576355, "num_tokens": 779401295.0, "step": 20524 }, { "epoch": 2.61099096807022, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.751507759094238, "learning_rate": 5e-05, "loss": 0.8234, "mean_token_accuracy": 0.8657736778259277, "num_tokens": 779440375.0, "step": 20525 }, { "epoch": 2.6111181783488107, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.291107654571533, "learning_rate": 5e-05, "loss": 0.7397, "mean_token_accuracy": 0.8762911558151245, "num_tokens": 779475881.0, "step": 20526 }, { "epoch": 2.611245388627401, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.89798641204834, "learning_rate": 5e-05, "loss": 0.8148, "mean_token_accuracy": 0.8706389665603638, "num_tokens": 779520155.0, "step": 20527 }, { "epoch": 2.6113725989059917, "ewc_loss": 0.375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.058567523956299, "learning_rate": 5e-05, "loss": 0.7255, "mean_token_accuracy": 0.8780653476715088, "num_tokens": 779563952.0, "step": 20528 }, { "epoch": 2.6114998091845822, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.184122085571289, "learning_rate": 5e-05, "loss": 0.7665, "mean_token_accuracy": 0.8905221819877625, "num_tokens": 779601618.0, "step": 20529 }, { "epoch": 2.6116270194631728, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.730297088623047, "learning_rate": 5e-05, "loss": 0.7144, "mean_token_accuracy": 0.8726997375488281, "num_tokens": 779645228.0, "step": 20530 }, { "epoch": 2.6117542297417633, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.628687858581543, "learning_rate": 5e-05, "loss": 0.8336, "mean_token_accuracy": 0.8790964484214783, "num_tokens": 779681780.0, "step": 20531 }, { "epoch": 2.611881440020354, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.346988201141357, "learning_rate": 5e-05, "loss": 0.6738, "mean_token_accuracy": 0.8812426328659058, "num_tokens": 779717783.0, "step": 20532 }, { "epoch": 2.612008650298944, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.23531436920166, "learning_rate": 5e-05, "loss": 0.8875, "mean_token_accuracy": 0.887187123298645, "num_tokens": 779752542.0, "step": 20533 }, { "epoch": 2.612135860577535, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.086770534515381, "learning_rate": 5e-05, "loss": 0.6981, "mean_token_accuracy": 0.8805074095726013, "num_tokens": 779795134.0, "step": 20534 }, { "epoch": 2.612263070856125, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.842314720153809, "learning_rate": 5e-05, "loss": 0.9715, "mean_token_accuracy": 0.8845551013946533, "num_tokens": 779828589.0, "step": 20535 }, { "epoch": 2.612390281134716, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 6.596614837646484, "learning_rate": 5e-05, "loss": 0.8345, "mean_token_accuracy": 0.8701000213623047, "num_tokens": 779865904.0, "step": 20536 }, { "epoch": 2.612517491413306, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.781339645385742, "learning_rate": 5e-05, "loss": 1.0206, "mean_token_accuracy": 0.880752444267273, "num_tokens": 779910204.0, "step": 20537 }, { "epoch": 2.6126447016918966, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 7.824126720428467, "learning_rate": 5e-05, "loss": 0.955, "mean_token_accuracy": 0.8884312510490417, "num_tokens": 779951032.0, "step": 20538 }, { "epoch": 2.612771911970487, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.537064552307129, "learning_rate": 5e-05, "loss": 0.9086, "mean_token_accuracy": 0.8907458782196045, "num_tokens": 779983316.0, "step": 20539 }, { "epoch": 2.6128991222490776, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.692366600036621, "learning_rate": 5e-05, "loss": 0.994, "mean_token_accuracy": 0.879676103591919, "num_tokens": 780021167.0, "step": 20540 }, { "epoch": 2.613026332527668, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.126213073730469, "learning_rate": 5e-05, "loss": 0.7953, "mean_token_accuracy": 0.8798744678497314, "num_tokens": 780061276.0, "step": 20541 }, { "epoch": 2.6131535428062587, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.194191932678223, "learning_rate": 5e-05, "loss": 0.957, "mean_token_accuracy": 0.8838177919387817, "num_tokens": 780097204.0, "step": 20542 }, { "epoch": 2.613280753084849, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 6.883622646331787, "learning_rate": 5e-05, "loss": 0.8145, "mean_token_accuracy": 0.8713816404342651, "num_tokens": 780137208.0, "step": 20543 }, { "epoch": 2.6134079633634397, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.92130184173584, "learning_rate": 5e-05, "loss": 0.9671, "mean_token_accuracy": 0.8730672597885132, "num_tokens": 780177815.0, "step": 20544 }, { "epoch": 2.6135351736420303, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.444905757904053, "learning_rate": 5e-05, "loss": 0.848, "mean_token_accuracy": 0.8947241902351379, "num_tokens": 780215798.0, "step": 20545 }, { "epoch": 2.613662383920621, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.089213371276855, "learning_rate": 5e-05, "loss": 0.897, "mean_token_accuracy": 0.864983081817627, "num_tokens": 780252252.0, "step": 20546 }, { "epoch": 2.6137895941992113, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 7.916666507720947, "learning_rate": 5e-05, "loss": 0.8422, "mean_token_accuracy": 0.8968534469604492, "num_tokens": 780288100.0, "step": 20547 }, { "epoch": 2.613916804477802, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.489555358886719, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8799189329147339, "num_tokens": 780331922.0, "step": 20548 }, { "epoch": 2.6140440147563924, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.068038940429688, "learning_rate": 5e-05, "loss": 0.8743, "mean_token_accuracy": 0.8730461001396179, "num_tokens": 780369930.0, "step": 20549 }, { "epoch": 2.614171225034983, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.279718399047852, "learning_rate": 5e-05, "loss": 0.7671, "mean_token_accuracy": 0.881168782711029, "num_tokens": 780403893.0, "step": 20550 }, { "epoch": 2.6142984353135734, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.063559532165527, "learning_rate": 5e-05, "loss": 0.8009, "mean_token_accuracy": 0.8914088606834412, "num_tokens": 780438426.0, "step": 20551 }, { "epoch": 2.614425645592164, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.214179992675781, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.8835029006004333, "num_tokens": 780483267.0, "step": 20552 }, { "epoch": 2.6145528558707545, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.069348335266113, "learning_rate": 5e-05, "loss": 0.8077, "mean_token_accuracy": 0.8833311200141907, "num_tokens": 780520355.0, "step": 20553 }, { "epoch": 2.614680066149345, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.141916275024414, "learning_rate": 5e-05, "loss": 0.7373, "mean_token_accuracy": 0.8806885480880737, "num_tokens": 780563449.0, "step": 20554 }, { "epoch": 2.6148072764279355, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.111091613769531, "learning_rate": 5e-05, "loss": 0.7597, "mean_token_accuracy": 0.8911871910095215, "num_tokens": 780598607.0, "step": 20555 }, { "epoch": 2.6149344867065256, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.043593883514404, "learning_rate": 5e-05, "loss": 0.738, "mean_token_accuracy": 0.8818681240081787, "num_tokens": 780638425.0, "step": 20556 }, { "epoch": 2.6150616969851166, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.211249351501465, "learning_rate": 5e-05, "loss": 0.8357, "mean_token_accuracy": 0.8713914155960083, "num_tokens": 780683441.0, "step": 20557 }, { "epoch": 2.6151889072637067, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.894941806793213, "learning_rate": 5e-05, "loss": 0.675, "mean_token_accuracy": 0.8944934606552124, "num_tokens": 780720281.0, "step": 20558 }, { "epoch": 2.6153161175422976, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.396190643310547, "learning_rate": 5e-05, "loss": 0.7949, "mean_token_accuracy": 0.8912594318389893, "num_tokens": 780761038.0, "step": 20559 }, { "epoch": 2.6154433278208877, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.7107462882995605, "learning_rate": 5e-05, "loss": 0.7794, "mean_token_accuracy": 0.8573294878005981, "num_tokens": 780802358.0, "step": 20560 }, { "epoch": 2.6155705380994787, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.668540000915527, "learning_rate": 5e-05, "loss": 0.839, "mean_token_accuracy": 0.885281503200531, "num_tokens": 780839820.0, "step": 20561 }, { "epoch": 2.615697748378069, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.52750825881958, "learning_rate": 5e-05, "loss": 0.7104, "mean_token_accuracy": 0.8796412944793701, "num_tokens": 780879969.0, "step": 20562 }, { "epoch": 2.6158249586566593, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.971678733825684, "learning_rate": 5e-05, "loss": 0.9067, "mean_token_accuracy": 0.8807210922241211, "num_tokens": 780916119.0, "step": 20563 }, { "epoch": 2.61595216893525, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.537348747253418, "learning_rate": 5e-05, "loss": 0.726, "mean_token_accuracy": 0.8802062273025513, "num_tokens": 780957311.0, "step": 20564 }, { "epoch": 2.6160793792138404, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.091447830200195, "learning_rate": 5e-05, "loss": 0.956, "mean_token_accuracy": 0.8762304782867432, "num_tokens": 780998322.0, "step": 20565 }, { "epoch": 2.616206589492431, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 6.959190368652344, "learning_rate": 5e-05, "loss": 0.8568, "mean_token_accuracy": 0.8659919500350952, "num_tokens": 781036249.0, "step": 20566 }, { "epoch": 2.6163337997710214, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.618245124816895, "learning_rate": 5e-05, "loss": 0.9166, "mean_token_accuracy": 0.8803349137306213, "num_tokens": 781075228.0, "step": 20567 }, { "epoch": 2.616461010049612, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.61227560043335, "learning_rate": 5e-05, "loss": 0.8463, "mean_token_accuracy": 0.887831449508667, "num_tokens": 781110907.0, "step": 20568 }, { "epoch": 2.6165882203282025, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.7432355880737305, "learning_rate": 5e-05, "loss": 0.8652, "mean_token_accuracy": 0.8712717294692993, "num_tokens": 781148615.0, "step": 20569 }, { "epoch": 2.616715430606793, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.136086463928223, "learning_rate": 5e-05, "loss": 0.8599, "mean_token_accuracy": 0.8859869241714478, "num_tokens": 781181159.0, "step": 20570 }, { "epoch": 2.6168426408853835, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.088647365570068, "learning_rate": 5e-05, "loss": 0.8193, "mean_token_accuracy": 0.8571299314498901, "num_tokens": 781221563.0, "step": 20571 }, { "epoch": 2.616969851163974, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.372209548950195, "learning_rate": 5e-05, "loss": 0.8962, "mean_token_accuracy": 0.8687084317207336, "num_tokens": 781257606.0, "step": 20572 }, { "epoch": 2.6170970614425646, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.938694000244141, "learning_rate": 5e-05, "loss": 0.7416, "mean_token_accuracy": 0.8776772618293762, "num_tokens": 781296653.0, "step": 20573 }, { "epoch": 2.617224271721155, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.321892738342285, "learning_rate": 5e-05, "loss": 0.8003, "mean_token_accuracy": 0.8941217660903931, "num_tokens": 781331596.0, "step": 20574 }, { "epoch": 2.6173514819997457, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.050266265869141, "learning_rate": 5e-05, "loss": 0.6796, "mean_token_accuracy": 0.9025386571884155, "num_tokens": 781365599.0, "step": 20575 }, { "epoch": 2.617478692278336, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.116573333740234, "learning_rate": 5e-05, "loss": 0.8197, "mean_token_accuracy": 0.8796616792678833, "num_tokens": 781406169.0, "step": 20576 }, { "epoch": 2.6176059025569267, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.197901725769043, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8729151487350464, "num_tokens": 781444570.0, "step": 20577 }, { "epoch": 2.6177331128355172, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.98150110244751, "learning_rate": 5e-05, "loss": 0.8228, "mean_token_accuracy": 0.8688808679580688, "num_tokens": 781482752.0, "step": 20578 }, { "epoch": 2.6178603231141078, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.228928089141846, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8691026568412781, "num_tokens": 781524549.0, "step": 20579 }, { "epoch": 2.6179875333926983, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.935427188873291, "learning_rate": 5e-05, "loss": 0.8374, "mean_token_accuracy": 0.8681826591491699, "num_tokens": 781560152.0, "step": 20580 }, { "epoch": 2.6181147436712884, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.183925628662109, "learning_rate": 5e-05, "loss": 0.7462, "mean_token_accuracy": 0.8782362341880798, "num_tokens": 781606399.0, "step": 20581 }, { "epoch": 2.6182419539498794, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.97897481918335, "learning_rate": 5e-05, "loss": 0.8034, "mean_token_accuracy": 0.8801990747451782, "num_tokens": 781646386.0, "step": 20582 }, { "epoch": 2.6183691642284694, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.087496757507324, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.8734729886054993, "num_tokens": 781685629.0, "step": 20583 }, { "epoch": 2.6184963745070604, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.076119422912598, "learning_rate": 5e-05, "loss": 0.736, "mean_token_accuracy": 0.8959802389144897, "num_tokens": 781717198.0, "step": 20584 }, { "epoch": 2.6186235847856505, "ewc_loss": 0.375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.93363094329834, "learning_rate": 5e-05, "loss": 0.7537, "mean_token_accuracy": 0.8664588928222656, "num_tokens": 781749281.0, "step": 20585 }, { "epoch": 2.6187507950642415, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.283914566040039, "learning_rate": 5e-05, "loss": 0.8005, "mean_token_accuracy": 0.8835152387619019, "num_tokens": 781791544.0, "step": 20586 }, { "epoch": 2.6188780053428315, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.719449520111084, "learning_rate": 5e-05, "loss": 0.7201, "mean_token_accuracy": 0.8741388916969299, "num_tokens": 781832864.0, "step": 20587 }, { "epoch": 2.619005215621422, "ewc_loss": 0.5, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.605814933776855, "learning_rate": 5e-05, "loss": 0.8552, "mean_token_accuracy": 0.8807955384254456, "num_tokens": 781875757.0, "step": 20588 }, { "epoch": 2.6191324259000126, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.4917707443237305, "learning_rate": 5e-05, "loss": 0.6852, "mean_token_accuracy": 0.8855171799659729, "num_tokens": 781911200.0, "step": 20589 }, { "epoch": 2.619259636178603, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.945862770080566, "learning_rate": 5e-05, "loss": 0.8665, "mean_token_accuracy": 0.8922473788261414, "num_tokens": 781948372.0, "step": 20590 }, { "epoch": 2.6193868464571937, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.514241695404053, "learning_rate": 5e-05, "loss": 0.7301, "mean_token_accuracy": 0.8768218755722046, "num_tokens": 781985360.0, "step": 20591 }, { "epoch": 2.619514056735784, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.990592956542969, "learning_rate": 5e-05, "loss": 0.9032, "mean_token_accuracy": 0.8834928274154663, "num_tokens": 782019439.0, "step": 20592 }, { "epoch": 2.6196412670143747, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.001699924468994, "learning_rate": 5e-05, "loss": 0.7371, "mean_token_accuracy": 0.9047824144363403, "num_tokens": 782055805.0, "step": 20593 }, { "epoch": 2.6197684772929652, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.369826316833496, "learning_rate": 5e-05, "loss": 0.9016, "mean_token_accuracy": 0.8753541111946106, "num_tokens": 782092879.0, "step": 20594 }, { "epoch": 2.6198956875715558, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.763556003570557, "learning_rate": 5e-05, "loss": 0.8845, "mean_token_accuracy": 0.87531977891922, "num_tokens": 782133123.0, "step": 20595 }, { "epoch": 2.6200228978501463, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.29659366607666, "learning_rate": 5e-05, "loss": 0.8089, "mean_token_accuracy": 0.8665764331817627, "num_tokens": 782169999.0, "step": 20596 }, { "epoch": 2.620150108128737, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.516324996948242, "learning_rate": 5e-05, "loss": 0.8789, "mean_token_accuracy": 0.8858687877655029, "num_tokens": 782208269.0, "step": 20597 }, { "epoch": 2.6202773184073274, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.499176502227783, "learning_rate": 5e-05, "loss": 0.7142, "mean_token_accuracy": 0.8781164288520813, "num_tokens": 782242927.0, "step": 20598 }, { "epoch": 2.620404528685918, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.159741401672363, "learning_rate": 5e-05, "loss": 0.9563, "mean_token_accuracy": 0.8739408254623413, "num_tokens": 782281018.0, "step": 20599 }, { "epoch": 2.6205317389645084, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.275203227996826, "learning_rate": 5e-05, "loss": 0.7967, "mean_token_accuracy": 0.8566486239433289, "num_tokens": 782315336.0, "step": 20600 }, { "epoch": 2.620658949243099, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.45723819732666, "learning_rate": 5e-05, "loss": 0.9457, "mean_token_accuracy": 0.893247127532959, "num_tokens": 782351163.0, "step": 20601 }, { "epoch": 2.6207861595216895, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 6.826815605163574, "learning_rate": 5e-05, "loss": 0.8751, "mean_token_accuracy": 0.8644726276397705, "num_tokens": 782396955.0, "step": 20602 }, { "epoch": 2.62091336980028, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.058392524719238, "learning_rate": 5e-05, "loss": 0.9367, "mean_token_accuracy": 0.8930652737617493, "num_tokens": 782432005.0, "step": 20603 }, { "epoch": 2.6210405800788705, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.75591516494751, "learning_rate": 5e-05, "loss": 0.9369, "mean_token_accuracy": 0.8844746351242065, "num_tokens": 782467659.0, "step": 20604 }, { "epoch": 2.621167790357461, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.944049835205078, "learning_rate": 5e-05, "loss": 0.8816, "mean_token_accuracy": 0.8789116740226746, "num_tokens": 782502621.0, "step": 20605 }, { "epoch": 2.621295000636051, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.372182846069336, "learning_rate": 5e-05, "loss": 0.9233, "mean_token_accuracy": 0.8871124386787415, "num_tokens": 782535803.0, "step": 20606 }, { "epoch": 2.621422210914642, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.171804904937744, "learning_rate": 5e-05, "loss": 0.7764, "mean_token_accuracy": 0.8793998956680298, "num_tokens": 782574720.0, "step": 20607 }, { "epoch": 2.621549421193232, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.389617919921875, "learning_rate": 5e-05, "loss": 0.8713, "mean_token_accuracy": 0.8827517032623291, "num_tokens": 782615477.0, "step": 20608 }, { "epoch": 2.621676631471823, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.304574012756348, "learning_rate": 5e-05, "loss": 0.8293, "mean_token_accuracy": 0.8706830739974976, "num_tokens": 782653153.0, "step": 20609 }, { "epoch": 2.6218038417504133, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.800426483154297, "learning_rate": 5e-05, "loss": 0.8639, "mean_token_accuracy": 0.8684588670730591, "num_tokens": 782691024.0, "step": 20610 }, { "epoch": 2.621931052029004, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.804412364959717, "learning_rate": 5e-05, "loss": 0.8264, "mean_token_accuracy": 0.8825967907905579, "num_tokens": 782732031.0, "step": 20611 }, { "epoch": 2.6220582623075943, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.188437461853027, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.8786247372627258, "num_tokens": 782773430.0, "step": 20612 }, { "epoch": 2.622185472586185, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.164193153381348, "learning_rate": 5e-05, "loss": 0.8488, "mean_token_accuracy": 0.8749009370803833, "num_tokens": 782815852.0, "step": 20613 }, { "epoch": 2.6223126828647754, "ewc_loss": 0.375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.879059791564941, "learning_rate": 5e-05, "loss": 0.6991, "mean_token_accuracy": 0.8858610391616821, "num_tokens": 782851327.0, "step": 20614 }, { "epoch": 2.622439893143366, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.450886726379395, "learning_rate": 5e-05, "loss": 0.8816, "mean_token_accuracy": 0.8698259592056274, "num_tokens": 782887382.0, "step": 20615 }, { "epoch": 2.6225671034219564, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.623345375061035, "learning_rate": 5e-05, "loss": 0.6732, "mean_token_accuracy": 0.8919793367385864, "num_tokens": 782923667.0, "step": 20616 }, { "epoch": 2.622694313700547, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.888559341430664, "learning_rate": 5e-05, "loss": 0.8958, "mean_token_accuracy": 0.886684000492096, "num_tokens": 782961200.0, "step": 20617 }, { "epoch": 2.6228215239791375, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.410285472869873, "learning_rate": 5e-05, "loss": 0.7202, "mean_token_accuracy": 0.8808766007423401, "num_tokens": 782995832.0, "step": 20618 }, { "epoch": 2.622948734257728, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.307367324829102, "learning_rate": 5e-05, "loss": 0.9624, "mean_token_accuracy": 0.8788489103317261, "num_tokens": 783040880.0, "step": 20619 }, { "epoch": 2.6230759445363185, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.6490912437438965, "learning_rate": 5e-05, "loss": 0.7523, "mean_token_accuracy": 0.8871056437492371, "num_tokens": 783075068.0, "step": 20620 }, { "epoch": 2.623203154814909, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.309906959533691, "learning_rate": 5e-05, "loss": 0.9747, "mean_token_accuracy": 0.8839760422706604, "num_tokens": 783109321.0, "step": 20621 }, { "epoch": 2.6233303650934996, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.298172473907471, "learning_rate": 5e-05, "loss": 0.8813, "mean_token_accuracy": 0.8849503993988037, "num_tokens": 783145222.0, "step": 20622 }, { "epoch": 2.62345757537209, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.733896255493164, "learning_rate": 5e-05, "loss": 0.9694, "mean_token_accuracy": 0.8748542070388794, "num_tokens": 783185017.0, "step": 20623 }, { "epoch": 2.6235847856506807, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 7.812559127807617, "learning_rate": 5e-05, "loss": 0.9529, "mean_token_accuracy": 0.875128448009491, "num_tokens": 783228137.0, "step": 20624 }, { "epoch": 2.623711995929271, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.044046401977539, "learning_rate": 5e-05, "loss": 0.8946, "mean_token_accuracy": 0.8770817518234253, "num_tokens": 783266525.0, "step": 20625 }, { "epoch": 2.6238392062078617, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.835330009460449, "learning_rate": 5e-05, "loss": 0.8913, "mean_token_accuracy": 0.8766566514968872, "num_tokens": 783301090.0, "step": 20626 }, { "epoch": 2.6239664164864522, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.914564609527588, "learning_rate": 5e-05, "loss": 0.8615, "mean_token_accuracy": 0.8811732530593872, "num_tokens": 783336816.0, "step": 20627 }, { "epoch": 2.6240936267650428, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.371658802032471, "learning_rate": 5e-05, "loss": 0.7712, "mean_token_accuracy": 0.8900365829467773, "num_tokens": 783376766.0, "step": 20628 }, { "epoch": 2.6242208370436333, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.355481147766113, "learning_rate": 5e-05, "loss": 0.8414, "mean_token_accuracy": 0.8907451033592224, "num_tokens": 783416780.0, "step": 20629 }, { "epoch": 2.624348047322224, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.848965644836426, "learning_rate": 5e-05, "loss": 0.7133, "mean_token_accuracy": 0.8914801478385925, "num_tokens": 783449957.0, "step": 20630 }, { "epoch": 2.624475257600814, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.766793251037598, "learning_rate": 5e-05, "loss": 0.8642, "mean_token_accuracy": 0.8954458236694336, "num_tokens": 783480488.0, "step": 20631 }, { "epoch": 2.624602467879405, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.815452575683594, "learning_rate": 5e-05, "loss": 0.7674, "mean_token_accuracy": 0.8766049146652222, "num_tokens": 783521296.0, "step": 20632 }, { "epoch": 2.624729678157995, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.717049598693848, "learning_rate": 5e-05, "loss": 0.9471, "mean_token_accuracy": 0.8727624416351318, "num_tokens": 783561937.0, "step": 20633 }, { "epoch": 2.624856888436586, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.201922416687012, "learning_rate": 5e-05, "loss": 0.8018, "mean_token_accuracy": 0.8867882490158081, "num_tokens": 783598928.0, "step": 20634 }, { "epoch": 2.624984098715176, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.230825424194336, "learning_rate": 5e-05, "loss": 0.8638, "mean_token_accuracy": 0.8850481510162354, "num_tokens": 783635538.0, "step": 20635 }, { "epoch": 2.6251113089937665, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.632381916046143, "learning_rate": 5e-05, "loss": 0.8154, "mean_token_accuracy": 0.8913620710372925, "num_tokens": 783675712.0, "step": 20636 }, { "epoch": 2.625238519272357, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.658278942108154, "learning_rate": 5e-05, "loss": 0.8184, "mean_token_accuracy": 0.8797166347503662, "num_tokens": 783715091.0, "step": 20637 }, { "epoch": 2.6253657295509476, "ewc_loss": 0.5, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.8659281730651855, "learning_rate": 5e-05, "loss": 0.8097, "mean_token_accuracy": 0.8903380632400513, "num_tokens": 783753263.0, "step": 20638 }, { "epoch": 2.625492939829538, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.347522735595703, "learning_rate": 5e-05, "loss": 0.7705, "mean_token_accuracy": 0.8761049509048462, "num_tokens": 783786251.0, "step": 20639 }, { "epoch": 2.6256201501081287, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.877682209014893, "learning_rate": 5e-05, "loss": 0.7977, "mean_token_accuracy": 0.8796396255493164, "num_tokens": 783829751.0, "step": 20640 }, { "epoch": 2.625747360386719, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.293119430541992, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.8741137981414795, "num_tokens": 783867786.0, "step": 20641 }, { "epoch": 2.6258745706653097, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.791788101196289, "learning_rate": 5e-05, "loss": 0.8091, "mean_token_accuracy": 0.8730289936065674, "num_tokens": 783904245.0, "step": 20642 }, { "epoch": 2.6260017809439002, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.286457538604736, "learning_rate": 5e-05, "loss": 0.6943, "mean_token_accuracy": 0.8920509815216064, "num_tokens": 783946892.0, "step": 20643 }, { "epoch": 2.6261289912224908, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.794137477874756, "learning_rate": 5e-05, "loss": 0.8153, "mean_token_accuracy": 0.8713697195053101, "num_tokens": 783988132.0, "step": 20644 }, { "epoch": 2.6262562015010813, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.196575164794922, "learning_rate": 5e-05, "loss": 0.7684, "mean_token_accuracy": 0.8694108128547668, "num_tokens": 784024321.0, "step": 20645 }, { "epoch": 2.626383411779672, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.893925189971924, "learning_rate": 5e-05, "loss": 0.8034, "mean_token_accuracy": 0.8733738660812378, "num_tokens": 784061266.0, "step": 20646 }, { "epoch": 2.6265106220582624, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.0292863845825195, "learning_rate": 5e-05, "loss": 0.7197, "mean_token_accuracy": 0.8762223720550537, "num_tokens": 784101991.0, "step": 20647 }, { "epoch": 2.626637832336853, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.101078033447266, "learning_rate": 5e-05, "loss": 0.7922, "mean_token_accuracy": 0.8791701793670654, "num_tokens": 784137726.0, "step": 20648 }, { "epoch": 2.6267650426154434, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.783842086791992, "learning_rate": 5e-05, "loss": 0.7157, "mean_token_accuracy": 0.8761883974075317, "num_tokens": 784177773.0, "step": 20649 }, { "epoch": 2.626892252894034, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.462198257446289, "learning_rate": 5e-05, "loss": 0.831, "mean_token_accuracy": 0.8788349628448486, "num_tokens": 784217429.0, "step": 20650 }, { "epoch": 2.6270194631726245, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.472757816314697, "learning_rate": 5e-05, "loss": 0.6526, "mean_token_accuracy": 0.8904263377189636, "num_tokens": 784252289.0, "step": 20651 }, { "epoch": 2.627146673451215, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.93858814239502, "learning_rate": 5e-05, "loss": 0.9533, "mean_token_accuracy": 0.8544907569885254, "num_tokens": 784293472.0, "step": 20652 }, { "epoch": 2.6272738837298055, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.254851341247559, "learning_rate": 5e-05, "loss": 0.7036, "mean_token_accuracy": 0.8765913248062134, "num_tokens": 784339423.0, "step": 20653 }, { "epoch": 2.6274010940083956, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.369595527648926, "learning_rate": 5e-05, "loss": 0.9085, "mean_token_accuracy": 0.893269419670105, "num_tokens": 784367026.0, "step": 20654 }, { "epoch": 2.6275283042869866, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.61283016204834, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8780539035797119, "num_tokens": 784404004.0, "step": 20655 }, { "epoch": 2.6276555145655767, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.10188102722168, "learning_rate": 5e-05, "loss": 0.922, "mean_token_accuracy": 0.8914897441864014, "num_tokens": 784439335.0, "step": 20656 }, { "epoch": 2.6277827248441676, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.592838287353516, "learning_rate": 5e-05, "loss": 0.9093, "mean_token_accuracy": 0.8739819526672363, "num_tokens": 784480141.0, "step": 20657 }, { "epoch": 2.6279099351227577, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.829430103302002, "learning_rate": 5e-05, "loss": 0.8384, "mean_token_accuracy": 0.8841888308525085, "num_tokens": 784514785.0, "step": 20658 }, { "epoch": 2.6280371454013487, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.633585929870605, "learning_rate": 5e-05, "loss": 0.9588, "mean_token_accuracy": 0.8705410957336426, "num_tokens": 784549776.0, "step": 20659 }, { "epoch": 2.628164355679939, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.472844123840332, "learning_rate": 5e-05, "loss": 0.7689, "mean_token_accuracy": 0.8660003542900085, "num_tokens": 784578558.0, "step": 20660 }, { "epoch": 2.6282915659585293, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.636984825134277, "learning_rate": 5e-05, "loss": 0.9675, "mean_token_accuracy": 0.8839654922485352, "num_tokens": 784614056.0, "step": 20661 }, { "epoch": 2.62841877623712, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.168133735656738, "learning_rate": 5e-05, "loss": 0.7076, "mean_token_accuracy": 0.8823282718658447, "num_tokens": 784651663.0, "step": 20662 }, { "epoch": 2.6285459865157104, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.18630599975586, "learning_rate": 5e-05, "loss": 1.0717, "mean_token_accuracy": 0.8712040185928345, "num_tokens": 784690343.0, "step": 20663 }, { "epoch": 2.628673196794301, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 7.368066787719727, "learning_rate": 5e-05, "loss": 0.9361, "mean_token_accuracy": 0.8803249597549438, "num_tokens": 784728569.0, "step": 20664 }, { "epoch": 2.6288004070728914, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.404643058776855, "learning_rate": 5e-05, "loss": 0.9721, "mean_token_accuracy": 0.8862950801849365, "num_tokens": 784762376.0, "step": 20665 }, { "epoch": 2.628927617351482, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 8.575247764587402, "learning_rate": 5e-05, "loss": 1.0395, "mean_token_accuracy": 0.8812599778175354, "num_tokens": 784793154.0, "step": 20666 }, { "epoch": 2.6290548276300725, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.78926420211792, "learning_rate": 5e-05, "loss": 0.8491, "mean_token_accuracy": 0.8809026479721069, "num_tokens": 784829640.0, "step": 20667 }, { "epoch": 2.629182037908663, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.12997055053711, "learning_rate": 5e-05, "loss": 0.9747, "mean_token_accuracy": 0.8887819051742554, "num_tokens": 784867045.0, "step": 20668 }, { "epoch": 2.6293092481872535, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.2066216468811035, "learning_rate": 5e-05, "loss": 0.8149, "mean_token_accuracy": 0.8820971250534058, "num_tokens": 784907157.0, "step": 20669 }, { "epoch": 2.629436458465844, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.924891471862793, "learning_rate": 5e-05, "loss": 0.9556, "mean_token_accuracy": 0.8795981407165527, "num_tokens": 784942247.0, "step": 20670 }, { "epoch": 2.6295636687444346, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.706236362457275, "learning_rate": 5e-05, "loss": 0.9142, "mean_token_accuracy": 0.8790931701660156, "num_tokens": 784976910.0, "step": 20671 }, { "epoch": 2.629690879023025, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.8558831214904785, "learning_rate": 5e-05, "loss": 0.8777, "mean_token_accuracy": 0.8622759580612183, "num_tokens": 785017969.0, "step": 20672 }, { "epoch": 2.6298180893016156, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.334908485412598, "learning_rate": 5e-05, "loss": 0.9104, "mean_token_accuracy": 0.882554829120636, "num_tokens": 785056182.0, "step": 20673 }, { "epoch": 2.629945299580206, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.0694580078125, "learning_rate": 5e-05, "loss": 0.7258, "mean_token_accuracy": 0.8875997066497803, "num_tokens": 785091996.0, "step": 20674 }, { "epoch": 2.6300725098587967, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.724343299865723, "learning_rate": 5e-05, "loss": 0.9057, "mean_token_accuracy": 0.8847662210464478, "num_tokens": 785129268.0, "step": 20675 }, { "epoch": 2.6301997201373872, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.863701820373535, "learning_rate": 5e-05, "loss": 0.719, "mean_token_accuracy": 0.8898571729660034, "num_tokens": 785168418.0, "step": 20676 }, { "epoch": 2.6303269304159778, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.92546272277832, "learning_rate": 5e-05, "loss": 0.9136, "mean_token_accuracy": 0.8859966397285461, "num_tokens": 785209981.0, "step": 20677 }, { "epoch": 2.6304541406945683, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.954892635345459, "learning_rate": 5e-05, "loss": 0.7954, "mean_token_accuracy": 0.8827324509620667, "num_tokens": 785250354.0, "step": 20678 }, { "epoch": 2.6305813509731584, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.889315605163574, "learning_rate": 5e-05, "loss": 0.9288, "mean_token_accuracy": 0.8791025876998901, "num_tokens": 785288594.0, "step": 20679 }, { "epoch": 2.6307085612517493, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.120353698730469, "learning_rate": 5e-05, "loss": 0.7839, "mean_token_accuracy": 0.892623782157898, "num_tokens": 785323885.0, "step": 20680 }, { "epoch": 2.6308357715303394, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.748086929321289, "learning_rate": 5e-05, "loss": 0.8709, "mean_token_accuracy": 0.8942282795906067, "num_tokens": 785361950.0, "step": 20681 }, { "epoch": 2.6309629818089304, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.137938499450684, "learning_rate": 5e-05, "loss": 0.8508, "mean_token_accuracy": 0.8727570176124573, "num_tokens": 785398650.0, "step": 20682 }, { "epoch": 2.6310901920875205, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.735594749450684, "learning_rate": 5e-05, "loss": 0.9435, "mean_token_accuracy": 0.874048113822937, "num_tokens": 785451273.0, "step": 20683 }, { "epoch": 2.6312174023661115, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 6.9812164306640625, "learning_rate": 5e-05, "loss": 0.805, "mean_token_accuracy": 0.8770360946655273, "num_tokens": 785491244.0, "step": 20684 }, { "epoch": 2.6313446126447015, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.974546432495117, "learning_rate": 5e-05, "loss": 0.9343, "mean_token_accuracy": 0.8815416097640991, "num_tokens": 785536558.0, "step": 20685 }, { "epoch": 2.631471822923292, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 6.902042388916016, "learning_rate": 5e-05, "loss": 0.8053, "mean_token_accuracy": 0.8756279349327087, "num_tokens": 785572635.0, "step": 20686 }, { "epoch": 2.6315990332018826, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.132580757141113, "learning_rate": 5e-05, "loss": 0.9915, "mean_token_accuracy": 0.8778687715530396, "num_tokens": 785615765.0, "step": 20687 }, { "epoch": 2.631726243480473, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.212437152862549, "learning_rate": 5e-05, "loss": 0.8421, "mean_token_accuracy": 0.8802117109298706, "num_tokens": 785652934.0, "step": 20688 }, { "epoch": 2.6318534537590637, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.831073760986328, "learning_rate": 5e-05, "loss": 0.9694, "mean_token_accuracy": 0.8757568001747131, "num_tokens": 785689018.0, "step": 20689 }, { "epoch": 2.631980664037654, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.667608261108398, "learning_rate": 5e-05, "loss": 0.9007, "mean_token_accuracy": 0.8728916645050049, "num_tokens": 785728257.0, "step": 20690 }, { "epoch": 2.6321078743162447, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.171008110046387, "learning_rate": 5e-05, "loss": 0.903, "mean_token_accuracy": 0.8720770478248596, "num_tokens": 785768381.0, "step": 20691 }, { "epoch": 2.6322350845948352, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.917746067047119, "learning_rate": 5e-05, "loss": 0.8719, "mean_token_accuracy": 0.8769843578338623, "num_tokens": 785806700.0, "step": 20692 }, { "epoch": 2.6323622948734258, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.648097991943359, "learning_rate": 5e-05, "loss": 0.8591, "mean_token_accuracy": 0.8651674389839172, "num_tokens": 785848602.0, "step": 20693 }, { "epoch": 2.6324895051520163, "ewc_loss": 0.5, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.8931965827941895, "learning_rate": 5e-05, "loss": 0.8359, "mean_token_accuracy": 0.8823438882827759, "num_tokens": 785888714.0, "step": 20694 }, { "epoch": 2.632616715430607, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.534989833831787, "learning_rate": 5e-05, "loss": 0.7535, "mean_token_accuracy": 0.8907986879348755, "num_tokens": 785925163.0, "step": 20695 }, { "epoch": 2.6327439257091974, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.669960021972656, "learning_rate": 5e-05, "loss": 0.8273, "mean_token_accuracy": 0.8641033172607422, "num_tokens": 785963566.0, "step": 20696 }, { "epoch": 2.632871135987788, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.680332660675049, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.8706549406051636, "num_tokens": 785997654.0, "step": 20697 }, { "epoch": 2.6329983462663784, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00077056884765625, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.403504848480225, "learning_rate": 5e-05, "loss": 0.7612, "mean_token_accuracy": 0.8749541640281677, "num_tokens": 786038761.0, "step": 20698 }, { "epoch": 2.633125556544969, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.805628776550293, "learning_rate": 5e-05, "loss": 0.7676, "mean_token_accuracy": 0.883433997631073, "num_tokens": 786072380.0, "step": 20699 }, { "epoch": 2.6332527668235595, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.3032660484313965, "learning_rate": 5e-05, "loss": 0.7318, "mean_token_accuracy": 0.8792344331741333, "num_tokens": 786112751.0, "step": 20700 }, { "epoch": 2.63337997710215, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.790440082550049, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.8801319599151611, "num_tokens": 786156841.0, "step": 20701 }, { "epoch": 2.6335071873807405, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.285637855529785, "learning_rate": 5e-05, "loss": 0.7446, "mean_token_accuracy": 0.8738764524459839, "num_tokens": 786191015.0, "step": 20702 }, { "epoch": 2.633634397659331, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.802679061889648, "learning_rate": 5e-05, "loss": 0.8076, "mean_token_accuracy": 0.8625730276107788, "num_tokens": 786231420.0, "step": 20703 }, { "epoch": 2.633761607937921, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.1906962394714355, "learning_rate": 5e-05, "loss": 0.7005, "mean_token_accuracy": 0.8842541575431824, "num_tokens": 786266586.0, "step": 20704 }, { "epoch": 2.633888818216512, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.93772029876709, "learning_rate": 5e-05, "loss": 0.7726, "mean_token_accuracy": 0.8782093524932861, "num_tokens": 786307489.0, "step": 20705 }, { "epoch": 2.634016028495102, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.994388103485107, "learning_rate": 5e-05, "loss": 0.6878, "mean_token_accuracy": 0.8842763900756836, "num_tokens": 786340960.0, "step": 20706 }, { "epoch": 2.634143238773693, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.18553638458252, "learning_rate": 5e-05, "loss": 0.8015, "mean_token_accuracy": 0.8761641979217529, "num_tokens": 786383086.0, "step": 20707 }, { "epoch": 2.6342704490522832, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.878966808319092, "learning_rate": 5e-05, "loss": 0.6719, "mean_token_accuracy": 0.8857150077819824, "num_tokens": 786421093.0, "step": 20708 }, { "epoch": 2.6343976593308738, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.215728759765625, "learning_rate": 5e-05, "loss": 0.7999, "mean_token_accuracy": 0.8788283467292786, "num_tokens": 786462712.0, "step": 20709 }, { "epoch": 2.6345248696094643, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.835495948791504, "learning_rate": 5e-05, "loss": 0.7065, "mean_token_accuracy": 0.8771530389785767, "num_tokens": 786501607.0, "step": 20710 }, { "epoch": 2.634652079888055, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.382928848266602, "learning_rate": 5e-05, "loss": 0.8088, "mean_token_accuracy": 0.8827617168426514, "num_tokens": 786537163.0, "step": 20711 }, { "epoch": 2.6347792901666454, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.649440765380859, "learning_rate": 5e-05, "loss": 0.6706, "mean_token_accuracy": 0.8871728777885437, "num_tokens": 786572880.0, "step": 20712 }, { "epoch": 2.634906500445236, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.694287300109863, "learning_rate": 5e-05, "loss": 0.847, "mean_token_accuracy": 0.881462574005127, "num_tokens": 786610082.0, "step": 20713 }, { "epoch": 2.6350337107238264, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.5267109870910645, "learning_rate": 5e-05, "loss": 0.7076, "mean_token_accuracy": 0.8788372278213501, "num_tokens": 786648875.0, "step": 20714 }, { "epoch": 2.635160921002417, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.883399963378906, "learning_rate": 5e-05, "loss": 0.9173, "mean_token_accuracy": 0.8758647441864014, "num_tokens": 786688076.0, "step": 20715 }, { "epoch": 2.6352881312810075, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.757846832275391, "learning_rate": 5e-05, "loss": 0.7438, "mean_token_accuracy": 0.8754189014434814, "num_tokens": 786723969.0, "step": 20716 }, { "epoch": 2.635415341559598, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.538759231567383, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.8759424686431885, "num_tokens": 786764018.0, "step": 20717 }, { "epoch": 2.6355425518381885, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.41691255569458, "learning_rate": 5e-05, "loss": 0.7797, "mean_token_accuracy": 0.8845646381378174, "num_tokens": 786796586.0, "step": 20718 }, { "epoch": 2.635669762116779, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.607738971710205, "learning_rate": 5e-05, "loss": 0.8194, "mean_token_accuracy": 0.8683594465255737, "num_tokens": 786833797.0, "step": 20719 }, { "epoch": 2.6357969723953696, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.211807250976562, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8694819211959839, "num_tokens": 786869971.0, "step": 20720 }, { "epoch": 2.63592418267396, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.643189907073975, "learning_rate": 5e-05, "loss": 0.7043, "mean_token_accuracy": 0.8803343772888184, "num_tokens": 786905854.0, "step": 20721 }, { "epoch": 2.6360513929525506, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 9.0488862991333, "learning_rate": 5e-05, "loss": 0.8944, "mean_token_accuracy": 0.8796196579933167, "num_tokens": 786940535.0, "step": 20722 }, { "epoch": 2.636178603231141, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.054510593414307, "learning_rate": 5e-05, "loss": 0.7045, "mean_token_accuracy": 0.8697012066841125, "num_tokens": 786974976.0, "step": 20723 }, { "epoch": 2.6363058135097317, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.818473815917969, "learning_rate": 5e-05, "loss": 0.9894, "mean_token_accuracy": 0.8763967752456665, "num_tokens": 787013389.0, "step": 20724 }, { "epoch": 2.6364330237883222, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.352116107940674, "learning_rate": 5e-05, "loss": 0.7021, "mean_token_accuracy": 0.8936880826950073, "num_tokens": 787044463.0, "step": 20725 }, { "epoch": 2.6365602340669128, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.952409744262695, "learning_rate": 5e-05, "loss": 1.0262, "mean_token_accuracy": 0.8839613795280457, "num_tokens": 787075123.0, "step": 20726 }, { "epoch": 2.636687444345503, "ewc_loss": 0.625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 7.752017021179199, "learning_rate": 5e-05, "loss": 0.9775, "mean_token_accuracy": 0.8770160675048828, "num_tokens": 787114790.0, "step": 20727 }, { "epoch": 2.636814654624094, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.663107872009277, "learning_rate": 5e-05, "loss": 0.909, "mean_token_accuracy": 0.8875648975372314, "num_tokens": 787153263.0, "step": 20728 }, { "epoch": 2.636941864902684, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.816802978515625, "learning_rate": 5e-05, "loss": 0.9855, "mean_token_accuracy": 0.886369526386261, "num_tokens": 787191955.0, "step": 20729 }, { "epoch": 2.637069075181275, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.049583911895752, "learning_rate": 5e-05, "loss": 0.7704, "mean_token_accuracy": 0.883071780204773, "num_tokens": 787228330.0, "step": 20730 }, { "epoch": 2.637196285459865, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.593326568603516, "learning_rate": 5e-05, "loss": 1.0614, "mean_token_accuracy": 0.8602017164230347, "num_tokens": 787271927.0, "step": 20731 }, { "epoch": 2.637323495738456, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 6.897177219390869, "learning_rate": 5e-05, "loss": 0.785, "mean_token_accuracy": 0.8921401500701904, "num_tokens": 787310515.0, "step": 20732 }, { "epoch": 2.637450706017046, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.335543632507324, "learning_rate": 5e-05, "loss": 0.9885, "mean_token_accuracy": 0.8771751523017883, "num_tokens": 787346720.0, "step": 20733 }, { "epoch": 2.6375779162956365, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.779611110687256, "learning_rate": 5e-05, "loss": 0.9421, "mean_token_accuracy": 0.8776358366012573, "num_tokens": 787387591.0, "step": 20734 }, { "epoch": 2.637705126574227, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.102875709533691, "learning_rate": 5e-05, "loss": 0.8716, "mean_token_accuracy": 0.8771278262138367, "num_tokens": 787427762.0, "step": 20735 }, { "epoch": 2.6378323368528176, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.448760986328125, "learning_rate": 5e-05, "loss": 0.9783, "mean_token_accuracy": 0.8683561682701111, "num_tokens": 787463047.0, "step": 20736 }, { "epoch": 2.637959547131408, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.259662628173828, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8910425901412964, "num_tokens": 787500862.0, "step": 20737 }, { "epoch": 2.6380867574099987, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.653423309326172, "learning_rate": 5e-05, "loss": 0.9106, "mean_token_accuracy": 0.8885335922241211, "num_tokens": 787529625.0, "step": 20738 }, { "epoch": 2.638213967688589, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.211669921875, "learning_rate": 5e-05, "loss": 0.7878, "mean_token_accuracy": 0.8838970065116882, "num_tokens": 787567248.0, "step": 20739 }, { "epoch": 2.6383411779671797, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.364014625549316, "learning_rate": 5e-05, "loss": 0.8886, "mean_token_accuracy": 0.8750099539756775, "num_tokens": 787606763.0, "step": 20740 }, { "epoch": 2.6384683882457702, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.426067352294922, "learning_rate": 5e-05, "loss": 0.8309, "mean_token_accuracy": 0.8775943517684937, "num_tokens": 787638795.0, "step": 20741 }, { "epoch": 2.6385955985243608, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.992137432098389, "learning_rate": 5e-05, "loss": 0.8376, "mean_token_accuracy": 0.8735097646713257, "num_tokens": 787679495.0, "step": 20742 }, { "epoch": 2.6387228088029513, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.500455856323242, "learning_rate": 5e-05, "loss": 0.7984, "mean_token_accuracy": 0.8819056749343872, "num_tokens": 787718583.0, "step": 20743 }, { "epoch": 2.638850019081542, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.803011894226074, "learning_rate": 5e-05, "loss": 0.7362, "mean_token_accuracy": 0.8955515623092651, "num_tokens": 787756558.0, "step": 20744 }, { "epoch": 2.6389772293601323, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.457780838012695, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.8817583918571472, "num_tokens": 787789796.0, "step": 20745 }, { "epoch": 2.639104439638723, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.786647319793701, "learning_rate": 5e-05, "loss": 0.8052, "mean_token_accuracy": 0.8705800175666809, "num_tokens": 787832054.0, "step": 20746 }, { "epoch": 2.6392316499173134, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.358033180236816, "learning_rate": 5e-05, "loss": 0.7533, "mean_token_accuracy": 0.8784345984458923, "num_tokens": 787876723.0, "step": 20747 }, { "epoch": 2.639358860195904, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.791067600250244, "learning_rate": 5e-05, "loss": 0.7943, "mean_token_accuracy": 0.8676780462265015, "num_tokens": 787912183.0, "step": 20748 }, { "epoch": 2.6394860704744945, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.294796943664551, "learning_rate": 5e-05, "loss": 0.7683, "mean_token_accuracy": 0.8681154251098633, "num_tokens": 787946865.0, "step": 20749 }, { "epoch": 2.639613280753085, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.7942962646484375, "learning_rate": 5e-05, "loss": 0.7384, "mean_token_accuracy": 0.8894640803337097, "num_tokens": 787984286.0, "step": 20750 }, { "epoch": 2.6397404910316755, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.224582672119141, "learning_rate": 5e-05, "loss": 0.6984, "mean_token_accuracy": 0.8920167088508606, "num_tokens": 788020565.0, "step": 20751 }, { "epoch": 2.6398677013102656, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.87227201461792, "learning_rate": 5e-05, "loss": 0.7435, "mean_token_accuracy": 0.8878794312477112, "num_tokens": 788051833.0, "step": 20752 }, { "epoch": 2.6399949115888566, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.070361137390137, "learning_rate": 5e-05, "loss": 0.6664, "mean_token_accuracy": 0.892134964466095, "num_tokens": 788092780.0, "step": 20753 }, { "epoch": 2.6401221218674467, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.07043170928955, "learning_rate": 5e-05, "loss": 0.7879, "mean_token_accuracy": 0.8722090125083923, "num_tokens": 788140088.0, "step": 20754 }, { "epoch": 2.6402493321460376, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.823335647583008, "learning_rate": 5e-05, "loss": 0.6815, "mean_token_accuracy": 0.8847204446792603, "num_tokens": 788177951.0, "step": 20755 }, { "epoch": 2.6403765424246277, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.40742301940918, "learning_rate": 5e-05, "loss": 0.7861, "mean_token_accuracy": 0.8865193724632263, "num_tokens": 788212124.0, "step": 20756 }, { "epoch": 2.6405037527032187, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.4296417236328125, "learning_rate": 5e-05, "loss": 0.6779, "mean_token_accuracy": 0.8776025176048279, "num_tokens": 788255860.0, "step": 20757 }, { "epoch": 2.6406309629818088, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 9.076910018920898, "learning_rate": 5e-05, "loss": 0.8354, "mean_token_accuracy": 0.8883062601089478, "num_tokens": 788296481.0, "step": 20758 }, { "epoch": 2.6407581732603993, "ewc_loss": 0.326171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 5.9225006103515625, "learning_rate": 5e-05, "loss": 0.6324, "mean_token_accuracy": 0.8897871971130371, "num_tokens": 788328818.0, "step": 20759 }, { "epoch": 2.64088538353899, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.998064041137695, "learning_rate": 5e-05, "loss": 0.9243, "mean_token_accuracy": 0.8990035057067871, "num_tokens": 788370359.0, "step": 20760 }, { "epoch": 2.6410125938175804, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.032681941986084, "learning_rate": 5e-05, "loss": 0.7195, "mean_token_accuracy": 0.8797063827514648, "num_tokens": 788412425.0, "step": 20761 }, { "epoch": 2.641139804096171, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 10.672183990478516, "learning_rate": 5e-05, "loss": 1.1572, "mean_token_accuracy": 0.8651107549667358, "num_tokens": 788445137.0, "step": 20762 }, { "epoch": 2.6412670143747614, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 7.701144218444824, "learning_rate": 5e-05, "loss": 0.9706, "mean_token_accuracy": 0.8880645036697388, "num_tokens": 788478185.0, "step": 20763 }, { "epoch": 2.641394224653352, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.787425994873047, "learning_rate": 5e-05, "loss": 1.0279, "mean_token_accuracy": 0.8774857521057129, "num_tokens": 788514329.0, "step": 20764 }, { "epoch": 2.6415214349319425, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 8.96910285949707, "learning_rate": 5e-05, "loss": 1.1309, "mean_token_accuracy": 0.8841404914855957, "num_tokens": 788554234.0, "step": 20765 }, { "epoch": 2.641648645210533, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.818397521972656, "learning_rate": 5e-05, "loss": 0.8212, "mean_token_accuracy": 0.8820600509643555, "num_tokens": 788591684.0, "step": 20766 }, { "epoch": 2.6417758554891235, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 10.02283763885498, "learning_rate": 5e-05, "loss": 1.1, "mean_token_accuracy": 0.8811382055282593, "num_tokens": 788626802.0, "step": 20767 }, { "epoch": 2.641903065767714, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 6.991887092590332, "learning_rate": 5e-05, "loss": 0.8214, "mean_token_accuracy": 0.8837526440620422, "num_tokens": 788673826.0, "step": 20768 }, { "epoch": 2.6420302760463046, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 10.926950454711914, "learning_rate": 5e-05, "loss": 1.1204, "mean_token_accuracy": 0.8889564871788025, "num_tokens": 788708393.0, "step": 20769 }, { "epoch": 2.642157486324895, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 8.303452491760254, "learning_rate": 5e-05, "loss": 1.043, "mean_token_accuracy": 0.8866949081420898, "num_tokens": 788746480.0, "step": 20770 }, { "epoch": 2.6422846966034856, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.22506332397461, "learning_rate": 5e-05, "loss": 1.0337, "mean_token_accuracy": 0.8783913254737854, "num_tokens": 788785295.0, "step": 20771 }, { "epoch": 2.642411906882076, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 8.910113334655762, "learning_rate": 5e-05, "loss": 1.1243, "mean_token_accuracy": 0.886154055595398, "num_tokens": 788818803.0, "step": 20772 }, { "epoch": 2.6425391171606667, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.058206558227539, "learning_rate": 5e-05, "loss": 0.9562, "mean_token_accuracy": 0.876439094543457, "num_tokens": 788857266.0, "step": 20773 }, { "epoch": 2.6426663274392572, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.75629997253418, "learning_rate": 5e-05, "loss": 1.05, "mean_token_accuracy": 0.8612724542617798, "num_tokens": 788898433.0, "step": 20774 }, { "epoch": 2.6427935377178478, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.320533752441406, "learning_rate": 5e-05, "loss": 0.947, "mean_token_accuracy": 0.8743748664855957, "num_tokens": 788942259.0, "step": 20775 }, { "epoch": 2.6429207479964383, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.375213623046875, "learning_rate": 5e-05, "loss": 0.952, "mean_token_accuracy": 0.8705698847770691, "num_tokens": 788981484.0, "step": 20776 }, { "epoch": 2.6430479582750284, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.101667404174805, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8818891048431396, "num_tokens": 789019863.0, "step": 20777 }, { "epoch": 2.6431751685536193, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.784745216369629, "learning_rate": 5e-05, "loss": 0.8156, "mean_token_accuracy": 0.8820429444313049, "num_tokens": 789059139.0, "step": 20778 }, { "epoch": 2.6433023788322094, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.207537651062012, "learning_rate": 5e-05, "loss": 0.847, "mean_token_accuracy": 0.8848292827606201, "num_tokens": 789100429.0, "step": 20779 }, { "epoch": 2.6434295891108004, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.272481918334961, "learning_rate": 5e-05, "loss": 0.756, "mean_token_accuracy": 0.8825548887252808, "num_tokens": 789136219.0, "step": 20780 }, { "epoch": 2.6435567993893905, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.462379455566406, "learning_rate": 5e-05, "loss": 0.8552, "mean_token_accuracy": 0.8836669921875, "num_tokens": 789177541.0, "step": 20781 }, { "epoch": 2.6436840096679814, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.850785255432129, "learning_rate": 5e-05, "loss": 0.6764, "mean_token_accuracy": 0.8975754976272583, "num_tokens": 789217165.0, "step": 20782 }, { "epoch": 2.6438112199465715, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.926486015319824, "learning_rate": 5e-05, "loss": 0.8828, "mean_token_accuracy": 0.8841204047203064, "num_tokens": 789255384.0, "step": 20783 }, { "epoch": 2.643938430225162, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.5675764083862305, "learning_rate": 5e-05, "loss": 0.7019, "mean_token_accuracy": 0.8850290775299072, "num_tokens": 789294153.0, "step": 20784 }, { "epoch": 2.6440656405037526, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.424166679382324, "learning_rate": 5e-05, "loss": 1.0145, "mean_token_accuracy": 0.8631256818771362, "num_tokens": 789334933.0, "step": 20785 }, { "epoch": 2.644192850782343, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 6.77043342590332, "learning_rate": 5e-05, "loss": 0.848, "mean_token_accuracy": 0.8628758788108826, "num_tokens": 789370068.0, "step": 20786 }, { "epoch": 2.6443200610609336, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.409911155700684, "learning_rate": 5e-05, "loss": 0.9555, "mean_token_accuracy": 0.8852822780609131, "num_tokens": 789413887.0, "step": 20787 }, { "epoch": 2.644447271339524, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 7.454572677612305, "learning_rate": 5e-05, "loss": 0.8754, "mean_token_accuracy": 0.8904492855072021, "num_tokens": 789452319.0, "step": 20788 }, { "epoch": 2.6445744816181147, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.72504997253418, "learning_rate": 5e-05, "loss": 0.8891, "mean_token_accuracy": 0.8894393444061279, "num_tokens": 789488988.0, "step": 20789 }, { "epoch": 2.6447016918967052, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 7.9398956298828125, "learning_rate": 5e-05, "loss": 0.9601, "mean_token_accuracy": 0.873665452003479, "num_tokens": 789529849.0, "step": 20790 }, { "epoch": 2.6448289021752958, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.921120643615723, "learning_rate": 5e-05, "loss": 0.8301, "mean_token_accuracy": 0.8781170845031738, "num_tokens": 789572877.0, "step": 20791 }, { "epoch": 2.6449561124538863, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.123174667358398, "learning_rate": 5e-05, "loss": 0.9252, "mean_token_accuracy": 0.8689281344413757, "num_tokens": 789605189.0, "step": 20792 }, { "epoch": 2.645083322732477, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.442048072814941, "learning_rate": 5e-05, "loss": 0.7655, "mean_token_accuracy": 0.8865969181060791, "num_tokens": 789641589.0, "step": 20793 }, { "epoch": 2.6452105330110673, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.164632797241211, "learning_rate": 5e-05, "loss": 0.8727, "mean_token_accuracy": 0.8703217506408691, "num_tokens": 789673869.0, "step": 20794 }, { "epoch": 2.645337743289658, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.303216934204102, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8807782530784607, "num_tokens": 789710605.0, "step": 20795 }, { "epoch": 2.6454649535682484, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.061776161193848, "learning_rate": 5e-05, "loss": 0.7875, "mean_token_accuracy": 0.8888098001480103, "num_tokens": 789747641.0, "step": 20796 }, { "epoch": 2.645592163846839, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.317906379699707, "learning_rate": 5e-05, "loss": 0.7172, "mean_token_accuracy": 0.8942089080810547, "num_tokens": 789781345.0, "step": 20797 }, { "epoch": 2.6457193741254295, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.993617057800293, "learning_rate": 5e-05, "loss": 0.8018, "mean_token_accuracy": 0.8778643012046814, "num_tokens": 789819582.0, "step": 20798 }, { "epoch": 2.64584658440402, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.220992565155029, "learning_rate": 5e-05, "loss": 0.7035, "mean_token_accuracy": 0.8907157182693481, "num_tokens": 789856281.0, "step": 20799 }, { "epoch": 2.6459737946826105, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.077661514282227, "learning_rate": 5e-05, "loss": 0.7563, "mean_token_accuracy": 0.8896053433418274, "num_tokens": 789895711.0, "step": 20800 }, { "epoch": 2.646101004961201, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.076303958892822, "learning_rate": 5e-05, "loss": 0.695, "mean_token_accuracy": 0.8932746052742004, "num_tokens": 789931751.0, "step": 20801 }, { "epoch": 2.646228215239791, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.152748107910156, "learning_rate": 5e-05, "loss": 0.8093, "mean_token_accuracy": 0.8759834170341492, "num_tokens": 789969207.0, "step": 20802 }, { "epoch": 2.646355425518382, "ewc_loss": 0.375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.000533103942871, "learning_rate": 5e-05, "loss": 0.7224, "mean_token_accuracy": 0.8778815865516663, "num_tokens": 790005706.0, "step": 20803 }, { "epoch": 2.646482635796972, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.252028465270996, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8773089647293091, "num_tokens": 790047389.0, "step": 20804 }, { "epoch": 2.646609846075563, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.839661598205566, "learning_rate": 5e-05, "loss": 0.7124, "mean_token_accuracy": 0.8818742632865906, "num_tokens": 790085430.0, "step": 20805 }, { "epoch": 2.6467370563541532, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.547795295715332, "learning_rate": 5e-05, "loss": 0.8486, "mean_token_accuracy": 0.8771048784255981, "num_tokens": 790126620.0, "step": 20806 }, { "epoch": 2.6468642666327438, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.581861972808838, "learning_rate": 5e-05, "loss": 0.6828, "mean_token_accuracy": 0.8839635252952576, "num_tokens": 790166033.0, "step": 20807 }, { "epoch": 2.6469914769113343, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.912433624267578, "learning_rate": 5e-05, "loss": 0.8446, "mean_token_accuracy": 0.8902365565299988, "num_tokens": 790202463.0, "step": 20808 }, { "epoch": 2.647118687189925, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.518617630004883, "learning_rate": 5e-05, "loss": 0.6945, "mean_token_accuracy": 0.8851515650749207, "num_tokens": 790241600.0, "step": 20809 }, { "epoch": 2.6472458974685154, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.110640525817871, "learning_rate": 5e-05, "loss": 0.9793, "mean_token_accuracy": 0.8593863248825073, "num_tokens": 790277669.0, "step": 20810 }, { "epoch": 2.647373107747106, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.861476421356201, "learning_rate": 5e-05, "loss": 0.7839, "mean_token_accuracy": 0.8727555274963379, "num_tokens": 790317453.0, "step": 20811 }, { "epoch": 2.6475003180256964, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.73729133605957, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.8816303014755249, "num_tokens": 790354460.0, "step": 20812 }, { "epoch": 2.647627528304287, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.562768936157227, "learning_rate": 5e-05, "loss": 0.8506, "mean_token_accuracy": 0.8809698820114136, "num_tokens": 790393933.0, "step": 20813 }, { "epoch": 2.6477547385828775, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.776526927947998, "learning_rate": 5e-05, "loss": 0.8043, "mean_token_accuracy": 0.8818086385726929, "num_tokens": 790425477.0, "step": 20814 }, { "epoch": 2.647881948861468, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.250414848327637, "learning_rate": 5e-05, "loss": 0.9197, "mean_token_accuracy": 0.8599890470504761, "num_tokens": 790470501.0, "step": 20815 }, { "epoch": 2.6480091591400585, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.859493255615234, "learning_rate": 5e-05, "loss": 0.7305, "mean_token_accuracy": 0.8758867979049683, "num_tokens": 790508869.0, "step": 20816 }, { "epoch": 2.648136369418649, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.849759101867676, "learning_rate": 5e-05, "loss": 0.8417, "mean_token_accuracy": 0.8954880833625793, "num_tokens": 790545675.0, "step": 20817 }, { "epoch": 2.6482635796972396, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.480649948120117, "learning_rate": 5e-05, "loss": 0.6946, "mean_token_accuracy": 0.8842031955718994, "num_tokens": 790586832.0, "step": 20818 }, { "epoch": 2.64839078997583, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.1696195602417, "learning_rate": 5e-05, "loss": 0.9567, "mean_token_accuracy": 0.878176748752594, "num_tokens": 790633587.0, "step": 20819 }, { "epoch": 2.6485180002544206, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.732054233551025, "learning_rate": 5e-05, "loss": 0.7871, "mean_token_accuracy": 0.8721657991409302, "num_tokens": 790668768.0, "step": 20820 }, { "epoch": 2.648645210533011, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.946891784667969, "learning_rate": 5e-05, "loss": 0.9129, "mean_token_accuracy": 0.8880959749221802, "num_tokens": 790708393.0, "step": 20821 }, { "epoch": 2.6487724208116017, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.374983787536621, "learning_rate": 5e-05, "loss": 0.8856, "mean_token_accuracy": 0.871121883392334, "num_tokens": 790742632.0, "step": 20822 }, { "epoch": 2.648899631090192, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.224160194396973, "learning_rate": 5e-05, "loss": 0.9103, "mean_token_accuracy": 0.8685301542282104, "num_tokens": 790779740.0, "step": 20823 }, { "epoch": 2.6490268413687827, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.863392353057861, "learning_rate": 5e-05, "loss": 0.8426, "mean_token_accuracy": 0.8890501260757446, "num_tokens": 790819040.0, "step": 20824 }, { "epoch": 2.649154051647373, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.61000394821167, "learning_rate": 5e-05, "loss": 0.7559, "mean_token_accuracy": 0.8914132118225098, "num_tokens": 790853382.0, "step": 20825 }, { "epoch": 2.649281261925964, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.9044928550720215, "learning_rate": 5e-05, "loss": 0.8499, "mean_token_accuracy": 0.8712956309318542, "num_tokens": 790890606.0, "step": 20826 }, { "epoch": 2.649408472204554, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.613827228546143, "learning_rate": 5e-05, "loss": 0.774, "mean_token_accuracy": 0.8870188593864441, "num_tokens": 790929469.0, "step": 20827 }, { "epoch": 2.649535682483145, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.460732460021973, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8796250224113464, "num_tokens": 790970745.0, "step": 20828 }, { "epoch": 2.649662892761735, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.00037670135498, "learning_rate": 5e-05, "loss": 0.831, "mean_token_accuracy": 0.8707779049873352, "num_tokens": 791010365.0, "step": 20829 }, { "epoch": 2.649790103040326, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.98561954498291, "learning_rate": 5e-05, "loss": 0.7187, "mean_token_accuracy": 0.8827707171440125, "num_tokens": 791047012.0, "step": 20830 }, { "epoch": 2.649917313318916, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.396145820617676, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.8840593099594116, "num_tokens": 791092836.0, "step": 20831 }, { "epoch": 2.6500445235975065, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.648481845855713, "learning_rate": 5e-05, "loss": 0.6708, "mean_token_accuracy": 0.8911410570144653, "num_tokens": 791127725.0, "step": 20832 }, { "epoch": 2.650171733876097, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.775243759155273, "learning_rate": 5e-05, "loss": 0.8364, "mean_token_accuracy": 0.8889822363853455, "num_tokens": 791164530.0, "step": 20833 }, { "epoch": 2.6502989441546876, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.453607559204102, "learning_rate": 5e-05, "loss": 0.6569, "mean_token_accuracy": 0.893929123878479, "num_tokens": 791202384.0, "step": 20834 }, { "epoch": 2.650426154433278, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.150330543518066, "learning_rate": 5e-05, "loss": 0.8805, "mean_token_accuracy": 0.8946155309677124, "num_tokens": 791237998.0, "step": 20835 }, { "epoch": 2.6505533647118686, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.553575038909912, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8593177795410156, "num_tokens": 791277755.0, "step": 20836 }, { "epoch": 2.650680574990459, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.226079940795898, "learning_rate": 5e-05, "loss": 0.9856, "mean_token_accuracy": 0.8695294857025146, "num_tokens": 791315779.0, "step": 20837 }, { "epoch": 2.6508077852690497, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.137871265411377, "learning_rate": 5e-05, "loss": 0.8503, "mean_token_accuracy": 0.8750782012939453, "num_tokens": 791359212.0, "step": 20838 }, { "epoch": 2.6509349955476402, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.679564476013184, "learning_rate": 5e-05, "loss": 0.8743, "mean_token_accuracy": 0.8964685201644897, "num_tokens": 791388308.0, "step": 20839 }, { "epoch": 2.6510622058262308, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 7.833387851715088, "learning_rate": 5e-05, "loss": 0.9163, "mean_token_accuracy": 0.8759720325469971, "num_tokens": 791425645.0, "step": 20840 }, { "epoch": 2.6511894161048213, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.705592632293701, "learning_rate": 5e-05, "loss": 0.8137, "mean_token_accuracy": 0.8792929649353027, "num_tokens": 791462632.0, "step": 20841 }, { "epoch": 2.651316626383412, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.254096984863281, "learning_rate": 5e-05, "loss": 0.8226, "mean_token_accuracy": 0.8979607820510864, "num_tokens": 791502834.0, "step": 20842 }, { "epoch": 2.6514438366620023, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.1439080238342285, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8774848580360413, "num_tokens": 791538973.0, "step": 20843 }, { "epoch": 2.651571046940593, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.339006423950195, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8808521032333374, "num_tokens": 791577804.0, "step": 20844 }, { "epoch": 2.6516982572191834, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.040563106536865, "learning_rate": 5e-05, "loss": 0.6928, "mean_token_accuracy": 0.8929978609085083, "num_tokens": 791618100.0, "step": 20845 }, { "epoch": 2.651825467497774, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.326401710510254, "learning_rate": 5e-05, "loss": 0.8666, "mean_token_accuracy": 0.8767290115356445, "num_tokens": 791660112.0, "step": 20846 }, { "epoch": 2.6519526777763645, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.965975284576416, "learning_rate": 5e-05, "loss": 0.7364, "mean_token_accuracy": 0.8778462409973145, "num_tokens": 791698133.0, "step": 20847 }, { "epoch": 2.652079888054955, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.511521339416504, "learning_rate": 5e-05, "loss": 0.8522, "mean_token_accuracy": 0.8869086503982544, "num_tokens": 791737810.0, "step": 20848 }, { "epoch": 2.6522070983335455, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.784805774688721, "learning_rate": 5e-05, "loss": 0.7331, "mean_token_accuracy": 0.8766171932220459, "num_tokens": 791775187.0, "step": 20849 }, { "epoch": 2.6523343086121356, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.831811904907227, "learning_rate": 5e-05, "loss": 0.884, "mean_token_accuracy": 0.8869434595108032, "num_tokens": 791810673.0, "step": 20850 }, { "epoch": 2.6524615188907266, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.745373249053955, "learning_rate": 5e-05, "loss": 0.6806, "mean_token_accuracy": 0.9014830589294434, "num_tokens": 791848144.0, "step": 20851 }, { "epoch": 2.6525887291693167, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.921785354614258, "learning_rate": 5e-05, "loss": 0.9234, "mean_token_accuracy": 0.8812688589096069, "num_tokens": 791885025.0, "step": 20852 }, { "epoch": 2.6527159394479076, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.05508279800415, "learning_rate": 5e-05, "loss": 0.8038, "mean_token_accuracy": 0.8787679076194763, "num_tokens": 791922885.0, "step": 20853 }, { "epoch": 2.6528431497264977, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.598427772521973, "learning_rate": 5e-05, "loss": 0.9652, "mean_token_accuracy": 0.8607975244522095, "num_tokens": 791959791.0, "step": 20854 }, { "epoch": 2.6529703600050887, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.484216213226318, "learning_rate": 5e-05, "loss": 0.8427, "mean_token_accuracy": 0.8810277581214905, "num_tokens": 791998559.0, "step": 20855 }, { "epoch": 2.6530975702836788, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.057936668395996, "learning_rate": 5e-05, "loss": 0.8573, "mean_token_accuracy": 0.8803211450576782, "num_tokens": 792038428.0, "step": 20856 }, { "epoch": 2.6532247805622693, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.719386577606201, "learning_rate": 5e-05, "loss": 0.8851, "mean_token_accuracy": 0.869451642036438, "num_tokens": 792074405.0, "step": 20857 }, { "epoch": 2.65335199084086, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.727277755737305, "learning_rate": 5e-05, "loss": 0.8275, "mean_token_accuracy": 0.871178388595581, "num_tokens": 792112831.0, "step": 20858 }, { "epoch": 2.6534792011194503, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.600010395050049, "learning_rate": 5e-05, "loss": 0.8138, "mean_token_accuracy": 0.8757057785987854, "num_tokens": 792154122.0, "step": 20859 }, { "epoch": 2.653606411398041, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.8619256019592285, "learning_rate": 5e-05, "loss": 0.8374, "mean_token_accuracy": 0.8644817471504211, "num_tokens": 792191818.0, "step": 20860 }, { "epoch": 2.6537336216766314, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.201741695404053, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8807006478309631, "num_tokens": 792226753.0, "step": 20861 }, { "epoch": 2.653860831955222, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.258785247802734, "learning_rate": 5e-05, "loss": 0.811, "mean_token_accuracy": 0.8812926411628723, "num_tokens": 792263941.0, "step": 20862 }, { "epoch": 2.6539880422338125, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.804744243621826, "learning_rate": 5e-05, "loss": 0.7013, "mean_token_accuracy": 0.8840916752815247, "num_tokens": 792301433.0, "step": 20863 }, { "epoch": 2.654115252512403, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.612170219421387, "learning_rate": 5e-05, "loss": 0.8594, "mean_token_accuracy": 0.880912184715271, "num_tokens": 792340188.0, "step": 20864 }, { "epoch": 2.6542424627909935, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.654924392700195, "learning_rate": 5e-05, "loss": 0.7248, "mean_token_accuracy": 0.8759526014328003, "num_tokens": 792383498.0, "step": 20865 }, { "epoch": 2.654369673069584, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.794116973876953, "learning_rate": 5e-05, "loss": 0.8539, "mean_token_accuracy": 0.8903852701187134, "num_tokens": 792417160.0, "step": 20866 }, { "epoch": 2.6544968833481746, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.825376510620117, "learning_rate": 5e-05, "loss": 0.7512, "mean_token_accuracy": 0.876658022403717, "num_tokens": 792451609.0, "step": 20867 }, { "epoch": 2.654624093626765, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.58808422088623, "learning_rate": 5e-05, "loss": 0.9308, "mean_token_accuracy": 0.8697792887687683, "num_tokens": 792498309.0, "step": 20868 }, { "epoch": 2.6547513039053556, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.319633483886719, "learning_rate": 5e-05, "loss": 0.8394, "mean_token_accuracy": 0.870644211769104, "num_tokens": 792538954.0, "step": 20869 }, { "epoch": 2.654878514183946, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.8918657302856445, "learning_rate": 5e-05, "loss": 0.8562, "mean_token_accuracy": 0.8654215931892395, "num_tokens": 792577037.0, "step": 20870 }, { "epoch": 2.6550057244625367, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.896595478057861, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.866775393486023, "num_tokens": 792609538.0, "step": 20871 }, { "epoch": 2.655132934741127, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.12772274017334, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.8799495697021484, "num_tokens": 792647793.0, "step": 20872 }, { "epoch": 2.6552601450197177, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.388850212097168, "learning_rate": 5e-05, "loss": 0.8191, "mean_token_accuracy": 0.8832449913024902, "num_tokens": 792686746.0, "step": 20873 }, { "epoch": 2.6553873552983083, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.65983772277832, "learning_rate": 5e-05, "loss": 0.68, "mean_token_accuracy": 0.8823771476745605, "num_tokens": 792716324.0, "step": 20874 }, { "epoch": 2.6555145655768984, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.744178771972656, "learning_rate": 5e-05, "loss": 0.8453, "mean_token_accuracy": 0.8854817152023315, "num_tokens": 792757242.0, "step": 20875 }, { "epoch": 2.6556417758554893, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.483951091766357, "learning_rate": 5e-05, "loss": 0.7113, "mean_token_accuracy": 0.8759141564369202, "num_tokens": 792795545.0, "step": 20876 }, { "epoch": 2.6557689861340794, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.982100486755371, "learning_rate": 5e-05, "loss": 0.917, "mean_token_accuracy": 0.8765016794204712, "num_tokens": 792835772.0, "step": 20877 }, { "epoch": 2.6558961964126704, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.6336140632629395, "learning_rate": 5e-05, "loss": 0.7588, "mean_token_accuracy": 0.8697152137756348, "num_tokens": 792877901.0, "step": 20878 }, { "epoch": 2.6560234066912605, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.902298927307129, "learning_rate": 5e-05, "loss": 0.9285, "mean_token_accuracy": 0.8785682320594788, "num_tokens": 792918528.0, "step": 20879 }, { "epoch": 2.6561506169698514, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.131088733673096, "learning_rate": 5e-05, "loss": 0.7508, "mean_token_accuracy": 0.8920928239822388, "num_tokens": 792954251.0, "step": 20880 }, { "epoch": 2.6562778272484415, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.315857887268066, "learning_rate": 5e-05, "loss": 0.884, "mean_token_accuracy": 0.8837070465087891, "num_tokens": 792991452.0, "step": 20881 }, { "epoch": 2.656405037527032, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.703742027282715, "learning_rate": 5e-05, "loss": 0.8634, "mean_token_accuracy": 0.873841404914856, "num_tokens": 793024488.0, "step": 20882 }, { "epoch": 2.6565322478056226, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.610832691192627, "learning_rate": 5e-05, "loss": 0.7807, "mean_token_accuracy": 0.8874458074569702, "num_tokens": 793064070.0, "step": 20883 }, { "epoch": 2.656659458084213, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.9638848304748535, "learning_rate": 5e-05, "loss": 0.856, "mean_token_accuracy": 0.8707250356674194, "num_tokens": 793103088.0, "step": 20884 }, { "epoch": 2.6567866683628036, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.399939060211182, "learning_rate": 5e-05, "loss": 0.806, "mean_token_accuracy": 0.8684287667274475, "num_tokens": 793146415.0, "step": 20885 }, { "epoch": 2.656913878641394, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.725972652435303, "learning_rate": 5e-05, "loss": 0.79, "mean_token_accuracy": 0.880703866481781, "num_tokens": 793181588.0, "step": 20886 }, { "epoch": 2.6570410889199847, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.680964946746826, "learning_rate": 5e-05, "loss": 0.8117, "mean_token_accuracy": 0.8638530969619751, "num_tokens": 793217764.0, "step": 20887 }, { "epoch": 2.6571682991985752, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.214316368103027, "learning_rate": 5e-05, "loss": 0.7128, "mean_token_accuracy": 0.8846573829650879, "num_tokens": 793255945.0, "step": 20888 }, { "epoch": 2.6572955094771658, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.143628120422363, "learning_rate": 5e-05, "loss": 0.8213, "mean_token_accuracy": 0.8767540454864502, "num_tokens": 793288981.0, "step": 20889 }, { "epoch": 2.6574227197557563, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.76023006439209, "learning_rate": 5e-05, "loss": 0.6893, "mean_token_accuracy": 0.8827643394470215, "num_tokens": 793327635.0, "step": 20890 }, { "epoch": 2.657549930034347, "ewc_loss": 0.5, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.591887474060059, "learning_rate": 5e-05, "loss": 0.8349, "mean_token_accuracy": 0.8818169236183167, "num_tokens": 793357077.0, "step": 20891 }, { "epoch": 2.6576771403129373, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.471528053283691, "learning_rate": 5e-05, "loss": 0.6477, "mean_token_accuracy": 0.894911527633667, "num_tokens": 793389664.0, "step": 20892 }, { "epoch": 2.657804350591528, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.953422546386719, "learning_rate": 5e-05, "loss": 0.8512, "mean_token_accuracy": 0.8926047086715698, "num_tokens": 793429951.0, "step": 20893 }, { "epoch": 2.6579315608701184, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.447421073913574, "learning_rate": 5e-05, "loss": 0.736, "mean_token_accuracy": 0.869342029094696, "num_tokens": 793471404.0, "step": 20894 }, { "epoch": 2.658058771148709, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.107511520385742, "learning_rate": 5e-05, "loss": 0.9437, "mean_token_accuracy": 0.8768740892410278, "num_tokens": 793510140.0, "step": 20895 }, { "epoch": 2.6581859814272994, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.809118270874023, "learning_rate": 5e-05, "loss": 0.7671, "mean_token_accuracy": 0.8813289999961853, "num_tokens": 793553927.0, "step": 20896 }, { "epoch": 2.65831319170589, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.783612251281738, "learning_rate": 5e-05, "loss": 0.9205, "mean_token_accuracy": 0.885587215423584, "num_tokens": 793590413.0, "step": 20897 }, { "epoch": 2.6584404019844805, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.452491283416748, "learning_rate": 5e-05, "loss": 0.9041, "mean_token_accuracy": 0.8608112931251526, "num_tokens": 793626320.0, "step": 20898 }, { "epoch": 2.658567612263071, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.9798150062561035, "learning_rate": 5e-05, "loss": 0.8359, "mean_token_accuracy": 0.8821569681167603, "num_tokens": 793665733.0, "step": 20899 }, { "epoch": 2.658694822541661, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.987335205078125, "learning_rate": 5e-05, "loss": 0.8573, "mean_token_accuracy": 0.8811664581298828, "num_tokens": 793705593.0, "step": 20900 }, { "epoch": 2.658822032820252, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.262380123138428, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8841678500175476, "num_tokens": 793744874.0, "step": 20901 }, { "epoch": 2.658949243098842, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.247674942016602, "learning_rate": 5e-05, "loss": 0.8882, "mean_token_accuracy": 0.8665302395820618, "num_tokens": 793786598.0, "step": 20902 }, { "epoch": 2.659076453377433, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.0131683349609375, "learning_rate": 5e-05, "loss": 0.7249, "mean_token_accuracy": 0.879367470741272, "num_tokens": 793826548.0, "step": 20903 }, { "epoch": 2.6592036636560232, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.256433486938477, "learning_rate": 5e-05, "loss": 0.8804, "mean_token_accuracy": 0.8652936220169067, "num_tokens": 793868411.0, "step": 20904 }, { "epoch": 2.6593308739346138, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.957673072814941, "learning_rate": 5e-05, "loss": 0.7455, "mean_token_accuracy": 0.8746151924133301, "num_tokens": 793905931.0, "step": 20905 }, { "epoch": 2.6594580842132043, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.36844253540039, "learning_rate": 5e-05, "loss": 0.8476, "mean_token_accuracy": 0.8783976435661316, "num_tokens": 793939711.0, "step": 20906 }, { "epoch": 2.659585294491795, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.773440361022949, "learning_rate": 5e-05, "loss": 0.7198, "mean_token_accuracy": 0.8816015720367432, "num_tokens": 793976746.0, "step": 20907 }, { "epoch": 2.6597125047703853, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.733641624450684, "learning_rate": 5e-05, "loss": 0.9361, "mean_token_accuracy": 0.8629366159439087, "num_tokens": 794014132.0, "step": 20908 }, { "epoch": 2.659839715048976, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000774383544921875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 15.117121696472168, "learning_rate": 5e-05, "loss": 1.2359, "mean_token_accuracy": 0.841741681098938, "num_tokens": 794051958.0, "step": 20909 }, { "epoch": 2.6599669253275664, "ewc_loss": 0.5, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.38680648803711, "learning_rate": 5e-05, "loss": 0.8241, "mean_token_accuracy": 0.8820118308067322, "num_tokens": 794087494.0, "step": 20910 }, { "epoch": 2.660094135606157, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 12.202727317810059, "learning_rate": 5e-05, "loss": 1.179, "mean_token_accuracy": 0.8685851097106934, "num_tokens": 794123156.0, "step": 20911 }, { "epoch": 2.6602213458847475, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.456587314605713, "learning_rate": 5e-05, "loss": 0.8759, "mean_token_accuracy": 0.8666654825210571, "num_tokens": 794169023.0, "step": 20912 }, { "epoch": 2.660348556163338, "ewc_loss": 1.4453125, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.00136566162109375, "grad_norm": 17.150766372680664, "learning_rate": 5e-05, "loss": 1.7496, "mean_token_accuracy": 0.8889927864074707, "num_tokens": 794202145.0, "step": 20913 }, { "epoch": 2.6604757664419285, "ewc_loss": 1.328125, "ewc_loss_diag": 0.000820159912109375, "ewc_loss_parallel": 0.001251220703125, "grad_norm": 14.84247875213623, "learning_rate": 5e-05, "loss": 1.6438, "mean_token_accuracy": 0.89200758934021, "num_tokens": 794239791.0, "step": 20914 }, { "epoch": 2.660602976720519, "ewc_loss": 0.90625, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.000823974609375, "grad_norm": 12.400333404541016, "learning_rate": 5e-05, "loss": 1.2323, "mean_token_accuracy": 0.8857567310333252, "num_tokens": 794285184.0, "step": 20915 }, { "epoch": 2.6607301869991096, "ewc_loss": 0.875, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 12.36309814453125, "learning_rate": 5e-05, "loss": 1.1824, "mean_token_accuracy": 0.8930602073669434, "num_tokens": 794324705.0, "step": 20916 }, { "epoch": 2.6608573972777, "ewc_loss": 0.9453125, "ewc_loss_diag": 0.00086212158203125, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 12.164031028747559, "learning_rate": 5e-05, "loss": 1.3072, "mean_token_accuracy": 0.8707417249679565, "num_tokens": 794360712.0, "step": 20917 }, { "epoch": 2.6609846075562906, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.0008697509765625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.709783554077148, "learning_rate": 5e-05, "loss": 1.0246, "mean_token_accuracy": 0.8824663758277893, "num_tokens": 794401634.0, "step": 20918 }, { "epoch": 2.661111817834881, "ewc_loss": 0.8828125, "ewc_loss_diag": 0.000873565673828125, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 12.433370590209961, "learning_rate": 5e-05, "loss": 1.2097, "mean_token_accuracy": 0.8837764263153076, "num_tokens": 794433460.0, "step": 20919 }, { "epoch": 2.6612390281134717, "ewc_loss": 0.703125, "ewc_loss_diag": 0.00087738037109375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.241411209106445, "learning_rate": 5e-05, "loss": 0.9948, "mean_token_accuracy": 0.8938993811607361, "num_tokens": 794468258.0, "step": 20920 }, { "epoch": 2.661366238392062, "ewc_loss": 0.921875, "ewc_loss_diag": 0.00087738037109375, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 13.103324890136719, "learning_rate": 5e-05, "loss": 1.2828, "mean_token_accuracy": 0.8750091791152954, "num_tokens": 794505493.0, "step": 20921 }, { "epoch": 2.6614934486706527, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000885009765625, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.514711380004883, "learning_rate": 5e-05, "loss": 0.9879, "mean_token_accuracy": 0.900850236415863, "num_tokens": 794539370.0, "step": 20922 }, { "epoch": 2.661620658949243, "ewc_loss": 0.984375, "ewc_loss_diag": 0.000888824462890625, "ewc_loss_parallel": 0.00089263916015625, "grad_norm": 13.729715347290039, "learning_rate": 5e-05, "loss": 1.3316, "mean_token_accuracy": 0.8767044544219971, "num_tokens": 794580820.0, "step": 20923 }, { "epoch": 2.661747869227834, "ewc_loss": 0.8125, "ewc_loss_diag": 0.0009002685546875, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 11.7568941116333, "learning_rate": 5e-05, "loss": 1.1384, "mean_token_accuracy": 0.8863677978515625, "num_tokens": 794617503.0, "step": 20924 }, { "epoch": 2.661875079506424, "ewc_loss": 0.86328125, "ewc_loss_diag": 0.0009002685546875, "ewc_loss_parallel": 0.000774383544921875, "grad_norm": 12.845197677612305, "learning_rate": 5e-05, "loss": 1.2064, "mean_token_accuracy": 0.8794286251068115, "num_tokens": 794653701.0, "step": 20925 }, { "epoch": 2.662002289785015, "ewc_loss": 0.765625, "ewc_loss_diag": 0.0009002685546875, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 11.675297737121582, "learning_rate": 5e-05, "loss": 1.1111, "mean_token_accuracy": 0.876309871673584, "num_tokens": 794689938.0, "step": 20926 }, { "epoch": 2.662129500063605, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.0009002685546875, "ewc_loss_parallel": 0.0006866455078125, "grad_norm": 12.107120513916016, "learning_rate": 5e-05, "loss": 1.1429, "mean_token_accuracy": 0.86934494972229, "num_tokens": 794730460.0, "step": 20927 }, { "epoch": 2.662256710342196, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000904083251953125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 10.74039363861084, "learning_rate": 5e-05, "loss": 0.9692, "mean_token_accuracy": 0.8884474039077759, "num_tokens": 794770686.0, "step": 20928 }, { "epoch": 2.662383920620786, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.000904083251953125, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 12.280983924865723, "learning_rate": 5e-05, "loss": 1.1636, "mean_token_accuracy": 0.8665429353713989, "num_tokens": 794814396.0, "step": 20929 }, { "epoch": 2.6625111308993765, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000904083251953125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 10.366777420043945, "learning_rate": 5e-05, "loss": 0.9553, "mean_token_accuracy": 0.8827104568481445, "num_tokens": 794857425.0, "step": 20930 }, { "epoch": 2.662638341177967, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000904083251953125, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 12.729721069335938, "learning_rate": 5e-05, "loss": 1.1158, "mean_token_accuracy": 0.8858159184455872, "num_tokens": 794894115.0, "step": 20931 }, { "epoch": 2.6627655514565576, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.0009002685546875, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 10.782546043395996, "learning_rate": 5e-05, "loss": 0.9978, "mean_token_accuracy": 0.8757878541946411, "num_tokens": 794929423.0, "step": 20932 }, { "epoch": 2.662892761735148, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.0009002685546875, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 12.656243324279785, "learning_rate": 5e-05, "loss": 1.1341, "mean_token_accuracy": 0.8790630102157593, "num_tokens": 794966019.0, "step": 20933 }, { "epoch": 2.6630199720137386, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.0009002685546875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 11.150164604187012, "learning_rate": 5e-05, "loss": 1.0106, "mean_token_accuracy": 0.8758642673492432, "num_tokens": 795006501.0, "step": 20934 }, { "epoch": 2.663147182292329, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.0009002685546875, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 12.07197380065918, "learning_rate": 5e-05, "loss": 1.0804, "mean_token_accuracy": 0.8812990188598633, "num_tokens": 795048068.0, "step": 20935 }, { "epoch": 2.6632743925709197, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.0009002685546875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 10.759422302246094, "learning_rate": 5e-05, "loss": 0.9926, "mean_token_accuracy": 0.8672316074371338, "num_tokens": 795084918.0, "step": 20936 }, { "epoch": 2.66340160284951, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.0009002685546875, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 11.866960525512695, "learning_rate": 5e-05, "loss": 1.0565, "mean_token_accuracy": 0.8839566111564636, "num_tokens": 795119261.0, "step": 20937 }, { "epoch": 2.6635288131281007, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00089263916015625, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 10.302042961120605, "learning_rate": 5e-05, "loss": 0.9062, "mean_token_accuracy": 0.8893904089927673, "num_tokens": 795162559.0, "step": 20938 }, { "epoch": 2.6636560234066913, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.00089263916015625, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 12.06177806854248, "learning_rate": 5e-05, "loss": 1.0646, "mean_token_accuracy": 0.884291410446167, "num_tokens": 795200853.0, "step": 20939 }, { "epoch": 2.663783233685282, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00089263916015625, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 10.296989440917969, "learning_rate": 5e-05, "loss": 0.9104, "mean_token_accuracy": 0.8878346681594849, "num_tokens": 795238405.0, "step": 20940 }, { "epoch": 2.6639104439638723, "ewc_loss": 0.75, "ewc_loss_diag": 0.00089263916015625, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 12.170943260192871, "learning_rate": 5e-05, "loss": 1.1047, "mean_token_accuracy": 0.8755092620849609, "num_tokens": 795278341.0, "step": 20941 }, { "epoch": 2.664037654242463, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00089263916015625, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 10.575045585632324, "learning_rate": 5e-05, "loss": 0.9637, "mean_token_accuracy": 0.8744980096817017, "num_tokens": 795319108.0, "step": 20942 }, { "epoch": 2.6641648645210534, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.00089263916015625, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 11.830809593200684, "learning_rate": 5e-05, "loss": 1.0491, "mean_token_accuracy": 0.8783544301986694, "num_tokens": 795354641.0, "step": 20943 }, { "epoch": 2.664292074799644, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000885009765625, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 10.480772972106934, "learning_rate": 5e-05, "loss": 0.9888, "mean_token_accuracy": 0.8671103715896606, "num_tokens": 795399098.0, "step": 20944 }, { "epoch": 2.6644192850782344, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000885009765625, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 11.487780570983887, "learning_rate": 5e-05, "loss": 0.9995, "mean_token_accuracy": 0.8870775699615479, "num_tokens": 795444191.0, "step": 20945 }, { "epoch": 2.664546495356825, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000881195068359375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 10.097959518432617, "learning_rate": 5e-05, "loss": 0.8895, "mean_token_accuracy": 0.8792909979820251, "num_tokens": 795476450.0, "step": 20946 }, { "epoch": 2.6646737056354155, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00087738037109375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 11.474111557006836, "learning_rate": 5e-05, "loss": 1.0072, "mean_token_accuracy": 0.8794715404510498, "num_tokens": 795518740.0, "step": 20947 }, { "epoch": 2.6648009159140056, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000873565673828125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 9.86237621307373, "learning_rate": 5e-05, "loss": 0.8948, "mean_token_accuracy": 0.8761146068572998, "num_tokens": 795558944.0, "step": 20948 }, { "epoch": 2.6649281261925966, "ewc_loss": 0.6875, "ewc_loss_diag": 0.0008697509765625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 11.628961563110352, "learning_rate": 5e-05, "loss": 1.0125, "mean_token_accuracy": 0.8815831542015076, "num_tokens": 795592685.0, "step": 20949 }, { "epoch": 2.6650553364711866, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0008697509765625, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 10.019043922424316, "learning_rate": 5e-05, "loss": 0.9061, "mean_token_accuracy": 0.8758184313774109, "num_tokens": 795627878.0, "step": 20950 }, { "epoch": 2.6651825467497776, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000865936279296875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 11.471637725830078, "learning_rate": 5e-05, "loss": 1.0433, "mean_token_accuracy": 0.868558406829834, "num_tokens": 795666784.0, "step": 20951 }, { "epoch": 2.6653097570283677, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000865936279296875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 10.080519676208496, "learning_rate": 5e-05, "loss": 0.8914, "mean_token_accuracy": 0.8809557557106018, "num_tokens": 795701921.0, "step": 20952 }, { "epoch": 2.6654369673069587, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000865936279296875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 11.089982032775879, "learning_rate": 5e-05, "loss": 0.9988, "mean_token_accuracy": 0.8757441639900208, "num_tokens": 795744497.0, "step": 20953 }, { "epoch": 2.6655641775855488, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00086212158203125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 9.805560111999512, "learning_rate": 5e-05, "loss": 0.9079, "mean_token_accuracy": 0.8672024607658386, "num_tokens": 795786871.0, "step": 20954 }, { "epoch": 2.6656913878641393, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00086212158203125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 10.908123970031738, "learning_rate": 5e-05, "loss": 0.9537, "mean_token_accuracy": 0.888744592666626, "num_tokens": 795829403.0, "step": 20955 }, { "epoch": 2.66581859814273, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00086212158203125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 9.476327896118164, "learning_rate": 5e-05, "loss": 0.8957, "mean_token_accuracy": 0.866537868976593, "num_tokens": 795867352.0, "step": 20956 }, { "epoch": 2.6659458084213203, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00086212158203125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 10.964800834655762, "learning_rate": 5e-05, "loss": 1.0153, "mean_token_accuracy": 0.8730592131614685, "num_tokens": 795907209.0, "step": 20957 }, { "epoch": 2.666073018699911, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00086212158203125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 9.411957740783691, "learning_rate": 5e-05, "loss": 0.853, "mean_token_accuracy": 0.8765326142311096, "num_tokens": 795946515.0, "step": 20958 }, { "epoch": 2.6662002289785014, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000858306884765625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 10.959648132324219, "learning_rate": 5e-05, "loss": 0.9668, "mean_token_accuracy": 0.8889169096946716, "num_tokens": 795983328.0, "step": 20959 }, { "epoch": 2.666327439257092, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000858306884765625, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 9.526597023010254, "learning_rate": 5e-05, "loss": 0.8783, "mean_token_accuracy": 0.8654907941818237, "num_tokens": 796018609.0, "step": 20960 }, { "epoch": 2.6664546495356825, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000858306884765625, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 10.668121337890625, "learning_rate": 5e-05, "loss": 0.9679, "mean_token_accuracy": 0.874567985534668, "num_tokens": 796059676.0, "step": 20961 }, { "epoch": 2.666581859814273, "ewc_loss": 0.5, "ewc_loss_diag": 0.0008544921875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 9.432957649230957, "learning_rate": 5e-05, "loss": 0.8408, "mean_token_accuracy": 0.8776426315307617, "num_tokens": 796100896.0, "step": 20962 }, { "epoch": 2.6667090700928635, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.0008544921875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 10.322408676147461, "learning_rate": 5e-05, "loss": 0.8994, "mean_token_accuracy": 0.8950443863868713, "num_tokens": 796140513.0, "step": 20963 }, { "epoch": 2.666836280371454, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.0008544921875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 9.116094589233398, "learning_rate": 5e-05, "loss": 0.7784, "mean_token_accuracy": 0.8893623352050781, "num_tokens": 796174339.0, "step": 20964 }, { "epoch": 2.6669634906500446, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.0008544921875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 10.196796417236328, "learning_rate": 5e-05, "loss": 0.9254, "mean_token_accuracy": 0.8829635977745056, "num_tokens": 796216624.0, "step": 20965 }, { "epoch": 2.667090700928635, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000850677490234375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.847200393676758, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.8673356771469116, "num_tokens": 796250681.0, "step": 20966 }, { "epoch": 2.6672179112072256, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000850677490234375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 10.244292259216309, "learning_rate": 5e-05, "loss": 0.9181, "mean_token_accuracy": 0.8835834264755249, "num_tokens": 796284860.0, "step": 20967 }, { "epoch": 2.667345121485816, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00084686279296875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.820374488830566, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.8930163383483887, "num_tokens": 796325833.0, "step": 20968 }, { "epoch": 2.6674723317644067, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00084686279296875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 10.161168098449707, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8894864320755005, "num_tokens": 796354955.0, "step": 20969 }, { "epoch": 2.667599542042997, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00084686279296875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.903277397155762, "learning_rate": 5e-05, "loss": 0.7843, "mean_token_accuracy": 0.8788271546363831, "num_tokens": 796390505.0, "step": 20970 }, { "epoch": 2.6677267523215877, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00084686279296875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 9.815532684326172, "learning_rate": 5e-05, "loss": 0.8586, "mean_token_accuracy": 0.8885922431945801, "num_tokens": 796428003.0, "step": 20971 }, { "epoch": 2.6678539626001783, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000843048095703125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.763433456420898, "learning_rate": 5e-05, "loss": 0.7448, "mean_token_accuracy": 0.8895968794822693, "num_tokens": 796467396.0, "step": 20972 }, { "epoch": 2.6679811728787683, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000843048095703125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 9.49921989440918, "learning_rate": 5e-05, "loss": 0.855, "mean_token_accuracy": 0.877358078956604, "num_tokens": 796505320.0, "step": 20973 }, { "epoch": 2.6681083831573593, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 8.459294319152832, "learning_rate": 5e-05, "loss": 0.7334, "mean_token_accuracy": 0.88735431432724, "num_tokens": 796542109.0, "step": 20974 }, { "epoch": 2.6682355934359494, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 9.37759780883789, "learning_rate": 5e-05, "loss": 0.8347, "mean_token_accuracy": 0.879005491733551, "num_tokens": 796588523.0, "step": 20975 }, { "epoch": 2.6683628037145404, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 8.213186264038086, "learning_rate": 5e-05, "loss": 0.7779, "mean_token_accuracy": 0.8638968467712402, "num_tokens": 796624493.0, "step": 20976 }, { "epoch": 2.6684900139931305, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 9.407906532287598, "learning_rate": 5e-05, "loss": 0.803, "mean_token_accuracy": 0.8901211023330688, "num_tokens": 796662147.0, "step": 20977 }, { "epoch": 2.668617224271721, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 8.143425941467285, "learning_rate": 5e-05, "loss": 0.7308, "mean_token_accuracy": 0.8787409663200378, "num_tokens": 796700024.0, "step": 20978 }, { "epoch": 2.6687444345503115, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 9.380582809448242, "learning_rate": 5e-05, "loss": 0.8056, "mean_token_accuracy": 0.8844079971313477, "num_tokens": 796735058.0, "step": 20979 }, { "epoch": 2.668871644828902, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 8.235503196716309, "learning_rate": 5e-05, "loss": 0.747, "mean_token_accuracy": 0.874015212059021, "num_tokens": 796778419.0, "step": 20980 }, { "epoch": 2.6689988551074926, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 9.09471607208252, "learning_rate": 5e-05, "loss": 0.7802, "mean_token_accuracy": 0.8866889476776123, "num_tokens": 796817495.0, "step": 20981 }, { "epoch": 2.669126065386083, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 8.230273246765137, "learning_rate": 5e-05, "loss": 0.7477, "mean_token_accuracy": 0.8774359226226807, "num_tokens": 796850432.0, "step": 20982 }, { "epoch": 2.6692532756646736, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.716821670532227, "learning_rate": 5e-05, "loss": 0.7986, "mean_token_accuracy": 0.874401330947876, "num_tokens": 796889584.0, "step": 20983 }, { "epoch": 2.669380485943264, "ewc_loss": 0.375, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 8.02136516571045, "learning_rate": 5e-05, "loss": 0.706, "mean_token_accuracy": 0.8813875913619995, "num_tokens": 796924052.0, "step": 20984 }, { "epoch": 2.6695076962218547, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.542506217956543, "learning_rate": 5e-05, "loss": 0.7856, "mean_token_accuracy": 0.8717339038848877, "num_tokens": 796963795.0, "step": 20985 }, { "epoch": 2.669634906500445, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.755154609680176, "learning_rate": 5e-05, "loss": 0.6726, "mean_token_accuracy": 0.8902536630630493, "num_tokens": 797002621.0, "step": 20986 }, { "epoch": 2.6697621167790357, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 8.58277702331543, "learning_rate": 5e-05, "loss": 0.8243, "mean_token_accuracy": 0.8597611784934998, "num_tokens": 797047173.0, "step": 20987 }, { "epoch": 2.6698893270576263, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 7.6271281242370605, "learning_rate": 5e-05, "loss": 0.7041, "mean_token_accuracy": 0.877693772315979, "num_tokens": 797088394.0, "step": 20988 }, { "epoch": 2.670016537336217, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.0008392333984375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 8.582698822021484, "learning_rate": 5e-05, "loss": 0.745, "mean_token_accuracy": 0.8886334300041199, "num_tokens": 797126108.0, "step": 20989 }, { "epoch": 2.6701437476148073, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 7.6058454513549805, "learning_rate": 5e-05, "loss": 0.6797, "mean_token_accuracy": 0.8854050636291504, "num_tokens": 797162718.0, "step": 20990 }, { "epoch": 2.670270957893398, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 8.490089416503906, "learning_rate": 5e-05, "loss": 0.7794, "mean_token_accuracy": 0.869580090045929, "num_tokens": 797197552.0, "step": 20991 }, { "epoch": 2.6703981681719884, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.655487537384033, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8582947254180908, "num_tokens": 797232523.0, "step": 20992 }, { "epoch": 2.670525378450579, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 8.279440879821777, "learning_rate": 5e-05, "loss": 0.7236, "mean_token_accuracy": 0.8839970231056213, "num_tokens": 797266620.0, "step": 20993 }, { "epoch": 2.6706525887291694, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.668173313140869, "learning_rate": 5e-05, "loss": 0.6829, "mean_token_accuracy": 0.8860782980918884, "num_tokens": 797305378.0, "step": 20994 }, { "epoch": 2.67077979900776, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 8.102258682250977, "learning_rate": 5e-05, "loss": 0.7124, "mean_token_accuracy": 0.8846768140792847, "num_tokens": 797343269.0, "step": 20995 }, { "epoch": 2.6709070092863505, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.717009544372559, "learning_rate": 5e-05, "loss": 0.6309, "mean_token_accuracy": 0.9039831757545471, "num_tokens": 797375034.0, "step": 20996 }, { "epoch": 2.671034219564941, "ewc_loss": 0.375, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.873382568359375, "learning_rate": 5e-05, "loss": 0.7186, "mean_token_accuracy": 0.8767502903938293, "num_tokens": 797410869.0, "step": 20997 }, { "epoch": 2.671161429843531, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.738432884216309, "learning_rate": 5e-05, "loss": 0.6824, "mean_token_accuracy": 0.8896185159683228, "num_tokens": 797449189.0, "step": 20998 }, { "epoch": 2.671288640122122, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.740292072296143, "learning_rate": 5e-05, "loss": 0.6898, "mean_token_accuracy": 0.8843640089035034, "num_tokens": 797492182.0, "step": 20999 }, { "epoch": 2.671415850400712, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.722269535064697, "learning_rate": 5e-05, "loss": 0.6504, "mean_token_accuracy": 0.8983348608016968, "num_tokens": 797529043.0, "step": 21000 }, { "epoch": 2.671543060679303, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00083160400390625, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.6871185302734375, "learning_rate": 5e-05, "loss": 0.7197, "mean_token_accuracy": 0.87352454662323, "num_tokens": 797560630.0, "step": 21001 }, { "epoch": 2.6716702709578932, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000827789306640625, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.713027477264404, "learning_rate": 5e-05, "loss": 0.7292, "mean_token_accuracy": 0.8742470741271973, "num_tokens": 797598667.0, "step": 21002 }, { "epoch": 2.6717974812364838, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000827789306640625, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 7.600694179534912, "learning_rate": 5e-05, "loss": 0.6648, "mean_token_accuracy": 0.890516996383667, "num_tokens": 797637420.0, "step": 21003 }, { "epoch": 2.6719246915150743, "ewc_loss": 0.375, "ewc_loss_diag": 0.000827789306640625, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.764830589294434, "learning_rate": 5e-05, "loss": 0.7004, "mean_token_accuracy": 0.8853341937065125, "num_tokens": 797670969.0, "step": 21004 }, { "epoch": 2.672051901793665, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000827789306640625, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 7.516806125640869, "learning_rate": 5e-05, "loss": 0.6807, "mean_token_accuracy": 0.8836023807525635, "num_tokens": 797711915.0, "step": 21005 }, { "epoch": 2.6721791120722553, "ewc_loss": 0.375, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.778897285461426, "learning_rate": 5e-05, "loss": 0.7132, "mean_token_accuracy": 0.8793970942497253, "num_tokens": 797751036.0, "step": 21006 }, { "epoch": 2.672306322350846, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.5031938552856445, "learning_rate": 5e-05, "loss": 0.7122, "mean_token_accuracy": 0.8738627433776855, "num_tokens": 797793999.0, "step": 21007 }, { "epoch": 2.6724335326294364, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.742377281188965, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.8790575861930847, "num_tokens": 797834836.0, "step": 21008 }, { "epoch": 2.672560742908027, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.528383255004883, "learning_rate": 5e-05, "loss": 0.7358, "mean_token_accuracy": 0.8712654113769531, "num_tokens": 797872855.0, "step": 21009 }, { "epoch": 2.6726879531866174, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.680723190307617, "learning_rate": 5e-05, "loss": 0.7258, "mean_token_accuracy": 0.8762608766555786, "num_tokens": 797910925.0, "step": 21010 }, { "epoch": 2.672815163465208, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.57287073135376, "learning_rate": 5e-05, "loss": 0.7147, "mean_token_accuracy": 0.8737099766731262, "num_tokens": 797952201.0, "step": 21011 }, { "epoch": 2.6729423737437985, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.617575168609619, "learning_rate": 5e-05, "loss": 0.7613, "mean_token_accuracy": 0.8618353605270386, "num_tokens": 797988706.0, "step": 21012 }, { "epoch": 2.673069584022389, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.626230716705322, "learning_rate": 5e-05, "loss": 0.7252, "mean_token_accuracy": 0.8751843571662903, "num_tokens": 798028438.0, "step": 21013 }, { "epoch": 2.6731967943009796, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.498440265655518, "learning_rate": 5e-05, "loss": 0.6727, "mean_token_accuracy": 0.8891686201095581, "num_tokens": 798062864.0, "step": 21014 }, { "epoch": 2.67332400457957, "ewc_loss": 0.375, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.762621879577637, "learning_rate": 5e-05, "loss": 0.7033, "mean_token_accuracy": 0.8837461471557617, "num_tokens": 798098271.0, "step": 21015 }, { "epoch": 2.6734512148581606, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 7.310580253601074, "learning_rate": 5e-05, "loss": 0.6694, "mean_token_accuracy": 0.8853393793106079, "num_tokens": 798131286.0, "step": 21016 }, { "epoch": 2.673578425136751, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.996674537658691, "learning_rate": 5e-05, "loss": 0.733, "mean_token_accuracy": 0.8809362649917603, "num_tokens": 798164711.0, "step": 21017 }, { "epoch": 2.6737056354153417, "ewc_loss": 0.330078125, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 7.027337074279785, "learning_rate": 5e-05, "loss": 0.6701, "mean_token_accuracy": 0.8801178932189941, "num_tokens": 798205280.0, "step": 21018 }, { "epoch": 2.673832845693932, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 8.393989562988281, "learning_rate": 5e-05, "loss": 0.7618, "mean_token_accuracy": 0.8816365003585815, "num_tokens": 798246561.0, "step": 21019 }, { "epoch": 2.6739600559725227, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.00022411346435546875, "grad_norm": 6.54778528213501, "learning_rate": 5e-05, "loss": 0.6351, "mean_token_accuracy": 0.8819283246994019, "num_tokens": 798287494.0, "step": 21020 }, { "epoch": 2.674087266251113, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 9.134977340698242, "learning_rate": 5e-05, "loss": 0.8231, "mean_token_accuracy": 0.8819849491119385, "num_tokens": 798325305.0, "step": 21021 }, { "epoch": 2.674214476529704, "ewc_loss": 0.28125, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0001983642578125, "grad_norm": 5.873698711395264, "learning_rate": 5e-05, "loss": 0.5897, "mean_token_accuracy": 0.8889467120170593, "num_tokens": 798363083.0, "step": 21022 }, { "epoch": 2.674341686808294, "ewc_loss": 0.625, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 10.282295227050781, "learning_rate": 5e-05, "loss": 1.0262, "mean_token_accuracy": 0.8606696724891663, "num_tokens": 798403324.0, "step": 21023 }, { "epoch": 2.674468897086885, "ewc_loss": 0.3125, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0002307891845703125, "grad_norm": 5.774172306060791, "learning_rate": 5e-05, "loss": 0.6335, "mean_token_accuracy": 0.8857734203338623, "num_tokens": 798441548.0, "step": 21024 }, { "epoch": 2.674596107365475, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 11.19058895111084, "learning_rate": 5e-05, "loss": 1.1166, "mean_token_accuracy": 0.8771877288818359, "num_tokens": 798486038.0, "step": 21025 }, { "epoch": 2.674723317644066, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 7.983954906463623, "learning_rate": 5e-05, "loss": 0.8673, "mean_token_accuracy": 0.8883630037307739, "num_tokens": 798530657.0, "step": 21026 }, { "epoch": 2.674850527922656, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000823974609375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.899537086486816, "learning_rate": 5e-05, "loss": 0.9732, "mean_token_accuracy": 0.8823649287223816, "num_tokens": 798571027.0, "step": 21027 }, { "epoch": 2.6749777382012465, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.694476127624512, "learning_rate": 5e-05, "loss": 1.0127, "mean_token_accuracy": 0.887100100517273, "num_tokens": 798609874.0, "step": 21028 }, { "epoch": 2.675104948479837, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.257808685302734, "learning_rate": 5e-05, "loss": 0.766, "mean_token_accuracy": 0.8736966848373413, "num_tokens": 798645191.0, "step": 21029 }, { "epoch": 2.6752321587584276, "ewc_loss": 0.8828125, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.00080108642578125, "grad_norm": 12.179458618164062, "learning_rate": 5e-05, "loss": 1.2007, "mean_token_accuracy": 0.8853267431259155, "num_tokens": 798688445.0, "step": 21030 }, { "epoch": 2.675359369037018, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.53427791595459, "learning_rate": 5e-05, "loss": 0.9355, "mean_token_accuracy": 0.8814960718154907, "num_tokens": 798729582.0, "step": 21031 }, { "epoch": 2.6754865793156086, "ewc_loss": 0.9375, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 12.929169654846191, "learning_rate": 5e-05, "loss": 1.25, "mean_token_accuracy": 0.8861032724380493, "num_tokens": 798765178.0, "step": 21032 }, { "epoch": 2.675613789594199, "ewc_loss": 0.86328125, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 10.912394523620605, "learning_rate": 5e-05, "loss": 1.2162, "mean_token_accuracy": 0.879249095916748, "num_tokens": 798806048.0, "step": 21033 }, { "epoch": 2.6757409998727897, "ewc_loss": 0.765625, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 11.857542037963867, "learning_rate": 5e-05, "loss": 1.0939, "mean_token_accuracy": 0.8842804431915283, "num_tokens": 798845554.0, "step": 21034 }, { "epoch": 2.67586821015138, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.504437446594238, "learning_rate": 5e-05, "loss": 0.949, "mean_token_accuracy": 0.8948158621788025, "num_tokens": 798883124.0, "step": 21035 }, { "epoch": 2.6759954204299707, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 12.309003829956055, "learning_rate": 5e-05, "loss": 1.1743, "mean_token_accuracy": 0.8937479257583618, "num_tokens": 798916040.0, "step": 21036 }, { "epoch": 2.6761226307085613, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.9551362991333, "learning_rate": 5e-05, "loss": 0.9448, "mean_token_accuracy": 0.8789100050926208, "num_tokens": 798952562.0, "step": 21037 }, { "epoch": 2.676249840987152, "ewc_loss": 0.95703125, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.000873565673828125, "grad_norm": 13.034061431884766, "learning_rate": 5e-05, "loss": 1.3294, "mean_token_accuracy": 0.8728469610214233, "num_tokens": 798990689.0, "step": 21038 }, { "epoch": 2.6763770512657423, "ewc_loss": 0.78125, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.77659797668457, "learning_rate": 5e-05, "loss": 1.1052, "mean_token_accuracy": 0.8849292397499084, "num_tokens": 799029398.0, "step": 21039 }, { "epoch": 2.676504261544333, "ewc_loss": 0.796875, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 11.879985809326172, "learning_rate": 5e-05, "loss": 1.1012, "mean_token_accuracy": 0.892835259437561, "num_tokens": 799073200.0, "step": 21040 }, { "epoch": 2.6766314718229234, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.705615043640137, "learning_rate": 5e-05, "loss": 1.0632, "mean_token_accuracy": 0.8859004378318787, "num_tokens": 799115020.0, "step": 21041 }, { "epoch": 2.676758682101514, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 11.127723693847656, "learning_rate": 5e-05, "loss": 1.035, "mean_token_accuracy": 0.8920601606369019, "num_tokens": 799152348.0, "step": 21042 }, { "epoch": 2.6768858923801044, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.799388885498047, "learning_rate": 5e-05, "loss": 0.9804, "mean_token_accuracy": 0.877242386341095, "num_tokens": 799191120.0, "step": 21043 }, { "epoch": 2.677013102658695, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.864900588989258, "learning_rate": 5e-05, "loss": 1.037, "mean_token_accuracy": 0.8811873197555542, "num_tokens": 799229044.0, "step": 21044 }, { "epoch": 2.6771403129372855, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.168270111083984, "learning_rate": 5e-05, "loss": 0.894, "mean_token_accuracy": 0.8863626718521118, "num_tokens": 799269148.0, "step": 21045 }, { "epoch": 2.6772675232158756, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.93023681640625, "learning_rate": 5e-05, "loss": 1.0057, "mean_token_accuracy": 0.8870984315872192, "num_tokens": 799305188.0, "step": 21046 }, { "epoch": 2.6773947334944665, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.884634971618652, "learning_rate": 5e-05, "loss": 0.9146, "mean_token_accuracy": 0.8723694682121277, "num_tokens": 799344156.0, "step": 21047 }, { "epoch": 2.6775219437730566, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 11.311057090759277, "learning_rate": 5e-05, "loss": 1.0991, "mean_token_accuracy": 0.8692128658294678, "num_tokens": 799374792.0, "step": 21048 }, { "epoch": 2.6776491540516476, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00081634521484375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.196296691894531, "learning_rate": 5e-05, "loss": 0.9137, "mean_token_accuracy": 0.8812668323516846, "num_tokens": 799408425.0, "step": 21049 }, { "epoch": 2.6777763643302377, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 11.373871803283691, "learning_rate": 5e-05, "loss": 1.0728, "mean_token_accuracy": 0.8858178853988647, "num_tokens": 799448362.0, "step": 21050 }, { "epoch": 2.6779035746088287, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.684646606445312, "learning_rate": 5e-05, "loss": 0.954, "mean_token_accuracy": 0.8864503502845764, "num_tokens": 799482225.0, "step": 21051 }, { "epoch": 2.6780307848874187, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.812530517578125, "learning_rate": 5e-05, "loss": 1.0763, "mean_token_accuracy": 0.8684406280517578, "num_tokens": 799523053.0, "step": 21052 }, { "epoch": 2.6781579951660093, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.62530517578125, "learning_rate": 5e-05, "loss": 0.9474, "mean_token_accuracy": 0.8809773921966553, "num_tokens": 799563070.0, "step": 21053 }, { "epoch": 2.6782852054446, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 10.15880298614502, "learning_rate": 5e-05, "loss": 0.9607, "mean_token_accuracy": 0.8899579048156738, "num_tokens": 799595465.0, "step": 21054 }, { "epoch": 2.6784124157231903, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 9.08402156829834, "learning_rate": 5e-05, "loss": 0.8657, "mean_token_accuracy": 0.8900413513183594, "num_tokens": 799632308.0, "step": 21055 }, { "epoch": 2.678539626001781, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.933619499206543, "learning_rate": 5e-05, "loss": 0.9416, "mean_token_accuracy": 0.8941661715507507, "num_tokens": 799669563.0, "step": 21056 }, { "epoch": 2.6786668362803714, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.493078231811523, "learning_rate": 5e-05, "loss": 0.8316, "mean_token_accuracy": 0.8777370452880859, "num_tokens": 799714575.0, "step": 21057 }, { "epoch": 2.678794046558962, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 10.182300567626953, "learning_rate": 5e-05, "loss": 0.9931, "mean_token_accuracy": 0.8825253248214722, "num_tokens": 799752730.0, "step": 21058 }, { "epoch": 2.6789212568375524, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.297334671020508, "learning_rate": 5e-05, "loss": 0.8482, "mean_token_accuracy": 0.8691983222961426, "num_tokens": 799788110.0, "step": 21059 }, { "epoch": 2.679048467116143, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.569879531860352, "learning_rate": 5e-05, "loss": 1.0307, "mean_token_accuracy": 0.890596330165863, "num_tokens": 799825561.0, "step": 21060 }, { "epoch": 2.6791756773947335, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.795087814331055, "learning_rate": 5e-05, "loss": 0.8511, "mean_token_accuracy": 0.8864459991455078, "num_tokens": 799865412.0, "step": 21061 }, { "epoch": 2.679302887673324, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 10.261642456054688, "learning_rate": 5e-05, "loss": 1.0305, "mean_token_accuracy": 0.8699836134910583, "num_tokens": 799900238.0, "step": 21062 }, { "epoch": 2.6794300979519146, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.133167266845703, "learning_rate": 5e-05, "loss": 0.916, "mean_token_accuracy": 0.8750122785568237, "num_tokens": 799936724.0, "step": 21063 }, { "epoch": 2.679557308230505, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.411951065063477, "learning_rate": 5e-05, "loss": 0.9486, "mean_token_accuracy": 0.8774346709251404, "num_tokens": 799978347.0, "step": 21064 }, { "epoch": 2.6796845185090956, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.758830070495605, "learning_rate": 5e-05, "loss": 0.877, "mean_token_accuracy": 0.8773373365402222, "num_tokens": 800015419.0, "step": 21065 }, { "epoch": 2.679811728787686, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.939553260803223, "learning_rate": 5e-05, "loss": 0.8217, "mean_token_accuracy": 0.901766836643219, "num_tokens": 800048435.0, "step": 21066 }, { "epoch": 2.6799389390662767, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.107794761657715, "learning_rate": 5e-05, "loss": 0.7704, "mean_token_accuracy": 0.8814957141876221, "num_tokens": 800083317.0, "step": 21067 }, { "epoch": 2.680066149344867, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 9.015175819396973, "learning_rate": 5e-05, "loss": 0.8892, "mean_token_accuracy": 0.8780263066291809, "num_tokens": 800123303.0, "step": 21068 }, { "epoch": 2.6801933596234577, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.622342109680176, "learning_rate": 5e-05, "loss": 0.74, "mean_token_accuracy": 0.8842077255249023, "num_tokens": 800165157.0, "step": 21069 }, { "epoch": 2.6803205699020483, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.299718856811523, "learning_rate": 5e-05, "loss": 0.9396, "mean_token_accuracy": 0.8730335235595703, "num_tokens": 800203657.0, "step": 21070 }, { "epoch": 2.6804477801806383, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.538538455963135, "learning_rate": 5e-05, "loss": 0.7292, "mean_token_accuracy": 0.884579062461853, "num_tokens": 800245008.0, "step": 21071 }, { "epoch": 2.6805749904592293, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.505481719970703, "learning_rate": 5e-05, "loss": 0.9347, "mean_token_accuracy": 0.8823812007904053, "num_tokens": 800285438.0, "step": 21072 }, { "epoch": 2.6807022007378194, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000812530517578125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.856043815612793, "learning_rate": 5e-05, "loss": 0.7798, "mean_token_accuracy": 0.8783267736434937, "num_tokens": 800321345.0, "step": 21073 }, { "epoch": 2.6808294110164104, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.270729064941406, "learning_rate": 5e-05, "loss": 0.9388, "mean_token_accuracy": 0.8726333379745483, "num_tokens": 800361657.0, "step": 21074 }, { "epoch": 2.6809566212950005, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.2554931640625, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8625335097312927, "num_tokens": 800393809.0, "step": 21075 }, { "epoch": 2.681083831573591, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.515091896057129, "learning_rate": 5e-05, "loss": 0.841, "mean_token_accuracy": 0.8830920457839966, "num_tokens": 800432541.0, "step": 21076 }, { "epoch": 2.6812110418521815, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.263330459594727, "learning_rate": 5e-05, "loss": 0.7949, "mean_token_accuracy": 0.8768047094345093, "num_tokens": 800468552.0, "step": 21077 }, { "epoch": 2.681338252130772, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.913129806518555, "learning_rate": 5e-05, "loss": 0.7705, "mean_token_accuracy": 0.8875713348388672, "num_tokens": 800503911.0, "step": 21078 }, { "epoch": 2.6814654624093626, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.099647521972656, "learning_rate": 5e-05, "loss": 0.7723, "mean_token_accuracy": 0.8809424042701721, "num_tokens": 800549755.0, "step": 21079 }, { "epoch": 2.681592672687953, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.668708324432373, "learning_rate": 5e-05, "loss": 0.7134, "mean_token_accuracy": 0.8895247578620911, "num_tokens": 800583240.0, "step": 21080 }, { "epoch": 2.6817198829665436, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 8.100533485412598, "learning_rate": 5e-05, "loss": 0.8086, "mean_token_accuracy": 0.8674001693725586, "num_tokens": 800624016.0, "step": 21081 }, { "epoch": 2.681847093245134, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.437079429626465, "learning_rate": 5e-05, "loss": 0.7642, "mean_token_accuracy": 0.8654923439025879, "num_tokens": 800661052.0, "step": 21082 }, { "epoch": 2.6819743035237247, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 8.296022415161133, "learning_rate": 5e-05, "loss": 0.7693, "mean_token_accuracy": 0.882757306098938, "num_tokens": 800701549.0, "step": 21083 }, { "epoch": 2.682101513802315, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.146071434020996, "learning_rate": 5e-05, "loss": 0.7006, "mean_token_accuracy": 0.8783022165298462, "num_tokens": 800738650.0, "step": 21084 }, { "epoch": 2.6822287240809057, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.619470596313477, "learning_rate": 5e-05, "loss": 0.8028, "mean_token_accuracy": 0.8838177919387817, "num_tokens": 800773687.0, "step": 21085 }, { "epoch": 2.6823559343594963, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.9393720626831055, "learning_rate": 5e-05, "loss": 0.6744, "mean_token_accuracy": 0.8894520998001099, "num_tokens": 800812523.0, "step": 21086 }, { "epoch": 2.682483144638087, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.842120170593262, "learning_rate": 5e-05, "loss": 0.8511, "mean_token_accuracy": 0.8826401233673096, "num_tokens": 800850838.0, "step": 21087 }, { "epoch": 2.6826103549166773, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.018935680389404, "learning_rate": 5e-05, "loss": 0.6695, "mean_token_accuracy": 0.8890727758407593, "num_tokens": 800886734.0, "step": 21088 }, { "epoch": 2.682737565195268, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.811707496643066, "learning_rate": 5e-05, "loss": 0.8609, "mean_token_accuracy": 0.876915454864502, "num_tokens": 800930816.0, "step": 21089 }, { "epoch": 2.6828647754738584, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.327971458435059, "learning_rate": 5e-05, "loss": 0.7259, "mean_token_accuracy": 0.8800064921379089, "num_tokens": 800962982.0, "step": 21090 }, { "epoch": 2.682991985752449, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.441969871520996, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.8694444894790649, "num_tokens": 801003608.0, "step": 21091 }, { "epoch": 2.6831191960310394, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.721728801727295, "learning_rate": 5e-05, "loss": 0.7546, "mean_token_accuracy": 0.8745793104171753, "num_tokens": 801032889.0, "step": 21092 }, { "epoch": 2.68324640630963, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.901555061340332, "learning_rate": 5e-05, "loss": 0.7866, "mean_token_accuracy": 0.87380051612854, "num_tokens": 801068717.0, "step": 21093 }, { "epoch": 2.6833736165882205, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.9261627197265625, "learning_rate": 5e-05, "loss": 0.7644, "mean_token_accuracy": 0.8784093856811523, "num_tokens": 801105120.0, "step": 21094 }, { "epoch": 2.683500826866811, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.667699813842773, "learning_rate": 5e-05, "loss": 0.759, "mean_token_accuracy": 0.8749846816062927, "num_tokens": 801140301.0, "step": 21095 }, { "epoch": 2.683628037145401, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.843583106994629, "learning_rate": 5e-05, "loss": 0.6795, "mean_token_accuracy": 0.9008064270019531, "num_tokens": 801176314.0, "step": 21096 }, { "epoch": 2.683755247423992, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.814884662628174, "learning_rate": 5e-05, "loss": 0.7406, "mean_token_accuracy": 0.8846853971481323, "num_tokens": 801212906.0, "step": 21097 }, { "epoch": 2.683882457702582, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.522942543029785, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.868880033493042, "num_tokens": 801243084.0, "step": 21098 }, { "epoch": 2.684009667981173, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 8.11709213256836, "learning_rate": 5e-05, "loss": 0.7705, "mean_token_accuracy": 0.8785910606384277, "num_tokens": 801282701.0, "step": 21099 }, { "epoch": 2.684136878259763, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.121186256408691, "learning_rate": 5e-05, "loss": 0.6846, "mean_token_accuracy": 0.8878962993621826, "num_tokens": 801315363.0, "step": 21100 }, { "epoch": 2.6842640885383537, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.568075180053711, "learning_rate": 5e-05, "loss": 0.8239, "mean_token_accuracy": 0.875670850276947, "num_tokens": 801354103.0, "step": 21101 }, { "epoch": 2.6843912988169443, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.722718715667725, "learning_rate": 5e-05, "loss": 0.6824, "mean_token_accuracy": 0.8789846897125244, "num_tokens": 801393351.0, "step": 21102 }, { "epoch": 2.684518509095535, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 9.063631057739258, "learning_rate": 5e-05, "loss": 0.8551, "mean_token_accuracy": 0.8760412931442261, "num_tokens": 801432829.0, "step": 21103 }, { "epoch": 2.6846457193741253, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.5595879554748535, "learning_rate": 5e-05, "loss": 0.6936, "mean_token_accuracy": 0.8788483738899231, "num_tokens": 801468735.0, "step": 21104 }, { "epoch": 2.684772929652716, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 9.233163833618164, "learning_rate": 5e-05, "loss": 0.8587, "mean_token_accuracy": 0.8901610374450684, "num_tokens": 801507238.0, "step": 21105 }, { "epoch": 2.6849001399313064, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.029794692993164, "learning_rate": 5e-05, "loss": 0.6996, "mean_token_accuracy": 0.8859571218490601, "num_tokens": 801543040.0, "step": 21106 }, { "epoch": 2.685027350209897, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.552997589111328, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8804349899291992, "num_tokens": 801584466.0, "step": 21107 }, { "epoch": 2.6851545604884874, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.9472126960754395, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8874770402908325, "num_tokens": 801617466.0, "step": 21108 }, { "epoch": 2.685281770767078, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.3441853523254395, "learning_rate": 5e-05, "loss": 0.7664, "mean_token_accuracy": 0.8696864247322083, "num_tokens": 801653408.0, "step": 21109 }, { "epoch": 2.6854089810456685, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.830450057983398, "learning_rate": 5e-05, "loss": 0.8512, "mean_token_accuracy": 0.8768555521965027, "num_tokens": 801687200.0, "step": 21110 }, { "epoch": 2.685536191324259, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.465980529785156, "learning_rate": 5e-05, "loss": 0.6961, "mean_token_accuracy": 0.8774043321609497, "num_tokens": 801728547.0, "step": 21111 }, { "epoch": 2.6856634016028496, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.631455421447754, "learning_rate": 5e-05, "loss": 0.9686, "mean_token_accuracy": 0.8730008602142334, "num_tokens": 801765425.0, "step": 21112 }, { "epoch": 2.68579061188144, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.228630542755127, "learning_rate": 5e-05, "loss": 0.6699, "mean_token_accuracy": 0.8835381269454956, "num_tokens": 801800694.0, "step": 21113 }, { "epoch": 2.6859178221600306, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 10.114317893981934, "learning_rate": 5e-05, "loss": 0.9777, "mean_token_accuracy": 0.8910998106002808, "num_tokens": 801835748.0, "step": 21114 }, { "epoch": 2.686045032438621, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.180851459503174, "learning_rate": 5e-05, "loss": 0.8121, "mean_token_accuracy": 0.8730658888816833, "num_tokens": 801873527.0, "step": 21115 }, { "epoch": 2.6861722427172117, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.472319602966309, "learning_rate": 5e-05, "loss": 0.9774, "mean_token_accuracy": 0.8896878957748413, "num_tokens": 801910630.0, "step": 21116 }, { "epoch": 2.686299452995802, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.41901969909668, "learning_rate": 5e-05, "loss": 0.9253, "mean_token_accuracy": 0.8858514428138733, "num_tokens": 801946136.0, "step": 21117 }, { "epoch": 2.6864266632743927, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.996390342712402, "learning_rate": 5e-05, "loss": 0.8581, "mean_token_accuracy": 0.8773493766784668, "num_tokens": 801982283.0, "step": 21118 }, { "epoch": 2.686553873552983, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.867841720581055, "learning_rate": 5e-05, "loss": 0.9366, "mean_token_accuracy": 0.8757697939872742, "num_tokens": 802022263.0, "step": 21119 }, { "epoch": 2.686681083831574, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.467217922210693, "learning_rate": 5e-05, "loss": 0.9035, "mean_token_accuracy": 0.844671905040741, "num_tokens": 802061837.0, "step": 21120 }, { "epoch": 2.686808294110164, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.35348129272461, "learning_rate": 5e-05, "loss": 0.8282, "mean_token_accuracy": 0.8858864307403564, "num_tokens": 802094528.0, "step": 21121 }, { "epoch": 2.686935504388755, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.983226299285889, "learning_rate": 5e-05, "loss": 0.8476, "mean_token_accuracy": 0.8766064047813416, "num_tokens": 802135157.0, "step": 21122 }, { "epoch": 2.687062714667345, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.402709484100342, "learning_rate": 5e-05, "loss": 0.7042, "mean_token_accuracy": 0.891616702079773, "num_tokens": 802168818.0, "step": 21123 }, { "epoch": 2.687189924945936, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.604230880737305, "learning_rate": 5e-05, "loss": 0.8733, "mean_token_accuracy": 0.875474750995636, "num_tokens": 802209486.0, "step": 21124 }, { "epoch": 2.687317135224526, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.909587860107422, "learning_rate": 5e-05, "loss": 0.7429, "mean_token_accuracy": 0.8678374290466309, "num_tokens": 802248534.0, "step": 21125 }, { "epoch": 2.6874443455031165, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.979920387268066, "learning_rate": 5e-05, "loss": 0.8848, "mean_token_accuracy": 0.880398690700531, "num_tokens": 802289330.0, "step": 21126 }, { "epoch": 2.687571555781707, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.7820844650268555, "learning_rate": 5e-05, "loss": 0.7111, "mean_token_accuracy": 0.8803349137306213, "num_tokens": 802322065.0, "step": 21127 }, { "epoch": 2.6876987660602976, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.155820846557617, "learning_rate": 5e-05, "loss": 0.9476, "mean_token_accuracy": 0.8753252625465393, "num_tokens": 802361820.0, "step": 21128 }, { "epoch": 2.687825976338888, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.938337802886963, "learning_rate": 5e-05, "loss": 0.7305, "mean_token_accuracy": 0.8811980485916138, "num_tokens": 802397748.0, "step": 21129 }, { "epoch": 2.6879531866174786, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.162643432617188, "learning_rate": 5e-05, "loss": 0.9879, "mean_token_accuracy": 0.8664120435714722, "num_tokens": 802432152.0, "step": 21130 }, { "epoch": 2.688080396896069, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.190634727478027, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8850226402282715, "num_tokens": 802471464.0, "step": 21131 }, { "epoch": 2.6882076071746597, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.988327026367188, "learning_rate": 5e-05, "loss": 0.9342, "mean_token_accuracy": 0.8809250593185425, "num_tokens": 802503096.0, "step": 21132 }, { "epoch": 2.68833481745325, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.459178447723389, "learning_rate": 5e-05, "loss": 0.8581, "mean_token_accuracy": 0.8716599941253662, "num_tokens": 802540036.0, "step": 21133 }, { "epoch": 2.6884620277318407, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.705564498901367, "learning_rate": 5e-05, "loss": 0.9346, "mean_token_accuracy": 0.8792430758476257, "num_tokens": 802580798.0, "step": 21134 }, { "epoch": 2.6885892380104313, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.5708770751953125, "learning_rate": 5e-05, "loss": 0.8364, "mean_token_accuracy": 0.8739117383956909, "num_tokens": 802619472.0, "step": 21135 }, { "epoch": 2.688716448289022, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.469006538391113, "learning_rate": 5e-05, "loss": 0.8915, "mean_token_accuracy": 0.8812910318374634, "num_tokens": 802656834.0, "step": 21136 }, { "epoch": 2.6888436585676123, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.621280193328857, "learning_rate": 5e-05, "loss": 0.7567, "mean_token_accuracy": 0.8993597030639648, "num_tokens": 802690269.0, "step": 21137 }, { "epoch": 2.688970868846203, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.2119722366333, "learning_rate": 5e-05, "loss": 0.8323, "mean_token_accuracy": 0.8901680111885071, "num_tokens": 802726822.0, "step": 21138 }, { "epoch": 2.6890980791247934, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.702768802642822, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.8888717293739319, "num_tokens": 802764479.0, "step": 21139 }, { "epoch": 2.689225289403384, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.9163007736206055, "learning_rate": 5e-05, "loss": 0.7974, "mean_token_accuracy": 0.8843996524810791, "num_tokens": 802800938.0, "step": 21140 }, { "epoch": 2.6893524996819744, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.835309028625488, "learning_rate": 5e-05, "loss": 0.7824, "mean_token_accuracy": 0.8848177194595337, "num_tokens": 802841593.0, "step": 21141 }, { "epoch": 2.689479709960565, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.670507431030273, "learning_rate": 5e-05, "loss": 0.7858, "mean_token_accuracy": 0.8783332109451294, "num_tokens": 802875847.0, "step": 21142 }, { "epoch": 2.6896069202391555, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.922275543212891, "learning_rate": 5e-05, "loss": 0.7714, "mean_token_accuracy": 0.8838618993759155, "num_tokens": 802915786.0, "step": 21143 }, { "epoch": 2.6897341305177456, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.502658843994141, "learning_rate": 5e-05, "loss": 0.7509, "mean_token_accuracy": 0.8772134780883789, "num_tokens": 802958774.0, "step": 21144 }, { "epoch": 2.6898613407963365, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.023978233337402, "learning_rate": 5e-05, "loss": 0.755, "mean_token_accuracy": 0.8939712643623352, "num_tokens": 802991360.0, "step": 21145 }, { "epoch": 2.6899885510749266, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.322011470794678, "learning_rate": 5e-05, "loss": 0.7109, "mean_token_accuracy": 0.8875943422317505, "num_tokens": 803030852.0, "step": 21146 }, { "epoch": 2.6901157613535176, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.20202922821045, "learning_rate": 5e-05, "loss": 0.7882, "mean_token_accuracy": 0.8848170638084412, "num_tokens": 803068965.0, "step": 21147 }, { "epoch": 2.6902429716321077, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.111827850341797, "learning_rate": 5e-05, "loss": 0.7271, "mean_token_accuracy": 0.8734833598136902, "num_tokens": 803103331.0, "step": 21148 }, { "epoch": 2.6903701819106987, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.429191589355469, "learning_rate": 5e-05, "loss": 0.8115, "mean_token_accuracy": 0.8837820291519165, "num_tokens": 803144242.0, "step": 21149 }, { "epoch": 2.6904973921892887, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.874278545379639, "learning_rate": 5e-05, "loss": 0.75, "mean_token_accuracy": 0.8632517457008362, "num_tokens": 803178355.0, "step": 21150 }, { "epoch": 2.6906246024678793, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.729856491088867, "learning_rate": 5e-05, "loss": 0.8649, "mean_token_accuracy": 0.8780341744422913, "num_tokens": 803218783.0, "step": 21151 }, { "epoch": 2.69075181274647, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.687094211578369, "learning_rate": 5e-05, "loss": 0.7665, "mean_token_accuracy": 0.8618930578231812, "num_tokens": 803256391.0, "step": 21152 }, { "epoch": 2.6908790230250603, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 9.057283401489258, "learning_rate": 5e-05, "loss": 0.8885, "mean_token_accuracy": 0.878860354423523, "num_tokens": 803285708.0, "step": 21153 }, { "epoch": 2.691006233303651, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.626195907592773, "learning_rate": 5e-05, "loss": 0.709, "mean_token_accuracy": 0.8799981474876404, "num_tokens": 803319711.0, "step": 21154 }, { "epoch": 2.6911334435822414, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.192377090454102, "learning_rate": 5e-05, "loss": 0.9773, "mean_token_accuracy": 0.8648570775985718, "num_tokens": 803353916.0, "step": 21155 }, { "epoch": 2.691260653860832, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.965137481689453, "learning_rate": 5e-05, "loss": 0.7247, "mean_token_accuracy": 0.8899168968200684, "num_tokens": 803388579.0, "step": 21156 }, { "epoch": 2.6913878641394224, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.823716163635254, "learning_rate": 5e-05, "loss": 0.8932, "mean_token_accuracy": 0.88918137550354, "num_tokens": 803427715.0, "step": 21157 }, { "epoch": 2.691515074418013, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.622408390045166, "learning_rate": 5e-05, "loss": 0.8697, "mean_token_accuracy": 0.871912956237793, "num_tokens": 803470363.0, "step": 21158 }, { "epoch": 2.6916422846966035, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.943733215332031, "learning_rate": 5e-05, "loss": 0.8596, "mean_token_accuracy": 0.8747182488441467, "num_tokens": 803515009.0, "step": 21159 }, { "epoch": 2.691769494975194, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.219849586486816, "learning_rate": 5e-05, "loss": 0.852, "mean_token_accuracy": 0.881968080997467, "num_tokens": 803554773.0, "step": 21160 }, { "epoch": 2.6918967052537845, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.251269340515137, "learning_rate": 5e-05, "loss": 0.7305, "mean_token_accuracy": 0.8849163055419922, "num_tokens": 803590404.0, "step": 21161 }, { "epoch": 2.692023915532375, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.460909843444824, "learning_rate": 5e-05, "loss": 0.8367, "mean_token_accuracy": 0.8870918154716492, "num_tokens": 803634388.0, "step": 21162 }, { "epoch": 2.6921511258109656, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.098100662231445, "learning_rate": 5e-05, "loss": 0.7241, "mean_token_accuracy": 0.8799909353256226, "num_tokens": 803673164.0, "step": 21163 }, { "epoch": 2.692278336089556, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.382231712341309, "learning_rate": 5e-05, "loss": 0.8404, "mean_token_accuracy": 0.879624605178833, "num_tokens": 803715271.0, "step": 21164 }, { "epoch": 2.6924055463681467, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.247682571411133, "learning_rate": 5e-05, "loss": 0.7221, "mean_token_accuracy": 0.884127140045166, "num_tokens": 803748295.0, "step": 21165 }, { "epoch": 2.692532756646737, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.120508193969727, "learning_rate": 5e-05, "loss": 0.8409, "mean_token_accuracy": 0.8706293106079102, "num_tokens": 803787591.0, "step": 21166 }, { "epoch": 2.6926599669253277, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.470122814178467, "learning_rate": 5e-05, "loss": 0.7422, "mean_token_accuracy": 0.8813870549201965, "num_tokens": 803817640.0, "step": 21167 }, { "epoch": 2.6927871772039182, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.857810974121094, "learning_rate": 5e-05, "loss": 0.8024, "mean_token_accuracy": 0.8720102310180664, "num_tokens": 803847932.0, "step": 21168 }, { "epoch": 2.6929143874825083, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.60813045501709, "learning_rate": 5e-05, "loss": 0.7689, "mean_token_accuracy": 0.8711380958557129, "num_tokens": 803884206.0, "step": 21169 }, { "epoch": 2.6930415977610993, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.6931681632995605, "learning_rate": 5e-05, "loss": 0.7502, "mean_token_accuracy": 0.8833367228507996, "num_tokens": 803923225.0, "step": 21170 }, { "epoch": 2.6931688080396894, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.693678379058838, "learning_rate": 5e-05, "loss": 0.7944, "mean_token_accuracy": 0.8667380213737488, "num_tokens": 803961880.0, "step": 21171 }, { "epoch": 2.6932960183182804, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.583842754364014, "learning_rate": 5e-05, "loss": 0.7415, "mean_token_accuracy": 0.8817610740661621, "num_tokens": 804004676.0, "step": 21172 }, { "epoch": 2.6934232285968704, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.766634941101074, "learning_rate": 5e-05, "loss": 0.7361, "mean_token_accuracy": 0.8827944993972778, "num_tokens": 804047875.0, "step": 21173 }, { "epoch": 2.693550438875461, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.513218879699707, "learning_rate": 5e-05, "loss": 0.674, "mean_token_accuracy": 0.8988642692565918, "num_tokens": 804083467.0, "step": 21174 }, { "epoch": 2.6936776491540515, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.779698371887207, "learning_rate": 5e-05, "loss": 0.7202, "mean_token_accuracy": 0.8885082602500916, "num_tokens": 804124653.0, "step": 21175 }, { "epoch": 2.693804859432642, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.476040840148926, "learning_rate": 5e-05, "loss": 0.6999, "mean_token_accuracy": 0.8876984119415283, "num_tokens": 804159081.0, "step": 21176 }, { "epoch": 2.6939320697112326, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.79764461517334, "learning_rate": 5e-05, "loss": 0.7559, "mean_token_accuracy": 0.8722072243690491, "num_tokens": 804200687.0, "step": 21177 }, { "epoch": 2.694059279989823, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.425565242767334, "learning_rate": 5e-05, "loss": 0.7279, "mean_token_accuracy": 0.8783361911773682, "num_tokens": 804237192.0, "step": 21178 }, { "epoch": 2.6941864902684136, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.845796585083008, "learning_rate": 5e-05, "loss": 0.7272, "mean_token_accuracy": 0.8847339749336243, "num_tokens": 804275078.0, "step": 21179 }, { "epoch": 2.694313700547004, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.337771892547607, "learning_rate": 5e-05, "loss": 0.7301, "mean_token_accuracy": 0.8741095066070557, "num_tokens": 804308551.0, "step": 21180 }, { "epoch": 2.6944409108255947, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.966172695159912, "learning_rate": 5e-05, "loss": 0.752, "mean_token_accuracy": 0.8787081241607666, "num_tokens": 804347530.0, "step": 21181 }, { "epoch": 2.694568121104185, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.147115230560303, "learning_rate": 5e-05, "loss": 0.7277, "mean_token_accuracy": 0.8733206391334534, "num_tokens": 804388663.0, "step": 21182 }, { "epoch": 2.6946953313827757, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.266864776611328, "learning_rate": 5e-05, "loss": 0.7595, "mean_token_accuracy": 0.8866018056869507, "num_tokens": 804423956.0, "step": 21183 }, { "epoch": 2.6948225416613663, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 6.735738754272461, "learning_rate": 5e-05, "loss": 0.6879, "mean_token_accuracy": 0.8768248558044434, "num_tokens": 804466736.0, "step": 21184 }, { "epoch": 2.694949751939957, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.905269622802734, "learning_rate": 5e-05, "loss": 0.9211, "mean_token_accuracy": 0.8551892042160034, "num_tokens": 804505101.0, "step": 21185 }, { "epoch": 2.6950769622185473, "ewc_loss": 0.306640625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00022602081298828125, "grad_norm": 6.023288249969482, "learning_rate": 5e-05, "loss": 0.6259, "mean_token_accuracy": 0.8866292834281921, "num_tokens": 804542520.0, "step": 21186 }, { "epoch": 2.695204172497138, "ewc_loss": 0.625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 10.124960899353027, "learning_rate": 5e-05, "loss": 0.9431, "mean_token_accuracy": 0.8861441612243652, "num_tokens": 804577749.0, "step": 21187 }, { "epoch": 2.6953313827757284, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 5.407750129699707, "learning_rate": 5e-05, "loss": 0.6101, "mean_token_accuracy": 0.8952107429504395, "num_tokens": 804611370.0, "step": 21188 }, { "epoch": 2.695458593054319, "ewc_loss": 0.9375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 11.830513000488281, "learning_rate": 5e-05, "loss": 1.2391, "mean_token_accuracy": 0.8894715309143066, "num_tokens": 804643696.0, "step": 21189 }, { "epoch": 2.6955858033329094, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.384669780731201, "learning_rate": 5e-05, "loss": 0.9356, "mean_token_accuracy": 0.895154595375061, "num_tokens": 804678343.0, "step": 21190 }, { "epoch": 2.6957130136115, "ewc_loss": 0.99609375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 12.264474868774414, "learning_rate": 5e-05, "loss": 1.2934, "mean_token_accuracy": 0.891075849533081, "num_tokens": 804711303.0, "step": 21191 }, { "epoch": 2.6958402238900905, "ewc_loss": 1.1171875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00103759765625, "grad_norm": 10.084982872009277, "learning_rate": 5e-05, "loss": 1.5237, "mean_token_accuracy": 0.8578305244445801, "num_tokens": 804749016.0, "step": 21192 }, { "epoch": 2.695967434168681, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.524096488952637, "learning_rate": 5e-05, "loss": 1.0462, "mean_token_accuracy": 0.8735790252685547, "num_tokens": 804789600.0, "step": 21193 }, { "epoch": 2.696094644447271, "ewc_loss": 0.9765625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000896453857421875, "grad_norm": 9.999598503112793, "learning_rate": 5e-05, "loss": 1.3042, "mean_token_accuracy": 0.8857040405273438, "num_tokens": 804829033.0, "step": 21194 }, { "epoch": 2.696221854725862, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.123725891113281, "learning_rate": 5e-05, "loss": 1.0373, "mean_token_accuracy": 0.8695460557937622, "num_tokens": 804865114.0, "step": 21195 }, { "epoch": 2.696349065004452, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00072479248046875, "grad_norm": 10.266298294067383, "learning_rate": 5e-05, "loss": 1.1946, "mean_token_accuracy": 0.8660577535629272, "num_tokens": 804907604.0, "step": 21196 }, { "epoch": 2.696476275283043, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.422780990600586, "learning_rate": 5e-05, "loss": 1.0286, "mean_token_accuracy": 0.8715205192565918, "num_tokens": 804942361.0, "step": 21197 }, { "epoch": 2.696603485561633, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.256657600402832, "learning_rate": 5e-05, "loss": 1.0507, "mean_token_accuracy": 0.8802217841148376, "num_tokens": 804980452.0, "step": 21198 }, { "epoch": 2.6967306958402237, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 8.488018035888672, "learning_rate": 5e-05, "loss": 1.0081, "mean_token_accuracy": 0.8867874145507812, "num_tokens": 805017969.0, "step": 21199 }, { "epoch": 2.6968579061188143, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.48992919921875, "learning_rate": 5e-05, "loss": 0.974, "mean_token_accuracy": 0.8860570788383484, "num_tokens": 805057053.0, "step": 21200 }, { "epoch": 2.696985116397405, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.636027336120605, "learning_rate": 5e-05, "loss": 1.0123, "mean_token_accuracy": 0.8825167417526245, "num_tokens": 805098928.0, "step": 21201 }, { "epoch": 2.6971123266759953, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.26742935180664, "learning_rate": 5e-05, "loss": 0.8766, "mean_token_accuracy": 0.8809549808502197, "num_tokens": 805140729.0, "step": 21202 }, { "epoch": 2.697239536954586, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.213094711303711, "learning_rate": 5e-05, "loss": 0.989, "mean_token_accuracy": 0.8810579180717468, "num_tokens": 805177150.0, "step": 21203 }, { "epoch": 2.6973667472331764, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.044274806976318, "learning_rate": 5e-05, "loss": 0.7642, "mean_token_accuracy": 0.8800392746925354, "num_tokens": 805221233.0, "step": 21204 }, { "epoch": 2.697493957511767, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.490992546081543, "learning_rate": 5e-05, "loss": 1.0441, "mean_token_accuracy": 0.8934531807899475, "num_tokens": 805255075.0, "step": 21205 }, { "epoch": 2.6976211677903574, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.155378818511963, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8751835823059082, "num_tokens": 805292468.0, "step": 21206 }, { "epoch": 2.697748378068948, "ewc_loss": 0.84375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000762939453125, "grad_norm": 11.417619705200195, "learning_rate": 5e-05, "loss": 1.1707, "mean_token_accuracy": 0.8823437690734863, "num_tokens": 805330648.0, "step": 21207 }, { "epoch": 2.6978755883475385, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 8.886245727539062, "learning_rate": 5e-05, "loss": 1.1401, "mean_token_accuracy": 0.8838637471199036, "num_tokens": 805368137.0, "step": 21208 }, { "epoch": 2.698002798626129, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.670618057250977, "learning_rate": 5e-05, "loss": 1.0562, "mean_token_accuracy": 0.8815364241600037, "num_tokens": 805399898.0, "step": 21209 }, { "epoch": 2.6981300089047195, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 9.178739547729492, "learning_rate": 5e-05, "loss": 1.149, "mean_token_accuracy": 0.8870612382888794, "num_tokens": 805439292.0, "step": 21210 }, { "epoch": 2.69825721918331, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.841936111450195, "learning_rate": 5e-05, "loss": 1.0636, "mean_token_accuracy": 0.8646560907363892, "num_tokens": 805479464.0, "step": 21211 }, { "epoch": 2.6983844294619006, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 8.590738296508789, "learning_rate": 5e-05, "loss": 0.9316, "mean_token_accuracy": 0.9032647609710693, "num_tokens": 805518844.0, "step": 21212 }, { "epoch": 2.698511639740491, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.500414848327637, "learning_rate": 5e-05, "loss": 0.9583, "mean_token_accuracy": 0.8901976943016052, "num_tokens": 805552192.0, "step": 21213 }, { "epoch": 2.6986388500190817, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.730253219604492, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.8849202394485474, "num_tokens": 805587343.0, "step": 21214 }, { "epoch": 2.698766060297672, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.950952529907227, "learning_rate": 5e-05, "loss": 1.0157, "mean_token_accuracy": 0.8823100924491882, "num_tokens": 805623256.0, "step": 21215 }, { "epoch": 2.6988932705762627, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.274138927459717, "learning_rate": 5e-05, "loss": 0.8418, "mean_token_accuracy": 0.8779035806655884, "num_tokens": 805664627.0, "step": 21216 }, { "epoch": 2.699020480854853, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.81078052520752, "learning_rate": 5e-05, "loss": 1.1418, "mean_token_accuracy": 0.8717905879020691, "num_tokens": 805698808.0, "step": 21217 }, { "epoch": 2.6991476911334438, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.328022003173828, "learning_rate": 5e-05, "loss": 0.9975, "mean_token_accuracy": 0.8826274275779724, "num_tokens": 805738468.0, "step": 21218 }, { "epoch": 2.699274901412034, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.359467506408691, "learning_rate": 5e-05, "loss": 1.1, "mean_token_accuracy": 0.8697502613067627, "num_tokens": 805776233.0, "step": 21219 }, { "epoch": 2.699402111690625, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 9.274718284606934, "learning_rate": 5e-05, "loss": 1.0302, "mean_token_accuracy": 0.8969593048095703, "num_tokens": 805818459.0, "step": 21220 }, { "epoch": 2.699529321969215, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.832925796508789, "learning_rate": 5e-05, "loss": 0.896, "mean_token_accuracy": 0.8875173330307007, "num_tokens": 805858340.0, "step": 21221 }, { "epoch": 2.699656532247806, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000804901123046875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.268381118774414, "learning_rate": 5e-05, "loss": 1.0176, "mean_token_accuracy": 0.8769195675849915, "num_tokens": 805895920.0, "step": 21222 }, { "epoch": 2.699783742526396, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.78244161605835, "learning_rate": 5e-05, "loss": 0.822, "mean_token_accuracy": 0.8856918811798096, "num_tokens": 805937692.0, "step": 21223 }, { "epoch": 2.6999109528049865, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.134255409240723, "learning_rate": 5e-05, "loss": 0.9653, "mean_token_accuracy": 0.8754744529724121, "num_tokens": 805974573.0, "step": 21224 }, { "epoch": 2.700038163083577, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.58172082901001, "learning_rate": 5e-05, "loss": 0.7989, "mean_token_accuracy": 0.8809693455696106, "num_tokens": 806013643.0, "step": 21225 }, { "epoch": 2.7001653733621676, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.750630378723145, "learning_rate": 5e-05, "loss": 0.9003, "mean_token_accuracy": 0.8822511434555054, "num_tokens": 806048027.0, "step": 21226 }, { "epoch": 2.700292583640758, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.687203407287598, "learning_rate": 5e-05, "loss": 0.8292, "mean_token_accuracy": 0.8715184926986694, "num_tokens": 806087269.0, "step": 21227 }, { "epoch": 2.7004197939193486, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.36970329284668, "learning_rate": 5e-05, "loss": 0.8781, "mean_token_accuracy": 0.8744672536849976, "num_tokens": 806126924.0, "step": 21228 }, { "epoch": 2.700547004197939, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.626873970031738, "learning_rate": 5e-05, "loss": 0.7876, "mean_token_accuracy": 0.880699634552002, "num_tokens": 806162510.0, "step": 21229 }, { "epoch": 2.7006742144765297, "ewc_loss": 0.5, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.286049842834473, "learning_rate": 5e-05, "loss": 0.8096, "mean_token_accuracy": 0.8875730633735657, "num_tokens": 806196519.0, "step": 21230 }, { "epoch": 2.70080142475512, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.3911967277526855, "learning_rate": 5e-05, "loss": 0.792, "mean_token_accuracy": 0.8658860921859741, "num_tokens": 806233995.0, "step": 21231 }, { "epoch": 2.7009286350337107, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.453980445861816, "learning_rate": 5e-05, "loss": 0.8415, "mean_token_accuracy": 0.8839522004127502, "num_tokens": 806268660.0, "step": 21232 }, { "epoch": 2.7010558453123013, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.073391914367676, "learning_rate": 5e-05, "loss": 0.7459, "mean_token_accuracy": 0.8738923668861389, "num_tokens": 806304187.0, "step": 21233 }, { "epoch": 2.701183055590892, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.816987991333008, "learning_rate": 5e-05, "loss": 0.8825, "mean_token_accuracy": 0.8781419396400452, "num_tokens": 806341524.0, "step": 21234 }, { "epoch": 2.7013102658694823, "ewc_loss": 0.375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.786808013916016, "learning_rate": 5e-05, "loss": 0.7463, "mean_token_accuracy": 0.8724530935287476, "num_tokens": 806384983.0, "step": 21235 }, { "epoch": 2.701437476148073, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.231517791748047, "learning_rate": 5e-05, "loss": 0.9279, "mean_token_accuracy": 0.8853787183761597, "num_tokens": 806429923.0, "step": 21236 }, { "epoch": 2.7015646864266634, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.830873489379883, "learning_rate": 5e-05, "loss": 0.7218, "mean_token_accuracy": 0.8871762156486511, "num_tokens": 806467562.0, "step": 21237 }, { "epoch": 2.701691896705254, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.288416862487793, "learning_rate": 5e-05, "loss": 0.9652, "mean_token_accuracy": 0.8801881670951843, "num_tokens": 806508634.0, "step": 21238 }, { "epoch": 2.7018191069838444, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.378670692443848, "learning_rate": 5e-05, "loss": 0.8492, "mean_token_accuracy": 0.8713189959526062, "num_tokens": 806547972.0, "step": 21239 }, { "epoch": 2.701946317262435, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.73572826385498, "learning_rate": 5e-05, "loss": 0.9396, "mean_token_accuracy": 0.8717777729034424, "num_tokens": 806591461.0, "step": 21240 }, { "epoch": 2.7020735275410255, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.943849563598633, "learning_rate": 5e-05, "loss": 0.8336, "mean_token_accuracy": 0.8940593004226685, "num_tokens": 806623276.0, "step": 21241 }, { "epoch": 2.7022007378196156, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.942914962768555, "learning_rate": 5e-05, "loss": 0.8164, "mean_token_accuracy": 0.8831560015678406, "num_tokens": 806657732.0, "step": 21242 }, { "epoch": 2.7023279480982065, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.214250564575195, "learning_rate": 5e-05, "loss": 0.8329, "mean_token_accuracy": 0.8878200650215149, "num_tokens": 806695018.0, "step": 21243 }, { "epoch": 2.7024551583767966, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.464073181152344, "learning_rate": 5e-05, "loss": 0.7069, "mean_token_accuracy": 0.8982739448547363, "num_tokens": 806735344.0, "step": 21244 }, { "epoch": 2.7025823686553876, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.312128067016602, "learning_rate": 5e-05, "loss": 0.8587, "mean_token_accuracy": 0.8777941465377808, "num_tokens": 806776854.0, "step": 21245 }, { "epoch": 2.7027095789339777, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.288086891174316, "learning_rate": 5e-05, "loss": 0.7145, "mean_token_accuracy": 0.8874466419219971, "num_tokens": 806811594.0, "step": 21246 }, { "epoch": 2.7028367892125686, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.312966346740723, "learning_rate": 5e-05, "loss": 0.8699, "mean_token_accuracy": 0.8735528588294983, "num_tokens": 806845249.0, "step": 21247 }, { "epoch": 2.7029639994911587, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.252944469451904, "learning_rate": 5e-05, "loss": 0.7537, "mean_token_accuracy": 0.8761910200119019, "num_tokens": 806885110.0, "step": 21248 }, { "epoch": 2.7030912097697493, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.297276496887207, "learning_rate": 5e-05, "loss": 0.823, "mean_token_accuracy": 0.8799477815628052, "num_tokens": 806924809.0, "step": 21249 }, { "epoch": 2.70321842004834, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.178067684173584, "learning_rate": 5e-05, "loss": 0.675, "mean_token_accuracy": 0.8954780101776123, "num_tokens": 806962386.0, "step": 21250 }, { "epoch": 2.7033456303269303, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.390738487243652, "learning_rate": 5e-05, "loss": 0.8295, "mean_token_accuracy": 0.8803966045379639, "num_tokens": 806998754.0, "step": 21251 }, { "epoch": 2.703472840605521, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.095278263092041, "learning_rate": 5e-05, "loss": 0.7121, "mean_token_accuracy": 0.8819738626480103, "num_tokens": 807036140.0, "step": 21252 }, { "epoch": 2.7036000508841114, "ewc_loss": 0.5, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.456555366516113, "learning_rate": 5e-05, "loss": 0.8494, "mean_token_accuracy": 0.8753618001937866, "num_tokens": 807077144.0, "step": 21253 }, { "epoch": 2.703727261162702, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.06771993637085, "learning_rate": 5e-05, "loss": 0.7171, "mean_token_accuracy": 0.8794459700584412, "num_tokens": 807114928.0, "step": 21254 }, { "epoch": 2.7038544714412924, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.47647476196289, "learning_rate": 5e-05, "loss": 0.8383, "mean_token_accuracy": 0.884650707244873, "num_tokens": 807151067.0, "step": 21255 }, { "epoch": 2.703981681719883, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.029682636260986, "learning_rate": 5e-05, "loss": 0.7542, "mean_token_accuracy": 0.8701488375663757, "num_tokens": 807185992.0, "step": 21256 }, { "epoch": 2.7041088919984735, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.582249641418457, "learning_rate": 5e-05, "loss": 0.8248, "mean_token_accuracy": 0.8913633823394775, "num_tokens": 807229122.0, "step": 21257 }, { "epoch": 2.704236102277064, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.015076160430908, "learning_rate": 5e-05, "loss": 0.7392, "mean_token_accuracy": 0.8768368363380432, "num_tokens": 807264118.0, "step": 21258 }, { "epoch": 2.7043633125556545, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.585107803344727, "learning_rate": 5e-05, "loss": 0.8505, "mean_token_accuracy": 0.8848068118095398, "num_tokens": 807298087.0, "step": 21259 }, { "epoch": 2.704490522834245, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.120146751403809, "learning_rate": 5e-05, "loss": 0.7579, "mean_token_accuracy": 0.8825820684432983, "num_tokens": 807338367.0, "step": 21260 }, { "epoch": 2.7046177331128356, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.431095123291016, "learning_rate": 5e-05, "loss": 0.8206, "mean_token_accuracy": 0.8907086253166199, "num_tokens": 807376858.0, "step": 21261 }, { "epoch": 2.704744943391426, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.344019412994385, "learning_rate": 5e-05, "loss": 0.7948, "mean_token_accuracy": 0.8706696033477783, "num_tokens": 807410840.0, "step": 21262 }, { "epoch": 2.7048721536700167, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.096553802490234, "learning_rate": 5e-05, "loss": 0.767, "mean_token_accuracy": 0.893243670463562, "num_tokens": 807445128.0, "step": 21263 }, { "epoch": 2.704999363948607, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.593233108520508, "learning_rate": 5e-05, "loss": 0.795, "mean_token_accuracy": 0.8739626407623291, "num_tokens": 807478884.0, "step": 21264 }, { "epoch": 2.7051265742271977, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.784578323364258, "learning_rate": 5e-05, "loss": 0.7549, "mean_token_accuracy": 0.8892046213150024, "num_tokens": 807517020.0, "step": 21265 }, { "epoch": 2.7052537845057882, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.766599655151367, "learning_rate": 5e-05, "loss": 0.8195, "mean_token_accuracy": 0.8680920004844666, "num_tokens": 807557037.0, "step": 21266 }, { "epoch": 2.7053809947843783, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.559171676635742, "learning_rate": 5e-05, "loss": 0.704, "mean_token_accuracy": 0.8943302631378174, "num_tokens": 807596876.0, "step": 21267 }, { "epoch": 2.7055082050629693, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.864076137542725, "learning_rate": 5e-05, "loss": 0.8011, "mean_token_accuracy": 0.8711855411529541, "num_tokens": 807638107.0, "step": 21268 }, { "epoch": 2.7056354153415594, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.453402996063232, "learning_rate": 5e-05, "loss": 0.7646, "mean_token_accuracy": 0.8686951398849487, "num_tokens": 807678751.0, "step": 21269 }, { "epoch": 2.7057626256201504, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.870782852172852, "learning_rate": 5e-05, "loss": 0.8285, "mean_token_accuracy": 0.860348105430603, "num_tokens": 807711447.0, "step": 21270 }, { "epoch": 2.7058898358987404, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.411610126495361, "learning_rate": 5e-05, "loss": 0.7112, "mean_token_accuracy": 0.8839638233184814, "num_tokens": 807747157.0, "step": 21271 }, { "epoch": 2.706017046177331, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.869932651519775, "learning_rate": 5e-05, "loss": 0.7691, "mean_token_accuracy": 0.8749298453330994, "num_tokens": 807782695.0, "step": 21272 }, { "epoch": 2.7061442564559215, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.396185874938965, "learning_rate": 5e-05, "loss": 0.731, "mean_token_accuracy": 0.8821811079978943, "num_tokens": 807816850.0, "step": 21273 }, { "epoch": 2.706271466734512, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.881720066070557, "learning_rate": 5e-05, "loss": 0.7784, "mean_token_accuracy": 0.8696073293685913, "num_tokens": 807856246.0, "step": 21274 }, { "epoch": 2.7063986770131025, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.355586528778076, "learning_rate": 5e-05, "loss": 0.6918, "mean_token_accuracy": 0.8864151239395142, "num_tokens": 807891417.0, "step": 21275 }, { "epoch": 2.706525887291693, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.921332359313965, "learning_rate": 5e-05, "loss": 0.7249, "mean_token_accuracy": 0.8868352174758911, "num_tokens": 807924831.0, "step": 21276 }, { "epoch": 2.7066530975702836, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.263326168060303, "learning_rate": 5e-05, "loss": 0.6697, "mean_token_accuracy": 0.8942874670028687, "num_tokens": 807955427.0, "step": 21277 }, { "epoch": 2.706780307848874, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.052059173583984, "learning_rate": 5e-05, "loss": 0.7681, "mean_token_accuracy": 0.8806427717208862, "num_tokens": 807990799.0, "step": 21278 }, { "epoch": 2.7069075181274647, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 7.0838398933410645, "learning_rate": 5e-05, "loss": 0.7358, "mean_token_accuracy": 0.8689340353012085, "num_tokens": 808027028.0, "step": 21279 }, { "epoch": 2.707034728406055, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.285350799560547, "learning_rate": 5e-05, "loss": 0.7796, "mean_token_accuracy": 0.8788532614707947, "num_tokens": 808056750.0, "step": 21280 }, { "epoch": 2.7071619386846457, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.78722620010376, "learning_rate": 5e-05, "loss": 0.6889, "mean_token_accuracy": 0.8784048557281494, "num_tokens": 808096372.0, "step": 21281 }, { "epoch": 2.7072891489632362, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.73219108581543, "learning_rate": 5e-05, "loss": 0.8207, "mean_token_accuracy": 0.879500150680542, "num_tokens": 808136783.0, "step": 21282 }, { "epoch": 2.7074163592418268, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.340182781219482, "learning_rate": 5e-05, "loss": 0.6636, "mean_token_accuracy": 0.8785784244537354, "num_tokens": 808172310.0, "step": 21283 }, { "epoch": 2.7075435695204173, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.45942497253418, "learning_rate": 5e-05, "loss": 0.8945, "mean_token_accuracy": 0.8865622878074646, "num_tokens": 808207745.0, "step": 21284 }, { "epoch": 2.707670779799008, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.91988468170166, "learning_rate": 5e-05, "loss": 0.6915, "mean_token_accuracy": 0.8790061473846436, "num_tokens": 808243135.0, "step": 21285 }, { "epoch": 2.7077979900775984, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.281562805175781, "learning_rate": 5e-05, "loss": 1.0232, "mean_token_accuracy": 0.879187285900116, "num_tokens": 808280601.0, "step": 21286 }, { "epoch": 2.707925200356189, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.522237300872803, "learning_rate": 5e-05, "loss": 0.7516, "mean_token_accuracy": 0.8886202573776245, "num_tokens": 808315541.0, "step": 21287 }, { "epoch": 2.7080524106347794, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.161489486694336, "learning_rate": 5e-05, "loss": 1.0788, "mean_token_accuracy": 0.8746155500411987, "num_tokens": 808355670.0, "step": 21288 }, { "epoch": 2.70817962091337, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 8.289163589477539, "learning_rate": 5e-05, "loss": 1.0197, "mean_token_accuracy": 0.8867132663726807, "num_tokens": 808394189.0, "step": 21289 }, { "epoch": 2.7083068311919605, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.185397148132324, "learning_rate": 5e-05, "loss": 0.8466, "mean_token_accuracy": 0.8902273178100586, "num_tokens": 808432861.0, "step": 21290 }, { "epoch": 2.708434041470551, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 9.879611015319824, "learning_rate": 5e-05, "loss": 1.1067, "mean_token_accuracy": 0.876798152923584, "num_tokens": 808462423.0, "step": 21291 }, { "epoch": 2.708561251749141, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.285452842712402, "learning_rate": 5e-05, "loss": 0.6779, "mean_token_accuracy": 0.8933984637260437, "num_tokens": 808494112.0, "step": 21292 }, { "epoch": 2.708688462027732, "ewc_loss": 0.9921875, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 12.239158630371094, "learning_rate": 5e-05, "loss": 1.3285, "mean_token_accuracy": 0.8839447498321533, "num_tokens": 808538399.0, "step": 21293 }, { "epoch": 2.708815672306322, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 8.585685729980469, "learning_rate": 5e-05, "loss": 1.1122, "mean_token_accuracy": 0.8804226517677307, "num_tokens": 808575857.0, "step": 21294 }, { "epoch": 2.708942882584913, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 12.236420631408691, "learning_rate": 5e-05, "loss": 1.2542, "mean_token_accuracy": 0.8750988245010376, "num_tokens": 808615529.0, "step": 21295 }, { "epoch": 2.709070092863503, "ewc_loss": 1.0234375, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000946044921875, "grad_norm": 10.177007675170898, "learning_rate": 5e-05, "loss": 1.3307, "mean_token_accuracy": 0.8919133543968201, "num_tokens": 808653899.0, "step": 21296 }, { "epoch": 2.7091973031420937, "ewc_loss": 0.796875, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 11.305632591247559, "learning_rate": 5e-05, "loss": 1.1625, "mean_token_accuracy": 0.8697889447212219, "num_tokens": 808697372.0, "step": 21297 }, { "epoch": 2.7093245134206843, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 9.342057228088379, "learning_rate": 5e-05, "loss": 1.1158, "mean_token_accuracy": 0.8756242394447327, "num_tokens": 808733854.0, "step": 21298 }, { "epoch": 2.709451723699275, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 10.92978286743164, "learning_rate": 5e-05, "loss": 1.1679, "mean_token_accuracy": 0.8800809979438782, "num_tokens": 808777026.0, "step": 21299 }, { "epoch": 2.7095789339778653, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.72392749786377, "learning_rate": 5e-05, "loss": 0.9607, "mean_token_accuracy": 0.8903533220291138, "num_tokens": 808814142.0, "step": 21300 }, { "epoch": 2.709706144256456, "ewc_loss": 0.8125, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.000732421875, "grad_norm": 10.885392189025879, "learning_rate": 5e-05, "loss": 1.1813, "mean_token_accuracy": 0.8716546893119812, "num_tokens": 808853355.0, "step": 21301 }, { "epoch": 2.7098333545350464, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.748297691345215, "learning_rate": 5e-05, "loss": 0.9905, "mean_token_accuracy": 0.8803899884223938, "num_tokens": 808892782.0, "step": 21302 }, { "epoch": 2.709960564813637, "ewc_loss": 0.796875, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 10.862106323242188, "learning_rate": 5e-05, "loss": 1.1593, "mean_token_accuracy": 0.872764527797699, "num_tokens": 808931935.0, "step": 21303 }, { "epoch": 2.7100877750922274, "ewc_loss": 0.703125, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.072068214416504, "learning_rate": 5e-05, "loss": 1.0411, "mean_token_accuracy": 0.879525363445282, "num_tokens": 808973525.0, "step": 21304 }, { "epoch": 2.710214985370818, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.42908000946045, "learning_rate": 5e-05, "loss": 1.1107, "mean_token_accuracy": 0.869634211063385, "num_tokens": 809008198.0, "step": 21305 }, { "epoch": 2.7103421956494085, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.962303161621094, "learning_rate": 5e-05, "loss": 0.949, "mean_token_accuracy": 0.8984825015068054, "num_tokens": 809038267.0, "step": 21306 }, { "epoch": 2.710469405927999, "ewc_loss": 0.71875, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.084303855895996, "learning_rate": 5e-05, "loss": 1.0662, "mean_token_accuracy": 0.8750396370887756, "num_tokens": 809073380.0, "step": 21307 }, { "epoch": 2.7105966162065895, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.574203491210938, "learning_rate": 5e-05, "loss": 0.9224, "mean_token_accuracy": 0.8886966109275818, "num_tokens": 809114635.0, "step": 21308 }, { "epoch": 2.71072382648518, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.09832763671875, "learning_rate": 5e-05, "loss": 1.0024, "mean_token_accuracy": 0.9010716080665588, "num_tokens": 809149441.0, "step": 21309 }, { "epoch": 2.7108510367637706, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.369155883789062, "learning_rate": 5e-05, "loss": 0.9055, "mean_token_accuracy": 0.8843291401863098, "num_tokens": 809190902.0, "step": 21310 }, { "epoch": 2.710978247042361, "ewc_loss": 0.75, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.211983680725098, "learning_rate": 5e-05, "loss": 1.0843, "mean_token_accuracy": 0.8838785886764526, "num_tokens": 809239297.0, "step": 21311 }, { "epoch": 2.7111054573209517, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.541285514831543, "learning_rate": 5e-05, "loss": 0.9406, "mean_token_accuracy": 0.8831655383110046, "num_tokens": 809277304.0, "step": 21312 }, { "epoch": 2.711232667599542, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.040026664733887, "learning_rate": 5e-05, "loss": 1.1048, "mean_token_accuracy": 0.8667142391204834, "num_tokens": 809313975.0, "step": 21313 }, { "epoch": 2.7113598778781327, "ewc_loss": 0.625, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.70858383178711, "learning_rate": 5e-05, "loss": 0.9111, "mean_token_accuracy": 0.8967357873916626, "num_tokens": 809354214.0, "step": 21314 }, { "epoch": 2.711487088156723, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.58180046081543, "learning_rate": 5e-05, "loss": 0.9928, "mean_token_accuracy": 0.884880781173706, "num_tokens": 809388248.0, "step": 21315 }, { "epoch": 2.7116142984353138, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.568689346313477, "learning_rate": 5e-05, "loss": 0.9226, "mean_token_accuracy": 0.8794274926185608, "num_tokens": 809428256.0, "step": 21316 }, { "epoch": 2.711741508713904, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00080108642578125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.230552673339844, "learning_rate": 5e-05, "loss": 0.9564, "mean_token_accuracy": 0.8850424289703369, "num_tokens": 809463706.0, "step": 21317 }, { "epoch": 2.711868718992495, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.17493724822998, "learning_rate": 5e-05, "loss": 0.8856, "mean_token_accuracy": 0.8817989230155945, "num_tokens": 809503172.0, "step": 21318 }, { "epoch": 2.711995929271085, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.274317741394043, "learning_rate": 5e-05, "loss": 0.9962, "mean_token_accuracy": 0.8765065670013428, "num_tokens": 809539894.0, "step": 21319 }, { "epoch": 2.712123139549676, "ewc_loss": 0.5, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.830465793609619, "learning_rate": 5e-05, "loss": 0.8923, "mean_token_accuracy": 0.8642309904098511, "num_tokens": 809582781.0, "step": 21320 }, { "epoch": 2.712250349828266, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.516328811645508, "learning_rate": 5e-05, "loss": 0.9943, "mean_token_accuracy": 0.8840800523757935, "num_tokens": 809627077.0, "step": 21321 }, { "epoch": 2.7123775601068565, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.9144487380981445, "learning_rate": 5e-05, "loss": 0.8684, "mean_token_accuracy": 0.8759718537330627, "num_tokens": 809663063.0, "step": 21322 }, { "epoch": 2.712504770385447, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.466643333435059, "learning_rate": 5e-05, "loss": 1.0481, "mean_token_accuracy": 0.8667809963226318, "num_tokens": 809702386.0, "step": 21323 }, { "epoch": 2.7126319806640375, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.185995101928711, "learning_rate": 5e-05, "loss": 0.8928, "mean_token_accuracy": 0.8745052814483643, "num_tokens": 809745927.0, "step": 21324 }, { "epoch": 2.712759190942628, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.050158500671387, "learning_rate": 5e-05, "loss": 0.9354, "mean_token_accuracy": 0.8855414390563965, "num_tokens": 809777796.0, "step": 21325 }, { "epoch": 2.7128864012212186, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.200202941894531, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.8874597549438477, "num_tokens": 809813450.0, "step": 21326 }, { "epoch": 2.713013611499809, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.665623664855957, "learning_rate": 5e-05, "loss": 0.9677, "mean_token_accuracy": 0.8685721755027771, "num_tokens": 809850740.0, "step": 21327 }, { "epoch": 2.7131408217783997, "ewc_loss": 0.5, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.953787326812744, "learning_rate": 5e-05, "loss": 0.8213, "mean_token_accuracy": 0.8873172402381897, "num_tokens": 809886170.0, "step": 21328 }, { "epoch": 2.71326803205699, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.550990104675293, "learning_rate": 5e-05, "loss": 0.8886, "mean_token_accuracy": 0.8858433961868286, "num_tokens": 809922078.0, "step": 21329 }, { "epoch": 2.7133952423355807, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.662248611450195, "learning_rate": 5e-05, "loss": 0.7999, "mean_token_accuracy": 0.8783077001571655, "num_tokens": 809962495.0, "step": 21330 }, { "epoch": 2.7135224526141712, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.620997428894043, "learning_rate": 5e-05, "loss": 0.9045, "mean_token_accuracy": 0.8771904110908508, "num_tokens": 810002278.0, "step": 21331 }, { "epoch": 2.7136496628927618, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.461033821105957, "learning_rate": 5e-05, "loss": 0.7827, "mean_token_accuracy": 0.8839244842529297, "num_tokens": 810041236.0, "step": 21332 }, { "epoch": 2.7137768731713523, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.719743728637695, "learning_rate": 5e-05, "loss": 0.8608, "mean_token_accuracy": 0.8963718414306641, "num_tokens": 810078649.0, "step": 21333 }, { "epoch": 2.713904083449943, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.4478840827941895, "learning_rate": 5e-05, "loss": 0.8195, "mean_token_accuracy": 0.8706386685371399, "num_tokens": 810118246.0, "step": 21334 }, { "epoch": 2.7140312937285334, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.638788223266602, "learning_rate": 5e-05, "loss": 0.8704, "mean_token_accuracy": 0.8929252624511719, "num_tokens": 810153407.0, "step": 21335 }, { "epoch": 2.714158504007124, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.549786567687988, "learning_rate": 5e-05, "loss": 0.7814, "mean_token_accuracy": 0.8804258108139038, "num_tokens": 810188110.0, "step": 21336 }, { "epoch": 2.7142857142857144, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.432642936706543, "learning_rate": 5e-05, "loss": 0.8803, "mean_token_accuracy": 0.874446451663971, "num_tokens": 810225290.0, "step": 21337 }, { "epoch": 2.714412924564305, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.633094310760498, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8857815265655518, "num_tokens": 810259801.0, "step": 21338 }, { "epoch": 2.7145401348428955, "ewc_loss": 0.5, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.208155632019043, "learning_rate": 5e-05, "loss": 0.8097, "mean_token_accuracy": 0.8883957266807556, "num_tokens": 810296556.0, "step": 21339 }, { "epoch": 2.7146673451214856, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.683809280395508, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8825860023498535, "num_tokens": 810332282.0, "step": 21340 }, { "epoch": 2.7147945554000765, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.952652454376221, "learning_rate": 5e-05, "loss": 0.7977, "mean_token_accuracy": 0.8841112852096558, "num_tokens": 810368017.0, "step": 21341 }, { "epoch": 2.7149217656786666, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.750593662261963, "learning_rate": 5e-05, "loss": 0.8275, "mean_token_accuracy": 0.8629688024520874, "num_tokens": 810412239.0, "step": 21342 }, { "epoch": 2.7150489759572576, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.7785139083862305, "learning_rate": 5e-05, "loss": 0.7828, "mean_token_accuracy": 0.8748906850814819, "num_tokens": 810452319.0, "step": 21343 }, { "epoch": 2.7151761862358477, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.779255390167236, "learning_rate": 5e-05, "loss": 0.7777, "mean_token_accuracy": 0.8734408020973206, "num_tokens": 810490936.0, "step": 21344 }, { "epoch": 2.7153033965144386, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.676392555236816, "learning_rate": 5e-05, "loss": 0.7508, "mean_token_accuracy": 0.8815978765487671, "num_tokens": 810533565.0, "step": 21345 }, { "epoch": 2.7154306067930287, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.787875175476074, "learning_rate": 5e-05, "loss": 0.7653, "mean_token_accuracy": 0.8769546151161194, "num_tokens": 810574623.0, "step": 21346 }, { "epoch": 2.7155578170716193, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.582800388336182, "learning_rate": 5e-05, "loss": 0.7125, "mean_token_accuracy": 0.8861255645751953, "num_tokens": 810609423.0, "step": 21347 }, { "epoch": 2.71568502735021, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.813313007354736, "learning_rate": 5e-05, "loss": 0.7494, "mean_token_accuracy": 0.8843013048171997, "num_tokens": 810643814.0, "step": 21348 }, { "epoch": 2.7158122376288003, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.4993815422058105, "learning_rate": 5e-05, "loss": 0.7264, "mean_token_accuracy": 0.8787079453468323, "num_tokens": 810682393.0, "step": 21349 }, { "epoch": 2.715939447907391, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.870023250579834, "learning_rate": 5e-05, "loss": 0.7135, "mean_token_accuracy": 0.8919956088066101, "num_tokens": 810718604.0, "step": 21350 }, { "epoch": 2.7160666581859814, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.445789337158203, "learning_rate": 5e-05, "loss": 0.7167, "mean_token_accuracy": 0.8851767778396606, "num_tokens": 810751259.0, "step": 21351 }, { "epoch": 2.716193868464572, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.86124849319458, "learning_rate": 5e-05, "loss": 0.7548, "mean_token_accuracy": 0.8807893395423889, "num_tokens": 810781741.0, "step": 21352 }, { "epoch": 2.7163210787431624, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.459929466247559, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8544676899909973, "num_tokens": 810820258.0, "step": 21353 }, { "epoch": 2.716448289021753, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.834750652313232, "learning_rate": 5e-05, "loss": 0.7243, "mean_token_accuracy": 0.886726975440979, "num_tokens": 810861341.0, "step": 21354 }, { "epoch": 2.7165754993003435, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.464475154876709, "learning_rate": 5e-05, "loss": 0.7322, "mean_token_accuracy": 0.8764456510543823, "num_tokens": 810897963.0, "step": 21355 }, { "epoch": 2.716702709578934, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.808073997497559, "learning_rate": 5e-05, "loss": 0.739, "mean_token_accuracy": 0.8834032416343689, "num_tokens": 810938042.0, "step": 21356 }, { "epoch": 2.7168299198575245, "ewc_loss": 0.375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.4871015548706055, "learning_rate": 5e-05, "loss": 0.7055, "mean_token_accuracy": 0.8828176856040955, "num_tokens": 810969905.0, "step": 21357 }, { "epoch": 2.716957130136115, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.752843856811523, "learning_rate": 5e-05, "loss": 0.7044, "mean_token_accuracy": 0.8867462873458862, "num_tokens": 811006818.0, "step": 21358 }, { "epoch": 2.7170843404147056, "ewc_loss": 0.375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.517696380615234, "learning_rate": 5e-05, "loss": 0.7148, "mean_token_accuracy": 0.8799991607666016, "num_tokens": 811045867.0, "step": 21359 }, { "epoch": 2.717211550693296, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.702866554260254, "learning_rate": 5e-05, "loss": 0.7174, "mean_token_accuracy": 0.8857304453849792, "num_tokens": 811084207.0, "step": 21360 }, { "epoch": 2.7173387609718866, "ewc_loss": 0.375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.5674896240234375, "learning_rate": 5e-05, "loss": 0.742, "mean_token_accuracy": 0.8686569929122925, "num_tokens": 811123093.0, "step": 21361 }, { "epoch": 2.717465971250477, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.636643886566162, "learning_rate": 5e-05, "loss": 0.7068, "mean_token_accuracy": 0.8831143379211426, "num_tokens": 811157951.0, "step": 21362 }, { "epoch": 2.7175931815290677, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.613770008087158, "learning_rate": 5e-05, "loss": 0.6596, "mean_token_accuracy": 0.9001514911651611, "num_tokens": 811192743.0, "step": 21363 }, { "epoch": 2.7177203918076582, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.595093250274658, "learning_rate": 5e-05, "loss": 0.7018, "mean_token_accuracy": 0.8826946020126343, "num_tokens": 811230042.0, "step": 21364 }, { "epoch": 2.7178476020862483, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.648130893707275, "learning_rate": 5e-05, "loss": 0.721, "mean_token_accuracy": 0.878778338432312, "num_tokens": 811270175.0, "step": 21365 }, { "epoch": 2.7179748123648393, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.508174419403076, "learning_rate": 5e-05, "loss": 0.6942, "mean_token_accuracy": 0.8822188377380371, "num_tokens": 811302926.0, "step": 21366 }, { "epoch": 2.7181020226434294, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.723831653594971, "learning_rate": 5e-05, "loss": 0.6638, "mean_token_accuracy": 0.8997945785522461, "num_tokens": 811334503.0, "step": 21367 }, { "epoch": 2.7182292329220203, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.414505958557129, "learning_rate": 5e-05, "loss": 0.6919, "mean_token_accuracy": 0.8811018466949463, "num_tokens": 811371480.0, "step": 21368 }, { "epoch": 2.7183564432006104, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.8617072105407715, "learning_rate": 5e-05, "loss": 0.7399, "mean_token_accuracy": 0.8781251311302185, "num_tokens": 811410044.0, "step": 21369 }, { "epoch": 2.718483653479201, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 7.211935043334961, "learning_rate": 5e-05, "loss": 0.7593, "mean_token_accuracy": 0.8546727895736694, "num_tokens": 811447815.0, "step": 21370 }, { "epoch": 2.7186108637577915, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 8.141302108764648, "learning_rate": 5e-05, "loss": 0.7815, "mean_token_accuracy": 0.8718990683555603, "num_tokens": 811484530.0, "step": 21371 }, { "epoch": 2.718738074036382, "ewc_loss": 0.328125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000247955322265625, "grad_norm": 6.834063529968262, "learning_rate": 5e-05, "loss": 0.6545, "mean_token_accuracy": 0.8841493725776672, "num_tokens": 811523010.0, "step": 21372 }, { "epoch": 2.7188652843149725, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.715585708618164, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8822914361953735, "num_tokens": 811559031.0, "step": 21373 }, { "epoch": 2.718992494593563, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00021457672119140625, "grad_norm": 6.132629871368408, "learning_rate": 5e-05, "loss": 0.6852, "mean_token_accuracy": 0.8642475605010986, "num_tokens": 811596501.0, "step": 21374 }, { "epoch": 2.7191197048721536, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.914083480834961, "learning_rate": 5e-05, "loss": 0.965, "mean_token_accuracy": 0.8628610372543335, "num_tokens": 811634619.0, "step": 21375 }, { "epoch": 2.719246915150744, "ewc_loss": 0.26953125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00019073486328125, "grad_norm": 5.1873979568481445, "learning_rate": 5e-05, "loss": 0.651, "mean_token_accuracy": 0.8657231330871582, "num_tokens": 811678142.0, "step": 21376 }, { "epoch": 2.7193741254293347, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 12.026007652282715, "learning_rate": 5e-05, "loss": 1.2649, "mean_token_accuracy": 0.8751304149627686, "num_tokens": 811711072.0, "step": 21377 }, { "epoch": 2.719501335707925, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 6.651637554168701, "learning_rate": 5e-05, "loss": 0.825, "mean_token_accuracy": 0.8940014839172363, "num_tokens": 811746372.0, "step": 21378 }, { "epoch": 2.7196285459865157, "ewc_loss": 1.1953125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0011138916015625, "grad_norm": 13.847277641296387, "learning_rate": 5e-05, "loss": 1.5715, "mean_token_accuracy": 0.8687329292297363, "num_tokens": 811788763.0, "step": 21379 }, { "epoch": 2.7197557562651062, "ewc_loss": 1.2421875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00116729736328125, "grad_norm": 11.145186424255371, "learning_rate": 5e-05, "loss": 1.5554, "mean_token_accuracy": 0.8847335577011108, "num_tokens": 811820866.0, "step": 21380 }, { "epoch": 2.7198829665436968, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00077056884765625, "grad_norm": 12.116843223571777, "learning_rate": 5e-05, "loss": 1.1842, "mean_token_accuracy": 0.8822944164276123, "num_tokens": 811866684.0, "step": 21381 }, { "epoch": 2.7200101768222873, "ewc_loss": 0.9921875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 10.2275972366333, "learning_rate": 5e-05, "loss": 1.3252, "mean_token_accuracy": 0.8818879723548889, "num_tokens": 811902728.0, "step": 21382 }, { "epoch": 2.720137387100878, "ewc_loss": 0.94140625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00086212158203125, "grad_norm": 11.97573471069336, "learning_rate": 5e-05, "loss": 1.2665, "mean_token_accuracy": 0.883440375328064, "num_tokens": 811946051.0, "step": 21383 }, { "epoch": 2.7202645973794684, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 8.994089126586914, "learning_rate": 5e-05, "loss": 1.0469, "mean_token_accuracy": 0.87873375415802, "num_tokens": 811984148.0, "step": 21384 }, { "epoch": 2.720391807658059, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 11.766216278076172, "learning_rate": 5e-05, "loss": 1.3024, "mean_token_accuracy": 0.8637250661849976, "num_tokens": 812020173.0, "step": 21385 }, { "epoch": 2.7205190179366494, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.019564628601074, "learning_rate": 5e-05, "loss": 1.0176, "mean_token_accuracy": 0.8886922597885132, "num_tokens": 812054509.0, "step": 21386 }, { "epoch": 2.72064622821524, "ewc_loss": 0.859375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 11.703583717346191, "learning_rate": 5e-05, "loss": 1.1833, "mean_token_accuracy": 0.8838428258895874, "num_tokens": 812088717.0, "step": 21387 }, { "epoch": 2.7207734384938305, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 9.462667465209961, "learning_rate": 5e-05, "loss": 1.1235, "mean_token_accuracy": 0.8749768733978271, "num_tokens": 812126230.0, "step": 21388 }, { "epoch": 2.720900648772421, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 11.383813858032227, "learning_rate": 5e-05, "loss": 1.1084, "mean_token_accuracy": 0.8892328143119812, "num_tokens": 812157385.0, "step": 21389 }, { "epoch": 2.721027859051011, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 9.512614250183105, "learning_rate": 5e-05, "loss": 1.148, "mean_token_accuracy": 0.8629716634750366, "num_tokens": 812184653.0, "step": 21390 }, { "epoch": 2.721155069329602, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 10.93616771697998, "learning_rate": 5e-05, "loss": 1.065, "mean_token_accuracy": 0.8897113800048828, "num_tokens": 812226784.0, "step": 21391 }, { "epoch": 2.721282279608192, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.220052719116211, "learning_rate": 5e-05, "loss": 1.0264, "mean_token_accuracy": 0.8770984411239624, "num_tokens": 812261088.0, "step": 21392 }, { "epoch": 2.721409489886783, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.643717765808105, "learning_rate": 5e-05, "loss": 1.0424, "mean_token_accuracy": 0.8905991911888123, "num_tokens": 812296859.0, "step": 21393 }, { "epoch": 2.721536700165373, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.885588645935059, "learning_rate": 5e-05, "loss": 0.9511, "mean_token_accuracy": 0.8855639696121216, "num_tokens": 812334027.0, "step": 21394 }, { "epoch": 2.7216639104439637, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.584187507629395, "learning_rate": 5e-05, "loss": 1.0777, "mean_token_accuracy": 0.8792521953582764, "num_tokens": 812370425.0, "step": 21395 }, { "epoch": 2.7217911207225542, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.766180992126465, "learning_rate": 5e-05, "loss": 0.9524, "mean_token_accuracy": 0.8803945779800415, "num_tokens": 812402067.0, "step": 21396 }, { "epoch": 2.7219183310011448, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.569497108459473, "learning_rate": 5e-05, "loss": 1.037, "mean_token_accuracy": 0.8946109414100647, "num_tokens": 812436482.0, "step": 21397 }, { "epoch": 2.7220455412797353, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.833520889282227, "learning_rate": 5e-05, "loss": 1.0042, "mean_token_accuracy": 0.867559015750885, "num_tokens": 812476059.0, "step": 21398 }, { "epoch": 2.722172751558326, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.416929244995117, "learning_rate": 5e-05, "loss": 1.0259, "mean_token_accuracy": 0.8912690877914429, "num_tokens": 812513846.0, "step": 21399 }, { "epoch": 2.7222999618369164, "ewc_loss": 0.625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.925946235656738, "learning_rate": 5e-05, "loss": 0.9542, "mean_token_accuracy": 0.8869116306304932, "num_tokens": 812550000.0, "step": 21400 }, { "epoch": 2.722427172115507, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.066092491149902, "learning_rate": 5e-05, "loss": 1.0269, "mean_token_accuracy": 0.881053626537323, "num_tokens": 812589539.0, "step": 21401 }, { "epoch": 2.7225543823940974, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.808563232421875, "learning_rate": 5e-05, "loss": 0.9113, "mean_token_accuracy": 0.8896455764770508, "num_tokens": 812622699.0, "step": 21402 }, { "epoch": 2.722681592672688, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.729516983032227, "learning_rate": 5e-05, "loss": 0.9633, "mean_token_accuracy": 0.8889524936676025, "num_tokens": 812662945.0, "step": 21403 }, { "epoch": 2.7228088029512785, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.501310348510742, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8852770328521729, "num_tokens": 812700650.0, "step": 21404 }, { "epoch": 2.722936013229869, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.639139175415039, "learning_rate": 5e-05, "loss": 0.9521, "mean_token_accuracy": 0.8945545554161072, "num_tokens": 812742651.0, "step": 21405 }, { "epoch": 2.7230632235084595, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.224570274353027, "learning_rate": 5e-05, "loss": 0.8878, "mean_token_accuracy": 0.873069703578949, "num_tokens": 812783792.0, "step": 21406 }, { "epoch": 2.72319043378705, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.730209350585938, "learning_rate": 5e-05, "loss": 1.0197, "mean_token_accuracy": 0.8738019466400146, "num_tokens": 812829059.0, "step": 21407 }, { "epoch": 2.7233176440656406, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.192651748657227, "learning_rate": 5e-05, "loss": 0.8371, "mean_token_accuracy": 0.8847122192382812, "num_tokens": 812859088.0, "step": 21408 }, { "epoch": 2.723444854344231, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.743260383605957, "learning_rate": 5e-05, "loss": 1.0476, "mean_token_accuracy": 0.8709948658943176, "num_tokens": 812899057.0, "step": 21409 }, { "epoch": 2.7235720646228216, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.349687576293945, "learning_rate": 5e-05, "loss": 0.9022, "mean_token_accuracy": 0.8744411468505859, "num_tokens": 812940711.0, "step": 21410 }, { "epoch": 2.723699274901412, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.439476013183594, "learning_rate": 5e-05, "loss": 1.0456, "mean_token_accuracy": 0.8593504428863525, "num_tokens": 812986967.0, "step": 21411 }, { "epoch": 2.7238264851800027, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.394644737243652, "learning_rate": 5e-05, "loss": 0.8724, "mean_token_accuracy": 0.8823689818382263, "num_tokens": 813024396.0, "step": 21412 }, { "epoch": 2.723953695458593, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.021086692810059, "learning_rate": 5e-05, "loss": 0.9574, "mean_token_accuracy": 0.8739403486251831, "num_tokens": 813058204.0, "step": 21413 }, { "epoch": 2.7240809057371838, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.145943641662598, "learning_rate": 5e-05, "loss": 0.8776, "mean_token_accuracy": 0.8692882061004639, "num_tokens": 813098477.0, "step": 21414 }, { "epoch": 2.724208116015774, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.880221366882324, "learning_rate": 5e-05, "loss": 0.9521, "mean_token_accuracy": 0.8790074586868286, "num_tokens": 813138091.0, "step": 21415 }, { "epoch": 2.724335326294365, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.809451580047607, "learning_rate": 5e-05, "loss": 0.8235, "mean_token_accuracy": 0.8744453191757202, "num_tokens": 813170280.0, "step": 21416 }, { "epoch": 2.724462536572955, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.958012580871582, "learning_rate": 5e-05, "loss": 0.9742, "mean_token_accuracy": 0.8708121180534363, "num_tokens": 813209198.0, "step": 21417 }, { "epoch": 2.724589746851546, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.63103723526001, "learning_rate": 5e-05, "loss": 0.7218, "mean_token_accuracy": 0.8995422124862671, "num_tokens": 813248836.0, "step": 21418 }, { "epoch": 2.724716957130136, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.038102149963379, "learning_rate": 5e-05, "loss": 0.9555, "mean_token_accuracy": 0.8760697841644287, "num_tokens": 813279421.0, "step": 21419 }, { "epoch": 2.7248441674087265, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.686862468719482, "learning_rate": 5e-05, "loss": 0.8285, "mean_token_accuracy": 0.8715038299560547, "num_tokens": 813321615.0, "step": 21420 }, { "epoch": 2.724971377687317, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.896417617797852, "learning_rate": 5e-05, "loss": 0.9256, "mean_token_accuracy": 0.8836336135864258, "num_tokens": 813357338.0, "step": 21421 }, { "epoch": 2.7250985879659075, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.7876296043396, "learning_rate": 5e-05, "loss": 0.836, "mean_token_accuracy": 0.877223789691925, "num_tokens": 813397774.0, "step": 21422 }, { "epoch": 2.725225798244498, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.623013496398926, "learning_rate": 5e-05, "loss": 0.8846, "mean_token_accuracy": 0.8878892660140991, "num_tokens": 813432373.0, "step": 21423 }, { "epoch": 2.7253530085230886, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.813263416290283, "learning_rate": 5e-05, "loss": 0.757, "mean_token_accuracy": 0.8985703587532043, "num_tokens": 813472354.0, "step": 21424 }, { "epoch": 2.725480218801679, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.353337287902832, "learning_rate": 5e-05, "loss": 0.8286, "mean_token_accuracy": 0.8890442848205566, "num_tokens": 813505498.0, "step": 21425 }, { "epoch": 2.7256074290802697, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.765727996826172, "learning_rate": 5e-05, "loss": 0.7658, "mean_token_accuracy": 0.8853944540023804, "num_tokens": 813546886.0, "step": 21426 }, { "epoch": 2.72573463935886, "ewc_loss": 0.5, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.160690307617188, "learning_rate": 5e-05, "loss": 0.8298, "mean_token_accuracy": 0.8836594820022583, "num_tokens": 813583753.0, "step": 21427 }, { "epoch": 2.7258618496374507, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.696831226348877, "learning_rate": 5e-05, "loss": 0.7721, "mean_token_accuracy": 0.8776484727859497, "num_tokens": 813617561.0, "step": 21428 }, { "epoch": 2.7259890599160412, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.037789344787598, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.8667099475860596, "num_tokens": 813652440.0, "step": 21429 }, { "epoch": 2.7261162701946318, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.6429762840271, "learning_rate": 5e-05, "loss": 0.7281, "mean_token_accuracy": 0.8861410617828369, "num_tokens": 813691118.0, "step": 21430 }, { "epoch": 2.7262434804732223, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.986824989318848, "learning_rate": 5e-05, "loss": 0.8084, "mean_token_accuracy": 0.8714752197265625, "num_tokens": 813733238.0, "step": 21431 }, { "epoch": 2.726370690751813, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.606039047241211, "learning_rate": 5e-05, "loss": 0.7305, "mean_token_accuracy": 0.885053277015686, "num_tokens": 813770002.0, "step": 21432 }, { "epoch": 2.7264979010304033, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.906495094299316, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.8772120475769043, "num_tokens": 813810778.0, "step": 21433 }, { "epoch": 2.726625111308994, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.596322059631348, "learning_rate": 5e-05, "loss": 0.7327, "mean_token_accuracy": 0.8799701929092407, "num_tokens": 813844004.0, "step": 21434 }, { "epoch": 2.7267523215875844, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.859287261962891, "learning_rate": 5e-05, "loss": 0.7411, "mean_token_accuracy": 0.8857691287994385, "num_tokens": 813884399.0, "step": 21435 }, { "epoch": 2.726879531866175, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.590631008148193, "learning_rate": 5e-05, "loss": 0.7053, "mean_token_accuracy": 0.8908490538597107, "num_tokens": 813920454.0, "step": 21436 }, { "epoch": 2.7270067421447655, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.835099220275879, "learning_rate": 5e-05, "loss": 0.7223, "mean_token_accuracy": 0.8871309757232666, "num_tokens": 813954629.0, "step": 21437 }, { "epoch": 2.7271339524233555, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.548393726348877, "learning_rate": 5e-05, "loss": 0.7122, "mean_token_accuracy": 0.8853244185447693, "num_tokens": 813992661.0, "step": 21438 }, { "epoch": 2.7272611627019465, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.855581283569336, "learning_rate": 5e-05, "loss": 0.7317, "mean_token_accuracy": 0.8833345770835876, "num_tokens": 814024401.0, "step": 21439 }, { "epoch": 2.7273883729805366, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.460599422454834, "learning_rate": 5e-05, "loss": 0.7247, "mean_token_accuracy": 0.8806338310241699, "num_tokens": 814064950.0, "step": 21440 }, { "epoch": 2.7275155832591276, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.9241814613342285, "learning_rate": 5e-05, "loss": 0.7465, "mean_token_accuracy": 0.884311318397522, "num_tokens": 814109503.0, "step": 21441 }, { "epoch": 2.7276427935377177, "ewc_loss": 0.375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.404723167419434, "learning_rate": 5e-05, "loss": 0.7345, "mean_token_accuracy": 0.8780971765518188, "num_tokens": 814143981.0, "step": 21442 }, { "epoch": 2.7277700038163086, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.95450496673584, "learning_rate": 5e-05, "loss": 0.7231, "mean_token_accuracy": 0.8893226981163025, "num_tokens": 814180781.0, "step": 21443 }, { "epoch": 2.7278972140948987, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.348453998565674, "learning_rate": 5e-05, "loss": 0.7109, "mean_token_accuracy": 0.8782109618186951, "num_tokens": 814224841.0, "step": 21444 }, { "epoch": 2.7280244243734892, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 8.01596450805664, "learning_rate": 5e-05, "loss": 0.7438, "mean_token_accuracy": 0.8835644125938416, "num_tokens": 814258618.0, "step": 21445 }, { "epoch": 2.7281516346520798, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.2528557777404785, "learning_rate": 5e-05, "loss": 0.6977, "mean_token_accuracy": 0.8833173513412476, "num_tokens": 814297286.0, "step": 21446 }, { "epoch": 2.7282788449306703, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 8.155974388122559, "learning_rate": 5e-05, "loss": 0.7691, "mean_token_accuracy": 0.8789904713630676, "num_tokens": 814331540.0, "step": 21447 }, { "epoch": 2.728406055209261, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 7.098093509674072, "learning_rate": 5e-05, "loss": 0.7308, "mean_token_accuracy": 0.8662012219429016, "num_tokens": 814371870.0, "step": 21448 }, { "epoch": 2.7285332654878514, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.313806533813477, "learning_rate": 5e-05, "loss": 0.7605, "mean_token_accuracy": 0.8847795724868774, "num_tokens": 814406561.0, "step": 21449 }, { "epoch": 2.728660475766442, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.962001323699951, "learning_rate": 5e-05, "loss": 0.6762, "mean_token_accuracy": 0.8834584355354309, "num_tokens": 814444176.0, "step": 21450 }, { "epoch": 2.7287876860450324, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.46939468383789, "learning_rate": 5e-05, "loss": 0.8086, "mean_token_accuracy": 0.8770376443862915, "num_tokens": 814485224.0, "step": 21451 }, { "epoch": 2.728914896323623, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.891598224639893, "learning_rate": 5e-05, "loss": 0.6642, "mean_token_accuracy": 0.8869165182113647, "num_tokens": 814518006.0, "step": 21452 }, { "epoch": 2.7290421066022135, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.516193389892578, "learning_rate": 5e-05, "loss": 0.8237, "mean_token_accuracy": 0.8766714334487915, "num_tokens": 814562883.0, "step": 21453 }, { "epoch": 2.729169316880804, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.956310272216797, "learning_rate": 5e-05, "loss": 0.6819, "mean_token_accuracy": 0.8875118494033813, "num_tokens": 814604619.0, "step": 21454 }, { "epoch": 2.7292965271593945, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.378607749938965, "learning_rate": 5e-05, "loss": 0.8057, "mean_token_accuracy": 0.8770388960838318, "num_tokens": 814641112.0, "step": 21455 }, { "epoch": 2.729423737437985, "ewc_loss": 0.375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.235810279846191, "learning_rate": 5e-05, "loss": 0.7095, "mean_token_accuracy": 0.8786320090293884, "num_tokens": 814674300.0, "step": 21456 }, { "epoch": 2.7295509477165756, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.934996604919434, "learning_rate": 5e-05, "loss": 0.7414, "mean_token_accuracy": 0.8840469121932983, "num_tokens": 814711013.0, "step": 21457 }, { "epoch": 2.729678157995166, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.75446081161499, "learning_rate": 5e-05, "loss": 0.7215, "mean_token_accuracy": 0.8855942487716675, "num_tokens": 814750789.0, "step": 21458 }, { "epoch": 2.7298053682737566, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.248170852661133, "learning_rate": 5e-05, "loss": 0.7088, "mean_token_accuracy": 0.8815063238143921, "num_tokens": 814791498.0, "step": 21459 }, { "epoch": 2.729932578552347, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.496658325195312, "learning_rate": 5e-05, "loss": 0.8003, "mean_token_accuracy": 0.8816713094711304, "num_tokens": 814829811.0, "step": 21460 }, { "epoch": 2.7300597888309377, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00023937225341796875, "grad_norm": 6.3873772621154785, "learning_rate": 5e-05, "loss": 0.6569, "mean_token_accuracy": 0.8800150752067566, "num_tokens": 814864878.0, "step": 21461 }, { "epoch": 2.7301869991095282, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.623169898986816, "learning_rate": 5e-05, "loss": 0.8927, "mean_token_accuracy": 0.8846522569656372, "num_tokens": 814903843.0, "step": 21462 }, { "epoch": 2.7303142093881183, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00021266937255859375, "grad_norm": 5.537108421325684, "learning_rate": 5e-05, "loss": 0.6118, "mean_token_accuracy": 0.8857284784317017, "num_tokens": 814943076.0, "step": 21463 }, { "epoch": 2.7304414196667093, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 11.191434860229492, "learning_rate": 5e-05, "loss": 1.0937, "mean_token_accuracy": 0.8878275156021118, "num_tokens": 814981257.0, "step": 21464 }, { "epoch": 2.7305686299452994, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 6.411618709564209, "learning_rate": 5e-05, "loss": 0.7655, "mean_token_accuracy": 0.8883531093597412, "num_tokens": 815019763.0, "step": 21465 }, { "epoch": 2.7306958402238903, "ewc_loss": 0.97265625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000896453857421875, "grad_norm": 12.016937255859375, "learning_rate": 5e-05, "loss": 1.3096, "mean_token_accuracy": 0.881359875202179, "num_tokens": 815061444.0, "step": 21466 }, { "epoch": 2.7308230505024804, "ewc_loss": 0.94140625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00086212158203125, "grad_norm": 9.513056755065918, "learning_rate": 5e-05, "loss": 1.2912, "mean_token_accuracy": 0.874835729598999, "num_tokens": 815105141.0, "step": 21467 }, { "epoch": 2.730950260781071, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.077486991882324, "learning_rate": 5e-05, "loss": 1.0101, "mean_token_accuracy": 0.8780614733695984, "num_tokens": 815142839.0, "step": 21468 }, { "epoch": 2.7310774710596615, "ewc_loss": 0.91796875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0008392333984375, "grad_norm": 10.186962127685547, "learning_rate": 5e-05, "loss": 1.2271, "mean_token_accuracy": 0.8890405893325806, "num_tokens": 815183077.0, "step": 21469 }, { "epoch": 2.731204681338252, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.27596664428711, "learning_rate": 5e-05, "loss": 0.8891, "mean_token_accuracy": 0.8805088996887207, "num_tokens": 815222601.0, "step": 21470 }, { "epoch": 2.7313318916168425, "ewc_loss": 0.91015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00083160400390625, "grad_norm": 11.427447319030762, "learning_rate": 5e-05, "loss": 1.2696, "mean_token_accuracy": 0.876992404460907, "num_tokens": 815264587.0, "step": 21471 }, { "epoch": 2.731459101895433, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 8.662480354309082, "learning_rate": 5e-05, "loss": 1.0302, "mean_token_accuracy": 0.877886176109314, "num_tokens": 815302575.0, "step": 21472 }, { "epoch": 2.7315863121740236, "ewc_loss": 0.82421875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 11.615517616271973, "learning_rate": 5e-05, "loss": 1.1671, "mean_token_accuracy": 0.8802487850189209, "num_tokens": 815342230.0, "step": 21473 }, { "epoch": 2.731713522452614, "ewc_loss": 0.8125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000736236572265625, "grad_norm": 9.323155403137207, "learning_rate": 5e-05, "loss": 1.1284, "mean_token_accuracy": 0.8875941038131714, "num_tokens": 815375906.0, "step": 21474 }, { "epoch": 2.7318407327312046, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.993941307067871, "learning_rate": 5e-05, "loss": 1.0727, "mean_token_accuracy": 0.8771743774414062, "num_tokens": 815415766.0, "step": 21475 }, { "epoch": 2.731967943009795, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 8.901792526245117, "learning_rate": 5e-05, "loss": 1.0395, "mean_token_accuracy": 0.8869086503982544, "num_tokens": 815449979.0, "step": 21476 }, { "epoch": 2.7320951532883857, "ewc_loss": 0.703125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.463034629821777, "learning_rate": 5e-05, "loss": 1.0641, "mean_token_accuracy": 0.8762362003326416, "num_tokens": 815489736.0, "step": 21477 }, { "epoch": 2.7322223635669762, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.41523551940918, "learning_rate": 5e-05, "loss": 0.9385, "mean_token_accuracy": 0.8874090909957886, "num_tokens": 815525713.0, "step": 21478 }, { "epoch": 2.7323495738455668, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.979090690612793, "learning_rate": 5e-05, "loss": 1.0336, "mean_token_accuracy": 0.8767352104187012, "num_tokens": 815569231.0, "step": 21479 }, { "epoch": 2.7324767841241573, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.129297256469727, "learning_rate": 5e-05, "loss": 0.9581, "mean_token_accuracy": 0.8735463619232178, "num_tokens": 815607958.0, "step": 21480 }, { "epoch": 2.732603994402748, "ewc_loss": 0.625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.56508731842041, "learning_rate": 5e-05, "loss": 0.9821, "mean_token_accuracy": 0.873465895652771, "num_tokens": 815645919.0, "step": 21481 }, { "epoch": 2.7327312046813383, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 7.994731426239014, "learning_rate": 5e-05, "loss": 0.923, "mean_token_accuracy": 0.8750253915786743, "num_tokens": 815688568.0, "step": 21482 }, { "epoch": 2.732858414959929, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.10575008392334, "learning_rate": 5e-05, "loss": 0.9431, "mean_token_accuracy": 0.8746621608734131, "num_tokens": 815723130.0, "step": 21483 }, { "epoch": 2.7329856252385194, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.921964168548584, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8788354396820068, "num_tokens": 815764806.0, "step": 21484 }, { "epoch": 2.73311283551711, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.615062713623047, "learning_rate": 5e-05, "loss": 0.9002, "mean_token_accuracy": 0.8717840909957886, "num_tokens": 815800226.0, "step": 21485 }, { "epoch": 2.7332400457957005, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.862709999084473, "learning_rate": 5e-05, "loss": 0.8102, "mean_token_accuracy": 0.8899279236793518, "num_tokens": 815833625.0, "step": 21486 }, { "epoch": 2.733367256074291, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.271817207336426, "learning_rate": 5e-05, "loss": 0.8126, "mean_token_accuracy": 0.8903316855430603, "num_tokens": 815870617.0, "step": 21487 }, { "epoch": 2.733494466352881, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.743412017822266, "learning_rate": 5e-05, "loss": 0.8213, "mean_token_accuracy": 0.8751200437545776, "num_tokens": 815910854.0, "step": 21488 }, { "epoch": 2.733621676631472, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.142334938049316, "learning_rate": 5e-05, "loss": 0.8062, "mean_token_accuracy": 0.8847469091415405, "num_tokens": 815946467.0, "step": 21489 }, { "epoch": 2.733748886910062, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.591012001037598, "learning_rate": 5e-05, "loss": 0.8339, "mean_token_accuracy": 0.8632138967514038, "num_tokens": 815986924.0, "step": 21490 }, { "epoch": 2.733876097188653, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.189806938171387, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.887211799621582, "num_tokens": 816023960.0, "step": 21491 }, { "epoch": 2.734003307467243, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.308016300201416, "learning_rate": 5e-05, "loss": 0.7493, "mean_token_accuracy": 0.8754441142082214, "num_tokens": 816063911.0, "step": 21492 }, { "epoch": 2.7341305177458337, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.545002937316895, "learning_rate": 5e-05, "loss": 0.9153, "mean_token_accuracy": 0.8600225448608398, "num_tokens": 816099979.0, "step": 21493 }, { "epoch": 2.7342577280244242, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.876994609832764, "learning_rate": 5e-05, "loss": 0.7279, "mean_token_accuracy": 0.8720249533653259, "num_tokens": 816135226.0, "step": 21494 }, { "epoch": 2.7343849383030148, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.191924095153809, "learning_rate": 5e-05, "loss": 0.9538, "mean_token_accuracy": 0.8723688721656799, "num_tokens": 816174031.0, "step": 21495 }, { "epoch": 2.7345121485816053, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.56292724609375, "learning_rate": 5e-05, "loss": 0.6892, "mean_token_accuracy": 0.8857978582382202, "num_tokens": 816211194.0, "step": 21496 }, { "epoch": 2.734639358860196, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.886087417602539, "learning_rate": 5e-05, "loss": 1.0084, "mean_token_accuracy": 0.8815063238143921, "num_tokens": 816253849.0, "step": 21497 }, { "epoch": 2.7347665691387864, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.094475269317627, "learning_rate": 5e-05, "loss": 0.8146, "mean_token_accuracy": 0.8797075748443604, "num_tokens": 816290081.0, "step": 21498 }, { "epoch": 2.734893779417377, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.74074649810791, "learning_rate": 5e-05, "loss": 0.9623, "mean_token_accuracy": 0.9010563492774963, "num_tokens": 816324818.0, "step": 21499 }, { "epoch": 2.7350209896959674, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.313353538513184, "learning_rate": 5e-05, "loss": 0.9969, "mean_token_accuracy": 0.8785498142242432, "num_tokens": 816369946.0, "step": 21500 }, { "epoch": 2.735148199974558, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.304122924804688, "learning_rate": 5e-05, "loss": 0.9282, "mean_token_accuracy": 0.866707444190979, "num_tokens": 816405541.0, "step": 21501 }, { "epoch": 2.7352754102531485, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.094101905822754, "learning_rate": 5e-05, "loss": 1.004, "mean_token_accuracy": 0.8755404949188232, "num_tokens": 816441232.0, "step": 21502 }, { "epoch": 2.735402620531739, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.8684468269348145, "learning_rate": 5e-05, "loss": 0.739, "mean_token_accuracy": 0.8773599863052368, "num_tokens": 816474538.0, "step": 21503 }, { "epoch": 2.7355298308103295, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.961174011230469, "learning_rate": 5e-05, "loss": 1.0281, "mean_token_accuracy": 0.8766474723815918, "num_tokens": 816512706.0, "step": 21504 }, { "epoch": 2.73565704108892, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.681911945343018, "learning_rate": 5e-05, "loss": 0.7287, "mean_token_accuracy": 0.8890399932861328, "num_tokens": 816548638.0, "step": 21505 }, { "epoch": 2.7357842513675106, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.497152328491211, "learning_rate": 5e-05, "loss": 1.1079, "mean_token_accuracy": 0.8729109764099121, "num_tokens": 816583721.0, "step": 21506 }, { "epoch": 2.735911461646101, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.969310760498047, "learning_rate": 5e-05, "loss": 0.9365, "mean_token_accuracy": 0.8819403648376465, "num_tokens": 816621873.0, "step": 21507 }, { "epoch": 2.7360386719246916, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.783249855041504, "learning_rate": 5e-05, "loss": 1.0606, "mean_token_accuracy": 0.8682219982147217, "num_tokens": 816659372.0, "step": 21508 }, { "epoch": 2.736165882203282, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 8.813434600830078, "learning_rate": 5e-05, "loss": 0.9875, "mean_token_accuracy": 0.8946380615234375, "num_tokens": 816696426.0, "step": 21509 }, { "epoch": 2.7362930924818727, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.545174598693848, "learning_rate": 5e-05, "loss": 0.8852, "mean_token_accuracy": 0.8843499422073364, "num_tokens": 816735344.0, "step": 21510 }, { "epoch": 2.7364203027604628, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.736217498779297, "learning_rate": 5e-05, "loss": 0.9451, "mean_token_accuracy": 0.8906025290489197, "num_tokens": 816773026.0, "step": 21511 }, { "epoch": 2.7365475130390537, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.073811531066895, "learning_rate": 5e-05, "loss": 0.8371, "mean_token_accuracy": 0.8849804401397705, "num_tokens": 816809777.0, "step": 21512 }, { "epoch": 2.736674723317644, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.196836471557617, "learning_rate": 5e-05, "loss": 0.9035, "mean_token_accuracy": 0.8680408000946045, "num_tokens": 816849018.0, "step": 21513 }, { "epoch": 2.736801933596235, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.135805130004883, "learning_rate": 5e-05, "loss": 0.8572, "mean_token_accuracy": 0.8761992454528809, "num_tokens": 816888867.0, "step": 21514 }, { "epoch": 2.736929143874825, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.708854675292969, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8857637047767639, "num_tokens": 816929221.0, "step": 21515 }, { "epoch": 2.737056354153416, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.16983413696289, "learning_rate": 5e-05, "loss": 0.8168, "mean_token_accuracy": 0.8867737054824829, "num_tokens": 816965628.0, "step": 21516 }, { "epoch": 2.737183564432006, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.516303539276123, "learning_rate": 5e-05, "loss": 0.7655, "mean_token_accuracy": 0.8757932186126709, "num_tokens": 816997908.0, "step": 21517 }, { "epoch": 2.7373107747105965, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.218241691589355, "learning_rate": 5e-05, "loss": 0.812, "mean_token_accuracy": 0.881504237651825, "num_tokens": 817034339.0, "step": 21518 }, { "epoch": 2.737437984989187, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.268512725830078, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.872809112071991, "num_tokens": 817073006.0, "step": 21519 }, { "epoch": 2.7375651952677775, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.542010307312012, "learning_rate": 5e-05, "loss": 0.8463, "mean_token_accuracy": 0.8756456971168518, "num_tokens": 817114517.0, "step": 21520 }, { "epoch": 2.737692405546368, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.859623908996582, "learning_rate": 5e-05, "loss": 0.6886, "mean_token_accuracy": 0.8855868577957153, "num_tokens": 817158249.0, "step": 21521 }, { "epoch": 2.7378196158249586, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 9.08051872253418, "learning_rate": 5e-05, "loss": 0.8837, "mean_token_accuracy": 0.8855286836624146, "num_tokens": 817197178.0, "step": 21522 }, { "epoch": 2.737946826103549, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.613662242889404, "learning_rate": 5e-05, "loss": 0.6829, "mean_token_accuracy": 0.8851481080055237, "num_tokens": 817231406.0, "step": 21523 }, { "epoch": 2.7380740363821396, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.421821594238281, "learning_rate": 5e-05, "loss": 0.9559, "mean_token_accuracy": 0.8826109766960144, "num_tokens": 817269984.0, "step": 21524 }, { "epoch": 2.73820124666073, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 6.873523235321045, "learning_rate": 5e-05, "loss": 0.7567, "mean_token_accuracy": 0.8850851655006409, "num_tokens": 817308997.0, "step": 21525 }, { "epoch": 2.7383284569393207, "ewc_loss": 0.625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.372998237609863, "learning_rate": 5e-05, "loss": 0.996, "mean_token_accuracy": 0.871407151222229, "num_tokens": 817349329.0, "step": 21526 }, { "epoch": 2.7384556672179112, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.486054420471191, "learning_rate": 5e-05, "loss": 0.853, "mean_token_accuracy": 0.8785876035690308, "num_tokens": 817384683.0, "step": 21527 }, { "epoch": 2.7385828774965018, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.812036514282227, "learning_rate": 5e-05, "loss": 0.913, "mean_token_accuracy": 0.8793364763259888, "num_tokens": 817423202.0, "step": 21528 }, { "epoch": 2.7387100877750923, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.987687110900879, "learning_rate": 5e-05, "loss": 0.8562, "mean_token_accuracy": 0.8912321925163269, "num_tokens": 817462048.0, "step": 21529 }, { "epoch": 2.738837298053683, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.994139671325684, "learning_rate": 5e-05, "loss": 0.8482, "mean_token_accuracy": 0.8757823705673218, "num_tokens": 817508820.0, "step": 21530 }, { "epoch": 2.7389645083322733, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.263280868530273, "learning_rate": 5e-05, "loss": 0.8706, "mean_token_accuracy": 0.8769651055335999, "num_tokens": 817554140.0, "step": 21531 }, { "epoch": 2.739091718610864, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.499536514282227, "learning_rate": 5e-05, "loss": 0.8184, "mean_token_accuracy": 0.8688573837280273, "num_tokens": 817601629.0, "step": 21532 }, { "epoch": 2.7392189288894544, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.295941352844238, "learning_rate": 5e-05, "loss": 0.8665, "mean_token_accuracy": 0.8771665096282959, "num_tokens": 817641900.0, "step": 21533 }, { "epoch": 2.739346139168045, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.380829811096191, "learning_rate": 5e-05, "loss": 0.7277, "mean_token_accuracy": 0.8905429840087891, "num_tokens": 817681843.0, "step": 21534 }, { "epoch": 2.7394733494466355, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.274866104125977, "learning_rate": 5e-05, "loss": 0.8745, "mean_token_accuracy": 0.8633731603622437, "num_tokens": 817728548.0, "step": 21535 }, { "epoch": 2.7396005597252255, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.258685111999512, "learning_rate": 5e-05, "loss": 0.7227, "mean_token_accuracy": 0.8831633925437927, "num_tokens": 817759087.0, "step": 21536 }, { "epoch": 2.7397277700038165, "ewc_loss": 0.5, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.4301176071167, "learning_rate": 5e-05, "loss": 0.8673, "mean_token_accuracy": 0.8723847270011902, "num_tokens": 817799056.0, "step": 21537 }, { "epoch": 2.7398549802824066, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.046572208404541, "learning_rate": 5e-05, "loss": 0.7139, "mean_token_accuracy": 0.8832147121429443, "num_tokens": 817831859.0, "step": 21538 }, { "epoch": 2.7399821905609976, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.677602767944336, "learning_rate": 5e-05, "loss": 0.8569, "mean_token_accuracy": 0.8800241947174072, "num_tokens": 817869699.0, "step": 21539 }, { "epoch": 2.7401094008395877, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.924687385559082, "learning_rate": 5e-05, "loss": 0.7468, "mean_token_accuracy": 0.8717170357704163, "num_tokens": 817909218.0, "step": 21540 }, { "epoch": 2.7402366111181786, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.813875198364258, "learning_rate": 5e-05, "loss": 0.8699, "mean_token_accuracy": 0.8865852952003479, "num_tokens": 817948398.0, "step": 21541 }, { "epoch": 2.7403638213967687, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.967975616455078, "learning_rate": 5e-05, "loss": 0.7563, "mean_token_accuracy": 0.8731461763381958, "num_tokens": 817987351.0, "step": 21542 }, { "epoch": 2.7404910316753592, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.786270141601562, "learning_rate": 5e-05, "loss": 0.8816, "mean_token_accuracy": 0.8864458799362183, "num_tokens": 818031306.0, "step": 21543 }, { "epoch": 2.7406182419539498, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.1591715812683105, "learning_rate": 5e-05, "loss": 0.7491, "mean_token_accuracy": 0.882841944694519, "num_tokens": 818067658.0, "step": 21544 }, { "epoch": 2.7407454522325403, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.575819969177246, "learning_rate": 5e-05, "loss": 0.9312, "mean_token_accuracy": 0.8671196103096008, "num_tokens": 818109766.0, "step": 21545 }, { "epoch": 2.740872662511131, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.402913570404053, "learning_rate": 5e-05, "loss": 0.8226, "mean_token_accuracy": 0.8695158362388611, "num_tokens": 818144793.0, "step": 21546 }, { "epoch": 2.7409998727897213, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.287992477416992, "learning_rate": 5e-05, "loss": 0.8751, "mean_token_accuracy": 0.874864399433136, "num_tokens": 818184140.0, "step": 21547 }, { "epoch": 2.741127083068312, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.542380332946777, "learning_rate": 5e-05, "loss": 0.7983, "mean_token_accuracy": 0.8752020001411438, "num_tokens": 818222216.0, "step": 21548 }, { "epoch": 2.7412542933469024, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.110974311828613, "learning_rate": 5e-05, "loss": 0.7883, "mean_token_accuracy": 0.8894495368003845, "num_tokens": 818256356.0, "step": 21549 }, { "epoch": 2.741381503625493, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.519385814666748, "learning_rate": 5e-05, "loss": 0.8034, "mean_token_accuracy": 0.8725792169570923, "num_tokens": 818300012.0, "step": 21550 }, { "epoch": 2.7415087139040835, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.132135391235352, "learning_rate": 5e-05, "loss": 0.8144, "mean_token_accuracy": 0.8760746717453003, "num_tokens": 818336090.0, "step": 21551 }, { "epoch": 2.741635924182674, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.357500076293945, "learning_rate": 5e-05, "loss": 0.7383, "mean_token_accuracy": 0.8800400495529175, "num_tokens": 818376475.0, "step": 21552 }, { "epoch": 2.7417631344612645, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.257513999938965, "learning_rate": 5e-05, "loss": 0.7977, "mean_token_accuracy": 0.8881621360778809, "num_tokens": 818417483.0, "step": 21553 }, { "epoch": 2.741890344739855, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.234719753265381, "learning_rate": 5e-05, "loss": 0.7378, "mean_token_accuracy": 0.879197359085083, "num_tokens": 818460323.0, "step": 21554 }, { "epoch": 2.7420175550184456, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.331905364990234, "learning_rate": 5e-05, "loss": 0.8341, "mean_token_accuracy": 0.879599392414093, "num_tokens": 818496989.0, "step": 21555 }, { "epoch": 2.742144765297036, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.188031196594238, "learning_rate": 5e-05, "loss": 0.686, "mean_token_accuracy": 0.8915573954582214, "num_tokens": 818532755.0, "step": 21556 }, { "epoch": 2.7422719755756266, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.300898551940918, "learning_rate": 5e-05, "loss": 0.8643, "mean_token_accuracy": 0.8638306856155396, "num_tokens": 818575237.0, "step": 21557 }, { "epoch": 2.742399185854217, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.249122619628906, "learning_rate": 5e-05, "loss": 0.7187, "mean_token_accuracy": 0.8855829238891602, "num_tokens": 818607865.0, "step": 21558 }, { "epoch": 2.7425263961328077, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.178668975830078, "learning_rate": 5e-05, "loss": 0.7779, "mean_token_accuracy": 0.8880329132080078, "num_tokens": 818647807.0, "step": 21559 }, { "epoch": 2.742653606411398, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.408724308013916, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.8803948163986206, "num_tokens": 818686951.0, "step": 21560 }, { "epoch": 2.7427808166899883, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.935220241546631, "learning_rate": 5e-05, "loss": 0.7829, "mean_token_accuracy": 0.8763560056686401, "num_tokens": 818724813.0, "step": 21561 }, { "epoch": 2.7429080269685793, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.607163906097412, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.8817929029464722, "num_tokens": 818757865.0, "step": 21562 }, { "epoch": 2.7430352372471694, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.680336952209473, "learning_rate": 5e-05, "loss": 0.7711, "mean_token_accuracy": 0.8699250817298889, "num_tokens": 818796994.0, "step": 21563 }, { "epoch": 2.7431624475257603, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.802885055541992, "learning_rate": 5e-05, "loss": 0.725, "mean_token_accuracy": 0.886335015296936, "num_tokens": 818836408.0, "step": 21564 }, { "epoch": 2.7432896578043504, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.442425727844238, "learning_rate": 5e-05, "loss": 0.7221, "mean_token_accuracy": 0.880961537361145, "num_tokens": 818876134.0, "step": 21565 }, { "epoch": 2.743416868082941, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.976248264312744, "learning_rate": 5e-05, "loss": 0.7478, "mean_token_accuracy": 0.8807517290115356, "num_tokens": 818911357.0, "step": 21566 }, { "epoch": 2.7435440783615315, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.255401611328125, "learning_rate": 5e-05, "loss": 0.7077, "mean_token_accuracy": 0.8777226209640503, "num_tokens": 818942743.0, "step": 21567 }, { "epoch": 2.743671288640122, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 8.13878059387207, "learning_rate": 5e-05, "loss": 0.7681, "mean_token_accuracy": 0.8785625696182251, "num_tokens": 818979217.0, "step": 21568 }, { "epoch": 2.7437984989187125, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 7.051867961883545, "learning_rate": 5e-05, "loss": 0.732, "mean_token_accuracy": 0.867813229560852, "num_tokens": 819019991.0, "step": 21569 }, { "epoch": 2.743925709197303, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.36668586730957, "learning_rate": 5e-05, "loss": 0.8154, "mean_token_accuracy": 0.8741751909255981, "num_tokens": 819059264.0, "step": 21570 }, { "epoch": 2.7440529194758936, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.820672035217285, "learning_rate": 5e-05, "loss": 0.6995, "mean_token_accuracy": 0.880574107170105, "num_tokens": 819099649.0, "step": 21571 }, { "epoch": 2.744180129754484, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.716964721679688, "learning_rate": 5e-05, "loss": 0.8394, "mean_token_accuracy": 0.875514030456543, "num_tokens": 819142434.0, "step": 21572 }, { "epoch": 2.7443073400330746, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.531306266784668, "learning_rate": 5e-05, "loss": 0.7449, "mean_token_accuracy": 0.8584364652633667, "num_tokens": 819183908.0, "step": 21573 }, { "epoch": 2.744434550311665, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 9.139896392822266, "learning_rate": 5e-05, "loss": 0.8444, "mean_token_accuracy": 0.8942965269088745, "num_tokens": 819218975.0, "step": 21574 }, { "epoch": 2.7445617605902557, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.345934867858887, "learning_rate": 5e-05, "loss": 0.6792, "mean_token_accuracy": 0.881773054599762, "num_tokens": 819264427.0, "step": 21575 }, { "epoch": 2.7446889708688462, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.454883575439453, "learning_rate": 5e-05, "loss": 0.971, "mean_token_accuracy": 0.871948778629303, "num_tokens": 819304947.0, "step": 21576 }, { "epoch": 2.7448161811474368, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.666685581207275, "learning_rate": 5e-05, "loss": 0.7151, "mean_token_accuracy": 0.8868682384490967, "num_tokens": 819345540.0, "step": 21577 }, { "epoch": 2.7449433914260273, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.101585388183594, "learning_rate": 5e-05, "loss": 0.9554, "mean_token_accuracy": 0.8797922730445862, "num_tokens": 819383007.0, "step": 21578 }, { "epoch": 2.745070601704618, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.616682052612305, "learning_rate": 5e-05, "loss": 0.9012, "mean_token_accuracy": 0.8742323517799377, "num_tokens": 819415640.0, "step": 21579 }, { "epoch": 2.7451978119832083, "ewc_loss": 0.5, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.839113712310791, "learning_rate": 5e-05, "loss": 0.8724, "mean_token_accuracy": 0.8682089447975159, "num_tokens": 819459497.0, "step": 21580 }, { "epoch": 2.745325022261799, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.679217338562012, "learning_rate": 5e-05, "loss": 0.9577, "mean_token_accuracy": 0.8681083917617798, "num_tokens": 819497010.0, "step": 21581 }, { "epoch": 2.7454522325403894, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.612213611602783, "learning_rate": 5e-05, "loss": 0.7139, "mean_token_accuracy": 0.8803843259811401, "num_tokens": 819535659.0, "step": 21582 }, { "epoch": 2.74557944281898, "ewc_loss": 0.625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.612177848815918, "learning_rate": 5e-05, "loss": 0.9758, "mean_token_accuracy": 0.8795406222343445, "num_tokens": 819572774.0, "step": 21583 }, { "epoch": 2.7457066530975704, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.268957614898682, "learning_rate": 5e-05, "loss": 0.6626, "mean_token_accuracy": 0.8916599750518799, "num_tokens": 819607461.0, "step": 21584 }, { "epoch": 2.745833863376161, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.987767219543457, "learning_rate": 5e-05, "loss": 1.0079, "mean_token_accuracy": 0.8862528800964355, "num_tokens": 819645213.0, "step": 21585 }, { "epoch": 2.745961073654751, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.185324192047119, "learning_rate": 5e-05, "loss": 0.8635, "mean_token_accuracy": 0.883538007736206, "num_tokens": 819679537.0, "step": 21586 }, { "epoch": 2.746088283933342, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.22354507446289, "learning_rate": 5e-05, "loss": 0.9764, "mean_token_accuracy": 0.8847206830978394, "num_tokens": 819717414.0, "step": 21587 }, { "epoch": 2.746215494211932, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.29178237915039, "learning_rate": 5e-05, "loss": 0.9614, "mean_token_accuracy": 0.8858761191368103, "num_tokens": 819757875.0, "step": 21588 }, { "epoch": 2.746342704490523, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.873530387878418, "learning_rate": 5e-05, "loss": 0.8692, "mean_token_accuracy": 0.8756301403045654, "num_tokens": 819796002.0, "step": 21589 }, { "epoch": 2.746469914769113, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.81404972076416, "learning_rate": 5e-05, "loss": 1.0136, "mean_token_accuracy": 0.8673542141914368, "num_tokens": 819834736.0, "step": 21590 }, { "epoch": 2.7465971250477037, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.346241474151611, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8934435844421387, "num_tokens": 819877566.0, "step": 21591 }, { "epoch": 2.7467243353262942, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.492142677307129, "learning_rate": 5e-05, "loss": 0.9001, "mean_token_accuracy": 0.8772244453430176, "num_tokens": 819914843.0, "step": 21592 }, { "epoch": 2.7468515456048848, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.772965908050537, "learning_rate": 5e-05, "loss": 0.865, "mean_token_accuracy": 0.8781202435493469, "num_tokens": 819957181.0, "step": 21593 }, { "epoch": 2.7469787558834753, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.639039039611816, "learning_rate": 5e-05, "loss": 0.7888, "mean_token_accuracy": 0.8839075565338135, "num_tokens": 819991291.0, "step": 21594 }, { "epoch": 2.747105966162066, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.307726860046387, "learning_rate": 5e-05, "loss": 0.8737, "mean_token_accuracy": 0.8739678263664246, "num_tokens": 820033096.0, "step": 21595 }, { "epoch": 2.7472331764406563, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.118951320648193, "learning_rate": 5e-05, "loss": 0.7038, "mean_token_accuracy": 0.8875445127487183, "num_tokens": 820070200.0, "step": 21596 }, { "epoch": 2.747360386719247, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.586182594299316, "learning_rate": 5e-05, "loss": 0.8395, "mean_token_accuracy": 0.890649676322937, "num_tokens": 820100989.0, "step": 21597 }, { "epoch": 2.7474875969978374, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.9442219734191895, "learning_rate": 5e-05, "loss": 0.7387, "mean_token_accuracy": 0.8755180835723877, "num_tokens": 820145052.0, "step": 21598 }, { "epoch": 2.747614807276428, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.785662651062012, "learning_rate": 5e-05, "loss": 0.8571, "mean_token_accuracy": 0.8900404572486877, "num_tokens": 820181463.0, "step": 21599 }, { "epoch": 2.7477420175550185, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.814935207366943, "learning_rate": 5e-05, "loss": 0.7403, "mean_token_accuracy": 0.8772157430648804, "num_tokens": 820219226.0, "step": 21600 }, { "epoch": 2.747869227833609, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.063617706298828, "learning_rate": 5e-05, "loss": 0.9418, "mean_token_accuracy": 0.8759273290634155, "num_tokens": 820255953.0, "step": 21601 }, { "epoch": 2.7479964381121995, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 6.740431785583496, "learning_rate": 5e-05, "loss": 0.7688, "mean_token_accuracy": 0.8751511573791504, "num_tokens": 820296587.0, "step": 21602 }, { "epoch": 2.74812364839079, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.30097770690918, "learning_rate": 5e-05, "loss": 0.9601, "mean_token_accuracy": 0.8820645809173584, "num_tokens": 820333815.0, "step": 21603 }, { "epoch": 2.7482508586693806, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 6.8780412673950195, "learning_rate": 5e-05, "loss": 0.7975, "mean_token_accuracy": 0.8829259872436523, "num_tokens": 820376401.0, "step": 21604 }, { "epoch": 2.748378068947971, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.340171813964844, "learning_rate": 5e-05, "loss": 0.9681, "mean_token_accuracy": 0.8839824199676514, "num_tokens": 820417355.0, "step": 21605 }, { "epoch": 2.7485052792265616, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.290302276611328, "learning_rate": 5e-05, "loss": 0.7973, "mean_token_accuracy": 0.9049121141433716, "num_tokens": 820456788.0, "step": 21606 }, { "epoch": 2.748632489505152, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.030535697937012, "learning_rate": 5e-05, "loss": 0.9351, "mean_token_accuracy": 0.8938757181167603, "num_tokens": 820499767.0, "step": 21607 }, { "epoch": 2.7487596997837427, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 7.708091735839844, "learning_rate": 5e-05, "loss": 0.8931, "mean_token_accuracy": 0.8865749835968018, "num_tokens": 820543052.0, "step": 21608 }, { "epoch": 2.7488869100623328, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.495378494262695, "learning_rate": 5e-05, "loss": 0.9021, "mean_token_accuracy": 0.8919713497161865, "num_tokens": 820586162.0, "step": 21609 }, { "epoch": 2.7490141203409237, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 7.861443996429443, "learning_rate": 5e-05, "loss": 0.928, "mean_token_accuracy": 0.8713918924331665, "num_tokens": 820626703.0, "step": 21610 }, { "epoch": 2.749141330619514, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.138938903808594, "learning_rate": 5e-05, "loss": 0.9184, "mean_token_accuracy": 0.8699867725372314, "num_tokens": 820663936.0, "step": 21611 }, { "epoch": 2.749268540898105, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.700882434844971, "learning_rate": 5e-05, "loss": 0.8479, "mean_token_accuracy": 0.8842436075210571, "num_tokens": 820703761.0, "step": 21612 }, { "epoch": 2.749395751176695, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.171626091003418, "learning_rate": 5e-05, "loss": 0.8996, "mean_token_accuracy": 0.8726759552955627, "num_tokens": 820742505.0, "step": 21613 }, { "epoch": 2.749522961455286, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.429995536804199, "learning_rate": 5e-05, "loss": 0.7765, "mean_token_accuracy": 0.8908672332763672, "num_tokens": 820784185.0, "step": 21614 }, { "epoch": 2.749650171733876, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.263871192932129, "learning_rate": 5e-05, "loss": 0.9086, "mean_token_accuracy": 0.8650486469268799, "num_tokens": 820826163.0, "step": 21615 }, { "epoch": 2.7497773820124665, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.3554606437683105, "learning_rate": 5e-05, "loss": 0.7977, "mean_token_accuracy": 0.8757899403572083, "num_tokens": 820861598.0, "step": 21616 }, { "epoch": 2.749904592291057, "ewc_loss": 0.5, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.174400329589844, "learning_rate": 5e-05, "loss": 0.8608, "mean_token_accuracy": 0.8744707703590393, "num_tokens": 820901201.0, "step": 21617 }, { "epoch": 2.7500318025696475, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.434363842010498, "learning_rate": 5e-05, "loss": 0.7575, "mean_token_accuracy": 0.8888054490089417, "num_tokens": 820941773.0, "step": 21618 }, { "epoch": 2.750159012848238, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.021758079528809, "learning_rate": 5e-05, "loss": 0.8773, "mean_token_accuracy": 0.8599010705947876, "num_tokens": 820984610.0, "step": 21619 }, { "epoch": 2.7502862231268286, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.474386692047119, "learning_rate": 5e-05, "loss": 0.7377, "mean_token_accuracy": 0.8884880542755127, "num_tokens": 821020814.0, "step": 21620 }, { "epoch": 2.750413433405419, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.945690631866455, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8892062306404114, "num_tokens": 821053073.0, "step": 21621 }, { "epoch": 2.7505406436840096, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.439905166625977, "learning_rate": 5e-05, "loss": 0.7748, "mean_token_accuracy": 0.8765350580215454, "num_tokens": 821088872.0, "step": 21622 }, { "epoch": 2.7506678539626, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.984817981719971, "learning_rate": 5e-05, "loss": 0.8212, "mean_token_accuracy": 0.8706177473068237, "num_tokens": 821124314.0, "step": 21623 }, { "epoch": 2.7507950642411907, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.349874496459961, "learning_rate": 5e-05, "loss": 0.737, "mean_token_accuracy": 0.8803769946098328, "num_tokens": 821152025.0, "step": 21624 }, { "epoch": 2.750922274519781, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.039987564086914, "learning_rate": 5e-05, "loss": 0.8312, "mean_token_accuracy": 0.8684298396110535, "num_tokens": 821198138.0, "step": 21625 }, { "epoch": 2.7510494847983717, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.281473159790039, "learning_rate": 5e-05, "loss": 0.7323, "mean_token_accuracy": 0.8787643909454346, "num_tokens": 821241256.0, "step": 21626 }, { "epoch": 2.7511766950769623, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.093095779418945, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.8857519030570984, "num_tokens": 821274932.0, "step": 21627 }, { "epoch": 2.751303905355553, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.230507850646973, "learning_rate": 5e-05, "loss": 0.7215, "mean_token_accuracy": 0.8809550404548645, "num_tokens": 821310769.0, "step": 21628 }, { "epoch": 2.7514311156341433, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.107503890991211, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8796212077140808, "num_tokens": 821346474.0, "step": 21629 }, { "epoch": 2.751558325912734, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.205418109893799, "learning_rate": 5e-05, "loss": 0.722, "mean_token_accuracy": 0.8791571855545044, "num_tokens": 821385293.0, "step": 21630 }, { "epoch": 2.7516855361913244, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.15272331237793, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.890625, "num_tokens": 821421161.0, "step": 21631 }, { "epoch": 2.751812746469915, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.133554458618164, "learning_rate": 5e-05, "loss": 0.6999, "mean_token_accuracy": 0.8892923593521118, "num_tokens": 821459937.0, "step": 21632 }, { "epoch": 2.7519399567485054, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.224239349365234, "learning_rate": 5e-05, "loss": 0.7936, "mean_token_accuracy": 0.8824895024299622, "num_tokens": 821503892.0, "step": 21633 }, { "epoch": 2.7520671670270955, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.094024181365967, "learning_rate": 5e-05, "loss": 0.7023, "mean_token_accuracy": 0.8878961801528931, "num_tokens": 821537781.0, "step": 21634 }, { "epoch": 2.7521943773056865, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.268675804138184, "learning_rate": 5e-05, "loss": 0.8414, "mean_token_accuracy": 0.8673136234283447, "num_tokens": 821576114.0, "step": 21635 }, { "epoch": 2.7523215875842766, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.085911750793457, "learning_rate": 5e-05, "loss": 0.7133, "mean_token_accuracy": 0.8857899308204651, "num_tokens": 821613587.0, "step": 21636 }, { "epoch": 2.7524487978628676, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.227252006530762, "learning_rate": 5e-05, "loss": 0.7341, "mean_token_accuracy": 0.9017626047134399, "num_tokens": 821650956.0, "step": 21637 }, { "epoch": 2.7525760081414576, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.184239864349365, "learning_rate": 5e-05, "loss": 0.7523, "mean_token_accuracy": 0.870863139629364, "num_tokens": 821692943.0, "step": 21638 }, { "epoch": 2.7527032184200486, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.072113037109375, "learning_rate": 5e-05, "loss": 0.8384, "mean_token_accuracy": 0.8645456433296204, "num_tokens": 821735344.0, "step": 21639 }, { "epoch": 2.7528304286986387, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.371264457702637, "learning_rate": 5e-05, "loss": 0.727, "mean_token_accuracy": 0.883243203163147, "num_tokens": 821772910.0, "step": 21640 }, { "epoch": 2.7529576389772292, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.780864238739014, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8873519897460938, "num_tokens": 821810211.0, "step": 21641 }, { "epoch": 2.7530848492558198, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.670043468475342, "learning_rate": 5e-05, "loss": 0.7433, "mean_token_accuracy": 0.8792594075202942, "num_tokens": 821840782.0, "step": 21642 }, { "epoch": 2.7532120595344103, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.373844623565674, "learning_rate": 5e-05, "loss": 0.7246, "mean_token_accuracy": 0.8785498142242432, "num_tokens": 821879160.0, "step": 21643 }, { "epoch": 2.753339269813001, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.062347412109375, "learning_rate": 5e-05, "loss": 0.7556, "mean_token_accuracy": 0.888114333152771, "num_tokens": 821915344.0, "step": 21644 }, { "epoch": 2.7534664800915913, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.904666900634766, "learning_rate": 5e-05, "loss": 0.7402, "mean_token_accuracy": 0.8595314621925354, "num_tokens": 821955628.0, "step": 21645 }, { "epoch": 2.753593690370182, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.61697769165039, "learning_rate": 5e-05, "loss": 0.8312, "mean_token_accuracy": 0.8745030164718628, "num_tokens": 821996917.0, "step": 21646 }, { "epoch": 2.7537209006487724, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00024127960205078125, "grad_norm": 6.324480056762695, "learning_rate": 5e-05, "loss": 0.6478, "mean_token_accuracy": 0.8836397528648376, "num_tokens": 822034222.0, "step": 21647 }, { "epoch": 2.753848110927363, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.46544075012207, "learning_rate": 5e-05, "loss": 0.9146, "mean_token_accuracy": 0.875080943107605, "num_tokens": 822072350.0, "step": 21648 }, { "epoch": 2.7539753212059535, "ewc_loss": 0.314453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00023555755615234375, "grad_norm": 5.7495036125183105, "learning_rate": 5e-05, "loss": 0.6676, "mean_token_accuracy": 0.8763737082481384, "num_tokens": 822111495.0, "step": 21649 }, { "epoch": 2.754102531484544, "ewc_loss": 0.703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.583662033081055, "learning_rate": 5e-05, "loss": 1.0497, "mean_token_accuracy": 0.875496506690979, "num_tokens": 822148132.0, "step": 21650 }, { "epoch": 2.7542297417631345, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.329519748687744, "learning_rate": 5e-05, "loss": 0.7925, "mean_token_accuracy": 0.8793806433677673, "num_tokens": 822187440.0, "step": 21651 }, { "epoch": 2.754356952041725, "ewc_loss": 0.859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000782012939453125, "grad_norm": 10.969718933105469, "learning_rate": 5e-05, "loss": 1.1965, "mean_token_accuracy": 0.8792903423309326, "num_tokens": 822225936.0, "step": 21652 }, { "epoch": 2.7544841623203156, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 8.595641136169434, "learning_rate": 5e-05, "loss": 1.1404, "mean_token_accuracy": 0.8773559331893921, "num_tokens": 822266187.0, "step": 21653 }, { "epoch": 2.754611372598906, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.122515678405762, "learning_rate": 5e-05, "loss": 0.9274, "mean_token_accuracy": 0.887301504611969, "num_tokens": 822302963.0, "step": 21654 }, { "epoch": 2.7547385828774966, "ewc_loss": 0.8828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000804901123046875, "grad_norm": 9.898877143859863, "learning_rate": 5e-05, "loss": 1.1934, "mean_token_accuracy": 0.8879851698875427, "num_tokens": 822336969.0, "step": 21655 }, { "epoch": 2.754865793156087, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.036910057067871, "learning_rate": 5e-05, "loss": 0.7812, "mean_token_accuracy": 0.8795102834701538, "num_tokens": 822376847.0, "step": 21656 }, { "epoch": 2.7549930034346777, "ewc_loss": 0.9921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 12.007173538208008, "learning_rate": 5e-05, "loss": 1.3513, "mean_token_accuracy": 0.8757851123809814, "num_tokens": 822417858.0, "step": 21657 }, { "epoch": 2.755120213713268, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 8.661725044250488, "learning_rate": 5e-05, "loss": 1.1511, "mean_token_accuracy": 0.8709136247634888, "num_tokens": 822459946.0, "step": 21658 }, { "epoch": 2.7552474239918583, "ewc_loss": 0.8671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000789642333984375, "grad_norm": 12.106170654296875, "learning_rate": 5e-05, "loss": 1.2345, "mean_token_accuracy": 0.8729006052017212, "num_tokens": 822498093.0, "step": 21659 }, { "epoch": 2.7553746342704493, "ewc_loss": 0.9921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00091552734375, "grad_norm": 9.908984184265137, "learning_rate": 5e-05, "loss": 1.3109, "mean_token_accuracy": 0.8872129321098328, "num_tokens": 822535918.0, "step": 21660 }, { "epoch": 2.7555018445490393, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 11.468276023864746, "learning_rate": 5e-05, "loss": 1.1387, "mean_token_accuracy": 0.8771984577178955, "num_tokens": 822574589.0, "step": 21661 }, { "epoch": 2.7556290548276303, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 9.051931381225586, "learning_rate": 5e-05, "loss": 1.0561, "mean_token_accuracy": 0.885985791683197, "num_tokens": 822612437.0, "step": 21662 }, { "epoch": 2.7557562651062204, "ewc_loss": 0.86328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 11.48635482788086, "learning_rate": 5e-05, "loss": 1.2068, "mean_token_accuracy": 0.8833574056625366, "num_tokens": 822654834.0, "step": 21663 }, { "epoch": 2.755883475384811, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.621562957763672, "learning_rate": 5e-05, "loss": 0.9601, "mean_token_accuracy": 0.8893338441848755, "num_tokens": 822689555.0, "step": 21664 }, { "epoch": 2.7560106856634015, "ewc_loss": 0.875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000797271728515625, "grad_norm": 11.854941368103027, "learning_rate": 5e-05, "loss": 1.2573, "mean_token_accuracy": 0.8672888875007629, "num_tokens": 822731609.0, "step": 21665 }, { "epoch": 2.756137895941992, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 9.347945213317871, "learning_rate": 5e-05, "loss": 1.0813, "mean_token_accuracy": 0.8720296621322632, "num_tokens": 822769881.0, "step": 21666 }, { "epoch": 2.7562651062205825, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 11.697850227355957, "learning_rate": 5e-05, "loss": 1.2193, "mean_token_accuracy": 0.8689144849777222, "num_tokens": 822810369.0, "step": 21667 }, { "epoch": 2.756392316499173, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 9.920632362365723, "learning_rate": 5e-05, "loss": 1.1425, "mean_token_accuracy": 0.8763319253921509, "num_tokens": 822852476.0, "step": 21668 }, { "epoch": 2.7565195267777636, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 10.984986305236816, "learning_rate": 5e-05, "loss": 1.1574, "mean_token_accuracy": 0.8659021854400635, "num_tokens": 822890745.0, "step": 21669 }, { "epoch": 2.756646737056354, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.641357421875, "learning_rate": 5e-05, "loss": 1.0374, "mean_token_accuracy": 0.8784126043319702, "num_tokens": 822931243.0, "step": 21670 }, { "epoch": 2.7567739473349446, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.367939949035645, "learning_rate": 5e-05, "loss": 1.0467, "mean_token_accuracy": 0.8773583173751831, "num_tokens": 822967560.0, "step": 21671 }, { "epoch": 2.756901157613535, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.129956245422363, "learning_rate": 5e-05, "loss": 0.9583, "mean_token_accuracy": 0.8813802003860474, "num_tokens": 823007129.0, "step": 21672 }, { "epoch": 2.7570283678921257, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.03026294708252, "learning_rate": 5e-05, "loss": 1.0255, "mean_token_accuracy": 0.8769749402999878, "num_tokens": 823043722.0, "step": 21673 }, { "epoch": 2.757155578170716, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.564544677734375, "learning_rate": 5e-05, "loss": 0.9629, "mean_token_accuracy": 0.864354133605957, "num_tokens": 823079937.0, "step": 21674 }, { "epoch": 2.7572827884493067, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.096689224243164, "learning_rate": 5e-05, "loss": 1.0278, "mean_token_accuracy": 0.877841055393219, "num_tokens": 823120455.0, "step": 21675 }, { "epoch": 2.7574099987278973, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.234468460083008, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8739715814590454, "num_tokens": 823151415.0, "step": 21676 }, { "epoch": 2.757537209006488, "ewc_loss": 0.71875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.440180778503418, "learning_rate": 5e-05, "loss": 1.0633, "mean_token_accuracy": 0.8794794082641602, "num_tokens": 823188281.0, "step": 21677 }, { "epoch": 2.7576644192850783, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.433845520019531, "learning_rate": 5e-05, "loss": 0.8842, "mean_token_accuracy": 0.884821355342865, "num_tokens": 823226040.0, "step": 21678 }, { "epoch": 2.757791629563669, "ewc_loss": 0.734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.547775268554688, "learning_rate": 5e-05, "loss": 1.0781, "mean_token_accuracy": 0.8774515390396118, "num_tokens": 823261246.0, "step": 21679 }, { "epoch": 2.7579188398422594, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.979645729064941, "learning_rate": 5e-05, "loss": 0.9599, "mean_token_accuracy": 0.881706953048706, "num_tokens": 823300034.0, "step": 21680 }, { "epoch": 2.75804605012085, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.93112850189209, "learning_rate": 5e-05, "loss": 0.9606, "mean_token_accuracy": 0.8944298028945923, "num_tokens": 823332152.0, "step": 21681 }, { "epoch": 2.7581732603994404, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.994422912597656, "learning_rate": 5e-05, "loss": 0.9486, "mean_token_accuracy": 0.8784347176551819, "num_tokens": 823373319.0, "step": 21682 }, { "epoch": 2.758300470678031, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.18800163269043, "learning_rate": 5e-05, "loss": 0.9087, "mean_token_accuracy": 0.8940324783325195, "num_tokens": 823409875.0, "step": 21683 }, { "epoch": 2.758427680956621, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.505718231201172, "learning_rate": 5e-05, "loss": 0.8279, "mean_token_accuracy": 0.8954077363014221, "num_tokens": 823450941.0, "step": 21684 }, { "epoch": 2.758554891235212, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.945812225341797, "learning_rate": 5e-05, "loss": 0.8934, "mean_token_accuracy": 0.8885033130645752, "num_tokens": 823493483.0, "step": 21685 }, { "epoch": 2.758682101513802, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.861769676208496, "learning_rate": 5e-05, "loss": 0.7864, "mean_token_accuracy": 0.8828945159912109, "num_tokens": 823531682.0, "step": 21686 }, { "epoch": 2.758809311792393, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.186766624450684, "learning_rate": 5e-05, "loss": 0.916, "mean_token_accuracy": 0.8892778158187866, "num_tokens": 823564089.0, "step": 21687 }, { "epoch": 2.758936522070983, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.482640266418457, "learning_rate": 5e-05, "loss": 0.7093, "mean_token_accuracy": 0.8988633155822754, "num_tokens": 823598700.0, "step": 21688 }, { "epoch": 2.7590637323495737, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.594993591308594, "learning_rate": 5e-05, "loss": 0.993, "mean_token_accuracy": 0.8763115406036377, "num_tokens": 823640248.0, "step": 21689 }, { "epoch": 2.7591909426281642, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.654007434844971, "learning_rate": 5e-05, "loss": 0.7915, "mean_token_accuracy": 0.8767199516296387, "num_tokens": 823676303.0, "step": 21690 }, { "epoch": 2.7593181529067548, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.67006778717041, "learning_rate": 5e-05, "loss": 1.0231, "mean_token_accuracy": 0.873111367225647, "num_tokens": 823717736.0, "step": 21691 }, { "epoch": 2.7594453631853453, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.220078468322754, "learning_rate": 5e-05, "loss": 0.8708, "mean_token_accuracy": 0.8786500692367554, "num_tokens": 823755602.0, "step": 21692 }, { "epoch": 2.759572573463936, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.046855926513672, "learning_rate": 5e-05, "loss": 0.9838, "mean_token_accuracy": 0.8673301935195923, "num_tokens": 823792257.0, "step": 21693 }, { "epoch": 2.7596997837425263, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.423809051513672, "learning_rate": 5e-05, "loss": 0.8655, "mean_token_accuracy": 0.8801280856132507, "num_tokens": 823835655.0, "step": 21694 }, { "epoch": 2.759826994021117, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.241750717163086, "learning_rate": 5e-05, "loss": 0.8653, "mean_token_accuracy": 0.8759750127792358, "num_tokens": 823874840.0, "step": 21695 }, { "epoch": 2.7599542042997074, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.200775146484375, "learning_rate": 5e-05, "loss": 0.8521, "mean_token_accuracy": 0.8706356883049011, "num_tokens": 823910844.0, "step": 21696 }, { "epoch": 2.760081414578298, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.943731784820557, "learning_rate": 5e-05, "loss": 0.7655, "mean_token_accuracy": 0.8926689624786377, "num_tokens": 823948472.0, "step": 21697 }, { "epoch": 2.7602086248568884, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.857921600341797, "learning_rate": 5e-05, "loss": 0.7496, "mean_token_accuracy": 0.8893519639968872, "num_tokens": 823997532.0, "step": 21698 }, { "epoch": 2.760335835135479, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.953808307647705, "learning_rate": 5e-05, "loss": 0.81, "mean_token_accuracy": 0.8723678588867188, "num_tokens": 824037426.0, "step": 21699 }, { "epoch": 2.7604630454140695, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.646109580993652, "learning_rate": 5e-05, "loss": 0.7669, "mean_token_accuracy": 0.875296950340271, "num_tokens": 824076731.0, "step": 21700 }, { "epoch": 2.76059025569266, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.935814380645752, "learning_rate": 5e-05, "loss": 0.814, "mean_token_accuracy": 0.8660386204719543, "num_tokens": 824119445.0, "step": 21701 }, { "epoch": 2.7607174659712506, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.590925216674805, "learning_rate": 5e-05, "loss": 0.7612, "mean_token_accuracy": 0.8707355260848999, "num_tokens": 824157372.0, "step": 21702 }, { "epoch": 2.760844676249841, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.859277725219727, "learning_rate": 5e-05, "loss": 0.7515, "mean_token_accuracy": 0.8790453672409058, "num_tokens": 824204312.0, "step": 21703 }, { "epoch": 2.7609718865284316, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.533535957336426, "learning_rate": 5e-05, "loss": 0.7073, "mean_token_accuracy": 0.8854069709777832, "num_tokens": 824245951.0, "step": 21704 }, { "epoch": 2.761099096807022, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.961719512939453, "learning_rate": 5e-05, "loss": 0.763, "mean_token_accuracy": 0.878952145576477, "num_tokens": 824288742.0, "step": 21705 }, { "epoch": 2.7612263070856127, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.310652732849121, "learning_rate": 5e-05, "loss": 0.7168, "mean_token_accuracy": 0.8745487928390503, "num_tokens": 824326435.0, "step": 21706 }, { "epoch": 2.7613535173642028, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.180033683776855, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.8790614008903503, "num_tokens": 824359442.0, "step": 21707 }, { "epoch": 2.7614807276427937, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.094850540161133, "learning_rate": 5e-05, "loss": 0.7318, "mean_token_accuracy": 0.8702385425567627, "num_tokens": 824392387.0, "step": 21708 }, { "epoch": 2.761607937921384, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.417265892028809, "learning_rate": 5e-05, "loss": 0.8168, "mean_token_accuracy": 0.8750303983688354, "num_tokens": 824427860.0, "step": 21709 }, { "epoch": 2.761735148199975, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.881345748901367, "learning_rate": 5e-05, "loss": 0.6916, "mean_token_accuracy": 0.879126250743866, "num_tokens": 824466011.0, "step": 21710 }, { "epoch": 2.761862358478565, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.734350204467773, "learning_rate": 5e-05, "loss": 0.7881, "mean_token_accuracy": 0.8922271728515625, "num_tokens": 824507235.0, "step": 21711 }, { "epoch": 2.761989568757156, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.696047306060791, "learning_rate": 5e-05, "loss": 0.6755, "mean_token_accuracy": 0.8853490352630615, "num_tokens": 824538807.0, "step": 21712 }, { "epoch": 2.762116779035746, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 9.034151077270508, "learning_rate": 5e-05, "loss": 0.8848, "mean_token_accuracy": 0.8796772956848145, "num_tokens": 824575047.0, "step": 21713 }, { "epoch": 2.7622439893143365, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.832342624664307, "learning_rate": 5e-05, "loss": 0.6493, "mean_token_accuracy": 0.8972525596618652, "num_tokens": 824604113.0, "step": 21714 }, { "epoch": 2.762371199592927, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.813458442687988, "learning_rate": 5e-05, "loss": 0.9015, "mean_token_accuracy": 0.8691859245300293, "num_tokens": 824640912.0, "step": 21715 }, { "epoch": 2.7624984098715175, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.44344425201416, "learning_rate": 5e-05, "loss": 0.7295, "mean_token_accuracy": 0.88511061668396, "num_tokens": 824682739.0, "step": 21716 }, { "epoch": 2.762625620150108, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.964482307434082, "learning_rate": 5e-05, "loss": 0.8427, "mean_token_accuracy": 0.8612213134765625, "num_tokens": 824719973.0, "step": 21717 }, { "epoch": 2.7627528304286986, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.17458724975586, "learning_rate": 5e-05, "loss": 0.7889, "mean_token_accuracy": 0.8843240141868591, "num_tokens": 824761214.0, "step": 21718 }, { "epoch": 2.762880040707289, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.032063007354736, "learning_rate": 5e-05, "loss": 0.7321, "mean_token_accuracy": 0.8703615665435791, "num_tokens": 824801114.0, "step": 21719 }, { "epoch": 2.7630072509858796, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.921992301940918, "learning_rate": 5e-05, "loss": 0.8686, "mean_token_accuracy": 0.8743347525596619, "num_tokens": 824832734.0, "step": 21720 }, { "epoch": 2.76313446126447, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.385081768035889, "learning_rate": 5e-05, "loss": 0.6837, "mean_token_accuracy": 0.8736345767974854, "num_tokens": 824867089.0, "step": 21721 }, { "epoch": 2.7632616715430607, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.68462085723877, "learning_rate": 5e-05, "loss": 0.9158, "mean_token_accuracy": 0.8913137912750244, "num_tokens": 824909173.0, "step": 21722 }, { "epoch": 2.763388881821651, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.429703712463379, "learning_rate": 5e-05, "loss": 0.6772, "mean_token_accuracy": 0.8884715437889099, "num_tokens": 824943815.0, "step": 21723 }, { "epoch": 2.7635160921002417, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 9.921788215637207, "learning_rate": 5e-05, "loss": 0.9608, "mean_token_accuracy": 0.8949202299118042, "num_tokens": 824978708.0, "step": 21724 }, { "epoch": 2.7636433023788323, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.64832067489624, "learning_rate": 5e-05, "loss": 0.913, "mean_token_accuracy": 0.8645291924476624, "num_tokens": 825015226.0, "step": 21725 }, { "epoch": 2.763770512657423, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.812515258789062, "learning_rate": 5e-05, "loss": 0.9585, "mean_token_accuracy": 0.8680518865585327, "num_tokens": 825054417.0, "step": 21726 }, { "epoch": 2.7638977229360133, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.716499328613281, "learning_rate": 5e-05, "loss": 0.9068, "mean_token_accuracy": 0.8923743963241577, "num_tokens": 825093757.0, "step": 21727 }, { "epoch": 2.764024933214604, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.373929500579834, "learning_rate": 5e-05, "loss": 0.8369, "mean_token_accuracy": 0.8635708689689636, "num_tokens": 825129955.0, "step": 21728 }, { "epoch": 2.7641521434931944, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.04981803894043, "learning_rate": 5e-05, "loss": 0.881, "mean_token_accuracy": 0.8889322280883789, "num_tokens": 825168426.0, "step": 21729 }, { "epoch": 2.764279353771785, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.3143439292907715, "learning_rate": 5e-05, "loss": 0.813, "mean_token_accuracy": 0.8668897151947021, "num_tokens": 825213475.0, "step": 21730 }, { "epoch": 2.7644065640503754, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.242903709411621, "learning_rate": 5e-05, "loss": 0.7985, "mean_token_accuracy": 0.888343334197998, "num_tokens": 825249146.0, "step": 21731 }, { "epoch": 2.7645337743289655, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.15789794921875, "learning_rate": 5e-05, "loss": 0.7818, "mean_token_accuracy": 0.8980926871299744, "num_tokens": 825286504.0, "step": 21732 }, { "epoch": 2.7646609846075565, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.056823253631592, "learning_rate": 5e-05, "loss": 0.7267, "mean_token_accuracy": 0.8774814605712891, "num_tokens": 825327820.0, "step": 21733 }, { "epoch": 2.7647881948861466, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 9.092637062072754, "learning_rate": 5e-05, "loss": 0.8674, "mean_token_accuracy": 0.8913135528564453, "num_tokens": 825365967.0, "step": 21734 }, { "epoch": 2.7649154051647375, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.4571428298950195, "learning_rate": 5e-05, "loss": 0.6998, "mean_token_accuracy": 0.8789777755737305, "num_tokens": 825406959.0, "step": 21735 }, { "epoch": 2.7650426154433276, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.770209312438965, "learning_rate": 5e-05, "loss": 1.0085, "mean_token_accuracy": 0.8702012896537781, "num_tokens": 825447355.0, "step": 21736 }, { "epoch": 2.7651698257219186, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.811456680297852, "learning_rate": 5e-05, "loss": 0.7811, "mean_token_accuracy": 0.8692716360092163, "num_tokens": 825485618.0, "step": 21737 }, { "epoch": 2.7652970360005087, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.692445755004883, "learning_rate": 5e-05, "loss": 0.9863, "mean_token_accuracy": 0.8852893710136414, "num_tokens": 825522287.0, "step": 21738 }, { "epoch": 2.765424246279099, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.721446514129639, "learning_rate": 5e-05, "loss": 0.8639, "mean_token_accuracy": 0.879694402217865, "num_tokens": 825566214.0, "step": 21739 }, { "epoch": 2.7655514565576897, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.022647857666016, "learning_rate": 5e-05, "loss": 0.9408, "mean_token_accuracy": 0.8806619644165039, "num_tokens": 825608169.0, "step": 21740 }, { "epoch": 2.7656786668362803, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.000123023986816, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8849582076072693, "num_tokens": 825643076.0, "step": 21741 }, { "epoch": 2.765805877114871, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.646811485290527, "learning_rate": 5e-05, "loss": 0.9131, "mean_token_accuracy": 0.8784344792366028, "num_tokens": 825679868.0, "step": 21742 }, { "epoch": 2.7659330873934613, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.544824600219727, "learning_rate": 5e-05, "loss": 0.8056, "mean_token_accuracy": 0.879195511341095, "num_tokens": 825719391.0, "step": 21743 }, { "epoch": 2.766060297672052, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.044195175170898, "learning_rate": 5e-05, "loss": 0.9442, "mean_token_accuracy": 0.8783114552497864, "num_tokens": 825761972.0, "step": 21744 }, { "epoch": 2.7661875079506424, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.076094150543213, "learning_rate": 5e-05, "loss": 0.7809, "mean_token_accuracy": 0.876897931098938, "num_tokens": 825802672.0, "step": 21745 }, { "epoch": 2.766314718229233, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.443952560424805, "learning_rate": 5e-05, "loss": 0.9965, "mean_token_accuracy": 0.8749482035636902, "num_tokens": 825840219.0, "step": 21746 }, { "epoch": 2.7664419285078234, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.438295364379883, "learning_rate": 5e-05, "loss": 0.83, "mean_token_accuracy": 0.8794770836830139, "num_tokens": 825876793.0, "step": 21747 }, { "epoch": 2.766569138786414, "ewc_loss": 0.625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.955814361572266, "learning_rate": 5e-05, "loss": 0.9876, "mean_token_accuracy": 0.8719468116760254, "num_tokens": 825913231.0, "step": 21748 }, { "epoch": 2.7666963490650045, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.074434280395508, "learning_rate": 5e-05, "loss": 0.9278, "mean_token_accuracy": 0.8673678636550903, "num_tokens": 825952600.0, "step": 21749 }, { "epoch": 2.766823559343595, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.053693771362305, "learning_rate": 5e-05, "loss": 0.8819, "mean_token_accuracy": 0.8780710697174072, "num_tokens": 825997239.0, "step": 21750 }, { "epoch": 2.7669507696221856, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.208948135375977, "learning_rate": 5e-05, "loss": 0.8544, "mean_token_accuracy": 0.8858399391174316, "num_tokens": 826037492.0, "step": 21751 }, { "epoch": 2.767077979900776, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.740584850311279, "learning_rate": 5e-05, "loss": 0.8059, "mean_token_accuracy": 0.8817314505577087, "num_tokens": 826077607.0, "step": 21752 }, { "epoch": 2.7672051901793666, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.922035217285156, "learning_rate": 5e-05, "loss": 0.8242, "mean_token_accuracy": 0.8755915760993958, "num_tokens": 826115503.0, "step": 21753 }, { "epoch": 2.767332400457957, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.841054439544678, "learning_rate": 5e-05, "loss": 0.8109, "mean_token_accuracy": 0.8755220770835876, "num_tokens": 826154526.0, "step": 21754 }, { "epoch": 2.7674596107365477, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.687445640563965, "learning_rate": 5e-05, "loss": 0.8107, "mean_token_accuracy": 0.870322585105896, "num_tokens": 826194945.0, "step": 21755 }, { "epoch": 2.767586821015138, "ewc_loss": 0.4375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.846789360046387, "learning_rate": 5e-05, "loss": 0.7641, "mean_token_accuracy": 0.8841674327850342, "num_tokens": 826231000.0, "step": 21756 }, { "epoch": 2.7677140312937283, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.672426223754883, "learning_rate": 5e-05, "loss": 0.7342, "mean_token_accuracy": 0.8911657333374023, "num_tokens": 826266453.0, "step": 21757 }, { "epoch": 2.7678412415723193, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.69995641708374, "learning_rate": 5e-05, "loss": 0.7401, "mean_token_accuracy": 0.8834769129753113, "num_tokens": 826310442.0, "step": 21758 }, { "epoch": 2.7679684518509093, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.787899971008301, "learning_rate": 5e-05, "loss": 0.7745, "mean_token_accuracy": 0.8807327747344971, "num_tokens": 826344595.0, "step": 21759 }, { "epoch": 2.7680956621295003, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.531803607940674, "learning_rate": 5e-05, "loss": 0.7476, "mean_token_accuracy": 0.8796751499176025, "num_tokens": 826384556.0, "step": 21760 }, { "epoch": 2.7682228724080904, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.8851847648620605, "learning_rate": 5e-05, "loss": 0.81, "mean_token_accuracy": 0.8600127696990967, "num_tokens": 826418344.0, "step": 21761 }, { "epoch": 2.768350082686681, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.402115821838379, "learning_rate": 5e-05, "loss": 0.7229, "mean_token_accuracy": 0.8810717463493347, "num_tokens": 826460027.0, "step": 21762 }, { "epoch": 2.7684772929652715, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.978199481964111, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8780350685119629, "num_tokens": 826497422.0, "step": 21763 }, { "epoch": 2.768604503243862, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.2979631423950195, "learning_rate": 5e-05, "loss": 0.7584, "mean_token_accuracy": 0.8717182874679565, "num_tokens": 826536624.0, "step": 21764 }, { "epoch": 2.7687317135224525, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.080026626586914, "learning_rate": 5e-05, "loss": 0.7764, "mean_token_accuracy": 0.8747520446777344, "num_tokens": 826572303.0, "step": 21765 }, { "epoch": 2.768858923801043, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.146980285644531, "learning_rate": 5e-05, "loss": 0.7536, "mean_token_accuracy": 0.8619222640991211, "num_tokens": 826613326.0, "step": 21766 }, { "epoch": 2.7689861340796336, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.291664123535156, "learning_rate": 5e-05, "loss": 0.7964, "mean_token_accuracy": 0.8765876293182373, "num_tokens": 826649059.0, "step": 21767 }, { "epoch": 2.769113344358224, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.890347003936768, "learning_rate": 5e-05, "loss": 0.6793, "mean_token_accuracy": 0.8831522464752197, "num_tokens": 826682983.0, "step": 21768 }, { "epoch": 2.7692405546368146, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.709091186523438, "learning_rate": 5e-05, "loss": 0.8355, "mean_token_accuracy": 0.8787834048271179, "num_tokens": 826715179.0, "step": 21769 }, { "epoch": 2.769367764915405, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.48131799697876, "learning_rate": 5e-05, "loss": 0.7012, "mean_token_accuracy": 0.8692681789398193, "num_tokens": 826748442.0, "step": 21770 }, { "epoch": 2.7694949751939957, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.336483001708984, "learning_rate": 5e-05, "loss": 0.8817, "mean_token_accuracy": 0.8877444863319397, "num_tokens": 826791014.0, "step": 21771 }, { "epoch": 2.769622185472586, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.187984943389893, "learning_rate": 5e-05, "loss": 0.6966, "mean_token_accuracy": 0.875109076499939, "num_tokens": 826827987.0, "step": 21772 }, { "epoch": 2.7697493957511767, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.889863967895508, "learning_rate": 5e-05, "loss": 0.9645, "mean_token_accuracy": 0.8839209675788879, "num_tokens": 826869315.0, "step": 21773 }, { "epoch": 2.7698766060297673, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 6.727694034576416, "learning_rate": 5e-05, "loss": 0.7701, "mean_token_accuracy": 0.8774948716163635, "num_tokens": 826912679.0, "step": 21774 }, { "epoch": 2.770003816308358, "ewc_loss": 0.65625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.545132637023926, "learning_rate": 5e-05, "loss": 0.9868, "mean_token_accuracy": 0.88158118724823, "num_tokens": 826949276.0, "step": 21775 }, { "epoch": 2.7701310265869483, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.055413246154785, "learning_rate": 5e-05, "loss": 0.9275, "mean_token_accuracy": 0.8854857683181763, "num_tokens": 826989441.0, "step": 21776 }, { "epoch": 2.770258236865539, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.8812127113342285, "learning_rate": 5e-05, "loss": 0.8711, "mean_token_accuracy": 0.8726060390472412, "num_tokens": 827025030.0, "step": 21777 }, { "epoch": 2.7703854471441294, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.276427268981934, "learning_rate": 5e-05, "loss": 0.986, "mean_token_accuracy": 0.8773848414421082, "num_tokens": 827060252.0, "step": 21778 }, { "epoch": 2.77051265742272, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.337461948394775, "learning_rate": 5e-05, "loss": 0.7032, "mean_token_accuracy": 0.8806202411651611, "num_tokens": 827100063.0, "step": 21779 }, { "epoch": 2.77063986770131, "ewc_loss": 0.7265625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00064849853515625, "grad_norm": 10.571401596069336, "learning_rate": 5e-05, "loss": 1.0321, "mean_token_accuracy": 0.8911102414131165, "num_tokens": 827131772.0, "step": 21780 }, { "epoch": 2.770767077979901, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 6.687252044677734, "learning_rate": 5e-05, "loss": 0.8422, "mean_token_accuracy": 0.8692995309829712, "num_tokens": 827167659.0, "step": 21781 }, { "epoch": 2.770894288258491, "ewc_loss": 0.828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 10.979060173034668, "learning_rate": 5e-05, "loss": 1.1907, "mean_token_accuracy": 0.8724923729896545, "num_tokens": 827211215.0, "step": 21782 }, { "epoch": 2.771021498537082, "ewc_loss": 0.7734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00069427490234375, "grad_norm": 8.784404754638672, "learning_rate": 5e-05, "loss": 1.1402, "mean_token_accuracy": 0.8714662790298462, "num_tokens": 827248075.0, "step": 21783 }, { "epoch": 2.771148708815672, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.415735244750977, "learning_rate": 5e-05, "loss": 0.959, "mean_token_accuracy": 0.8841257095336914, "num_tokens": 827283270.0, "step": 21784 }, { "epoch": 2.771275919094263, "ewc_loss": 0.75, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 9.449071884155273, "learning_rate": 5e-05, "loss": 1.0737, "mean_token_accuracy": 0.88470059633255, "num_tokens": 827317520.0, "step": 21785 }, { "epoch": 2.771403129372853, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.982855319976807, "learning_rate": 5e-05, "loss": 0.8906, "mean_token_accuracy": 0.8714364171028137, "num_tokens": 827355899.0, "step": 21786 }, { "epoch": 2.7715303396514437, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.57499885559082, "learning_rate": 5e-05, "loss": 1.0692, "mean_token_accuracy": 0.8620715737342834, "num_tokens": 827390038.0, "step": 21787 }, { "epoch": 2.771657549930034, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.00423526763916, "learning_rate": 5e-05, "loss": 0.9373, "mean_token_accuracy": 0.8764756321907043, "num_tokens": 827428563.0, "step": 21788 }, { "epoch": 2.7717847602086247, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.59325885772705, "learning_rate": 5e-05, "loss": 0.8756, "mean_token_accuracy": 0.8862854242324829, "num_tokens": 827469658.0, "step": 21789 }, { "epoch": 2.7719119704872153, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.598586082458496, "learning_rate": 5e-05, "loss": 0.9351, "mean_token_accuracy": 0.8817377090454102, "num_tokens": 827504277.0, "step": 21790 }, { "epoch": 2.772039180765806, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.388271331787109, "learning_rate": 5e-05, "loss": 0.7819, "mean_token_accuracy": 0.874449610710144, "num_tokens": 827544840.0, "step": 21791 }, { "epoch": 2.7721663910443963, "ewc_loss": 0.625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.283548355102539, "learning_rate": 5e-05, "loss": 1.012, "mean_token_accuracy": 0.865768313407898, "num_tokens": 827590607.0, "step": 21792 }, { "epoch": 2.772293601322987, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.015212059020996, "learning_rate": 5e-05, "loss": 0.7105, "mean_token_accuracy": 0.8948675394058228, "num_tokens": 827622039.0, "step": 21793 }, { "epoch": 2.7724208116015774, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.603697776794434, "learning_rate": 5e-05, "loss": 0.967, "mean_token_accuracy": 0.8858383297920227, "num_tokens": 827659234.0, "step": 21794 }, { "epoch": 2.772548021880168, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.351078033447266, "learning_rate": 5e-05, "loss": 0.8411, "mean_token_accuracy": 0.8819780349731445, "num_tokens": 827702044.0, "step": 21795 }, { "epoch": 2.7726752321587584, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.408474922180176, "learning_rate": 5e-05, "loss": 0.953, "mean_token_accuracy": 0.8837413787841797, "num_tokens": 827738189.0, "step": 21796 }, { "epoch": 2.772802442437349, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.694375038146973, "learning_rate": 5e-05, "loss": 0.9368, "mean_token_accuracy": 0.8695212602615356, "num_tokens": 827780483.0, "step": 21797 }, { "epoch": 2.7729296527159395, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.013205528259277, "learning_rate": 5e-05, "loss": 0.9247, "mean_token_accuracy": 0.8851661682128906, "num_tokens": 827817156.0, "step": 21798 }, { "epoch": 2.77305686299453, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.648991584777832, "learning_rate": 5e-05, "loss": 0.8641, "mean_token_accuracy": 0.8817729949951172, "num_tokens": 827851591.0, "step": 21799 }, { "epoch": 2.7731840732731206, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.831984519958496, "learning_rate": 5e-05, "loss": 0.9166, "mean_token_accuracy": 0.8851245641708374, "num_tokens": 827891289.0, "step": 21800 }, { "epoch": 2.773311283551711, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.316793918609619, "learning_rate": 5e-05, "loss": 0.7585, "mean_token_accuracy": 0.8988990187644958, "num_tokens": 827926466.0, "step": 21801 }, { "epoch": 2.7734384938303016, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.060808181762695, "learning_rate": 5e-05, "loss": 1.0085, "mean_token_accuracy": 0.8543123006820679, "num_tokens": 827964885.0, "step": 21802 }, { "epoch": 2.773565704108892, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.0530171394348145, "learning_rate": 5e-05, "loss": 0.8079, "mean_token_accuracy": 0.8702180981636047, "num_tokens": 828006063.0, "step": 21803 }, { "epoch": 2.7736929143874827, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.36230182647705, "learning_rate": 5e-05, "loss": 0.9615, "mean_token_accuracy": 0.8867571353912354, "num_tokens": 828037533.0, "step": 21804 }, { "epoch": 2.7738201246660728, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.26420259475708, "learning_rate": 5e-05, "loss": 0.8045, "mean_token_accuracy": 0.8855233788490295, "num_tokens": 828072298.0, "step": 21805 }, { "epoch": 2.7739473349446637, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.206207275390625, "learning_rate": 5e-05, "loss": 0.9576, "mean_token_accuracy": 0.887781023979187, "num_tokens": 828107900.0, "step": 21806 }, { "epoch": 2.774074545223254, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.73603630065918, "learning_rate": 5e-05, "loss": 0.8802, "mean_token_accuracy": 0.8826791048049927, "num_tokens": 828143323.0, "step": 21807 }, { "epoch": 2.774201755501845, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.669143676757812, "learning_rate": 5e-05, "loss": 0.8985, "mean_token_accuracy": 0.8936055898666382, "num_tokens": 828183439.0, "step": 21808 }, { "epoch": 2.774328965780435, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.995513439178467, "learning_rate": 5e-05, "loss": 0.8935, "mean_token_accuracy": 0.884170413017273, "num_tokens": 828218360.0, "step": 21809 }, { "epoch": 2.774456176059026, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.11953067779541, "learning_rate": 5e-05, "loss": 0.9029, "mean_token_accuracy": 0.8729593753814697, "num_tokens": 828254042.0, "step": 21810 }, { "epoch": 2.774583386337616, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.000956535339355, "learning_rate": 5e-05, "loss": 0.873, "mean_token_accuracy": 0.8778150081634521, "num_tokens": 828290992.0, "step": 21811 }, { "epoch": 2.7747105966162064, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.851536273956299, "learning_rate": 5e-05, "loss": 0.8131, "mean_token_accuracy": 0.8834809064865112, "num_tokens": 828323351.0, "step": 21812 }, { "epoch": 2.774837806894797, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.843986988067627, "learning_rate": 5e-05, "loss": 0.7964, "mean_token_accuracy": 0.8856977224349976, "num_tokens": 828354555.0, "step": 21813 }, { "epoch": 2.7749650171733875, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.893998622894287, "learning_rate": 5e-05, "loss": 0.7933, "mean_token_accuracy": 0.8846919536590576, "num_tokens": 828395641.0, "step": 21814 }, { "epoch": 2.775092227451978, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.6004319190979, "learning_rate": 5e-05, "loss": 0.7685, "mean_token_accuracy": 0.8812072277069092, "num_tokens": 828437871.0, "step": 21815 }, { "epoch": 2.7752194377305686, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.019983291625977, "learning_rate": 5e-05, "loss": 0.8388, "mean_token_accuracy": 0.8672282695770264, "num_tokens": 828480545.0, "step": 21816 }, { "epoch": 2.775346648009159, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.433572292327881, "learning_rate": 5e-05, "loss": 0.7363, "mean_token_accuracy": 0.8842618465423584, "num_tokens": 828513519.0, "step": 21817 }, { "epoch": 2.7754738582877496, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.130463600158691, "learning_rate": 5e-05, "loss": 0.814, "mean_token_accuracy": 0.8740663528442383, "num_tokens": 828549329.0, "step": 21818 }, { "epoch": 2.77560106856634, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.254546165466309, "learning_rate": 5e-05, "loss": 0.7703, "mean_token_accuracy": 0.8671243190765381, "num_tokens": 828592357.0, "step": 21819 }, { "epoch": 2.7757282788449307, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.342164039611816, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.8770633339881897, "num_tokens": 828627622.0, "step": 21820 }, { "epoch": 2.775855489123521, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.031426429748535, "learning_rate": 5e-05, "loss": 0.6894, "mean_token_accuracy": 0.8892597556114197, "num_tokens": 828668680.0, "step": 21821 }, { "epoch": 2.7759826994021117, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.599403381347656, "learning_rate": 5e-05, "loss": 0.849, "mean_token_accuracy": 0.8787053823471069, "num_tokens": 828705233.0, "step": 21822 }, { "epoch": 2.7761099096807023, "ewc_loss": 0.375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.897723197937012, "learning_rate": 5e-05, "loss": 0.7416, "mean_token_accuracy": 0.8705167174339294, "num_tokens": 828745183.0, "step": 21823 }, { "epoch": 2.776237119959293, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.750870704650879, "learning_rate": 5e-05, "loss": 0.8943, "mean_token_accuracy": 0.8747721910476685, "num_tokens": 828787769.0, "step": 21824 }, { "epoch": 2.7763643302378833, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.975510597229004, "learning_rate": 5e-05, "loss": 0.7296, "mean_token_accuracy": 0.8760099411010742, "num_tokens": 828826511.0, "step": 21825 }, { "epoch": 2.776491540516474, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.65778923034668, "learning_rate": 5e-05, "loss": 0.8748, "mean_token_accuracy": 0.8816565275192261, "num_tokens": 828868632.0, "step": 21826 }, { "epoch": 2.7766187507950644, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.2368998527526855, "learning_rate": 5e-05, "loss": 0.7864, "mean_token_accuracy": 0.8697952032089233, "num_tokens": 828904342.0, "step": 21827 }, { "epoch": 2.776745961073655, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.304850578308105, "learning_rate": 5e-05, "loss": 0.8929, "mean_token_accuracy": 0.8755563497543335, "num_tokens": 828947074.0, "step": 21828 }, { "epoch": 2.7768731713522454, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.628925323486328, "learning_rate": 5e-05, "loss": 0.7762, "mean_token_accuracy": 0.8838168382644653, "num_tokens": 828989583.0, "step": 21829 }, { "epoch": 2.7770003816308355, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.730238437652588, "learning_rate": 5e-05, "loss": 0.8605, "mean_token_accuracy": 0.8527834415435791, "num_tokens": 829028848.0, "step": 21830 }, { "epoch": 2.7771275919094265, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.050307273864746, "learning_rate": 5e-05, "loss": 0.8106, "mean_token_accuracy": 0.8781465888023376, "num_tokens": 829063904.0, "step": 21831 }, { "epoch": 2.7772548021880166, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.183554649353027, "learning_rate": 5e-05, "loss": 0.7611, "mean_token_accuracy": 0.8662960529327393, "num_tokens": 829102416.0, "step": 21832 }, { "epoch": 2.7773820124666075, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.439093589782715, "learning_rate": 5e-05, "loss": 0.8649, "mean_token_accuracy": 0.8731289505958557, "num_tokens": 829138870.0, "step": 21833 }, { "epoch": 2.7775092227451976, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.831813335418701, "learning_rate": 5e-05, "loss": 0.7389, "mean_token_accuracy": 0.8721184134483337, "num_tokens": 829176704.0, "step": 21834 }, { "epoch": 2.7776364330237886, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.816975593566895, "learning_rate": 5e-05, "loss": 0.8407, "mean_token_accuracy": 0.8855757117271423, "num_tokens": 829210170.0, "step": 21835 }, { "epoch": 2.7777636433023787, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.545810699462891, "learning_rate": 5e-05, "loss": 0.6576, "mean_token_accuracy": 0.8934244513511658, "num_tokens": 829249692.0, "step": 21836 }, { "epoch": 2.777890853580969, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.202942848205566, "learning_rate": 5e-05, "loss": 0.8971, "mean_token_accuracy": 0.8885630965232849, "num_tokens": 829285342.0, "step": 21837 }, { "epoch": 2.7780180638595597, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.536575794219971, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.8806219100952148, "num_tokens": 829320281.0, "step": 21838 }, { "epoch": 2.7781452741381503, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.324414253234863, "learning_rate": 5e-05, "loss": 0.9749, "mean_token_accuracy": 0.876189112663269, "num_tokens": 829359990.0, "step": 21839 }, { "epoch": 2.778272484416741, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.070817947387695, "learning_rate": 5e-05, "loss": 0.7567, "mean_token_accuracy": 0.888847827911377, "num_tokens": 829395330.0, "step": 21840 }, { "epoch": 2.7783996946953313, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.705954551696777, "learning_rate": 5e-05, "loss": 0.947, "mean_token_accuracy": 0.8774394989013672, "num_tokens": 829432127.0, "step": 21841 }, { "epoch": 2.778526904973922, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.855227470397949, "learning_rate": 5e-05, "loss": 0.8847, "mean_token_accuracy": 0.8753758668899536, "num_tokens": 829469735.0, "step": 21842 }, { "epoch": 2.7786541152525124, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.734470367431641, "learning_rate": 5e-05, "loss": 0.8245, "mean_token_accuracy": 0.8776164054870605, "num_tokens": 829502893.0, "step": 21843 }, { "epoch": 2.778781325531103, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.358319282531738, "learning_rate": 5e-05, "loss": 0.8691, "mean_token_accuracy": 0.8853849768638611, "num_tokens": 829538307.0, "step": 21844 }, { "epoch": 2.7789085358096934, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.242570877075195, "learning_rate": 5e-05, "loss": 0.7251, "mean_token_accuracy": 0.8878439664840698, "num_tokens": 829575992.0, "step": 21845 }, { "epoch": 2.779035746088284, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.322952270507812, "learning_rate": 5e-05, "loss": 0.8563, "mean_token_accuracy": 0.8735755681991577, "num_tokens": 829610167.0, "step": 21846 }, { "epoch": 2.7791629563668745, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.439763069152832, "learning_rate": 5e-05, "loss": 0.7872, "mean_token_accuracy": 0.8729972839355469, "num_tokens": 829646828.0, "step": 21847 }, { "epoch": 2.779290166645465, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.8206400871276855, "learning_rate": 5e-05, "loss": 0.8141, "mean_token_accuracy": 0.8722109198570251, "num_tokens": 829683593.0, "step": 21848 }, { "epoch": 2.7794173769240555, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.874434471130371, "learning_rate": 5e-05, "loss": 0.789, "mean_token_accuracy": 0.8771262168884277, "num_tokens": 829721080.0, "step": 21849 }, { "epoch": 2.779544587202646, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.297324180603027, "learning_rate": 5e-05, "loss": 0.6932, "mean_token_accuracy": 0.8930411338806152, "num_tokens": 829755871.0, "step": 21850 }, { "epoch": 2.7796717974812366, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.288166046142578, "learning_rate": 5e-05, "loss": 0.8113, "mean_token_accuracy": 0.8804287314414978, "num_tokens": 829793377.0, "step": 21851 }, { "epoch": 2.779799007759827, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.93712854385376, "learning_rate": 5e-05, "loss": 0.7036, "mean_token_accuracy": 0.8768950700759888, "num_tokens": 829833165.0, "step": 21852 }, { "epoch": 2.7799262180384177, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.65455436706543, "learning_rate": 5e-05, "loss": 0.8365, "mean_token_accuracy": 0.8828873038291931, "num_tokens": 829869558.0, "step": 21853 }, { "epoch": 2.780053428317008, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.6479716300964355, "learning_rate": 5e-05, "loss": 0.7053, "mean_token_accuracy": 0.8790308237075806, "num_tokens": 829906385.0, "step": 21854 }, { "epoch": 2.7801806385955983, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 9.04702091217041, "learning_rate": 5e-05, "loss": 0.9263, "mean_token_accuracy": 0.8685271143913269, "num_tokens": 829944216.0, "step": 21855 }, { "epoch": 2.7803078488741892, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.483413219451904, "learning_rate": 5e-05, "loss": 0.7721, "mean_token_accuracy": 0.8577967286109924, "num_tokens": 829976817.0, "step": 21856 }, { "epoch": 2.7804350591527793, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.34843635559082, "learning_rate": 5e-05, "loss": 0.9849, "mean_token_accuracy": 0.8719260096549988, "num_tokens": 830020434.0, "step": 21857 }, { "epoch": 2.7805622694313703, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.698063373565674, "learning_rate": 5e-05, "loss": 0.7449, "mean_token_accuracy": 0.880788266658783, "num_tokens": 830054469.0, "step": 21858 }, { "epoch": 2.7806894797099604, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.24041748046875, "learning_rate": 5e-05, "loss": 0.9733, "mean_token_accuracy": 0.8795807361602783, "num_tokens": 830091536.0, "step": 21859 }, { "epoch": 2.780816689988551, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.328855514526367, "learning_rate": 5e-05, "loss": 0.875, "mean_token_accuracy": 0.8715469241142273, "num_tokens": 830129425.0, "step": 21860 }, { "epoch": 2.7809439002671414, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.59738540649414, "learning_rate": 5e-05, "loss": 0.9463, "mean_token_accuracy": 0.8715737462043762, "num_tokens": 830171137.0, "step": 21861 }, { "epoch": 2.781071110545732, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.881353378295898, "learning_rate": 5e-05, "loss": 0.8577, "mean_token_accuracy": 0.8893568515777588, "num_tokens": 830211434.0, "step": 21862 }, { "epoch": 2.7811983208243225, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.884021282196045, "learning_rate": 5e-05, "loss": 0.8152, "mean_token_accuracy": 0.8932240009307861, "num_tokens": 830250843.0, "step": 21863 }, { "epoch": 2.781325531102913, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.033132553100586, "learning_rate": 5e-05, "loss": 0.8943, "mean_token_accuracy": 0.8689743280410767, "num_tokens": 830292131.0, "step": 21864 }, { "epoch": 2.7814527413815036, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.71074104309082, "learning_rate": 5e-05, "loss": 0.7789, "mean_token_accuracy": 0.8878366351127625, "num_tokens": 830329678.0, "step": 21865 }, { "epoch": 2.781579951660094, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.738714694976807, "learning_rate": 5e-05, "loss": 0.7879, "mean_token_accuracy": 0.8874858021736145, "num_tokens": 830367822.0, "step": 21866 }, { "epoch": 2.7817071619386846, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.93781042098999, "learning_rate": 5e-05, "loss": 0.8295, "mean_token_accuracy": 0.8776810765266418, "num_tokens": 830408170.0, "step": 21867 }, { "epoch": 2.781834372217275, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.40645170211792, "learning_rate": 5e-05, "loss": 0.7376, "mean_token_accuracy": 0.8849908113479614, "num_tokens": 830447402.0, "step": 21868 }, { "epoch": 2.7819615824958657, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.113986015319824, "learning_rate": 5e-05, "loss": 0.8849, "mean_token_accuracy": 0.8559038639068604, "num_tokens": 830485411.0, "step": 21869 }, { "epoch": 2.782088792774456, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.2646894454956055, "learning_rate": 5e-05, "loss": 0.7494, "mean_token_accuracy": 0.8783853650093079, "num_tokens": 830522562.0, "step": 21870 }, { "epoch": 2.7822160030530467, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.098039627075195, "learning_rate": 5e-05, "loss": 0.8435, "mean_token_accuracy": 0.8670323491096497, "num_tokens": 830565979.0, "step": 21871 }, { "epoch": 2.7823432133316373, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.305663108825684, "learning_rate": 5e-05, "loss": 0.7455, "mean_token_accuracy": 0.8762246966362, "num_tokens": 830602771.0, "step": 21872 }, { "epoch": 2.782470423610228, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.019574165344238, "learning_rate": 5e-05, "loss": 0.8169, "mean_token_accuracy": 0.8696348667144775, "num_tokens": 830639891.0, "step": 21873 }, { "epoch": 2.7825976338888183, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.3508830070495605, "learning_rate": 5e-05, "loss": 0.6855, "mean_token_accuracy": 0.8894361853599548, "num_tokens": 830676044.0, "step": 21874 }, { "epoch": 2.782724844167409, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.982319355010986, "learning_rate": 5e-05, "loss": 0.7941, "mean_token_accuracy": 0.876602292060852, "num_tokens": 830718252.0, "step": 21875 }, { "epoch": 2.7828520544459994, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.295971393585205, "learning_rate": 5e-05, "loss": 0.6937, "mean_token_accuracy": 0.8898835778236389, "num_tokens": 830757563.0, "step": 21876 }, { "epoch": 2.78297926472459, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 8.044358253479004, "learning_rate": 5e-05, "loss": 0.7933, "mean_token_accuracy": 0.8737369775772095, "num_tokens": 830797808.0, "step": 21877 }, { "epoch": 2.78310647500318, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.187431335449219, "learning_rate": 5e-05, "loss": 0.733, "mean_token_accuracy": 0.8755282163619995, "num_tokens": 830838691.0, "step": 21878 }, { "epoch": 2.783233685281771, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.20261001586914, "learning_rate": 5e-05, "loss": 0.7591, "mean_token_accuracy": 0.8917190432548523, "num_tokens": 830877016.0, "step": 21879 }, { "epoch": 2.783360895560361, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.00604772567749, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8806254863739014, "num_tokens": 830908841.0, "step": 21880 }, { "epoch": 2.783488105838952, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.433637619018555, "learning_rate": 5e-05, "loss": 0.8372, "mean_token_accuracy": 0.87485271692276, "num_tokens": 830947580.0, "step": 21881 }, { "epoch": 2.783615316117542, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.803142547607422, "learning_rate": 5e-05, "loss": 0.7098, "mean_token_accuracy": 0.8795293569564819, "num_tokens": 830984385.0, "step": 21882 }, { "epoch": 2.783742526396133, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.699766159057617, "learning_rate": 5e-05, "loss": 0.8275, "mean_token_accuracy": 0.8889929056167603, "num_tokens": 831022280.0, "step": 21883 }, { "epoch": 2.783869736674723, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.662695407867432, "learning_rate": 5e-05, "loss": 0.7032, "mean_token_accuracy": 0.8815754652023315, "num_tokens": 831055764.0, "step": 21884 }, { "epoch": 2.7839969469533137, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.894399642944336, "learning_rate": 5e-05, "loss": 0.9418, "mean_token_accuracy": 0.8691966533660889, "num_tokens": 831093397.0, "step": 21885 }, { "epoch": 2.784124157231904, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.745728492736816, "learning_rate": 5e-05, "loss": 0.7254, "mean_token_accuracy": 0.8786188364028931, "num_tokens": 831130720.0, "step": 21886 }, { "epoch": 2.7842513675104947, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.774215698242188, "learning_rate": 5e-05, "loss": 0.8966, "mean_token_accuracy": 0.8770948648452759, "num_tokens": 831164092.0, "step": 21887 }, { "epoch": 2.7843785777890853, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.197461128234863, "learning_rate": 5e-05, "loss": 0.7971, "mean_token_accuracy": 0.873686671257019, "num_tokens": 831202570.0, "step": 21888 }, { "epoch": 2.784505788067676, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.099947929382324, "learning_rate": 5e-05, "loss": 0.8759, "mean_token_accuracy": 0.8677622675895691, "num_tokens": 831242596.0, "step": 21889 }, { "epoch": 2.7846329983462663, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.914299011230469, "learning_rate": 5e-05, "loss": 0.7954, "mean_token_accuracy": 0.8913784623146057, "num_tokens": 831283305.0, "step": 21890 }, { "epoch": 2.784760208624857, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.126000881195068, "learning_rate": 5e-05, "loss": 0.7032, "mean_token_accuracy": 0.8911605477333069, "num_tokens": 831322355.0, "step": 21891 }, { "epoch": 2.7848874189034474, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.76213550567627, "learning_rate": 5e-05, "loss": 0.8942, "mean_token_accuracy": 0.8754213452339172, "num_tokens": 831361410.0, "step": 21892 }, { "epoch": 2.785014629182038, "ewc_loss": 0.34375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.278524875640869, "learning_rate": 5e-05, "loss": 0.6995, "mean_token_accuracy": 0.8741330504417419, "num_tokens": 831402493.0, "step": 21893 }, { "epoch": 2.7851418394606284, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.735109329223633, "learning_rate": 5e-05, "loss": 1.0302, "mean_token_accuracy": 0.8552263975143433, "num_tokens": 831443231.0, "step": 21894 }, { "epoch": 2.785269049739219, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 5.934291362762451, "learning_rate": 5e-05, "loss": 0.634, "mean_token_accuracy": 0.9027243852615356, "num_tokens": 831474716.0, "step": 21895 }, { "epoch": 2.7853962600178095, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 10.547005653381348, "learning_rate": 5e-05, "loss": 1.149, "mean_token_accuracy": 0.863808274269104, "num_tokens": 831519410.0, "step": 21896 }, { "epoch": 2.7855234702964, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 7.1307830810546875, "learning_rate": 5e-05, "loss": 0.9019, "mean_token_accuracy": 0.8837013244628906, "num_tokens": 831562040.0, "step": 21897 }, { "epoch": 2.7856506805749905, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.087197303771973, "learning_rate": 5e-05, "loss": 1.0929, "mean_token_accuracy": 0.8706014156341553, "num_tokens": 831604956.0, "step": 21898 }, { "epoch": 2.785777890853581, "ewc_loss": 0.78515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000705718994140625, "grad_norm": 8.831625938415527, "learning_rate": 5e-05, "loss": 1.1129, "mean_token_accuracy": 0.8806049823760986, "num_tokens": 831637270.0, "step": 21899 }, { "epoch": 2.7859051011321716, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.035316467285156, "learning_rate": 5e-05, "loss": 0.8503, "mean_token_accuracy": 0.8849358558654785, "num_tokens": 831671342.0, "step": 21900 }, { "epoch": 2.786032311410762, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 9.943902015686035, "learning_rate": 5e-05, "loss": 1.1111, "mean_token_accuracy": 0.8779685497283936, "num_tokens": 831709222.0, "step": 21901 }, { "epoch": 2.7861595216893527, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 6.855899333953857, "learning_rate": 5e-05, "loss": 0.7874, "mean_token_accuracy": 0.8873001337051392, "num_tokens": 831749690.0, "step": 21902 }, { "epoch": 2.7862867319679427, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 10.818686485290527, "learning_rate": 5e-05, "loss": 1.1993, "mean_token_accuracy": 0.8675248622894287, "num_tokens": 831787963.0, "step": 21903 }, { "epoch": 2.7864139422465337, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 8.007853507995605, "learning_rate": 5e-05, "loss": 0.997, "mean_token_accuracy": 0.884373664855957, "num_tokens": 831820925.0, "step": 21904 }, { "epoch": 2.786541152525124, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 10.057056427001953, "learning_rate": 5e-05, "loss": 1.0506, "mean_token_accuracy": 0.8718933463096619, "num_tokens": 831856190.0, "step": 21905 }, { "epoch": 2.7866683628037148, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 8.841712951660156, "learning_rate": 5e-05, "loss": 1.1513, "mean_token_accuracy": 0.8660563826560974, "num_tokens": 831900339.0, "step": 21906 }, { "epoch": 2.786795573082305, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.694232940673828, "learning_rate": 5e-05, "loss": 0.9385, "mean_token_accuracy": 0.8707776069641113, "num_tokens": 831932166.0, "step": 21907 }, { "epoch": 2.786922783360896, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.940774917602539, "learning_rate": 5e-05, "loss": 0.9947, "mean_token_accuracy": 0.881820023059845, "num_tokens": 831964796.0, "step": 21908 }, { "epoch": 2.787049993639486, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.052971839904785, "learning_rate": 5e-05, "loss": 0.9147, "mean_token_accuracy": 0.8755601644515991, "num_tokens": 832004603.0, "step": 21909 }, { "epoch": 2.7871772039180764, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.569101333618164, "learning_rate": 5e-05, "loss": 0.9014, "mean_token_accuracy": 0.8888199329376221, "num_tokens": 832042847.0, "step": 21910 }, { "epoch": 2.787304414196667, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.063026428222656, "learning_rate": 5e-05, "loss": 0.906, "mean_token_accuracy": 0.8796367645263672, "num_tokens": 832079746.0, "step": 21911 }, { "epoch": 2.7874316244752575, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.856337070465088, "learning_rate": 5e-05, "loss": 0.8499, "mean_token_accuracy": 0.8710081577301025, "num_tokens": 832116300.0, "step": 21912 }, { "epoch": 2.787558834753848, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.35997486114502, "learning_rate": 5e-05, "loss": 0.8665, "mean_token_accuracy": 0.8894765973091125, "num_tokens": 832151985.0, "step": 21913 }, { "epoch": 2.7876860450324386, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.269344806671143, "learning_rate": 5e-05, "loss": 0.7722, "mean_token_accuracy": 0.879618763923645, "num_tokens": 832181327.0, "step": 21914 }, { "epoch": 2.787813255311029, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.732858657836914, "learning_rate": 5e-05, "loss": 0.8895, "mean_token_accuracy": 0.8852593898773193, "num_tokens": 832217710.0, "step": 21915 }, { "epoch": 2.7879404655896196, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.888758659362793, "learning_rate": 5e-05, "loss": 0.7678, "mean_token_accuracy": 0.87378990650177, "num_tokens": 832255628.0, "step": 21916 }, { "epoch": 2.78806767586821, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.142569541931152, "learning_rate": 5e-05, "loss": 0.9149, "mean_token_accuracy": 0.8807449340820312, "num_tokens": 832294239.0, "step": 21917 }, { "epoch": 2.7881948861468007, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.774740695953369, "learning_rate": 5e-05, "loss": 0.7464, "mean_token_accuracy": 0.8852540254592896, "num_tokens": 832331270.0, "step": 21918 }, { "epoch": 2.788322096425391, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.349983215332031, "learning_rate": 5e-05, "loss": 0.976, "mean_token_accuracy": 0.8715890049934387, "num_tokens": 832372276.0, "step": 21919 }, { "epoch": 2.7884493067039817, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.032125473022461, "learning_rate": 5e-05, "loss": 0.8229, "mean_token_accuracy": 0.8872298002243042, "num_tokens": 832410352.0, "step": 21920 }, { "epoch": 2.7885765169825723, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.102653503417969, "learning_rate": 5e-05, "loss": 0.9734, "mean_token_accuracy": 0.8723184466362, "num_tokens": 832452633.0, "step": 21921 }, { "epoch": 2.788703727261163, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.418012619018555, "learning_rate": 5e-05, "loss": 0.8675, "mean_token_accuracy": 0.8845268487930298, "num_tokens": 832492100.0, "step": 21922 }, { "epoch": 2.7888309375397533, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.64698600769043, "learning_rate": 5e-05, "loss": 0.9215, "mean_token_accuracy": 0.8754317760467529, "num_tokens": 832529972.0, "step": 21923 }, { "epoch": 2.788958147818344, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.594347953796387, "learning_rate": 5e-05, "loss": 0.8443, "mean_token_accuracy": 0.8893125057220459, "num_tokens": 832571042.0, "step": 21924 }, { "epoch": 2.7890853580969344, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.34881591796875, "learning_rate": 5e-05, "loss": 0.9031, "mean_token_accuracy": 0.869905948638916, "num_tokens": 832610575.0, "step": 21925 }, { "epoch": 2.789212568375525, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.497498989105225, "learning_rate": 5e-05, "loss": 0.7992, "mean_token_accuracy": 0.8910727500915527, "num_tokens": 832649477.0, "step": 21926 }, { "epoch": 2.7893397786541154, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.284893989562988, "learning_rate": 5e-05, "loss": 0.8481, "mean_token_accuracy": 0.8854158520698547, "num_tokens": 832688958.0, "step": 21927 }, { "epoch": 2.7894669889327055, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.345478534698486, "learning_rate": 5e-05, "loss": 0.7646, "mean_token_accuracy": 0.8896925449371338, "num_tokens": 832724368.0, "step": 21928 }, { "epoch": 2.7895941992112965, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.33489990234375, "learning_rate": 5e-05, "loss": 0.8842, "mean_token_accuracy": 0.8755155801773071, "num_tokens": 832765950.0, "step": 21929 }, { "epoch": 2.7897214094898866, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.198944568634033, "learning_rate": 5e-05, "loss": 0.7124, "mean_token_accuracy": 0.8934952616691589, "num_tokens": 832800666.0, "step": 21930 }, { "epoch": 2.7898486197684775, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.486799240112305, "learning_rate": 5e-05, "loss": 0.9064, "mean_token_accuracy": 0.8685365319252014, "num_tokens": 832835830.0, "step": 21931 }, { "epoch": 2.7899758300470676, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.107504844665527, "learning_rate": 5e-05, "loss": 0.7226, "mean_token_accuracy": 0.8893307447433472, "num_tokens": 832866370.0, "step": 21932 }, { "epoch": 2.7901030403256586, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.518217086791992, "learning_rate": 5e-05, "loss": 0.883, "mean_token_accuracy": 0.8798650503158569, "num_tokens": 832914174.0, "step": 21933 }, { "epoch": 2.7902302506042487, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.164212703704834, "learning_rate": 5e-05, "loss": 0.7363, "mean_token_accuracy": 0.890411376953125, "num_tokens": 832951397.0, "step": 21934 }, { "epoch": 2.790357460882839, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.42021656036377, "learning_rate": 5e-05, "loss": 0.8631, "mean_token_accuracy": 0.8828496336936951, "num_tokens": 832986282.0, "step": 21935 }, { "epoch": 2.7904846711614297, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.316938877105713, "learning_rate": 5e-05, "loss": 0.7644, "mean_token_accuracy": 0.8873260021209717, "num_tokens": 833023288.0, "step": 21936 }, { "epoch": 2.7906118814400203, "ewc_loss": 0.5, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.190778732299805, "learning_rate": 5e-05, "loss": 0.8535, "mean_token_accuracy": 0.8775951266288757, "num_tokens": 833053744.0, "step": 21937 }, { "epoch": 2.790739091718611, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.51862907409668, "learning_rate": 5e-05, "loss": 0.7649, "mean_token_accuracy": 0.888871431350708, "num_tokens": 833096499.0, "step": 21938 }, { "epoch": 2.7908663019972013, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.8659234046936035, "learning_rate": 5e-05, "loss": 0.768, "mean_token_accuracy": 0.8888378739356995, "num_tokens": 833132405.0, "step": 21939 }, { "epoch": 2.790993512275792, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.718419551849365, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8796259760856628, "num_tokens": 833175581.0, "step": 21940 }, { "epoch": 2.7911207225543824, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.569357872009277, "learning_rate": 5e-05, "loss": 0.7915, "mean_token_accuracy": 0.8712543249130249, "num_tokens": 833213466.0, "step": 21941 }, { "epoch": 2.791247932832973, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.891623497009277, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8743125200271606, "num_tokens": 833250344.0, "step": 21942 }, { "epoch": 2.7913751431115634, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.3456807136535645, "learning_rate": 5e-05, "loss": 0.7337, "mean_token_accuracy": 0.8793420791625977, "num_tokens": 833291032.0, "step": 21943 }, { "epoch": 2.791502353390154, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.018239974975586, "learning_rate": 5e-05, "loss": 0.7732, "mean_token_accuracy": 0.8838819265365601, "num_tokens": 833331213.0, "step": 21944 }, { "epoch": 2.7916295636687445, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.185425758361816, "learning_rate": 5e-05, "loss": 0.6951, "mean_token_accuracy": 0.8863955736160278, "num_tokens": 833373906.0, "step": 21945 }, { "epoch": 2.791756773947335, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.16862964630127, "learning_rate": 5e-05, "loss": 0.804, "mean_token_accuracy": 0.8734217882156372, "num_tokens": 833411724.0, "step": 21946 }, { "epoch": 2.7918839842259255, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 7.023212909698486, "learning_rate": 5e-05, "loss": 0.6858, "mean_token_accuracy": 0.8871046304702759, "num_tokens": 833446337.0, "step": 21947 }, { "epoch": 2.792011194504516, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.347466468811035, "learning_rate": 5e-05, "loss": 0.8353, "mean_token_accuracy": 0.8695464134216309, "num_tokens": 833484444.0, "step": 21948 }, { "epoch": 2.7921384047831066, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.845201015472412, "learning_rate": 5e-05, "loss": 0.7017, "mean_token_accuracy": 0.8804895877838135, "num_tokens": 833521375.0, "step": 21949 }, { "epoch": 2.792265615061697, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.610025405883789, "learning_rate": 5e-05, "loss": 0.8143, "mean_token_accuracy": 0.8882192373275757, "num_tokens": 833557784.0, "step": 21950 }, { "epoch": 2.7923928253402877, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.621294975280762, "learning_rate": 5e-05, "loss": 0.6487, "mean_token_accuracy": 0.8983877897262573, "num_tokens": 833598872.0, "step": 21951 }, { "epoch": 2.792520035618878, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.935389518737793, "learning_rate": 5e-05, "loss": 0.8553, "mean_token_accuracy": 0.8873777389526367, "num_tokens": 833635799.0, "step": 21952 }, { "epoch": 2.7926472458974683, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.5089545249938965, "learning_rate": 5e-05, "loss": 0.6801, "mean_token_accuracy": 0.8864572048187256, "num_tokens": 833671814.0, "step": 21953 }, { "epoch": 2.7927744561760592, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.170738220214844, "learning_rate": 5e-05, "loss": 0.9663, "mean_token_accuracy": 0.8686468601226807, "num_tokens": 833718081.0, "step": 21954 }, { "epoch": 2.7929016664546493, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.74049711227417, "learning_rate": 5e-05, "loss": 0.7458, "mean_token_accuracy": 0.877164900302887, "num_tokens": 833759197.0, "step": 21955 }, { "epoch": 2.7930288767332403, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.919305801391602, "learning_rate": 5e-05, "loss": 0.9066, "mean_token_accuracy": 0.8869326114654541, "num_tokens": 833785582.0, "step": 21956 }, { "epoch": 2.7931560870118304, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.405981063842773, "learning_rate": 5e-05, "loss": 0.8293, "mean_token_accuracy": 0.8861526846885681, "num_tokens": 833826611.0, "step": 21957 }, { "epoch": 2.793283297290421, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.037620544433594, "learning_rate": 5e-05, "loss": 0.877, "mean_token_accuracy": 0.8756008148193359, "num_tokens": 833870657.0, "step": 21958 }, { "epoch": 2.7934105075690114, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.152992248535156, "learning_rate": 5e-05, "loss": 0.87, "mean_token_accuracy": 0.8793540000915527, "num_tokens": 833907621.0, "step": 21959 }, { "epoch": 2.793537717847602, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.083499431610107, "learning_rate": 5e-05, "loss": 0.7315, "mean_token_accuracy": 0.8842068314552307, "num_tokens": 833944364.0, "step": 21960 }, { "epoch": 2.7936649281261925, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.725543975830078, "learning_rate": 5e-05, "loss": 0.8587, "mean_token_accuracy": 0.8898119330406189, "num_tokens": 833983840.0, "step": 21961 }, { "epoch": 2.793792138404783, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.66326379776001, "learning_rate": 5e-05, "loss": 0.673, "mean_token_accuracy": 0.8908542394638062, "num_tokens": 834016367.0, "step": 21962 }, { "epoch": 2.7939193486833735, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.942625999450684, "learning_rate": 5e-05, "loss": 0.8737, "mean_token_accuracy": 0.8874085545539856, "num_tokens": 834050190.0, "step": 21963 }, { "epoch": 2.794046558961964, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.74411678314209, "learning_rate": 5e-05, "loss": 0.7626, "mean_token_accuracy": 0.8734562397003174, "num_tokens": 834087013.0, "step": 21964 }, { "epoch": 2.7941737692405546, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.787513732910156, "learning_rate": 5e-05, "loss": 0.9212, "mean_token_accuracy": 0.8772690296173096, "num_tokens": 834131771.0, "step": 21965 }, { "epoch": 2.794300979519145, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.1291937828063965, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.884347140789032, "num_tokens": 834169055.0, "step": 21966 }, { "epoch": 2.7944281897977357, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.338151931762695, "learning_rate": 5e-05, "loss": 0.8568, "mean_token_accuracy": 0.8858805894851685, "num_tokens": 834206081.0, "step": 21967 }, { "epoch": 2.794555400076326, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.538847923278809, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.8897467851638794, "num_tokens": 834240846.0, "step": 21968 }, { "epoch": 2.7946826103549167, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.844045639038086, "learning_rate": 5e-05, "loss": 0.8093, "mean_token_accuracy": 0.8780525922775269, "num_tokens": 834275037.0, "step": 21969 }, { "epoch": 2.7948098206335072, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.7522993087768555, "learning_rate": 5e-05, "loss": 0.8071, "mean_token_accuracy": 0.8805146217346191, "num_tokens": 834316408.0, "step": 21970 }, { "epoch": 2.7949370309120978, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.648078918457031, "learning_rate": 5e-05, "loss": 0.7483, "mean_token_accuracy": 0.8882558941841125, "num_tokens": 834358473.0, "step": 21971 }, { "epoch": 2.7950642411906883, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.698643207550049, "learning_rate": 5e-05, "loss": 0.7846, "mean_token_accuracy": 0.8802740573883057, "num_tokens": 834399982.0, "step": 21972 }, { "epoch": 2.795191451469279, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.717463970184326, "learning_rate": 5e-05, "loss": 0.7495, "mean_token_accuracy": 0.8872029781341553, "num_tokens": 834429731.0, "step": 21973 }, { "epoch": 2.7953186617478694, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.447287559509277, "learning_rate": 5e-05, "loss": 0.7624, "mean_token_accuracy": 0.8774357438087463, "num_tokens": 834476958.0, "step": 21974 }, { "epoch": 2.79544587202646, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.967404365539551, "learning_rate": 5e-05, "loss": 0.7794, "mean_token_accuracy": 0.885182797908783, "num_tokens": 834516820.0, "step": 21975 }, { "epoch": 2.79557308230505, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.1225738525390625, "learning_rate": 5e-05, "loss": 0.7252, "mean_token_accuracy": 0.8786588907241821, "num_tokens": 834552455.0, "step": 21976 }, { "epoch": 2.795700292583641, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.289905548095703, "learning_rate": 5e-05, "loss": 0.8142, "mean_token_accuracy": 0.8766064643859863, "num_tokens": 834593059.0, "step": 21977 }, { "epoch": 2.795827502862231, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.773711681365967, "learning_rate": 5e-05, "loss": 0.7341, "mean_token_accuracy": 0.8698787689208984, "num_tokens": 834640134.0, "step": 21978 }, { "epoch": 2.795954713140822, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.75001335144043, "learning_rate": 5e-05, "loss": 0.8623, "mean_token_accuracy": 0.8776504397392273, "num_tokens": 834679172.0, "step": 21979 }, { "epoch": 2.796081923419412, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.395894527435303, "learning_rate": 5e-05, "loss": 0.7195, "mean_token_accuracy": 0.8698094487190247, "num_tokens": 834713463.0, "step": 21980 }, { "epoch": 2.796209133698003, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.335217475891113, "learning_rate": 5e-05, "loss": 0.9849, "mean_token_accuracy": 0.8589329719543457, "num_tokens": 834757782.0, "step": 21981 }, { "epoch": 2.796336343976593, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.252363681793213, "learning_rate": 5e-05, "loss": 0.7377, "mean_token_accuracy": 0.8719433546066284, "num_tokens": 834791598.0, "step": 21982 }, { "epoch": 2.7964635542551837, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.70482349395752, "learning_rate": 5e-05, "loss": 1.0232, "mean_token_accuracy": 0.8693488240242004, "num_tokens": 834835205.0, "step": 21983 }, { "epoch": 2.796590764533774, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 6.900067329406738, "learning_rate": 5e-05, "loss": 0.7975, "mean_token_accuracy": 0.8891671895980835, "num_tokens": 834867450.0, "step": 21984 }, { "epoch": 2.7967179748123647, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.199159622192383, "learning_rate": 5e-05, "loss": 0.9963, "mean_token_accuracy": 0.8763933181762695, "num_tokens": 834911067.0, "step": 21985 }, { "epoch": 2.7968451850909553, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.004463195800781, "learning_rate": 5e-05, "loss": 0.9531, "mean_token_accuracy": 0.8790091276168823, "num_tokens": 834949117.0, "step": 21986 }, { "epoch": 2.796972395369546, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.735762119293213, "learning_rate": 5e-05, "loss": 0.8604, "mean_token_accuracy": 0.8770847320556641, "num_tokens": 834988909.0, "step": 21987 }, { "epoch": 2.7970996056481363, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.934585571289062, "learning_rate": 5e-05, "loss": 0.9661, "mean_token_accuracy": 0.8840088248252869, "num_tokens": 835031168.0, "step": 21988 }, { "epoch": 2.797226815926727, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.62332820892334, "learning_rate": 5e-05, "loss": 0.7088, "mean_token_accuracy": 0.886039137840271, "num_tokens": 835072110.0, "step": 21989 }, { "epoch": 2.7973540262053174, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.680583000183105, "learning_rate": 5e-05, "loss": 0.9647, "mean_token_accuracy": 0.8900129199028015, "num_tokens": 835104605.0, "step": 21990 }, { "epoch": 2.797481236483908, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 6.623164176940918, "learning_rate": 5e-05, "loss": 0.8325, "mean_token_accuracy": 0.8660564422607422, "num_tokens": 835141970.0, "step": 21991 }, { "epoch": 2.7976084467624984, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.759233474731445, "learning_rate": 5e-05, "loss": 1.0783, "mean_token_accuracy": 0.8654485940933228, "num_tokens": 835177469.0, "step": 21992 }, { "epoch": 2.797735657041089, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 7.5920186042785645, "learning_rate": 5e-05, "loss": 0.9649, "mean_token_accuracy": 0.87608802318573, "num_tokens": 835220286.0, "step": 21993 }, { "epoch": 2.7978628673196795, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.904852867126465, "learning_rate": 5e-05, "loss": 0.9654, "mean_token_accuracy": 0.8818466663360596, "num_tokens": 835254290.0, "step": 21994 }, { "epoch": 2.79799007759827, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.318521499633789, "learning_rate": 5e-05, "loss": 0.9877, "mean_token_accuracy": 0.8772754073143005, "num_tokens": 835286375.0, "step": 21995 }, { "epoch": 2.7981172878768605, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.877192497253418, "learning_rate": 5e-05, "loss": 0.8619, "mean_token_accuracy": 0.881234884262085, "num_tokens": 835315472.0, "step": 21996 }, { "epoch": 2.798244498155451, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.393528938293457, "learning_rate": 5e-05, "loss": 0.9136, "mean_token_accuracy": 0.8860734701156616, "num_tokens": 835355865.0, "step": 21997 }, { "epoch": 2.7983717084340416, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.667373180389404, "learning_rate": 5e-05, "loss": 0.8606, "mean_token_accuracy": 0.8798823952674866, "num_tokens": 835391251.0, "step": 21998 }, { "epoch": 2.798498918712632, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.858377456665039, "learning_rate": 5e-05, "loss": 0.821, "mean_token_accuracy": 0.8901203274726868, "num_tokens": 835424869.0, "step": 21999 }, { "epoch": 2.7986261289912227, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.013731956481934, "learning_rate": 5e-05, "loss": 0.8688, "mean_token_accuracy": 0.8812263607978821, "num_tokens": 835464916.0, "step": 22000 }, { "epoch": 2.7987533392698127, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.286370277404785, "learning_rate": 5e-05, "loss": 0.7967, "mean_token_accuracy": 0.8728950023651123, "num_tokens": 835503114.0, "step": 22001 }, { "epoch": 2.7988805495484037, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.340263366699219, "learning_rate": 5e-05, "loss": 0.909, "mean_token_accuracy": 0.8693549036979675, "num_tokens": 835541446.0, "step": 22002 }, { "epoch": 2.799007759826994, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.961691379547119, "learning_rate": 5e-05, "loss": 0.6911, "mean_token_accuracy": 0.8932652473449707, "num_tokens": 835579261.0, "step": 22003 }, { "epoch": 2.7991349701055848, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.649153709411621, "learning_rate": 5e-05, "loss": 0.8962, "mean_token_accuracy": 0.8758625388145447, "num_tokens": 835615847.0, "step": 22004 }, { "epoch": 2.799262180384175, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.687660217285156, "learning_rate": 5e-05, "loss": 0.7809, "mean_token_accuracy": 0.8610414266586304, "num_tokens": 835658461.0, "step": 22005 }, { "epoch": 2.799389390662766, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 9.124433517456055, "learning_rate": 5e-05, "loss": 0.911, "mean_token_accuracy": 0.885554850101471, "num_tokens": 835693489.0, "step": 22006 }, { "epoch": 2.799516600941356, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.521118640899658, "learning_rate": 5e-05, "loss": 0.6953, "mean_token_accuracy": 0.8911697268486023, "num_tokens": 835733925.0, "step": 22007 }, { "epoch": 2.7996438112199464, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.52847671508789, "learning_rate": 5e-05, "loss": 1.0121, "mean_token_accuracy": 0.8729851245880127, "num_tokens": 835775071.0, "step": 22008 }, { "epoch": 2.799771021498537, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 6.918507099151611, "learning_rate": 5e-05, "loss": 0.8443, "mean_token_accuracy": 0.8758841156959534, "num_tokens": 835813940.0, "step": 22009 }, { "epoch": 2.7998982317771275, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.35555362701416, "learning_rate": 5e-05, "loss": 1.0454, "mean_token_accuracy": 0.8663251996040344, "num_tokens": 835849667.0, "step": 22010 }, { "epoch": 2.800025442055718, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.68927526473999, "learning_rate": 5e-05, "loss": 0.9258, "mean_token_accuracy": 0.8831934332847595, "num_tokens": 835880928.0, "step": 22011 }, { "epoch": 2.8001526523343085, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.553418159484863, "learning_rate": 5e-05, "loss": 0.8849, "mean_token_accuracy": 0.8963377475738525, "num_tokens": 835917502.0, "step": 22012 }, { "epoch": 2.800279862612899, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.11923599243164, "learning_rate": 5e-05, "loss": 0.972, "mean_token_accuracy": 0.8699000477790833, "num_tokens": 835958547.0, "step": 22013 }, { "epoch": 2.8004070728914896, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.82205057144165, "learning_rate": 5e-05, "loss": 0.8742, "mean_token_accuracy": 0.8792287111282349, "num_tokens": 835997427.0, "step": 22014 }, { "epoch": 2.80053428317008, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.099027633666992, "learning_rate": 5e-05, "loss": 0.8745, "mean_token_accuracy": 0.8860713243484497, "num_tokens": 836039096.0, "step": 22015 }, { "epoch": 2.8006614934486707, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.667791366577148, "learning_rate": 5e-05, "loss": 0.8235, "mean_token_accuracy": 0.8795820474624634, "num_tokens": 836079613.0, "step": 22016 }, { "epoch": 2.800788703727261, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.8340067863464355, "learning_rate": 5e-05, "loss": 0.8275, "mean_token_accuracy": 0.8842829465866089, "num_tokens": 836117500.0, "step": 22017 }, { "epoch": 2.8009159140058517, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.687608242034912, "learning_rate": 5e-05, "loss": 0.7963, "mean_token_accuracy": 0.8799476623535156, "num_tokens": 836149680.0, "step": 22018 }, { "epoch": 2.8010431242844422, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.7294392585754395, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.8971608877182007, "num_tokens": 836179869.0, "step": 22019 }, { "epoch": 2.8011703345630328, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.53018856048584, "learning_rate": 5e-05, "loss": 0.784, "mean_token_accuracy": 0.8765318393707275, "num_tokens": 836222305.0, "step": 22020 }, { "epoch": 2.8012975448416233, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.920962333679199, "learning_rate": 5e-05, "loss": 0.8151, "mean_token_accuracy": 0.8744159936904907, "num_tokens": 836257687.0, "step": 22021 }, { "epoch": 2.801424755120214, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.193681716918945, "learning_rate": 5e-05, "loss": 0.7365, "mean_token_accuracy": 0.8797204494476318, "num_tokens": 836299105.0, "step": 22022 }, { "epoch": 2.8015519653988044, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.235870361328125, "learning_rate": 5e-05, "loss": 0.8111, "mean_token_accuracy": 0.8798521757125854, "num_tokens": 836337661.0, "step": 22023 }, { "epoch": 2.801679175677395, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.883701801300049, "learning_rate": 5e-05, "loss": 0.7419, "mean_token_accuracy": 0.8708741068840027, "num_tokens": 836378792.0, "step": 22024 }, { "epoch": 2.8018063859559854, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.59045124053955, "learning_rate": 5e-05, "loss": 0.8602, "mean_token_accuracy": 0.8765811920166016, "num_tokens": 836412754.0, "step": 22025 }, { "epoch": 2.8019335962345755, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.587059497833252, "learning_rate": 5e-05, "loss": 0.6847, "mean_token_accuracy": 0.8839490413665771, "num_tokens": 836452173.0, "step": 22026 }, { "epoch": 2.8020608065131665, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.047853469848633, "learning_rate": 5e-05, "loss": 0.8847, "mean_token_accuracy": 0.888590931892395, "num_tokens": 836491398.0, "step": 22027 }, { "epoch": 2.8021880167917566, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.449295520782471, "learning_rate": 5e-05, "loss": 0.744, "mean_token_accuracy": 0.8700233697891235, "num_tokens": 836525130.0, "step": 22028 }, { "epoch": 2.8023152270703475, "ewc_loss": 0.625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.380963325500488, "learning_rate": 5e-05, "loss": 0.9806, "mean_token_accuracy": 0.8765568733215332, "num_tokens": 836561288.0, "step": 22029 }, { "epoch": 2.8024424373489376, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 6.828190326690674, "learning_rate": 5e-05, "loss": 0.8394, "mean_token_accuracy": 0.863563060760498, "num_tokens": 836603595.0, "step": 22030 }, { "epoch": 2.8025696476275286, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.089613914489746, "learning_rate": 5e-05, "loss": 0.9817, "mean_token_accuracy": 0.8738172054290771, "num_tokens": 836640460.0, "step": 22031 }, { "epoch": 2.8026968579061187, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 7.664896011352539, "learning_rate": 5e-05, "loss": 0.9368, "mean_token_accuracy": 0.86875981092453, "num_tokens": 836678024.0, "step": 22032 }, { "epoch": 2.802824068184709, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.036105155944824, "learning_rate": 5e-05, "loss": 0.8322, "mean_token_accuracy": 0.8925894498825073, "num_tokens": 836718383.0, "step": 22033 }, { "epoch": 2.8029512784632997, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.353175163269043, "learning_rate": 5e-05, "loss": 0.9677, "mean_token_accuracy": 0.8657327890396118, "num_tokens": 836755564.0, "step": 22034 }, { "epoch": 2.8030784887418903, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.126818656921387, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8781230449676514, "num_tokens": 836792891.0, "step": 22035 }, { "epoch": 2.803205699020481, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.655190467834473, "learning_rate": 5e-05, "loss": 0.9189, "mean_token_accuracy": 0.8742652535438538, "num_tokens": 836830024.0, "step": 22036 }, { "epoch": 2.8033329092990713, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.92609167098999, "learning_rate": 5e-05, "loss": 0.774, "mean_token_accuracy": 0.8707002997398376, "num_tokens": 836871600.0, "step": 22037 }, { "epoch": 2.803460119577662, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.571435928344727, "learning_rate": 5e-05, "loss": 0.873, "mean_token_accuracy": 0.8870662450790405, "num_tokens": 836911495.0, "step": 22038 }, { "epoch": 2.8035873298562524, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.126672744750977, "learning_rate": 5e-05, "loss": 0.7713, "mean_token_accuracy": 0.8793901205062866, "num_tokens": 836952840.0, "step": 22039 }, { "epoch": 2.803714540134843, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.272982597351074, "learning_rate": 5e-05, "loss": 0.879, "mean_token_accuracy": 0.875583291053772, "num_tokens": 836988355.0, "step": 22040 }, { "epoch": 2.8038417504134334, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.359769821166992, "learning_rate": 5e-05, "loss": 0.7926, "mean_token_accuracy": 0.8802372217178345, "num_tokens": 837021559.0, "step": 22041 }, { "epoch": 2.803968960692024, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.022196769714355, "learning_rate": 5e-05, "loss": 0.8301, "mean_token_accuracy": 0.8728817105293274, "num_tokens": 837060401.0, "step": 22042 }, { "epoch": 2.8040961709706145, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.462804317474365, "learning_rate": 5e-05, "loss": 0.7554, "mean_token_accuracy": 0.8879302740097046, "num_tokens": 837094769.0, "step": 22043 }, { "epoch": 2.804223381249205, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.846865653991699, "learning_rate": 5e-05, "loss": 0.7566, "mean_token_accuracy": 0.892731785774231, "num_tokens": 837125820.0, "step": 22044 }, { "epoch": 2.8043505915277955, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.522515773773193, "learning_rate": 5e-05, "loss": 0.7534, "mean_token_accuracy": 0.8869134187698364, "num_tokens": 837167392.0, "step": 22045 }, { "epoch": 2.804477801806386, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.73950719833374, "learning_rate": 5e-05, "loss": 0.7704, "mean_token_accuracy": 0.8787119388580322, "num_tokens": 837207439.0, "step": 22046 }, { "epoch": 2.8046050120849766, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.519859790802002, "learning_rate": 5e-05, "loss": 0.7563, "mean_token_accuracy": 0.8832195997238159, "num_tokens": 837240038.0, "step": 22047 }, { "epoch": 2.804732222363567, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.716263771057129, "learning_rate": 5e-05, "loss": 0.7877, "mean_token_accuracy": 0.8732880353927612, "num_tokens": 837278842.0, "step": 22048 }, { "epoch": 2.8048594326421576, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.453042030334473, "learning_rate": 5e-05, "loss": 0.6926, "mean_token_accuracy": 0.8949881792068481, "num_tokens": 837317519.0, "step": 22049 }, { "epoch": 2.804986642920748, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.764822483062744, "learning_rate": 5e-05, "loss": 0.7359, "mean_token_accuracy": 0.8908051252365112, "num_tokens": 837349053.0, "step": 22050 }, { "epoch": 2.8051138531993383, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.331779479980469, "learning_rate": 5e-05, "loss": 0.7292, "mean_token_accuracy": 0.8743658065795898, "num_tokens": 837386018.0, "step": 22051 }, { "epoch": 2.8052410634779292, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.891425609588623, "learning_rate": 5e-05, "loss": 0.7822, "mean_token_accuracy": 0.8749927878379822, "num_tokens": 837423640.0, "step": 22052 }, { "epoch": 2.8053682737565193, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.146127700805664, "learning_rate": 5e-05, "loss": 0.7032, "mean_token_accuracy": 0.8801287412643433, "num_tokens": 837462198.0, "step": 22053 }, { "epoch": 2.8054954840351103, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.119620323181152, "learning_rate": 5e-05, "loss": 0.7867, "mean_token_accuracy": 0.8782325387001038, "num_tokens": 837492722.0, "step": 22054 }, { "epoch": 2.8056226943137004, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.855332374572754, "learning_rate": 5e-05, "loss": 0.6867, "mean_token_accuracy": 0.8848991394042969, "num_tokens": 837531598.0, "step": 22055 }, { "epoch": 2.805749904592291, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.52268123626709, "learning_rate": 5e-05, "loss": 0.8292, "mean_token_accuracy": 0.8721415996551514, "num_tokens": 837572580.0, "step": 22056 }, { "epoch": 2.8058771148708814, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 6.432783603668213, "learning_rate": 5e-05, "loss": 0.6561, "mean_token_accuracy": 0.8868051767349243, "num_tokens": 837610398.0, "step": 22057 }, { "epoch": 2.806004325149472, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 9.17080020904541, "learning_rate": 5e-05, "loss": 0.8974, "mean_token_accuracy": 0.8755390644073486, "num_tokens": 837653518.0, "step": 22058 }, { "epoch": 2.8061315354280625, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002574920654296875, "grad_norm": 5.99285364151001, "learning_rate": 5e-05, "loss": 0.6827, "mean_token_accuracy": 0.8785684704780579, "num_tokens": 837697397.0, "step": 22059 }, { "epoch": 2.806258745706653, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 10.005435943603516, "learning_rate": 5e-05, "loss": 0.9434, "mean_token_accuracy": 0.8940446376800537, "num_tokens": 837736565.0, "step": 22060 }, { "epoch": 2.8063859559852435, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.230710506439209, "learning_rate": 5e-05, "loss": 0.746, "mean_token_accuracy": 0.8794658780097961, "num_tokens": 837772514.0, "step": 22061 }, { "epoch": 2.806513166263834, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.289178848266602, "learning_rate": 5e-05, "loss": 1.0581, "mean_token_accuracy": 0.8872346878051758, "num_tokens": 837805147.0, "step": 22062 }, { "epoch": 2.8066403765424246, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 7.817325592041016, "learning_rate": 5e-05, "loss": 1.0314, "mean_token_accuracy": 0.8648271560668945, "num_tokens": 837846795.0, "step": 22063 }, { "epoch": 2.806767586821015, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.897828102111816, "learning_rate": 5e-05, "loss": 0.9453, "mean_token_accuracy": 0.880875289440155, "num_tokens": 837882703.0, "step": 22064 }, { "epoch": 2.8068947970996057, "ewc_loss": 0.734375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 9.164362907409668, "learning_rate": 5e-05, "loss": 1.1197, "mean_token_accuracy": 0.8652878999710083, "num_tokens": 837924698.0, "step": 22065 }, { "epoch": 2.807022007378196, "ewc_loss": 0.40625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 6.791018009185791, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.8817434310913086, "num_tokens": 837958553.0, "step": 22066 }, { "epoch": 2.8071492176567867, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 10.85694694519043, "learning_rate": 5e-05, "loss": 1.1106, "mean_token_accuracy": 0.9017378091812134, "num_tokens": 837998674.0, "step": 22067 }, { "epoch": 2.8072764279353772, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 6.958205699920654, "learning_rate": 5e-05, "loss": 0.8596, "mean_token_accuracy": 0.8801980018615723, "num_tokens": 838035111.0, "step": 22068 }, { "epoch": 2.8074036382139678, "ewc_loss": 0.9375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000858306884765625, "grad_norm": 11.813254356384277, "learning_rate": 5e-05, "loss": 1.2994, "mean_token_accuracy": 0.875295102596283, "num_tokens": 838075444.0, "step": 22069 }, { "epoch": 2.8075308484925583, "ewc_loss": 0.9609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000881195068359375, "grad_norm": 9.403684616088867, "learning_rate": 5e-05, "loss": 1.3729, "mean_token_accuracy": 0.8603885173797607, "num_tokens": 838117972.0, "step": 22070 }, { "epoch": 2.807658058771149, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.507474899291992, "learning_rate": 5e-05, "loss": 1.0682, "mean_token_accuracy": 0.8690975904464722, "num_tokens": 838158797.0, "step": 22071 }, { "epoch": 2.8077852690497394, "ewc_loss": 0.86328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00078582763671875, "grad_norm": 9.369214057922363, "learning_rate": 5e-05, "loss": 1.2276, "mean_token_accuracy": 0.8737844824790955, "num_tokens": 838198958.0, "step": 22072 }, { "epoch": 2.80791247932833, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.573409080505371, "learning_rate": 5e-05, "loss": 1.018, "mean_token_accuracy": 0.8773101568222046, "num_tokens": 838233936.0, "step": 22073 }, { "epoch": 2.80803968960692, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.75743579864502, "learning_rate": 5e-05, "loss": 1.0052, "mean_token_accuracy": 0.8747754693031311, "num_tokens": 838274333.0, "step": 22074 }, { "epoch": 2.808166899885511, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.06462287902832, "learning_rate": 5e-05, "loss": 0.9863, "mean_token_accuracy": 0.884433925151825, "num_tokens": 838304932.0, "step": 22075 }, { "epoch": 2.808294110164101, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.910648822784424, "learning_rate": 5e-05, "loss": 0.9013, "mean_token_accuracy": 0.8686777949333191, "num_tokens": 838345377.0, "step": 22076 }, { "epoch": 2.808421320442692, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.324101448059082, "learning_rate": 5e-05, "loss": 1.0121, "mean_token_accuracy": 0.8783169984817505, "num_tokens": 838383105.0, "step": 22077 }, { "epoch": 2.808548530721282, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.225735187530518, "learning_rate": 5e-05, "loss": 0.8461, "mean_token_accuracy": 0.8681520223617554, "num_tokens": 838421857.0, "step": 22078 }, { "epoch": 2.808675740999873, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.00903034210205, "learning_rate": 5e-05, "loss": 1.0716, "mean_token_accuracy": 0.8667528033256531, "num_tokens": 838459439.0, "step": 22079 }, { "epoch": 2.808802951278463, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.413553714752197, "learning_rate": 5e-05, "loss": 0.8978, "mean_token_accuracy": 0.876802921295166, "num_tokens": 838501176.0, "step": 22080 }, { "epoch": 2.8089301615570537, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 10.24285888671875, "learning_rate": 5e-05, "loss": 1.1018, "mean_token_accuracy": 0.8657269477844238, "num_tokens": 838538725.0, "step": 22081 }, { "epoch": 2.809057371835644, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 8.190406799316406, "learning_rate": 5e-05, "loss": 1.0045, "mean_token_accuracy": 0.8724219799041748, "num_tokens": 838577365.0, "step": 22082 }, { "epoch": 2.8091845821142347, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.716318130493164, "learning_rate": 5e-05, "loss": 1.0067, "mean_token_accuracy": 0.8865519165992737, "num_tokens": 838617872.0, "step": 22083 }, { "epoch": 2.8093117923928252, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.47885513305664, "learning_rate": 5e-05, "loss": 1.006, "mean_token_accuracy": 0.8750559687614441, "num_tokens": 838656599.0, "step": 22084 }, { "epoch": 2.8094390026714158, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.939032554626465, "learning_rate": 5e-05, "loss": 0.9825, "mean_token_accuracy": 0.8764880895614624, "num_tokens": 838693270.0, "step": 22085 }, { "epoch": 2.8095662129500063, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.325024604797363, "learning_rate": 5e-05, "loss": 0.9467, "mean_token_accuracy": 0.8716718554496765, "num_tokens": 838728493.0, "step": 22086 }, { "epoch": 2.809693423228597, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.460260391235352, "learning_rate": 5e-05, "loss": 0.9149, "mean_token_accuracy": 0.8811858296394348, "num_tokens": 838767136.0, "step": 22087 }, { "epoch": 2.8098206335071874, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.8297014236450195, "learning_rate": 5e-05, "loss": 0.8451, "mean_token_accuracy": 0.8780502676963806, "num_tokens": 838801894.0, "step": 22088 }, { "epoch": 2.809947843785778, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.66695499420166, "learning_rate": 5e-05, "loss": 0.8678, "mean_token_accuracy": 0.8998150825500488, "num_tokens": 838838732.0, "step": 22089 }, { "epoch": 2.8100750540643684, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.177671432495117, "learning_rate": 5e-05, "loss": 0.7065, "mean_token_accuracy": 0.9033851623535156, "num_tokens": 838876331.0, "step": 22090 }, { "epoch": 2.810202264342959, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.367262840270996, "learning_rate": 5e-05, "loss": 0.9734, "mean_token_accuracy": 0.8819079399108887, "num_tokens": 838909613.0, "step": 22091 }, { "epoch": 2.8103294746215495, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.054315567016602, "learning_rate": 5e-05, "loss": 0.8002, "mean_token_accuracy": 0.8709996938705444, "num_tokens": 838944694.0, "step": 22092 }, { "epoch": 2.81045668490014, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 9.81529712677002, "learning_rate": 5e-05, "loss": 1.0309, "mean_token_accuracy": 0.8833007216453552, "num_tokens": 838984595.0, "step": 22093 }, { "epoch": 2.8105838951787305, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.810868263244629, "learning_rate": 5e-05, "loss": 0.9069, "mean_token_accuracy": 0.8755606412887573, "num_tokens": 839018528.0, "step": 22094 }, { "epoch": 2.810711105457321, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.3015775680542, "learning_rate": 5e-05, "loss": 1.0032, "mean_token_accuracy": 0.8781782388687134, "num_tokens": 839056447.0, "step": 22095 }, { "epoch": 2.8108383157359116, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.526093482971191, "learning_rate": 5e-05, "loss": 0.9648, "mean_token_accuracy": 0.8861331939697266, "num_tokens": 839093612.0, "step": 22096 }, { "epoch": 2.810965526014502, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.015450477600098, "learning_rate": 5e-05, "loss": 0.8723, "mean_token_accuracy": 0.8776108026504517, "num_tokens": 839130192.0, "step": 22097 }, { "epoch": 2.8110927362930926, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.776564598083496, "learning_rate": 5e-05, "loss": 0.9519, "mean_token_accuracy": 0.8750523924827576, "num_tokens": 839167821.0, "step": 22098 }, { "epoch": 2.8112199465716827, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.1395087242126465, "learning_rate": 5e-05, "loss": 0.8037, "mean_token_accuracy": 0.8706264495849609, "num_tokens": 839200124.0, "step": 22099 }, { "epoch": 2.8113471568502737, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.851692199707031, "learning_rate": 5e-05, "loss": 0.9211, "mean_token_accuracy": 0.8761090636253357, "num_tokens": 839242787.0, "step": 22100 }, { "epoch": 2.811474367128864, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.051100254058838, "learning_rate": 5e-05, "loss": 0.7152, "mean_token_accuracy": 0.8948349952697754, "num_tokens": 839277517.0, "step": 22101 }, { "epoch": 2.8116015774074548, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.687896728515625, "learning_rate": 5e-05, "loss": 0.855, "mean_token_accuracy": 0.8923388719558716, "num_tokens": 839318609.0, "step": 22102 }, { "epoch": 2.811728787686045, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.243324279785156, "learning_rate": 5e-05, "loss": 0.7938, "mean_token_accuracy": 0.8755021095275879, "num_tokens": 839353964.0, "step": 22103 }, { "epoch": 2.811855997964636, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.432474136352539, "learning_rate": 5e-05, "loss": 0.8292, "mean_token_accuracy": 0.8950273990631104, "num_tokens": 839393926.0, "step": 22104 }, { "epoch": 2.811983208243226, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.35264253616333, "learning_rate": 5e-05, "loss": 0.784, "mean_token_accuracy": 0.8785423636436462, "num_tokens": 839437857.0, "step": 22105 }, { "epoch": 2.8121104185218164, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.242035865783691, "learning_rate": 5e-05, "loss": 0.8834, "mean_token_accuracy": 0.8682068586349487, "num_tokens": 839478444.0, "step": 22106 }, { "epoch": 2.812237628800407, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.422245502471924, "learning_rate": 5e-05, "loss": 0.7515, "mean_token_accuracy": 0.8826011419296265, "num_tokens": 839512752.0, "step": 22107 }, { "epoch": 2.8123648390789975, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.10025691986084, "learning_rate": 5e-05, "loss": 0.8715, "mean_token_accuracy": 0.8658827543258667, "num_tokens": 839550024.0, "step": 22108 }, { "epoch": 2.812492049357588, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.348946571350098, "learning_rate": 5e-05, "loss": 0.7189, "mean_token_accuracy": 0.8867363929748535, "num_tokens": 839582612.0, "step": 22109 }, { "epoch": 2.8126192596361785, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.2269868850708, "learning_rate": 5e-05, "loss": 0.8173, "mean_token_accuracy": 0.8803770542144775, "num_tokens": 839621791.0, "step": 22110 }, { "epoch": 2.812746469914769, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.062246799468994, "learning_rate": 5e-05, "loss": 0.7684, "mean_token_accuracy": 0.8655077219009399, "num_tokens": 839660829.0, "step": 22111 }, { "epoch": 2.8128736801933596, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.564375877380371, "learning_rate": 5e-05, "loss": 0.8644, "mean_token_accuracy": 0.8801565170288086, "num_tokens": 839701675.0, "step": 22112 }, { "epoch": 2.81300089047195, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.780060768127441, "learning_rate": 5e-05, "loss": 0.7409, "mean_token_accuracy": 0.8702632188796997, "num_tokens": 839742958.0, "step": 22113 }, { "epoch": 2.8131281007505406, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.90578556060791, "learning_rate": 5e-05, "loss": 0.9067, "mean_token_accuracy": 0.8799830675125122, "num_tokens": 839787552.0, "step": 22114 }, { "epoch": 2.813255311029131, "ewc_loss": 0.375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.72238826751709, "learning_rate": 5e-05, "loss": 0.7133, "mean_token_accuracy": 0.8779910802841187, "num_tokens": 839824630.0, "step": 22115 }, { "epoch": 2.8133825213077217, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.060070991516113, "learning_rate": 5e-05, "loss": 0.91, "mean_token_accuracy": 0.8904244303703308, "num_tokens": 839857671.0, "step": 22116 }, { "epoch": 2.8135097315863122, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.014011859893799, "learning_rate": 5e-05, "loss": 0.7765, "mean_token_accuracy": 0.8764340877532959, "num_tokens": 839896395.0, "step": 22117 }, { "epoch": 2.8136369418649028, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.740378379821777, "learning_rate": 5e-05, "loss": 0.9243, "mean_token_accuracy": 0.8784170150756836, "num_tokens": 839942188.0, "step": 22118 }, { "epoch": 2.8137641521434933, "ewc_loss": 0.5, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.580111980438232, "learning_rate": 5e-05, "loss": 0.7864, "mean_token_accuracy": 0.8949042558670044, "num_tokens": 839970510.0, "step": 22119 }, { "epoch": 2.813891362422084, "ewc_loss": 0.5, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.96682071685791, "learning_rate": 5e-05, "loss": 0.8355, "mean_token_accuracy": 0.8798652291297913, "num_tokens": 840007458.0, "step": 22120 }, { "epoch": 2.8140185727006743, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.044478416442871, "learning_rate": 5e-05, "loss": 0.8683, "mean_token_accuracy": 0.8766272068023682, "num_tokens": 840049729.0, "step": 22121 }, { "epoch": 2.814145782979265, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.373532772064209, "learning_rate": 5e-05, "loss": 0.7524, "mean_token_accuracy": 0.8812692761421204, "num_tokens": 840089883.0, "step": 22122 }, { "epoch": 2.8142729932578554, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.226731300354004, "learning_rate": 5e-05, "loss": 0.7974, "mean_token_accuracy": 0.8967851400375366, "num_tokens": 840121237.0, "step": 22123 }, { "epoch": 2.8144002035364455, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.2061333656311035, "learning_rate": 5e-05, "loss": 0.7588, "mean_token_accuracy": 0.8703306317329407, "num_tokens": 840157213.0, "step": 22124 }, { "epoch": 2.8145274138150365, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.135347366333008, "learning_rate": 5e-05, "loss": 0.7982, "mean_token_accuracy": 0.8832697868347168, "num_tokens": 840191404.0, "step": 22125 }, { "epoch": 2.8146546240936265, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.294663429260254, "learning_rate": 5e-05, "loss": 0.7714, "mean_token_accuracy": 0.8728991746902466, "num_tokens": 840227803.0, "step": 22126 }, { "epoch": 2.8147818343722175, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.963513374328613, "learning_rate": 5e-05, "loss": 0.8601, "mean_token_accuracy": 0.8584212064743042, "num_tokens": 840266686.0, "step": 22127 }, { "epoch": 2.8149090446508076, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.409975528717041, "learning_rate": 5e-05, "loss": 0.7178, "mean_token_accuracy": 0.8854324817657471, "num_tokens": 840310437.0, "step": 22128 }, { "epoch": 2.815036254929398, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.819985866546631, "learning_rate": 5e-05, "loss": 0.775, "mean_token_accuracy": 0.8766475915908813, "num_tokens": 840349599.0, "step": 22129 }, { "epoch": 2.8151634652079887, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.485598087310791, "learning_rate": 5e-05, "loss": 0.7303, "mean_token_accuracy": 0.8832119703292847, "num_tokens": 840389143.0, "step": 22130 }, { "epoch": 2.815290675486579, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.708901882171631, "learning_rate": 5e-05, "loss": 0.7354, "mean_token_accuracy": 0.8862860202789307, "num_tokens": 840429440.0, "step": 22131 }, { "epoch": 2.8154178857651697, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.561995506286621, "learning_rate": 5e-05, "loss": 0.7437, "mean_token_accuracy": 0.8786606788635254, "num_tokens": 840466056.0, "step": 22132 }, { "epoch": 2.8155450960437602, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.584950923919678, "learning_rate": 5e-05, "loss": 0.7395, "mean_token_accuracy": 0.8792607188224792, "num_tokens": 840503039.0, "step": 22133 }, { "epoch": 2.8156723063223508, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.641805648803711, "learning_rate": 5e-05, "loss": 0.724, "mean_token_accuracy": 0.8845645189285278, "num_tokens": 840546809.0, "step": 22134 }, { "epoch": 2.8157995166009413, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.489803314208984, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8762118220329285, "num_tokens": 840582617.0, "step": 22135 }, { "epoch": 2.815926726879532, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.701659202575684, "learning_rate": 5e-05, "loss": 0.7356, "mean_token_accuracy": 0.881381630897522, "num_tokens": 840620279.0, "step": 22136 }, { "epoch": 2.8160539371581224, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.426371097564697, "learning_rate": 5e-05, "loss": 0.711, "mean_token_accuracy": 0.8816812038421631, "num_tokens": 840655925.0, "step": 22137 }, { "epoch": 2.816181147436713, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.7259521484375, "learning_rate": 5e-05, "loss": 0.7295, "mean_token_accuracy": 0.8815663456916809, "num_tokens": 840692328.0, "step": 22138 }, { "epoch": 2.8163083577153034, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.382444381713867, "learning_rate": 5e-05, "loss": 0.7058, "mean_token_accuracy": 0.8801770210266113, "num_tokens": 840730674.0, "step": 22139 }, { "epoch": 2.816435567993894, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.772641658782959, "learning_rate": 5e-05, "loss": 0.7578, "mean_token_accuracy": 0.8725848197937012, "num_tokens": 840764791.0, "step": 22140 }, { "epoch": 2.8165627782724845, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.293522357940674, "learning_rate": 5e-05, "loss": 0.6949, "mean_token_accuracy": 0.8803941607475281, "num_tokens": 840803527.0, "step": 22141 }, { "epoch": 2.816689988551075, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.873304843902588, "learning_rate": 5e-05, "loss": 0.7675, "mean_token_accuracy": 0.8722116947174072, "num_tokens": 840843823.0, "step": 22142 }, { "epoch": 2.8168171988296655, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 7.127626895904541, "learning_rate": 5e-05, "loss": 0.7189, "mean_token_accuracy": 0.8685378432273865, "num_tokens": 840879648.0, "step": 22143 }, { "epoch": 2.816944409108256, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 8.102310180664062, "learning_rate": 5e-05, "loss": 0.7134, "mean_token_accuracy": 0.8938083648681641, "num_tokens": 840914979.0, "step": 22144 }, { "epoch": 2.8170716193868466, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.831892967224121, "learning_rate": 5e-05, "loss": 0.6702, "mean_token_accuracy": 0.8821764588356018, "num_tokens": 840951534.0, "step": 22145 }, { "epoch": 2.817198829665437, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.587516784667969, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.8724732398986816, "num_tokens": 840987416.0, "step": 22146 }, { "epoch": 2.8173260399440276, "ewc_loss": 0.3046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002269744873046875, "grad_norm": 6.244166851043701, "learning_rate": 5e-05, "loss": 0.6981, "mean_token_accuracy": 0.8622362017631531, "num_tokens": 841026734.0, "step": 22147 }, { "epoch": 2.817453250222618, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 9.544304847717285, "learning_rate": 5e-05, "loss": 0.8532, "mean_token_accuracy": 0.8927949666976929, "num_tokens": 841065333.0, "step": 22148 }, { "epoch": 2.8175804605012083, "ewc_loss": 0.29296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002155303955078125, "grad_norm": 5.4894700050354, "learning_rate": 5e-05, "loss": 0.6051, "mean_token_accuracy": 0.8872333765029907, "num_tokens": 841102257.0, "step": 22149 }, { "epoch": 2.8177076707797992, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 11.096275329589844, "learning_rate": 5e-05, "loss": 1.1042, "mean_token_accuracy": 0.8822184801101685, "num_tokens": 841140253.0, "step": 22150 }, { "epoch": 2.8178348810583893, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.16823673248291, "learning_rate": 5e-05, "loss": 0.7884, "mean_token_accuracy": 0.8819257616996765, "num_tokens": 841181675.0, "step": 22151 }, { "epoch": 2.8179620913369803, "ewc_loss": 1.0234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000946044921875, "grad_norm": 12.35357666015625, "learning_rate": 5e-05, "loss": 1.3501, "mean_token_accuracy": 0.8854082822799683, "num_tokens": 841219922.0, "step": 22152 }, { "epoch": 2.8180893016155704, "ewc_loss": 1.0, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000926971435546875, "grad_norm": 9.50771427154541, "learning_rate": 5e-05, "loss": 1.3244, "mean_token_accuracy": 0.8863795399665833, "num_tokens": 841256296.0, "step": 22153 }, { "epoch": 2.818216511894161, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.797813415527344, "learning_rate": 5e-05, "loss": 1.0215, "mean_token_accuracy": 0.8869096040725708, "num_tokens": 841296737.0, "step": 22154 }, { "epoch": 2.8183437221727514, "ewc_loss": 0.99609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000919342041015625, "grad_norm": 9.897551536560059, "learning_rate": 5e-05, "loss": 1.3886, "mean_token_accuracy": 0.8667069673538208, "num_tokens": 841329863.0, "step": 22155 }, { "epoch": 2.818470932451342, "ewc_loss": 0.640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.300809860229492, "learning_rate": 5e-05, "loss": 0.9939, "mean_token_accuracy": 0.8740350008010864, "num_tokens": 841365057.0, "step": 22156 }, { "epoch": 2.8185981427299325, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 10.126769065856934, "learning_rate": 5e-05, "loss": 1.1191, "mean_token_accuracy": 0.8827121257781982, "num_tokens": 841404885.0, "step": 22157 }, { "epoch": 2.818725353008523, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 8.330196380615234, "learning_rate": 5e-05, "loss": 0.9829, "mean_token_accuracy": 0.8783046007156372, "num_tokens": 841445067.0, "step": 22158 }, { "epoch": 2.8188525632871135, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.442439079284668, "learning_rate": 5e-05, "loss": 1.0302, "mean_token_accuracy": 0.8936986923217773, "num_tokens": 841480569.0, "step": 22159 }, { "epoch": 2.818979773565704, "ewc_loss": 0.6875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 8.475125312805176, "learning_rate": 5e-05, "loss": 1.0044, "mean_token_accuracy": 0.8860324621200562, "num_tokens": 841520540.0, "step": 22160 }, { "epoch": 2.8191069838442946, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.819397926330566, "learning_rate": 5e-05, "loss": 0.9921, "mean_token_accuracy": 0.887772262096405, "num_tokens": 841557765.0, "step": 22161 }, { "epoch": 2.819234194122885, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 8.646501541137695, "learning_rate": 5e-05, "loss": 1.0466, "mean_token_accuracy": 0.8769211173057556, "num_tokens": 841593345.0, "step": 22162 }, { "epoch": 2.8193614044014756, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.656859397888184, "learning_rate": 5e-05, "loss": 0.9303, "mean_token_accuracy": 0.871728777885437, "num_tokens": 841629977.0, "step": 22163 }, { "epoch": 2.819488614680066, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.986266136169434, "learning_rate": 5e-05, "loss": 1.0004, "mean_token_accuracy": 0.8806041479110718, "num_tokens": 841671901.0, "step": 22164 }, { "epoch": 2.8196158249586567, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.375274181365967, "learning_rate": 5e-05, "loss": 0.7915, "mean_token_accuracy": 0.8806777000427246, "num_tokens": 841706072.0, "step": 22165 }, { "epoch": 2.8197430352372472, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.087616920471191, "learning_rate": 5e-05, "loss": 1.0547, "mean_token_accuracy": 0.8709461688995361, "num_tokens": 841745707.0, "step": 22166 }, { "epoch": 2.8198702455158378, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 6.9750542640686035, "learning_rate": 5e-05, "loss": 0.8136, "mean_token_accuracy": 0.8822839260101318, "num_tokens": 841781468.0, "step": 22167 }, { "epoch": 2.8199974557944283, "ewc_loss": 0.8203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000743865966796875, "grad_norm": 11.27262020111084, "learning_rate": 5e-05, "loss": 1.135, "mean_token_accuracy": 0.8882008790969849, "num_tokens": 841821890.0, "step": 22168 }, { "epoch": 2.820124666073019, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 8.494543075561523, "learning_rate": 5e-05, "loss": 1.094, "mean_token_accuracy": 0.8689900636672974, "num_tokens": 841857668.0, "step": 22169 }, { "epoch": 2.8202518763516093, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.88923454284668, "learning_rate": 5e-05, "loss": 1.0577, "mean_token_accuracy": 0.8783023357391357, "num_tokens": 841892595.0, "step": 22170 }, { "epoch": 2.8203790866302, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000732421875, "grad_norm": 9.146053314208984, "learning_rate": 5e-05, "loss": 1.1123, "mean_token_accuracy": 0.8944768905639648, "num_tokens": 841928077.0, "step": 22171 }, { "epoch": 2.82050629690879, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 10.075136184692383, "learning_rate": 5e-05, "loss": 1.0015, "mean_token_accuracy": 0.880999743938446, "num_tokens": 841967847.0, "step": 22172 }, { "epoch": 2.820633507187381, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 8.625785827636719, "learning_rate": 5e-05, "loss": 1.0626, "mean_token_accuracy": 0.869243860244751, "num_tokens": 842008931.0, "step": 22173 }, { "epoch": 2.820760717465971, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.70358943939209, "learning_rate": 5e-05, "loss": 1.0087, "mean_token_accuracy": 0.8773536682128906, "num_tokens": 842042870.0, "step": 22174 }, { "epoch": 2.820887927744562, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 7.77731990814209, "learning_rate": 5e-05, "loss": 0.868, "mean_token_accuracy": 0.8800999522209167, "num_tokens": 842080193.0, "step": 22175 }, { "epoch": 2.821015138023152, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.068337440490723, "learning_rate": 5e-05, "loss": 0.9771, "mean_token_accuracy": 0.8941226005554199, "num_tokens": 842120953.0, "step": 22176 }, { "epoch": 2.821142348301743, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.355958461761475, "learning_rate": 5e-05, "loss": 0.8025, "mean_token_accuracy": 0.8910618424415588, "num_tokens": 842158466.0, "step": 22177 }, { "epoch": 2.821269558580333, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 10.865849494934082, "learning_rate": 5e-05, "loss": 1.1061, "mean_token_accuracy": 0.8776679039001465, "num_tokens": 842194164.0, "step": 22178 }, { "epoch": 2.8213967688589237, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.409183502197266, "learning_rate": 5e-05, "loss": 0.9907, "mean_token_accuracy": 0.8774494528770447, "num_tokens": 842232137.0, "step": 22179 }, { "epoch": 2.821523979137514, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.431001663208008, "learning_rate": 5e-05, "loss": 1.0878, "mean_token_accuracy": 0.8698903322219849, "num_tokens": 842263770.0, "step": 22180 }, { "epoch": 2.8216511894161047, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 9.297135353088379, "learning_rate": 5e-05, "loss": 1.0706, "mean_token_accuracy": 0.8736792802810669, "num_tokens": 842309420.0, "step": 22181 }, { "epoch": 2.8217783996946952, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.950328826904297, "learning_rate": 5e-05, "loss": 0.9571, "mean_token_accuracy": 0.8672977685928345, "num_tokens": 842350533.0, "step": 22182 }, { "epoch": 2.8219056099732858, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.193323135375977, "learning_rate": 5e-05, "loss": 1.0205, "mean_token_accuracy": 0.8694151043891907, "num_tokens": 842389734.0, "step": 22183 }, { "epoch": 2.8220328202518763, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.950889587402344, "learning_rate": 5e-05, "loss": 0.8472, "mean_token_accuracy": 0.8710395097732544, "num_tokens": 842424425.0, "step": 22184 }, { "epoch": 2.822160030530467, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.862737655639648, "learning_rate": 5e-05, "loss": 0.9115, "mean_token_accuracy": 0.8794010877609253, "num_tokens": 842460269.0, "step": 22185 }, { "epoch": 2.8222872408090574, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.7739410400390625, "learning_rate": 5e-05, "loss": 0.8121, "mean_token_accuracy": 0.8765115737915039, "num_tokens": 842497018.0, "step": 22186 }, { "epoch": 2.822414451087648, "ewc_loss": 0.5, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.307260513305664, "learning_rate": 5e-05, "loss": 0.8315, "mean_token_accuracy": 0.8825376033782959, "num_tokens": 842537563.0, "step": 22187 }, { "epoch": 2.8225416613662384, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.794130802154541, "learning_rate": 5e-05, "loss": 0.7324, "mean_token_accuracy": 0.8994338512420654, "num_tokens": 842573842.0, "step": 22188 }, { "epoch": 2.822668871644829, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.995116710662842, "learning_rate": 5e-05, "loss": 0.8196, "mean_token_accuracy": 0.8694552779197693, "num_tokens": 842614216.0, "step": 22189 }, { "epoch": 2.8227960819234195, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.638914108276367, "learning_rate": 5e-05, "loss": 0.83, "mean_token_accuracy": 0.861518919467926, "num_tokens": 842651823.0, "step": 22190 }, { "epoch": 2.82292329220201, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.021330833435059, "learning_rate": 5e-05, "loss": 0.8362, "mean_token_accuracy": 0.8626568913459778, "num_tokens": 842695113.0, "step": 22191 }, { "epoch": 2.8230505024806005, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.361268520355225, "learning_rate": 5e-05, "loss": 0.6845, "mean_token_accuracy": 0.8959404230117798, "num_tokens": 842731341.0, "step": 22192 }, { "epoch": 2.823177712759191, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.214059829711914, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.8710180521011353, "num_tokens": 842769611.0, "step": 22193 }, { "epoch": 2.8233049230377816, "ewc_loss": 0.375, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.122532367706299, "learning_rate": 5e-05, "loss": 0.7435, "mean_token_accuracy": 0.8709052801132202, "num_tokens": 842807106.0, "step": 22194 }, { "epoch": 2.823432133316372, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.407722473144531, "learning_rate": 5e-05, "loss": 0.8022, "mean_token_accuracy": 0.8838111162185669, "num_tokens": 842847957.0, "step": 22195 }, { "epoch": 2.8235593435949626, "ewc_loss": 0.359375, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.875720024108887, "learning_rate": 5e-05, "loss": 0.7374, "mean_token_accuracy": 0.867826521396637, "num_tokens": 842889222.0, "step": 22196 }, { "epoch": 2.8236865538735527, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.822107315063477, "learning_rate": 5e-05, "loss": 0.8395, "mean_token_accuracy": 0.8845459222793579, "num_tokens": 842925164.0, "step": 22197 }, { "epoch": 2.8238137641521437, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.5769572257995605, "learning_rate": 5e-05, "loss": 0.7356, "mean_token_accuracy": 0.8656209707260132, "num_tokens": 842963832.0, "step": 22198 }, { "epoch": 2.8239409744307338, "ewc_loss": 0.59375, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.336990356445312, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.8756605386734009, "num_tokens": 843004357.0, "step": 22199 }, { "epoch": 2.8240681847093247, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.657775402069092, "learning_rate": 5e-05, "loss": 0.7783, "mean_token_accuracy": 0.8603481650352478, "num_tokens": 843045241.0, "step": 22200 }, { "epoch": 2.824195394987915, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.411051750183105, "learning_rate": 5e-05, "loss": 0.9902, "mean_token_accuracy": 0.8714233636856079, "num_tokens": 843084473.0, "step": 22201 }, { "epoch": 2.824322605266506, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.357610702514648, "learning_rate": 5e-05, "loss": 0.83, "mean_token_accuracy": 0.8777415752410889, "num_tokens": 843122118.0, "step": 22202 }, { "epoch": 2.824449815545096, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.780600547790527, "learning_rate": 5e-05, "loss": 0.9565, "mean_token_accuracy": 0.8668798804283142, "num_tokens": 843159936.0, "step": 22203 }, { "epoch": 2.8245770258236864, "ewc_loss": 0.546875, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.037129402160645, "learning_rate": 5e-05, "loss": 0.9028, "mean_token_accuracy": 0.8753591775894165, "num_tokens": 843196212.0, "step": 22204 }, { "epoch": 2.824704236102277, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.806327819824219, "learning_rate": 5e-05, "loss": 0.8064, "mean_token_accuracy": 0.8831295371055603, "num_tokens": 843236870.0, "step": 22205 }, { "epoch": 2.8248314463808675, "ewc_loss": 0.53125, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.370092391967773, "learning_rate": 5e-05, "loss": 0.8568, "mean_token_accuracy": 0.883991003036499, "num_tokens": 843270701.0, "step": 22206 }, { "epoch": 2.824958656659458, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.0007781982421875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.25027322769165, "learning_rate": 5e-05, "loss": 0.7121, "mean_token_accuracy": 0.8848050832748413, "num_tokens": 843308547.0, "step": 22207 }, { "epoch": 2.8250858669380485, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.428572654724121, "learning_rate": 5e-05, "loss": 0.8154, "mean_token_accuracy": 0.8929287195205688, "num_tokens": 843354496.0, "step": 22208 }, { "epoch": 2.825213077216639, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.211700916290283, "learning_rate": 5e-05, "loss": 0.7456, "mean_token_accuracy": 0.8726457953453064, "num_tokens": 843389995.0, "step": 22209 }, { "epoch": 2.8253402874952296, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.276942253112793, "learning_rate": 5e-05, "loss": 0.8557, "mean_token_accuracy": 0.8768162727355957, "num_tokens": 843428444.0, "step": 22210 }, { "epoch": 2.82546749777382, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.248626708984375, "learning_rate": 5e-05, "loss": 0.723, "mean_token_accuracy": 0.8817869424819946, "num_tokens": 843466390.0, "step": 22211 }, { "epoch": 2.8255947080524106, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.27518081665039, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.8640159368515015, "num_tokens": 843503424.0, "step": 22212 }, { "epoch": 2.825721918331001, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.118198871612549, "learning_rate": 5e-05, "loss": 0.7508, "mean_token_accuracy": 0.8687527179718018, "num_tokens": 843544669.0, "step": 22213 }, { "epoch": 2.8258491286095917, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.41502857208252, "learning_rate": 5e-05, "loss": 0.7882, "mean_token_accuracy": 0.8948907852172852, "num_tokens": 843579777.0, "step": 22214 }, { "epoch": 2.8259763388881822, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.962771892547607, "learning_rate": 5e-05, "loss": 0.7029, "mean_token_accuracy": 0.8791185617446899, "num_tokens": 843613481.0, "step": 22215 }, { "epoch": 2.8261035491667728, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.570405960083008, "learning_rate": 5e-05, "loss": 0.8089, "mean_token_accuracy": 0.8932760953903198, "num_tokens": 843656870.0, "step": 22216 }, { "epoch": 2.8262307594453633, "ewc_loss": 0.375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.863441467285156, "learning_rate": 5e-05, "loss": 0.7003, "mean_token_accuracy": 0.8813215494155884, "num_tokens": 843693859.0, "step": 22217 }, { "epoch": 2.826357969723954, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.743224143981934, "learning_rate": 5e-05, "loss": 0.832, "mean_token_accuracy": 0.8953859806060791, "num_tokens": 843726999.0, "step": 22218 }, { "epoch": 2.8264851800025443, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.868350505828857, "learning_rate": 5e-05, "loss": 0.7513, "mean_token_accuracy": 0.8726531267166138, "num_tokens": 843767244.0, "step": 22219 }, { "epoch": 2.826612390281135, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.754779815673828, "learning_rate": 5e-05, "loss": 0.8878, "mean_token_accuracy": 0.878003716468811, "num_tokens": 843799828.0, "step": 22220 }, { "epoch": 2.8267396005597254, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.102534770965576, "learning_rate": 5e-05, "loss": 0.7546, "mean_token_accuracy": 0.8776684403419495, "num_tokens": 843837485.0, "step": 22221 }, { "epoch": 2.8268668108383155, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.426024436950684, "learning_rate": 5e-05, "loss": 0.8449, "mean_token_accuracy": 0.8847805261611938, "num_tokens": 843878796.0, "step": 22222 }, { "epoch": 2.8269940211169065, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.495437145233154, "learning_rate": 5e-05, "loss": 0.8175, "mean_token_accuracy": 0.871849775314331, "num_tokens": 843920200.0, "step": 22223 }, { "epoch": 2.8271212313954965, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.920839309692383, "learning_rate": 5e-05, "loss": 0.7835, "mean_token_accuracy": 0.8877071738243103, "num_tokens": 843961719.0, "step": 22224 }, { "epoch": 2.8272484416740875, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.781052589416504, "learning_rate": 5e-05, "loss": 0.8386, "mean_token_accuracy": 0.8705036640167236, "num_tokens": 844005085.0, "step": 22225 }, { "epoch": 2.8273756519526776, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.572910785675049, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8703950643539429, "num_tokens": 844036964.0, "step": 22226 }, { "epoch": 2.827502862231268, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.859089374542236, "learning_rate": 5e-05, "loss": 0.8264, "mean_token_accuracy": 0.8744922876358032, "num_tokens": 844078297.0, "step": 22227 }, { "epoch": 2.8276300725098586, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.478448390960693, "learning_rate": 5e-05, "loss": 0.7028, "mean_token_accuracy": 0.890117883682251, "num_tokens": 844113749.0, "step": 22228 }, { "epoch": 2.827757282788449, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.776279926300049, "learning_rate": 5e-05, "loss": 0.7864, "mean_token_accuracy": 0.8712455034255981, "num_tokens": 844155993.0, "step": 22229 }, { "epoch": 2.8278844930670397, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.574459075927734, "learning_rate": 5e-05, "loss": 0.7409, "mean_token_accuracy": 0.8788821697235107, "num_tokens": 844204302.0, "step": 22230 }, { "epoch": 2.8280117033456302, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.5856146812438965, "learning_rate": 5e-05, "loss": 0.7229, "mean_token_accuracy": 0.8835209608078003, "num_tokens": 844241490.0, "step": 22231 }, { "epoch": 2.8281389136242208, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.731942176818848, "learning_rate": 5e-05, "loss": 0.7336, "mean_token_accuracy": 0.8831373453140259, "num_tokens": 844282891.0, "step": 22232 }, { "epoch": 2.8282661239028113, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.377931594848633, "learning_rate": 5e-05, "loss": 0.733, "mean_token_accuracy": 0.8762789368629456, "num_tokens": 844324407.0, "step": 22233 }, { "epoch": 2.828393334181402, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.927197456359863, "learning_rate": 5e-05, "loss": 0.7502, "mean_token_accuracy": 0.8814505338668823, "num_tokens": 844356656.0, "step": 22234 }, { "epoch": 2.8285205444599923, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.14884090423584, "learning_rate": 5e-05, "loss": 0.7259, "mean_token_accuracy": 0.8722376823425293, "num_tokens": 844389789.0, "step": 22235 }, { "epoch": 2.828647754738583, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.164029121398926, "learning_rate": 5e-05, "loss": 0.7756, "mean_token_accuracy": 0.8797800540924072, "num_tokens": 844423599.0, "step": 22236 }, { "epoch": 2.8287749650171734, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.870218276977539, "learning_rate": 5e-05, "loss": 0.6838, "mean_token_accuracy": 0.8823579549789429, "num_tokens": 844460117.0, "step": 22237 }, { "epoch": 2.828902175295764, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.523762702941895, "learning_rate": 5e-05, "loss": 0.8296, "mean_token_accuracy": 0.8678434491157532, "num_tokens": 844501263.0, "step": 22238 }, { "epoch": 2.8290293855743545, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.5196733474731445, "learning_rate": 5e-05, "loss": 0.6433, "mean_token_accuracy": 0.8815091848373413, "num_tokens": 844541573.0, "step": 22239 }, { "epoch": 2.829156595852945, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 9.065046310424805, "learning_rate": 5e-05, "loss": 0.8615, "mean_token_accuracy": 0.8786830902099609, "num_tokens": 844583728.0, "step": 22240 }, { "epoch": 2.8292838061315355, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.190578460693359, "learning_rate": 5e-05, "loss": 0.6765, "mean_token_accuracy": 0.8821749687194824, "num_tokens": 844616611.0, "step": 22241 }, { "epoch": 2.829411016410126, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.614604949951172, "learning_rate": 5e-05, "loss": 1.0064, "mean_token_accuracy": 0.8625749349594116, "num_tokens": 844652067.0, "step": 22242 }, { "epoch": 2.8295382266887166, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.406123638153076, "learning_rate": 5e-05, "loss": 0.7351, "mean_token_accuracy": 0.8725642561912537, "num_tokens": 844688582.0, "step": 22243 }, { "epoch": 2.829665436967307, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.527563095092773, "learning_rate": 5e-05, "loss": 0.9735, "mean_token_accuracy": 0.8798166513442993, "num_tokens": 844727120.0, "step": 22244 }, { "epoch": 2.8297926472458976, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 7.521635055541992, "learning_rate": 5e-05, "loss": 0.8823, "mean_token_accuracy": 0.8803088068962097, "num_tokens": 844772328.0, "step": 22245 }, { "epoch": 2.829919857524488, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.176453590393066, "learning_rate": 5e-05, "loss": 0.8221, "mean_token_accuracy": 0.8935445547103882, "num_tokens": 844812528.0, "step": 22246 }, { "epoch": 2.8300470678030782, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.758830070495605, "learning_rate": 5e-05, "loss": 0.979, "mean_token_accuracy": 0.8727974891662598, "num_tokens": 844857208.0, "step": 22247 }, { "epoch": 2.830174278081669, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.533234119415283, "learning_rate": 5e-05, "loss": 0.7299, "mean_token_accuracy": 0.8705332279205322, "num_tokens": 844890362.0, "step": 22248 }, { "epoch": 2.8303014883602593, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 10.084281921386719, "learning_rate": 5e-05, "loss": 1.0466, "mean_token_accuracy": 0.8665019273757935, "num_tokens": 844928664.0, "step": 22249 }, { "epoch": 2.8304286986388503, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.242508411407471, "learning_rate": 5e-05, "loss": 0.6822, "mean_token_accuracy": 0.8971315026283264, "num_tokens": 844957043.0, "step": 22250 }, { "epoch": 2.8305559089174404, "ewc_loss": 0.796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 10.84361457824707, "learning_rate": 5e-05, "loss": 1.1545, "mean_token_accuracy": 0.8736847639083862, "num_tokens": 844991808.0, "step": 22251 }, { "epoch": 2.830683119196031, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 8.03070068359375, "learning_rate": 5e-05, "loss": 1.0171, "mean_token_accuracy": 0.8795048594474792, "num_tokens": 845025571.0, "step": 22252 }, { "epoch": 2.8308103294746214, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 9.820219993591309, "learning_rate": 5e-05, "loss": 1.0486, "mean_token_accuracy": 0.8697066307067871, "num_tokens": 845063154.0, "step": 22253 }, { "epoch": 2.830937539753212, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 9.110206604003906, "learning_rate": 5e-05, "loss": 1.0867, "mean_token_accuracy": 0.8870748281478882, "num_tokens": 845100256.0, "step": 22254 }, { "epoch": 2.8310647500318025, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.13990592956543, "learning_rate": 5e-05, "loss": 0.8722, "mean_token_accuracy": 0.8792752027511597, "num_tokens": 845138111.0, "step": 22255 }, { "epoch": 2.831191960310393, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.48786449432373, "learning_rate": 5e-05, "loss": 0.9924, "mean_token_accuracy": 0.8838826417922974, "num_tokens": 845175759.0, "step": 22256 }, { "epoch": 2.8313191705889835, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.707977294921875, "learning_rate": 5e-05, "loss": 0.8766, "mean_token_accuracy": 0.8781632781028748, "num_tokens": 845212232.0, "step": 22257 }, { "epoch": 2.831446380867574, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.013647079467773, "learning_rate": 5e-05, "loss": 0.8969, "mean_token_accuracy": 0.8959527611732483, "num_tokens": 845247141.0, "step": 22258 }, { "epoch": 2.8315735911461646, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.147515296936035, "learning_rate": 5e-05, "loss": 0.9242, "mean_token_accuracy": 0.8829163312911987, "num_tokens": 845287908.0, "step": 22259 }, { "epoch": 2.831700801424755, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.857823371887207, "learning_rate": 5e-05, "loss": 0.8022, "mean_token_accuracy": 0.8842885494232178, "num_tokens": 845328134.0, "step": 22260 }, { "epoch": 2.8318280117033456, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.681638717651367, "learning_rate": 5e-05, "loss": 0.9141, "mean_token_accuracy": 0.8860474824905396, "num_tokens": 845366096.0, "step": 22261 }, { "epoch": 2.831955221981936, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.184717655181885, "learning_rate": 5e-05, "loss": 0.7496, "mean_token_accuracy": 0.8796942234039307, "num_tokens": 845397737.0, "step": 22262 }, { "epoch": 2.8320824322605267, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.994776725769043, "learning_rate": 5e-05, "loss": 0.9299, "mean_token_accuracy": 0.881894588470459, "num_tokens": 845437015.0, "step": 22263 }, { "epoch": 2.8322096425391172, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.079728126525879, "learning_rate": 5e-05, "loss": 0.7549, "mean_token_accuracy": 0.8889445662498474, "num_tokens": 845473778.0, "step": 22264 }, { "epoch": 2.8323368528177078, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.059185981750488, "learning_rate": 5e-05, "loss": 0.892, "mean_token_accuracy": 0.8907618522644043, "num_tokens": 845509254.0, "step": 22265 }, { "epoch": 2.8324640630962983, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.217906951904297, "learning_rate": 5e-05, "loss": 0.8504, "mean_token_accuracy": 0.8733012080192566, "num_tokens": 845547682.0, "step": 22266 }, { "epoch": 2.832591273374889, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.878755569458008, "learning_rate": 5e-05, "loss": 0.8685, "mean_token_accuracy": 0.8900859355926514, "num_tokens": 845585315.0, "step": 22267 }, { "epoch": 2.8327184836534793, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.341926097869873, "learning_rate": 5e-05, "loss": 0.8266, "mean_token_accuracy": 0.8849986791610718, "num_tokens": 845625065.0, "step": 22268 }, { "epoch": 2.83284569393207, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.584213256835938, "learning_rate": 5e-05, "loss": 0.8607, "mean_token_accuracy": 0.8886485695838928, "num_tokens": 845664148.0, "step": 22269 }, { "epoch": 2.83297290421066, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.382366180419922, "learning_rate": 5e-05, "loss": 0.7952, "mean_token_accuracy": 0.8830164670944214, "num_tokens": 845695986.0, "step": 22270 }, { "epoch": 2.833100114489251, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.435693740844727, "learning_rate": 5e-05, "loss": 0.8696, "mean_token_accuracy": 0.8796266913414001, "num_tokens": 845732807.0, "step": 22271 }, { "epoch": 2.833227324767841, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.209089279174805, "learning_rate": 5e-05, "loss": 0.7633, "mean_token_accuracy": 0.8861856460571289, "num_tokens": 845763238.0, "step": 22272 }, { "epoch": 2.833354535046432, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.573709487915039, "learning_rate": 5e-05, "loss": 0.8357, "mean_token_accuracy": 0.892156183719635, "num_tokens": 845802218.0, "step": 22273 }, { "epoch": 2.833481745325022, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.986624717712402, "learning_rate": 5e-05, "loss": 0.8077, "mean_token_accuracy": 0.8597273826599121, "num_tokens": 845845056.0, "step": 22274 }, { "epoch": 2.833608955603613, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.77175521850586, "learning_rate": 5e-05, "loss": 0.9095, "mean_token_accuracy": 0.8762772679328918, "num_tokens": 845884909.0, "step": 22275 }, { "epoch": 2.833736165882203, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.911247730255127, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8782089352607727, "num_tokens": 845923894.0, "step": 22276 }, { "epoch": 2.8338633761607936, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.914770126342773, "learning_rate": 5e-05, "loss": 0.9823, "mean_token_accuracy": 0.8604334592819214, "num_tokens": 845962452.0, "step": 22277 }, { "epoch": 2.833990586439384, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.010917663574219, "learning_rate": 5e-05, "loss": 0.7723, "mean_token_accuracy": 0.879605770111084, "num_tokens": 845998031.0, "step": 22278 }, { "epoch": 2.8341177967179747, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.873519897460938, "learning_rate": 5e-05, "loss": 0.9444, "mean_token_accuracy": 0.8766698837280273, "num_tokens": 846036920.0, "step": 22279 }, { "epoch": 2.8342450069965652, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.319468975067139, "learning_rate": 5e-05, "loss": 0.8794, "mean_token_accuracy": 0.8661829829216003, "num_tokens": 846075978.0, "step": 22280 }, { "epoch": 2.8343722172751558, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.465943336486816, "learning_rate": 5e-05, "loss": 0.9064, "mean_token_accuracy": 0.8774577379226685, "num_tokens": 846111893.0, "step": 22281 }, { "epoch": 2.8344994275537463, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 7.714851379394531, "learning_rate": 5e-05, "loss": 0.8738, "mean_token_accuracy": 0.8784790635108948, "num_tokens": 846153417.0, "step": 22282 }, { "epoch": 2.834626637832337, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 7.838929653167725, "learning_rate": 5e-05, "loss": 0.8546, "mean_token_accuracy": 0.8739286661148071, "num_tokens": 846194375.0, "step": 22283 }, { "epoch": 2.8347538481109273, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.979045867919922, "learning_rate": 5e-05, "loss": 0.8265, "mean_token_accuracy": 0.8859298825263977, "num_tokens": 846231040.0, "step": 22284 }, { "epoch": 2.834881058389518, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.4483537673950195, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.8893091678619385, "num_tokens": 846270645.0, "step": 22285 }, { "epoch": 2.8350082686681084, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.06089973449707, "learning_rate": 5e-05, "loss": 0.7917, "mean_token_accuracy": 0.8853485584259033, "num_tokens": 846304038.0, "step": 22286 }, { "epoch": 2.835135478946699, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.314640998840332, "learning_rate": 5e-05, "loss": 0.7846, "mean_token_accuracy": 0.8709288239479065, "num_tokens": 846338136.0, "step": 22287 }, { "epoch": 2.8352626892252895, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.048959732055664, "learning_rate": 5e-05, "loss": 0.7985, "mean_token_accuracy": 0.8791240453720093, "num_tokens": 846372654.0, "step": 22288 }, { "epoch": 2.83538989950388, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.293361186981201, "learning_rate": 5e-05, "loss": 0.7414, "mean_token_accuracy": 0.8764437437057495, "num_tokens": 846409180.0, "step": 22289 }, { "epoch": 2.8355171097824705, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.014837265014648, "learning_rate": 5e-05, "loss": 0.8257, "mean_token_accuracy": 0.8663031458854675, "num_tokens": 846453410.0, "step": 22290 }, { "epoch": 2.835644320061061, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.2548675537109375, "learning_rate": 5e-05, "loss": 0.7053, "mean_token_accuracy": 0.8899649381637573, "num_tokens": 846488271.0, "step": 22291 }, { "epoch": 2.8357715303396516, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.020977973937988, "learning_rate": 5e-05, "loss": 0.7563, "mean_token_accuracy": 0.8869131207466125, "num_tokens": 846528730.0, "step": 22292 }, { "epoch": 2.835898740618242, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.215719223022461, "learning_rate": 5e-05, "loss": 0.6792, "mean_token_accuracy": 0.892215371131897, "num_tokens": 846567276.0, "step": 22293 }, { "epoch": 2.8360259508968326, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.056658744812012, "learning_rate": 5e-05, "loss": 0.767, "mean_token_accuracy": 0.8866496682167053, "num_tokens": 846608482.0, "step": 22294 }, { "epoch": 2.8361531611754227, "ewc_loss": 0.375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.147732734680176, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.8737005591392517, "num_tokens": 846648582.0, "step": 22295 }, { "epoch": 2.8362803714540137, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 8.14130973815918, "learning_rate": 5e-05, "loss": 0.7661, "mean_token_accuracy": 0.8878202438354492, "num_tokens": 846687392.0, "step": 22296 }, { "epoch": 2.8364075817326038, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 7.019828796386719, "learning_rate": 5e-05, "loss": 0.697, "mean_token_accuracy": 0.8836512565612793, "num_tokens": 846721542.0, "step": 22297 }, { "epoch": 2.8365347920111947, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.323864936828613, "learning_rate": 5e-05, "loss": 0.8444, "mean_token_accuracy": 0.869154691696167, "num_tokens": 846759262.0, "step": 22298 }, { "epoch": 2.836662002289785, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.830517292022705, "learning_rate": 5e-05, "loss": 0.7238, "mean_token_accuracy": 0.8736463785171509, "num_tokens": 846796979.0, "step": 22299 }, { "epoch": 2.836789212568376, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.610552787780762, "learning_rate": 5e-05, "loss": 0.8493, "mean_token_accuracy": 0.8768168687820435, "num_tokens": 846832417.0, "step": 22300 }, { "epoch": 2.836916422846966, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.61525821685791, "learning_rate": 5e-05, "loss": 0.7277, "mean_token_accuracy": 0.8737475275993347, "num_tokens": 846870289.0, "step": 22301 }, { "epoch": 2.8370436331255564, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.916299819946289, "learning_rate": 5e-05, "loss": 0.878, "mean_token_accuracy": 0.8803932666778564, "num_tokens": 846909821.0, "step": 22302 }, { "epoch": 2.837170843404147, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.556355953216553, "learning_rate": 5e-05, "loss": 0.7095, "mean_token_accuracy": 0.8808813095092773, "num_tokens": 846945512.0, "step": 22303 }, { "epoch": 2.8372980536827375, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.037779808044434, "learning_rate": 5e-05, "loss": 0.9147, "mean_token_accuracy": 0.8841818571090698, "num_tokens": 846980834.0, "step": 22304 }, { "epoch": 2.837425263961328, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 6.870123863220215, "learning_rate": 5e-05, "loss": 0.7518, "mean_token_accuracy": 0.8798806667327881, "num_tokens": 847025320.0, "step": 22305 }, { "epoch": 2.8375524742399185, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.610424041748047, "learning_rate": 5e-05, "loss": 0.8707, "mean_token_accuracy": 0.8893991708755493, "num_tokens": 847063637.0, "step": 22306 }, { "epoch": 2.837679684518509, "ewc_loss": 0.5, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.563745975494385, "learning_rate": 5e-05, "loss": 0.8109, "mean_token_accuracy": 0.8870284557342529, "num_tokens": 847102314.0, "step": 22307 }, { "epoch": 2.8378068947970996, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.664693832397461, "learning_rate": 5e-05, "loss": 0.7812, "mean_token_accuracy": 0.882074773311615, "num_tokens": 847141956.0, "step": 22308 }, { "epoch": 2.83793410507569, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.307517051696777, "learning_rate": 5e-05, "loss": 0.8912, "mean_token_accuracy": 0.870464563369751, "num_tokens": 847178855.0, "step": 22309 }, { "epoch": 2.8380613153542806, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.798309326171875, "learning_rate": 5e-05, "loss": 0.7187, "mean_token_accuracy": 0.8797394037246704, "num_tokens": 847214228.0, "step": 22310 }, { "epoch": 2.838188525632871, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.918623924255371, "learning_rate": 5e-05, "loss": 0.9005, "mean_token_accuracy": 0.8769491910934448, "num_tokens": 847255942.0, "step": 22311 }, { "epoch": 2.8383157359114617, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.411410331726074, "learning_rate": 5e-05, "loss": 0.7488, "mean_token_accuracy": 0.8647971153259277, "num_tokens": 847294742.0, "step": 22312 }, { "epoch": 2.838442946190052, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.304756164550781, "learning_rate": 5e-05, "loss": 0.9626, "mean_token_accuracy": 0.8757531642913818, "num_tokens": 847330105.0, "step": 22313 }, { "epoch": 2.8385701564686427, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.5485615730285645, "learning_rate": 5e-05, "loss": 0.7484, "mean_token_accuracy": 0.871957004070282, "num_tokens": 847373291.0, "step": 22314 }, { "epoch": 2.8386973667472333, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.196578979492188, "learning_rate": 5e-05, "loss": 0.9257, "mean_token_accuracy": 0.8904917240142822, "num_tokens": 847409059.0, "step": 22315 }, { "epoch": 2.838824577025824, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.289241790771484, "learning_rate": 5e-05, "loss": 0.8535, "mean_token_accuracy": 0.8839588165283203, "num_tokens": 847445869.0, "step": 22316 }, { "epoch": 2.8389517873044143, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.386075973510742, "learning_rate": 5e-05, "loss": 0.8935, "mean_token_accuracy": 0.8837791085243225, "num_tokens": 847486816.0, "step": 22317 }, { "epoch": 2.839078997583005, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.013007164001465, "learning_rate": 5e-05, "loss": 0.8906, "mean_token_accuracy": 0.8844101428985596, "num_tokens": 847527482.0, "step": 22318 }, { "epoch": 2.8392062078615954, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.451498985290527, "learning_rate": 5e-05, "loss": 0.8367, "mean_token_accuracy": 0.8671975135803223, "num_tokens": 847570850.0, "step": 22319 }, { "epoch": 2.8393334181401855, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.409339904785156, "learning_rate": 5e-05, "loss": 0.9, "mean_token_accuracy": 0.8759810924530029, "num_tokens": 847608872.0, "step": 22320 }, { "epoch": 2.8394606284187764, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.156476974487305, "learning_rate": 5e-05, "loss": 0.7784, "mean_token_accuracy": 0.873184323310852, "num_tokens": 847647813.0, "step": 22321 }, { "epoch": 2.8395878386973665, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.219714164733887, "learning_rate": 5e-05, "loss": 0.8347, "mean_token_accuracy": 0.8827075362205505, "num_tokens": 847685726.0, "step": 22322 }, { "epoch": 2.8397150489759575, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.432368278503418, "learning_rate": 5e-05, "loss": 0.7954, "mean_token_accuracy": 0.8733195066452026, "num_tokens": 847723029.0, "step": 22323 }, { "epoch": 2.8398422592545476, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.720318794250488, "learning_rate": 5e-05, "loss": 0.7698, "mean_token_accuracy": 0.8843743801116943, "num_tokens": 847761922.0, "step": 22324 }, { "epoch": 2.839969469533138, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.791493892669678, "learning_rate": 5e-05, "loss": 0.8081, "mean_token_accuracy": 0.8754300475120544, "num_tokens": 847806229.0, "step": 22325 }, { "epoch": 2.8400966798117286, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.33468770980835, "learning_rate": 5e-05, "loss": 0.7503, "mean_token_accuracy": 0.8778826594352722, "num_tokens": 847844448.0, "step": 22326 }, { "epoch": 2.840223890090319, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.042997360229492, "learning_rate": 5e-05, "loss": 0.8336, "mean_token_accuracy": 0.8698318004608154, "num_tokens": 847889387.0, "step": 22327 }, { "epoch": 2.8403511003689097, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.080450534820557, "learning_rate": 5e-05, "loss": 0.7238, "mean_token_accuracy": 0.876849353313446, "num_tokens": 847926993.0, "step": 22328 }, { "epoch": 2.8404783106475002, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.289164543151855, "learning_rate": 5e-05, "loss": 0.7901, "mean_token_accuracy": 0.8858031034469604, "num_tokens": 847958884.0, "step": 22329 }, { "epoch": 2.8406055209260908, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.824515342712402, "learning_rate": 5e-05, "loss": 0.7268, "mean_token_accuracy": 0.8763478398323059, "num_tokens": 847997894.0, "step": 22330 }, { "epoch": 2.8407327312046813, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.621298789978027, "learning_rate": 5e-05, "loss": 0.8544, "mean_token_accuracy": 0.8786681890487671, "num_tokens": 848037392.0, "step": 22331 }, { "epoch": 2.840859941483272, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.562796115875244, "learning_rate": 5e-05, "loss": 0.6962, "mean_token_accuracy": 0.8803454041481018, "num_tokens": 848072844.0, "step": 22332 }, { "epoch": 2.8409871517618623, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 9.045114517211914, "learning_rate": 5e-05, "loss": 0.9349, "mean_token_accuracy": 0.8687463998794556, "num_tokens": 848107477.0, "step": 22333 }, { "epoch": 2.841114362040453, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.381843566894531, "learning_rate": 5e-05, "loss": 0.7094, "mean_token_accuracy": 0.8793448209762573, "num_tokens": 848143124.0, "step": 22334 }, { "epoch": 2.8412415723190434, "ewc_loss": 0.625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.43620491027832, "learning_rate": 5e-05, "loss": 0.9681, "mean_token_accuracy": 0.8806461095809937, "num_tokens": 848181454.0, "step": 22335 }, { "epoch": 2.841368782597634, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.620819568634033, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.8866536617279053, "num_tokens": 848219300.0, "step": 22336 }, { "epoch": 2.8414959928762245, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.394418716430664, "learning_rate": 5e-05, "loss": 0.9841, "mean_token_accuracy": 0.8824201822280884, "num_tokens": 848256650.0, "step": 22337 }, { "epoch": 2.841623203154815, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 7.361462593078613, "learning_rate": 5e-05, "loss": 0.895, "mean_token_accuracy": 0.8715298175811768, "num_tokens": 848294836.0, "step": 22338 }, { "epoch": 2.8417504134334055, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.682286262512207, "learning_rate": 5e-05, "loss": 0.9394, "mean_token_accuracy": 0.8802269697189331, "num_tokens": 848326306.0, "step": 22339 }, { "epoch": 2.841877623711996, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.013020515441895, "learning_rate": 5e-05, "loss": 0.9031, "mean_token_accuracy": 0.8914057016372681, "num_tokens": 848360886.0, "step": 22340 }, { "epoch": 2.8420048339905866, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.756486415863037, "learning_rate": 5e-05, "loss": 0.8541, "mean_token_accuracy": 0.8787387609481812, "num_tokens": 848399364.0, "step": 22341 }, { "epoch": 2.842132044269177, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.205595970153809, "learning_rate": 5e-05, "loss": 0.874, "mean_token_accuracy": 0.885612428188324, "num_tokens": 848436978.0, "step": 22342 }, { "epoch": 2.8422592545477676, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.504645347595215, "learning_rate": 5e-05, "loss": 0.8485, "mean_token_accuracy": 0.8678820133209229, "num_tokens": 848474661.0, "step": 22343 }, { "epoch": 2.842386464826358, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 7.8198137283325195, "learning_rate": 5e-05, "loss": 0.7955, "mean_token_accuracy": 0.8906457424163818, "num_tokens": 848514469.0, "step": 22344 }, { "epoch": 2.8425136751049482, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.879936695098877, "learning_rate": 5e-05, "loss": 0.8019, "mean_token_accuracy": 0.8826795816421509, "num_tokens": 848548606.0, "step": 22345 }, { "epoch": 2.842640885383539, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.233035564422607, "learning_rate": 5e-05, "loss": 0.7675, "mean_token_accuracy": 0.8711668252944946, "num_tokens": 848585309.0, "step": 22346 }, { "epoch": 2.8427680956621293, "ewc_loss": 0.5, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.31523323059082, "learning_rate": 5e-05, "loss": 0.8381, "mean_token_accuracy": 0.8817825317382812, "num_tokens": 848625062.0, "step": 22347 }, { "epoch": 2.8428953059407203, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.890595436096191, "learning_rate": 5e-05, "loss": 0.7239, "mean_token_accuracy": 0.8773698806762695, "num_tokens": 848661914.0, "step": 22348 }, { "epoch": 2.8430225162193103, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.539351463317871, "learning_rate": 5e-05, "loss": 0.9414, "mean_token_accuracy": 0.8532874584197998, "num_tokens": 848704343.0, "step": 22349 }, { "epoch": 2.843149726497901, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.811336517333984, "learning_rate": 5e-05, "loss": 0.6756, "mean_token_accuracy": 0.8958276510238647, "num_tokens": 848737415.0, "step": 22350 }, { "epoch": 2.8432769367764914, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.631196022033691, "learning_rate": 5e-05, "loss": 0.8762, "mean_token_accuracy": 0.8764393329620361, "num_tokens": 848770795.0, "step": 22351 }, { "epoch": 2.843404147055082, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 6.864274501800537, "learning_rate": 5e-05, "loss": 0.7124, "mean_token_accuracy": 0.8854658603668213, "num_tokens": 848815086.0, "step": 22352 }, { "epoch": 2.8435313573336725, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.600163459777832, "learning_rate": 5e-05, "loss": 0.8423, "mean_token_accuracy": 0.8950073719024658, "num_tokens": 848853545.0, "step": 22353 }, { "epoch": 2.843658567612263, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.087451934814453, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8856298923492432, "num_tokens": 848886047.0, "step": 22354 }, { "epoch": 2.8437857778908535, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.320219039916992, "learning_rate": 5e-05, "loss": 0.8492, "mean_token_accuracy": 0.8850625157356262, "num_tokens": 848929147.0, "step": 22355 }, { "epoch": 2.843912988169444, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.3953022956848145, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.8765648603439331, "num_tokens": 848962912.0, "step": 22356 }, { "epoch": 2.8440401984480346, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.9051995277404785, "learning_rate": 5e-05, "loss": 0.8126, "mean_token_accuracy": 0.8782058954238892, "num_tokens": 848997954.0, "step": 22357 }, { "epoch": 2.844167408726625, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.645229816436768, "learning_rate": 5e-05, "loss": 0.8052, "mean_token_accuracy": 0.8750444054603577, "num_tokens": 849044418.0, "step": 22358 }, { "epoch": 2.8442946190052156, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.579222679138184, "learning_rate": 5e-05, "loss": 0.7757, "mean_token_accuracy": 0.8778988122940063, "num_tokens": 849085591.0, "step": 22359 }, { "epoch": 2.844421829283806, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.804849624633789, "learning_rate": 5e-05, "loss": 0.8032, "mean_token_accuracy": 0.876082181930542, "num_tokens": 849128046.0, "step": 22360 }, { "epoch": 2.8445490395623967, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.336714744567871, "learning_rate": 5e-05, "loss": 0.7607, "mean_token_accuracy": 0.8725678324699402, "num_tokens": 849165231.0, "step": 22361 }, { "epoch": 2.844676249840987, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.944531440734863, "learning_rate": 5e-05, "loss": 0.7874, "mean_token_accuracy": 0.8792921304702759, "num_tokens": 849203414.0, "step": 22362 }, { "epoch": 2.8448034601195777, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.14402437210083, "learning_rate": 5e-05, "loss": 0.707, "mean_token_accuracy": 0.882872998714447, "num_tokens": 849242463.0, "step": 22363 }, { "epoch": 2.8449306703981683, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.112433433532715, "learning_rate": 5e-05, "loss": 0.7845, "mean_token_accuracy": 0.8833537101745605, "num_tokens": 849280098.0, "step": 22364 }, { "epoch": 2.845057880676759, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.921775817871094, "learning_rate": 5e-05, "loss": 0.6891, "mean_token_accuracy": 0.8833248615264893, "num_tokens": 849317460.0, "step": 22365 }, { "epoch": 2.8451850909553493, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.385229110717773, "learning_rate": 5e-05, "loss": 0.7947, "mean_token_accuracy": 0.8838106393814087, "num_tokens": 849354852.0, "step": 22366 }, { "epoch": 2.84531230123394, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.626777648925781, "learning_rate": 5e-05, "loss": 0.7021, "mean_token_accuracy": 0.8784689903259277, "num_tokens": 849397132.0, "step": 22367 }, { "epoch": 2.84543951151253, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.831050872802734, "learning_rate": 5e-05, "loss": 0.8487, "mean_token_accuracy": 0.8814556002616882, "num_tokens": 849441751.0, "step": 22368 }, { "epoch": 2.845566721791121, "ewc_loss": 0.349609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.327266693115234, "learning_rate": 5e-05, "loss": 0.7164, "mean_token_accuracy": 0.8699578642845154, "num_tokens": 849480177.0, "step": 22369 }, { "epoch": 2.845693932069711, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.298873901367188, "learning_rate": 5e-05, "loss": 0.9684, "mean_token_accuracy": 0.8726158142089844, "num_tokens": 849520211.0, "step": 22370 }, { "epoch": 2.845821142348302, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.427001476287842, "learning_rate": 5e-05, "loss": 0.7209, "mean_token_accuracy": 0.8797434568405151, "num_tokens": 849551591.0, "step": 22371 }, { "epoch": 2.845948352626892, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.333745956420898, "learning_rate": 5e-05, "loss": 0.9909, "mean_token_accuracy": 0.8723925948143005, "num_tokens": 849592700.0, "step": 22372 }, { "epoch": 2.846075562905483, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.1951375007629395, "learning_rate": 5e-05, "loss": 0.8869, "mean_token_accuracy": 0.869991660118103, "num_tokens": 849631220.0, "step": 22373 }, { "epoch": 2.846202773184073, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.453680038452148, "learning_rate": 5e-05, "loss": 0.8957, "mean_token_accuracy": 0.883668839931488, "num_tokens": 849670349.0, "step": 22374 }, { "epoch": 2.8463299834626636, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.1843843460083, "learning_rate": 5e-05, "loss": 1.0106, "mean_token_accuracy": 0.8542736768722534, "num_tokens": 849711000.0, "step": 22375 }, { "epoch": 2.846457193741254, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.085788726806641, "learning_rate": 5e-05, "loss": 0.7874, "mean_token_accuracy": 0.8670400381088257, "num_tokens": 849751199.0, "step": 22376 }, { "epoch": 2.8465844040198447, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.033217430114746, "learning_rate": 5e-05, "loss": 0.8845, "mean_token_accuracy": 0.9028513431549072, "num_tokens": 849785416.0, "step": 22377 }, { "epoch": 2.8467116142984352, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.409677982330322, "learning_rate": 5e-05, "loss": 0.7414, "mean_token_accuracy": 0.8707647919654846, "num_tokens": 849818553.0, "step": 22378 }, { "epoch": 2.8468388245770258, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.489927291870117, "learning_rate": 5e-05, "loss": 0.9526, "mean_token_accuracy": 0.8861202597618103, "num_tokens": 849854994.0, "step": 22379 }, { "epoch": 2.8469660348556163, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.742905616760254, "learning_rate": 5e-05, "loss": 0.8323, "mean_token_accuracy": 0.8611288070678711, "num_tokens": 849896603.0, "step": 22380 }, { "epoch": 2.847093245134207, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.101675987243652, "learning_rate": 5e-05, "loss": 0.9966, "mean_token_accuracy": 0.868313193321228, "num_tokens": 849935999.0, "step": 22381 }, { "epoch": 2.8472204554127973, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.648202896118164, "learning_rate": 5e-05, "loss": 0.8991, "mean_token_accuracy": 0.8832998275756836, "num_tokens": 849976858.0, "step": 22382 }, { "epoch": 2.847347665691388, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.038331031799316, "learning_rate": 5e-05, "loss": 0.8358, "mean_token_accuracy": 0.8926382064819336, "num_tokens": 850012802.0, "step": 22383 }, { "epoch": 2.8474748759699784, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.278532028198242, "learning_rate": 5e-05, "loss": 0.9521, "mean_token_accuracy": 0.871979296207428, "num_tokens": 850047330.0, "step": 22384 }, { "epoch": 2.847602086248569, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.303864002227783, "learning_rate": 5e-05, "loss": 0.8034, "mean_token_accuracy": 0.8754775524139404, "num_tokens": 850087543.0, "step": 22385 }, { "epoch": 2.8477292965271594, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.378806114196777, "learning_rate": 5e-05, "loss": 0.8637, "mean_token_accuracy": 0.8835155963897705, "num_tokens": 850126864.0, "step": 22386 }, { "epoch": 2.84785650680575, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.306561470031738, "learning_rate": 5e-05, "loss": 0.7488, "mean_token_accuracy": 0.890863835811615, "num_tokens": 850162476.0, "step": 22387 }, { "epoch": 2.8479837170843405, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.998353004455566, "learning_rate": 5e-05, "loss": 0.8427, "mean_token_accuracy": 0.8764925003051758, "num_tokens": 850197606.0, "step": 22388 }, { "epoch": 2.848110927362931, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.620483875274658, "learning_rate": 5e-05, "loss": 0.8103, "mean_token_accuracy": 0.8785455226898193, "num_tokens": 850232223.0, "step": 22389 }, { "epoch": 2.8482381376415216, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.532736778259277, "learning_rate": 5e-05, "loss": 0.7456, "mean_token_accuracy": 0.8862065076828003, "num_tokens": 850268936.0, "step": 22390 }, { "epoch": 2.848365347920112, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.906604290008545, "learning_rate": 5e-05, "loss": 0.8346, "mean_token_accuracy": 0.8742906451225281, "num_tokens": 850301652.0, "step": 22391 }, { "epoch": 2.8484925581987026, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.2485127449035645, "learning_rate": 5e-05, "loss": 0.7132, "mean_token_accuracy": 0.8824458122253418, "num_tokens": 850335033.0, "step": 22392 }, { "epoch": 2.8486197684772927, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.091593742370605, "learning_rate": 5e-05, "loss": 0.8175, "mean_token_accuracy": 0.8765504360198975, "num_tokens": 850369677.0, "step": 22393 }, { "epoch": 2.8487469787558837, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.034422874450684, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.867043137550354, "num_tokens": 850411464.0, "step": 22394 }, { "epoch": 2.8488741890344738, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.333115577697754, "learning_rate": 5e-05, "loss": 0.8107, "mean_token_accuracy": 0.8825474381446838, "num_tokens": 850447578.0, "step": 22395 }, { "epoch": 2.8490013993130647, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.770317077636719, "learning_rate": 5e-05, "loss": 0.6799, "mean_token_accuracy": 0.8862001895904541, "num_tokens": 850489689.0, "step": 22396 }, { "epoch": 2.849128609591655, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.683669090270996, "learning_rate": 5e-05, "loss": 0.8274, "mean_token_accuracy": 0.8860230445861816, "num_tokens": 850528598.0, "step": 22397 }, { "epoch": 2.849255819870246, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.494968891143799, "learning_rate": 5e-05, "loss": 0.7004, "mean_token_accuracy": 0.881317138671875, "num_tokens": 850566953.0, "step": 22398 }, { "epoch": 2.849383030148836, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.138442039489746, "learning_rate": 5e-05, "loss": 0.8868, "mean_token_accuracy": 0.8881650567054749, "num_tokens": 850604286.0, "step": 22399 }, { "epoch": 2.8495102404274264, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.351457595825195, "learning_rate": 5e-05, "loss": 0.7351, "mean_token_accuracy": 0.8767479658126831, "num_tokens": 850640724.0, "step": 22400 }, { "epoch": 2.849637450706017, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.521641731262207, "learning_rate": 5e-05, "loss": 0.983, "mean_token_accuracy": 0.8785011768341064, "num_tokens": 850682377.0, "step": 22401 }, { "epoch": 2.8497646609846075, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 6.689957618713379, "learning_rate": 5e-05, "loss": 0.7849, "mean_token_accuracy": 0.8806076645851135, "num_tokens": 850721637.0, "step": 22402 }, { "epoch": 2.849891871263198, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.440396308898926, "learning_rate": 5e-05, "loss": 1.0126, "mean_token_accuracy": 0.8730827569961548, "num_tokens": 850754884.0, "step": 22403 }, { "epoch": 2.8500190815417885, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 7.515254020690918, "learning_rate": 5e-05, "loss": 0.9232, "mean_token_accuracy": 0.8740338087081909, "num_tokens": 850793706.0, "step": 22404 }, { "epoch": 2.850146291820379, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.613946914672852, "learning_rate": 5e-05, "loss": 0.9691, "mean_token_accuracy": 0.8673548698425293, "num_tokens": 850827708.0, "step": 22405 }, { "epoch": 2.8502735020989696, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.149954795837402, "learning_rate": 5e-05, "loss": 0.9646, "mean_token_accuracy": 0.8708100318908691, "num_tokens": 850870808.0, "step": 22406 }, { "epoch": 2.85040071237756, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.648210525512695, "learning_rate": 5e-05, "loss": 0.8274, "mean_token_accuracy": 0.8805650472640991, "num_tokens": 850906129.0, "step": 22407 }, { "epoch": 2.8505279226561506, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.307374954223633, "learning_rate": 5e-05, "loss": 0.9066, "mean_token_accuracy": 0.8795630931854248, "num_tokens": 850944021.0, "step": 22408 }, { "epoch": 2.850655132934741, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.404627323150635, "learning_rate": 5e-05, "loss": 0.7999, "mean_token_accuracy": 0.8831766843795776, "num_tokens": 850983040.0, "step": 22409 }, { "epoch": 2.8507823432133317, "ewc_loss": 0.5, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 7.921424865722656, "learning_rate": 5e-05, "loss": 0.8406, "mean_token_accuracy": 0.8784337639808655, "num_tokens": 851017400.0, "step": 22410 }, { "epoch": 2.850909553491922, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.735237121582031, "learning_rate": 5e-05, "loss": 0.8083, "mean_token_accuracy": 0.8802283406257629, "num_tokens": 851055559.0, "step": 22411 }, { "epoch": 2.8510367637705127, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.375703811645508, "learning_rate": 5e-05, "loss": 0.7404, "mean_token_accuracy": 0.8859859704971313, "num_tokens": 851088810.0, "step": 22412 }, { "epoch": 2.8511639740491033, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.080487251281738, "learning_rate": 5e-05, "loss": 0.8343, "mean_token_accuracy": 0.8791449069976807, "num_tokens": 851121531.0, "step": 22413 }, { "epoch": 2.851291184327694, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.065924167633057, "learning_rate": 5e-05, "loss": 0.7256, "mean_token_accuracy": 0.879504919052124, "num_tokens": 851160601.0, "step": 22414 }, { "epoch": 2.8514183946062843, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.24863052368164, "learning_rate": 5e-05, "loss": 0.8469, "mean_token_accuracy": 0.8693406581878662, "num_tokens": 851195887.0, "step": 22415 }, { "epoch": 2.851545604884875, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.954158306121826, "learning_rate": 5e-05, "loss": 0.6989, "mean_token_accuracy": 0.8844729661941528, "num_tokens": 851231761.0, "step": 22416 }, { "epoch": 2.8516728151634654, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.385546684265137, "learning_rate": 5e-05, "loss": 0.8813, "mean_token_accuracy": 0.8626344203948975, "num_tokens": 851270765.0, "step": 22417 }, { "epoch": 2.8518000254420555, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.845166206359863, "learning_rate": 5e-05, "loss": 0.7819, "mean_token_accuracy": 0.864183783531189, "num_tokens": 851312200.0, "step": 22418 }, { "epoch": 2.8519272357206464, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.542774200439453, "learning_rate": 5e-05, "loss": 0.8652, "mean_token_accuracy": 0.8745139241218567, "num_tokens": 851350854.0, "step": 22419 }, { "epoch": 2.8520544459992365, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.773110389709473, "learning_rate": 5e-05, "loss": 0.7551, "mean_token_accuracy": 0.8666917085647583, "num_tokens": 851389242.0, "step": 22420 }, { "epoch": 2.8521816562778275, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.679327964782715, "learning_rate": 5e-05, "loss": 0.8782, "mean_token_accuracy": 0.8779240250587463, "num_tokens": 851425450.0, "step": 22421 }, { "epoch": 2.8523088665564176, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 6.835702419281006, "learning_rate": 5e-05, "loss": 0.7453, "mean_token_accuracy": 0.8766740560531616, "num_tokens": 851459470.0, "step": 22422 }, { "epoch": 2.852436076835008, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.62012004852295, "learning_rate": 5e-05, "loss": 0.926, "mean_token_accuracy": 0.868772566318512, "num_tokens": 851500849.0, "step": 22423 }, { "epoch": 2.8525632871135986, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 7.0571699142456055, "learning_rate": 5e-05, "loss": 0.7815, "mean_token_accuracy": 0.8783650398254395, "num_tokens": 851540048.0, "step": 22424 }, { "epoch": 2.852690497392189, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.346131324768066, "learning_rate": 5e-05, "loss": 0.8888, "mean_token_accuracy": 0.8764931559562683, "num_tokens": 851573343.0, "step": 22425 }, { "epoch": 2.8528177076707797, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.383281230926514, "learning_rate": 5e-05, "loss": 0.8432, "mean_token_accuracy": 0.8655695915222168, "num_tokens": 851613658.0, "step": 22426 }, { "epoch": 2.85294491794937, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.890982151031494, "learning_rate": 5e-05, "loss": 0.7763, "mean_token_accuracy": 0.8889834880828857, "num_tokens": 851649961.0, "step": 22427 }, { "epoch": 2.8530721282279607, "ewc_loss": 0.46875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.701566696166992, "learning_rate": 5e-05, "loss": 0.8147, "mean_token_accuracy": 0.8760372400283813, "num_tokens": 851685465.0, "step": 22428 }, { "epoch": 2.8531993385065513, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.440618991851807, "learning_rate": 5e-05, "loss": 0.7194, "mean_token_accuracy": 0.8944412469863892, "num_tokens": 851718434.0, "step": 22429 }, { "epoch": 2.853326548785142, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.942322731018066, "learning_rate": 5e-05, "loss": 0.7909, "mean_token_accuracy": 0.8849166035652161, "num_tokens": 851756207.0, "step": 22430 }, { "epoch": 2.8534537590637323, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.167679786682129, "learning_rate": 5e-05, "loss": 0.7493, "mean_token_accuracy": 0.8748587965965271, "num_tokens": 851795398.0, "step": 22431 }, { "epoch": 2.853580969342323, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.122133255004883, "learning_rate": 5e-05, "loss": 0.8441, "mean_token_accuracy": 0.8679629564285278, "num_tokens": 851832569.0, "step": 22432 }, { "epoch": 2.8537081796209134, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.925209999084473, "learning_rate": 5e-05, "loss": 0.7077, "mean_token_accuracy": 0.8829808831214905, "num_tokens": 851870749.0, "step": 22433 }, { "epoch": 2.853835389899504, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.378124237060547, "learning_rate": 5e-05, "loss": 0.8166, "mean_token_accuracy": 0.8814748525619507, "num_tokens": 851912591.0, "step": 22434 }, { "epoch": 2.8539626001780944, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.697479724884033, "learning_rate": 5e-05, "loss": 0.7074, "mean_token_accuracy": 0.8785387873649597, "num_tokens": 851948792.0, "step": 22435 }, { "epoch": 2.854089810456685, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.7090482711792, "learning_rate": 5e-05, "loss": 0.8683, "mean_token_accuracy": 0.870978593826294, "num_tokens": 851988811.0, "step": 22436 }, { "epoch": 2.8542170207352755, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.497304439544678, "learning_rate": 5e-05, "loss": 0.6854, "mean_token_accuracy": 0.885292649269104, "num_tokens": 852024781.0, "step": 22437 }, { "epoch": 2.854344231013866, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.05717658996582, "learning_rate": 5e-05, "loss": 0.8949, "mean_token_accuracy": 0.8839508295059204, "num_tokens": 852056676.0, "step": 22438 }, { "epoch": 2.8544714412924566, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.547492980957031, "learning_rate": 5e-05, "loss": 0.7702, "mean_token_accuracy": 0.8684659600257874, "num_tokens": 852095500.0, "step": 22439 }, { "epoch": 2.854598651571047, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.075972557067871, "learning_rate": 5e-05, "loss": 0.9183, "mean_token_accuracy": 0.8899577856063843, "num_tokens": 852133658.0, "step": 22440 }, { "epoch": 2.8547258618496376, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.079822063446045, "learning_rate": 5e-05, "loss": 0.8581, "mean_token_accuracy": 0.8629114627838135, "num_tokens": 852173963.0, "step": 22441 }, { "epoch": 2.854853072128228, "ewc_loss": 0.5625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.457409858703613, "learning_rate": 5e-05, "loss": 0.9084, "mean_token_accuracy": 0.8771038055419922, "num_tokens": 852209957.0, "step": 22442 }, { "epoch": 2.8549802824068182, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.796432018280029, "learning_rate": 5e-05, "loss": 0.8725, "mean_token_accuracy": 0.879959225654602, "num_tokens": 852252281.0, "step": 22443 }, { "epoch": 2.855107492685409, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.500401020050049, "learning_rate": 5e-05, "loss": 0.7461, "mean_token_accuracy": 0.894731879234314, "num_tokens": 852285287.0, "step": 22444 }, { "epoch": 2.8552347029639993, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.293986320495605, "learning_rate": 5e-05, "loss": 0.9081, "mean_token_accuracy": 0.87109375, "num_tokens": 852326596.0, "step": 22445 }, { "epoch": 2.8553619132425903, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.022613048553467, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.8839163780212402, "num_tokens": 852358062.0, "step": 22446 }, { "epoch": 2.8554891235211803, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.310873985290527, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8701565265655518, "num_tokens": 852395828.0, "step": 22447 }, { "epoch": 2.855616333799771, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.1596198081970215, "learning_rate": 5e-05, "loss": 0.7766, "mean_token_accuracy": 0.8735741376876831, "num_tokens": 852439466.0, "step": 22448 }, { "epoch": 2.8557435440783614, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.872065544128418, "learning_rate": 5e-05, "loss": 0.7944, "mean_token_accuracy": 0.8806669116020203, "num_tokens": 852477992.0, "step": 22449 }, { "epoch": 2.855870754356952, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.592663288116455, "learning_rate": 5e-05, "loss": 0.7659, "mean_token_accuracy": 0.8842830657958984, "num_tokens": 852514814.0, "step": 22450 }, { "epoch": 2.8559979646355425, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.341822624206543, "learning_rate": 5e-05, "loss": 0.7287, "mean_token_accuracy": 0.8809316754341125, "num_tokens": 852553866.0, "step": 22451 }, { "epoch": 2.856125174914133, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.007031440734863, "learning_rate": 5e-05, "loss": 0.8305, "mean_token_accuracy": 0.8677630424499512, "num_tokens": 852594291.0, "step": 22452 }, { "epoch": 2.8562523851927235, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.938736915588379, "learning_rate": 5e-05, "loss": 0.6987, "mean_token_accuracy": 0.8832324147224426, "num_tokens": 852630586.0, "step": 22453 }, { "epoch": 2.856379595471314, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.377184867858887, "learning_rate": 5e-05, "loss": 0.8258, "mean_token_accuracy": 0.8763409852981567, "num_tokens": 852670013.0, "step": 22454 }, { "epoch": 2.8565068057499046, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.631496429443359, "learning_rate": 5e-05, "loss": 0.7002, "mean_token_accuracy": 0.8786124587059021, "num_tokens": 852704097.0, "step": 22455 }, { "epoch": 2.856634016028495, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.774365425109863, "learning_rate": 5e-05, "loss": 0.8696, "mean_token_accuracy": 0.873540997505188, "num_tokens": 852739771.0, "step": 22456 }, { "epoch": 2.8567612263070856, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.395026206970215, "learning_rate": 5e-05, "loss": 0.7344, "mean_token_accuracy": 0.8704146146774292, "num_tokens": 852778604.0, "step": 22457 }, { "epoch": 2.856888436585676, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.158246994018555, "learning_rate": 5e-05, "loss": 0.9982, "mean_token_accuracy": 0.8548150062561035, "num_tokens": 852814966.0, "step": 22458 }, { "epoch": 2.8570156468642667, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.429708957672119, "learning_rate": 5e-05, "loss": 0.743, "mean_token_accuracy": 0.8696498870849609, "num_tokens": 852849562.0, "step": 22459 }, { "epoch": 2.857142857142857, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.300649642944336, "learning_rate": 5e-05, "loss": 0.9458, "mean_token_accuracy": 0.8839781880378723, "num_tokens": 852884183.0, "step": 22460 }, { "epoch": 2.8572700674214477, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 6.946906566619873, "learning_rate": 5e-05, "loss": 0.863, "mean_token_accuracy": 0.8715009093284607, "num_tokens": 852925121.0, "step": 22461 }, { "epoch": 2.8573972777000383, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.836421012878418, "learning_rate": 5e-05, "loss": 0.9353, "mean_token_accuracy": 0.8845908045768738, "num_tokens": 852965962.0, "step": 22462 }, { "epoch": 2.857524487978629, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 7.651291370391846, "learning_rate": 5e-05, "loss": 0.9054, "mean_token_accuracy": 0.8770186901092529, "num_tokens": 853004886.0, "step": 22463 }, { "epoch": 2.8576516982572193, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 7.9773850440979, "learning_rate": 5e-05, "loss": 0.8777, "mean_token_accuracy": 0.8737456798553467, "num_tokens": 853038898.0, "step": 22464 }, { "epoch": 2.85777890853581, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.065387725830078, "learning_rate": 5e-05, "loss": 0.9315, "mean_token_accuracy": 0.8630764484405518, "num_tokens": 853079064.0, "step": 22465 }, { "epoch": 2.8579061188144, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.413137912750244, "learning_rate": 5e-05, "loss": 0.8523, "mean_token_accuracy": 0.86126309633255, "num_tokens": 853125704.0, "step": 22466 }, { "epoch": 2.858033329092991, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.000409126281738, "learning_rate": 5e-05, "loss": 0.8554, "mean_token_accuracy": 0.8730974793434143, "num_tokens": 853168760.0, "step": 22467 }, { "epoch": 2.858160539371581, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.536235332489014, "learning_rate": 5e-05, "loss": 0.7946, "mean_token_accuracy": 0.8770160675048828, "num_tokens": 853207146.0, "step": 22468 }, { "epoch": 2.858287749650172, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.494781017303467, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.8861809372901917, "num_tokens": 853248764.0, "step": 22469 }, { "epoch": 2.858414959928762, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.981879234313965, "learning_rate": 5e-05, "loss": 0.7885, "mean_token_accuracy": 0.8836785554885864, "num_tokens": 853283325.0, "step": 22470 }, { "epoch": 2.858542170207353, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.999676704406738, "learning_rate": 5e-05, "loss": 0.7388, "mean_token_accuracy": 0.8767600059509277, "num_tokens": 853316579.0, "step": 22471 }, { "epoch": 2.858669380485943, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.352338790893555, "learning_rate": 5e-05, "loss": 0.8423, "mean_token_accuracy": 0.8807176351547241, "num_tokens": 853353109.0, "step": 22472 }, { "epoch": 2.8587965907645336, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.750736236572266, "learning_rate": 5e-05, "loss": 0.716, "mean_token_accuracy": 0.8791574239730835, "num_tokens": 853394974.0, "step": 22473 }, { "epoch": 2.858923801043124, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.566268920898438, "learning_rate": 5e-05, "loss": 0.8049, "mean_token_accuracy": 0.8937103748321533, "num_tokens": 853433819.0, "step": 22474 }, { "epoch": 2.8590510113217147, "ewc_loss": 0.375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.6810078620910645, "learning_rate": 5e-05, "loss": 0.7167, "mean_token_accuracy": 0.8764280676841736, "num_tokens": 853471947.0, "step": 22475 }, { "epoch": 2.859178221600305, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.685836791992188, "learning_rate": 5e-05, "loss": 0.871, "mean_token_accuracy": 0.8784997463226318, "num_tokens": 853505936.0, "step": 22476 }, { "epoch": 2.8593054318788957, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.740849494934082, "learning_rate": 5e-05, "loss": 0.6793, "mean_token_accuracy": 0.8957067728042603, "num_tokens": 853540957.0, "step": 22477 }, { "epoch": 2.8594326421574863, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.67590045928955, "learning_rate": 5e-05, "loss": 0.8683, "mean_token_accuracy": 0.8900721073150635, "num_tokens": 853574117.0, "step": 22478 }, { "epoch": 2.859559852436077, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 6.992072582244873, "learning_rate": 5e-05, "loss": 0.7976, "mean_token_accuracy": 0.8717725872993469, "num_tokens": 853612424.0, "step": 22479 }, { "epoch": 2.8596870627146673, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.378623008728027, "learning_rate": 5e-05, "loss": 0.8927, "mean_token_accuracy": 0.8746123313903809, "num_tokens": 853652659.0, "step": 22480 }, { "epoch": 2.859814272993258, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.377642631530762, "learning_rate": 5e-05, "loss": 0.7957, "mean_token_accuracy": 0.8832280039787292, "num_tokens": 853690763.0, "step": 22481 }, { "epoch": 2.8599414832718484, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.855556488037109, "learning_rate": 5e-05, "loss": 0.7841, "mean_token_accuracy": 0.8894412517547607, "num_tokens": 853727437.0, "step": 22482 }, { "epoch": 2.860068693550439, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.741027355194092, "learning_rate": 5e-05, "loss": 0.8481, "mean_token_accuracy": 0.8705005049705505, "num_tokens": 853769179.0, "step": 22483 }, { "epoch": 2.8601959038290294, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.356167793273926, "learning_rate": 5e-05, "loss": 0.724, "mean_token_accuracy": 0.8909403085708618, "num_tokens": 853803374.0, "step": 22484 }, { "epoch": 2.86032311410762, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.02636432647705, "learning_rate": 5e-05, "loss": 0.8411, "mean_token_accuracy": 0.8709203004837036, "num_tokens": 853834286.0, "step": 22485 }, { "epoch": 2.8604503243862105, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.018266201019287, "learning_rate": 5e-05, "loss": 0.7107, "mean_token_accuracy": 0.882983386516571, "num_tokens": 853872369.0, "step": 22486 }, { "epoch": 2.860577534664801, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.26116943359375, "learning_rate": 5e-05, "loss": 0.8612, "mean_token_accuracy": 0.8688937425613403, "num_tokens": 853916470.0, "step": 22487 }, { "epoch": 2.8607047449433916, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.776018142700195, "learning_rate": 5e-05, "loss": 0.6559, "mean_token_accuracy": 0.8977636098861694, "num_tokens": 853956802.0, "step": 22488 }, { "epoch": 2.860831955221982, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.571039199829102, "learning_rate": 5e-05, "loss": 0.8391, "mean_token_accuracy": 0.8827573657035828, "num_tokens": 853990914.0, "step": 22489 }, { "epoch": 2.8609591655005726, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.507740020751953, "learning_rate": 5e-05, "loss": 0.6855, "mean_token_accuracy": 0.8840240836143494, "num_tokens": 854029707.0, "step": 22490 }, { "epoch": 2.8610863757791627, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.012077331542969, "learning_rate": 5e-05, "loss": 0.8457, "mean_token_accuracy": 0.9002028703689575, "num_tokens": 854065584.0, "step": 22491 }, { "epoch": 2.8612135860577537, "ewc_loss": 0.375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.388894081115723, "learning_rate": 5e-05, "loss": 0.7229, "mean_token_accuracy": 0.8774459362030029, "num_tokens": 854105937.0, "step": 22492 }, { "epoch": 2.8613407963363438, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.302760124206543, "learning_rate": 5e-05, "loss": 0.9454, "mean_token_accuracy": 0.8851640820503235, "num_tokens": 854146746.0, "step": 22493 }, { "epoch": 2.8614680066149347, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 6.831633567810059, "learning_rate": 5e-05, "loss": 0.7731, "mean_token_accuracy": 0.8901987671852112, "num_tokens": 854179953.0, "step": 22494 }, { "epoch": 2.861595216893525, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.923040390014648, "learning_rate": 5e-05, "loss": 0.9461, "mean_token_accuracy": 0.8817400932312012, "num_tokens": 854220765.0, "step": 22495 }, { "epoch": 2.861722427172116, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.689779281616211, "learning_rate": 5e-05, "loss": 0.9044, "mean_token_accuracy": 0.8827730417251587, "num_tokens": 854258544.0, "step": 22496 }, { "epoch": 2.861849637450706, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 7.804251670837402, "learning_rate": 5e-05, "loss": 0.8626, "mean_token_accuracy": 0.8795038461685181, "num_tokens": 854297250.0, "step": 22497 }, { "epoch": 2.8619768477292964, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.38128662109375, "learning_rate": 5e-05, "loss": 0.928, "mean_token_accuracy": 0.8768340349197388, "num_tokens": 854340283.0, "step": 22498 }, { "epoch": 2.862104058007887, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.933211803436279, "learning_rate": 5e-05, "loss": 0.7142, "mean_token_accuracy": 0.8924396634101868, "num_tokens": 854380145.0, "step": 22499 }, { "epoch": 2.8622312682864774, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 8.659178733825684, "learning_rate": 5e-05, "loss": 0.907, "mean_token_accuracy": 0.8787819147109985, "num_tokens": 854422809.0, "step": 22500 }, { "epoch": 2.862358478565068, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.886113166809082, "learning_rate": 5e-05, "loss": 0.7398, "mean_token_accuracy": 0.8856091499328613, "num_tokens": 854461926.0, "step": 22501 }, { "epoch": 2.8624856888436585, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.33316707611084, "learning_rate": 5e-05, "loss": 0.9002, "mean_token_accuracy": 0.868614912033081, "num_tokens": 854502431.0, "step": 22502 }, { "epoch": 2.862612899122249, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.371703147888184, "learning_rate": 5e-05, "loss": 0.7476, "mean_token_accuracy": 0.8931266665458679, "num_tokens": 854537440.0, "step": 22503 }, { "epoch": 2.8627401094008396, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.637025833129883, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.8849816918373108, "num_tokens": 854575121.0, "step": 22504 }, { "epoch": 2.86286731967943, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.886221408843994, "learning_rate": 5e-05, "loss": 0.7747, "mean_token_accuracy": 0.8938601016998291, "num_tokens": 854603624.0, "step": 22505 }, { "epoch": 2.8629945299580206, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.101731300354004, "learning_rate": 5e-05, "loss": 0.7777, "mean_token_accuracy": 0.8615631461143494, "num_tokens": 854644616.0, "step": 22506 }, { "epoch": 2.863121740236611, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.237585067749023, "learning_rate": 5e-05, "loss": 0.8519, "mean_token_accuracy": 0.876301646232605, "num_tokens": 854689772.0, "step": 22507 }, { "epoch": 2.8632489505152017, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.8283843994140625, "learning_rate": 5e-05, "loss": 0.6847, "mean_token_accuracy": 0.8881978988647461, "num_tokens": 854725549.0, "step": 22508 }, { "epoch": 2.863376160793792, "ewc_loss": 0.5, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.476288795471191, "learning_rate": 5e-05, "loss": 0.8295, "mean_token_accuracy": 0.8835237622261047, "num_tokens": 854768509.0, "step": 22509 }, { "epoch": 2.8635033710723827, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.681234836578369, "learning_rate": 5e-05, "loss": 0.7192, "mean_token_accuracy": 0.8746716976165771, "num_tokens": 854804822.0, "step": 22510 }, { "epoch": 2.8636305813509733, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.686066627502441, "learning_rate": 5e-05, "loss": 0.8584, "mean_token_accuracy": 0.8855279684066772, "num_tokens": 854837458.0, "step": 22511 }, { "epoch": 2.863757791629564, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.6322431564331055, "learning_rate": 5e-05, "loss": 0.7245, "mean_token_accuracy": 0.8798644542694092, "num_tokens": 854877996.0, "step": 22512 }, { "epoch": 2.8638850019081543, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.830792427062988, "learning_rate": 5e-05, "loss": 0.9332, "mean_token_accuracy": 0.8739542365074158, "num_tokens": 854913078.0, "step": 22513 }, { "epoch": 2.864012212186745, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.744077682495117, "learning_rate": 5e-05, "loss": 0.7277, "mean_token_accuracy": 0.8881299495697021, "num_tokens": 854952300.0, "step": 22514 }, { "epoch": 2.8641394224653354, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.80998420715332, "learning_rate": 5e-05, "loss": 0.8999, "mean_token_accuracy": 0.8851839303970337, "num_tokens": 854992657.0, "step": 22515 }, { "epoch": 2.8642666327439255, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 6.983026027679443, "learning_rate": 5e-05, "loss": 0.7916, "mean_token_accuracy": 0.8800308704376221, "num_tokens": 855028147.0, "step": 22516 }, { "epoch": 2.8643938430225164, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.603522300720215, "learning_rate": 5e-05, "loss": 0.917, "mean_token_accuracy": 0.8780745267868042, "num_tokens": 855069930.0, "step": 22517 }, { "epoch": 2.8645210533011065, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.2347331047058105, "learning_rate": 5e-05, "loss": 0.81, "mean_token_accuracy": 0.8820147514343262, "num_tokens": 855105224.0, "step": 22518 }, { "epoch": 2.8646482635796975, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.343655586242676, "learning_rate": 5e-05, "loss": 0.9241, "mean_token_accuracy": 0.8686002492904663, "num_tokens": 855139533.0, "step": 22519 }, { "epoch": 2.8647754738582876, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.337441444396973, "learning_rate": 5e-05, "loss": 0.8622, "mean_token_accuracy": 0.866645097732544, "num_tokens": 855178380.0, "step": 22520 }, { "epoch": 2.864902684136878, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.225679397583008, "learning_rate": 5e-05, "loss": 0.89, "mean_token_accuracy": 0.8752611875534058, "num_tokens": 855212700.0, "step": 22521 }, { "epoch": 2.8650298944154686, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.236421585083008, "learning_rate": 5e-05, "loss": 0.7738, "mean_token_accuracy": 0.8854933977127075, "num_tokens": 855250340.0, "step": 22522 }, { "epoch": 2.865157104694059, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.322230339050293, "learning_rate": 5e-05, "loss": 0.8747, "mean_token_accuracy": 0.8799573183059692, "num_tokens": 855285341.0, "step": 22523 }, { "epoch": 2.8652843149726497, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.057541370391846, "learning_rate": 5e-05, "loss": 0.7814, "mean_token_accuracy": 0.875382125377655, "num_tokens": 855331038.0, "step": 22524 }, { "epoch": 2.86541152525124, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.486873626708984, "learning_rate": 5e-05, "loss": 0.8693, "mean_token_accuracy": 0.8836911916732788, "num_tokens": 855363643.0, "step": 22525 }, { "epoch": 2.8655387355298307, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.034139633178711, "learning_rate": 5e-05, "loss": 0.75, "mean_token_accuracy": 0.8863734602928162, "num_tokens": 855397342.0, "step": 22526 }, { "epoch": 2.8656659458084213, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.42618179321289, "learning_rate": 5e-05, "loss": 0.8599, "mean_token_accuracy": 0.889546811580658, "num_tokens": 855432867.0, "step": 22527 }, { "epoch": 2.865793156087012, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.25531005859375, "learning_rate": 5e-05, "loss": 0.817, "mean_token_accuracy": 0.8745632767677307, "num_tokens": 855466091.0, "step": 22528 }, { "epoch": 2.8659203663656023, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.050481796264648, "learning_rate": 5e-05, "loss": 0.8921, "mean_token_accuracy": 0.867256760597229, "num_tokens": 855501200.0, "step": 22529 }, { "epoch": 2.866047576644193, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.576153755187988, "learning_rate": 5e-05, "loss": 0.7935, "mean_token_accuracy": 0.8883004784584045, "num_tokens": 855543625.0, "step": 22530 }, { "epoch": 2.8661747869227834, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.58951473236084, "learning_rate": 5e-05, "loss": 0.8185, "mean_token_accuracy": 0.87245112657547, "num_tokens": 855579653.0, "step": 22531 }, { "epoch": 2.866301997201374, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.80737829208374, "learning_rate": 5e-05, "loss": 0.8118, "mean_token_accuracy": 0.8807317018508911, "num_tokens": 855611081.0, "step": 22532 }, { "epoch": 2.8664292074799644, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.287436485290527, "learning_rate": 5e-05, "loss": 0.719, "mean_token_accuracy": 0.8921493887901306, "num_tokens": 855642389.0, "step": 22533 }, { "epoch": 2.866556417758555, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.884825706481934, "learning_rate": 5e-05, "loss": 0.8071, "mean_token_accuracy": 0.8753616213798523, "num_tokens": 855686240.0, "step": 22534 }, { "epoch": 2.8666836280371455, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.2088303565979, "learning_rate": 5e-05, "loss": 0.7409, "mean_token_accuracy": 0.8754819631576538, "num_tokens": 855727979.0, "step": 22535 }, { "epoch": 2.866810838315736, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.846738338470459, "learning_rate": 5e-05, "loss": 0.816, "mean_token_accuracy": 0.8700181245803833, "num_tokens": 855769119.0, "step": 22536 }, { "epoch": 2.8669380485943265, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.214809894561768, "learning_rate": 5e-05, "loss": 0.6845, "mean_token_accuracy": 0.8903583884239197, "num_tokens": 855811472.0, "step": 22537 }, { "epoch": 2.867065258872917, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.835818767547607, "learning_rate": 5e-05, "loss": 0.7573, "mean_token_accuracy": 0.882190465927124, "num_tokens": 855842739.0, "step": 22538 }, { "epoch": 2.8671924691515076, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.179344177246094, "learning_rate": 5e-05, "loss": 0.7252, "mean_token_accuracy": 0.8767974376678467, "num_tokens": 855874520.0, "step": 22539 }, { "epoch": 2.867319679430098, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.882081508636475, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.8878695964813232, "num_tokens": 855915151.0, "step": 22540 }, { "epoch": 2.867446889708688, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.0839996337890625, "learning_rate": 5e-05, "loss": 0.7539, "mean_token_accuracy": 0.8659560680389404, "num_tokens": 855948944.0, "step": 22541 }, { "epoch": 2.867574099987279, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.017101287841797, "learning_rate": 5e-05, "loss": 0.7728, "mean_token_accuracy": 0.878279447555542, "num_tokens": 855981604.0, "step": 22542 }, { "epoch": 2.8677013102658693, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.909601211547852, "learning_rate": 5e-05, "loss": 0.6724, "mean_token_accuracy": 0.89051753282547, "num_tokens": 856020407.0, "step": 22543 }, { "epoch": 2.8678285205444602, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.267481803894043, "learning_rate": 5e-05, "loss": 0.8296, "mean_token_accuracy": 0.8655396699905396, "num_tokens": 856056514.0, "step": 22544 }, { "epoch": 2.8679557308230503, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000270843505859375, "grad_norm": 6.644745826721191, "learning_rate": 5e-05, "loss": 0.7009, "mean_token_accuracy": 0.878460168838501, "num_tokens": 856091385.0, "step": 22545 }, { "epoch": 2.868082941101641, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.664207458496094, "learning_rate": 5e-05, "loss": 0.7955, "mean_token_accuracy": 0.8960707187652588, "num_tokens": 856133077.0, "step": 22546 }, { "epoch": 2.8682101513802314, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.357522010803223, "learning_rate": 5e-05, "loss": 0.6743, "mean_token_accuracy": 0.8842254877090454, "num_tokens": 856174682.0, "step": 22547 }, { "epoch": 2.868337361658822, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000484466552734375, "grad_norm": 9.125898361206055, "learning_rate": 5e-05, "loss": 0.9432, "mean_token_accuracy": 0.8643028140068054, "num_tokens": 856214486.0, "step": 22548 }, { "epoch": 2.8684645719374124, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.3325581550598145, "learning_rate": 5e-05, "loss": 0.7167, "mean_token_accuracy": 0.8723864555358887, "num_tokens": 856246816.0, "step": 22549 }, { "epoch": 2.868591782216003, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.259034156799316, "learning_rate": 5e-05, "loss": 0.9907, "mean_token_accuracy": 0.8727284669876099, "num_tokens": 856286232.0, "step": 22550 }, { "epoch": 2.8687189924945935, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 6.992232799530029, "learning_rate": 5e-05, "loss": 0.7871, "mean_token_accuracy": 0.8844102621078491, "num_tokens": 856326451.0, "step": 22551 }, { "epoch": 2.868846202773184, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.487852096557617, "learning_rate": 5e-05, "loss": 0.9293, "mean_token_accuracy": 0.8745933771133423, "num_tokens": 856367766.0, "step": 22552 }, { "epoch": 2.8689734130517746, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.040284156799316, "learning_rate": 5e-05, "loss": 0.9254, "mean_token_accuracy": 0.8737680912017822, "num_tokens": 856396085.0, "step": 22553 }, { "epoch": 2.869100623330365, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.95291805267334, "learning_rate": 5e-05, "loss": 0.7051, "mean_token_accuracy": 0.8929591774940491, "num_tokens": 856439835.0, "step": 22554 }, { "epoch": 2.8692278336089556, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.296013832092285, "learning_rate": 5e-05, "loss": 0.9431, "mean_token_accuracy": 0.8859080076217651, "num_tokens": 856470890.0, "step": 22555 }, { "epoch": 2.869355043887546, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 5.884531497955322, "learning_rate": 5e-05, "loss": 0.7276, "mean_token_accuracy": 0.8691827654838562, "num_tokens": 856510281.0, "step": 22556 }, { "epoch": 2.8694822541661367, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.559432029724121, "learning_rate": 5e-05, "loss": 1.1127, "mean_token_accuracy": 0.8655463457107544, "num_tokens": 856558194.0, "step": 22557 }, { "epoch": 2.869609464444727, "ewc_loss": 0.5, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 6.616390228271484, "learning_rate": 5e-05, "loss": 0.836, "mean_token_accuracy": 0.8808979988098145, "num_tokens": 856594001.0, "step": 22558 }, { "epoch": 2.8697366747233177, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 10.845664978027344, "learning_rate": 5e-05, "loss": 1.1416, "mean_token_accuracy": 0.8802317380905151, "num_tokens": 856636424.0, "step": 22559 }, { "epoch": 2.8698638850019083, "ewc_loss": 0.82421875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 8.698285102844238, "learning_rate": 5e-05, "loss": 1.1962, "mean_token_accuracy": 0.8680825233459473, "num_tokens": 856670561.0, "step": 22560 }, { "epoch": 2.869991095280499, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.249314308166504, "learning_rate": 5e-05, "loss": 0.9981, "mean_token_accuracy": 0.8707060217857361, "num_tokens": 856704852.0, "step": 22561 }, { "epoch": 2.8701183055590893, "ewc_loss": 0.796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 9.325919151306152, "learning_rate": 5e-05, "loss": 1.123, "mean_token_accuracy": 0.8849875926971436, "num_tokens": 856743943.0, "step": 22562 }, { "epoch": 2.87024551583768, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.77785587310791, "learning_rate": 5e-05, "loss": 0.8834, "mean_token_accuracy": 0.8744838833808899, "num_tokens": 856782944.0, "step": 22563 }, { "epoch": 2.87037272611627, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.804828643798828, "learning_rate": 5e-05, "loss": 1.0392, "mean_token_accuracy": 0.8789188861846924, "num_tokens": 856821240.0, "step": 22564 }, { "epoch": 2.870499936394861, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 7.70400333404541, "learning_rate": 5e-05, "loss": 0.9558, "mean_token_accuracy": 0.8760313987731934, "num_tokens": 856860113.0, "step": 22565 }, { "epoch": 2.870627146673451, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.362369537353516, "learning_rate": 5e-05, "loss": 0.9831, "mean_token_accuracy": 0.8803510665893555, "num_tokens": 856902501.0, "step": 22566 }, { "epoch": 2.870754356952042, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.14603042602539, "learning_rate": 5e-05, "loss": 0.9681, "mean_token_accuracy": 0.886081337928772, "num_tokens": 856940023.0, "step": 22567 }, { "epoch": 2.870881567230632, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.289288520812988, "learning_rate": 5e-05, "loss": 0.9325, "mean_token_accuracy": 0.8577492237091064, "num_tokens": 856980653.0, "step": 22568 }, { "epoch": 2.871008777509223, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.431888580322266, "learning_rate": 5e-05, "loss": 0.9331, "mean_token_accuracy": 0.8833593726158142, "num_tokens": 857008081.0, "step": 22569 }, { "epoch": 2.871135987787813, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.562505722045898, "learning_rate": 5e-05, "loss": 0.7716, "mean_token_accuracy": 0.8908407092094421, "num_tokens": 857040884.0, "step": 22570 }, { "epoch": 2.8712631980664036, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.51637077331543, "learning_rate": 5e-05, "loss": 0.9332, "mean_token_accuracy": 0.8755656480789185, "num_tokens": 857078519.0, "step": 22571 }, { "epoch": 2.871390408344994, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.327709674835205, "learning_rate": 5e-05, "loss": 0.813, "mean_token_accuracy": 0.8804458379745483, "num_tokens": 857119250.0, "step": 22572 }, { "epoch": 2.8715176186235847, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.428183555603027, "learning_rate": 5e-05, "loss": 0.8671, "mean_token_accuracy": 0.8813374638557434, "num_tokens": 857160003.0, "step": 22573 }, { "epoch": 2.871644828902175, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.307647705078125, "learning_rate": 5e-05, "loss": 0.8051, "mean_token_accuracy": 0.8808474540710449, "num_tokens": 857200918.0, "step": 22574 }, { "epoch": 2.8717720391807657, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.169361114501953, "learning_rate": 5e-05, "loss": 0.7991, "mean_token_accuracy": 0.8891781568527222, "num_tokens": 857241157.0, "step": 22575 }, { "epoch": 2.8718992494593563, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.431796550750732, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.878497838973999, "num_tokens": 857276237.0, "step": 22576 }, { "epoch": 2.872026459737947, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.787046432495117, "learning_rate": 5e-05, "loss": 0.737, "mean_token_accuracy": 0.8938537240028381, "num_tokens": 857311059.0, "step": 22577 }, { "epoch": 2.8721536700165373, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.700042247772217, "learning_rate": 5e-05, "loss": 0.8484, "mean_token_accuracy": 0.867626428604126, "num_tokens": 857350326.0, "step": 22578 }, { "epoch": 2.872280880295128, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.344052791595459, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8707716464996338, "num_tokens": 857387708.0, "step": 22579 }, { "epoch": 2.8724080905737184, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.102826118469238, "learning_rate": 5e-05, "loss": 0.8047, "mean_token_accuracy": 0.8824998140335083, "num_tokens": 857427658.0, "step": 22580 }, { "epoch": 2.872535300852309, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.886015892028809, "learning_rate": 5e-05, "loss": 0.7265, "mean_token_accuracy": 0.8754405379295349, "num_tokens": 857468529.0, "step": 22581 }, { "epoch": 2.8726625111308994, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.622563362121582, "learning_rate": 5e-05, "loss": 0.8433, "mean_token_accuracy": 0.8823715448379517, "num_tokens": 857509406.0, "step": 22582 }, { "epoch": 2.87278972140949, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.469847202301025, "learning_rate": 5e-05, "loss": 0.7004, "mean_token_accuracy": 0.8822149634361267, "num_tokens": 857548302.0, "step": 22583 }, { "epoch": 2.8729169316880805, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.188342094421387, "learning_rate": 5e-05, "loss": 0.8579, "mean_token_accuracy": 0.8972088098526001, "num_tokens": 857584022.0, "step": 22584 }, { "epoch": 2.873044141966671, "ewc_loss": 0.375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.341639995574951, "learning_rate": 5e-05, "loss": 0.7365, "mean_token_accuracy": 0.8738421201705933, "num_tokens": 857627334.0, "step": 22585 }, { "epoch": 2.8731713522452615, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.56965160369873, "learning_rate": 5e-05, "loss": 0.9826, "mean_token_accuracy": 0.8792872428894043, "num_tokens": 857663317.0, "step": 22586 }, { "epoch": 2.873298562523852, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 6.801325798034668, "learning_rate": 5e-05, "loss": 0.8567, "mean_token_accuracy": 0.8641849756240845, "num_tokens": 857697358.0, "step": 22587 }, { "epoch": 2.8734257728024426, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.329073905944824, "learning_rate": 5e-05, "loss": 0.9945, "mean_token_accuracy": 0.8763605952262878, "num_tokens": 857737686.0, "step": 22588 }, { "epoch": 2.8735529830810327, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.5720648765563965, "learning_rate": 5e-05, "loss": 0.9527, "mean_token_accuracy": 0.8715286254882812, "num_tokens": 857774730.0, "step": 22589 }, { "epoch": 2.8736801933596237, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.484153747558594, "learning_rate": 5e-05, "loss": 0.9303, "mean_token_accuracy": 0.8747771978378296, "num_tokens": 857811503.0, "step": 22590 }, { "epoch": 2.8738074036382137, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.015296936035156, "learning_rate": 5e-05, "loss": 0.9465, "mean_token_accuracy": 0.870672881603241, "num_tokens": 857850696.0, "step": 22591 }, { "epoch": 2.8739346139168047, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.74908971786499, "learning_rate": 5e-05, "loss": 0.8127, "mean_token_accuracy": 0.8846542239189148, "num_tokens": 857894849.0, "step": 22592 }, { "epoch": 2.874061824195395, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.010437965393066, "learning_rate": 5e-05, "loss": 0.8462, "mean_token_accuracy": 0.8828105926513672, "num_tokens": 857938876.0, "step": 22593 }, { "epoch": 2.8741890344739858, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.657562732696533, "learning_rate": 5e-05, "loss": 0.8034, "mean_token_accuracy": 0.8839725852012634, "num_tokens": 857973424.0, "step": 22594 }, { "epoch": 2.874316244752576, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.536348342895508, "learning_rate": 5e-05, "loss": 0.7993, "mean_token_accuracy": 0.8740000128746033, "num_tokens": 858008105.0, "step": 22595 }, { "epoch": 2.8744434550311664, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.086344718933105, "learning_rate": 5e-05, "loss": 0.8373, "mean_token_accuracy": 0.8814620971679688, "num_tokens": 858043898.0, "step": 22596 }, { "epoch": 2.874570665309757, "ewc_loss": 0.375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.932509422302246, "learning_rate": 5e-05, "loss": 0.7648, "mean_token_accuracy": 0.8638157248497009, "num_tokens": 858080671.0, "step": 22597 }, { "epoch": 2.8746978755883474, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.623598098754883, "learning_rate": 5e-05, "loss": 0.8328, "mean_token_accuracy": 0.8927867412567139, "num_tokens": 858118619.0, "step": 22598 }, { "epoch": 2.874825085866938, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.566623687744141, "learning_rate": 5e-05, "loss": 0.7188, "mean_token_accuracy": 0.8755454421043396, "num_tokens": 858160207.0, "step": 22599 }, { "epoch": 2.8749522961455285, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.035881996154785, "learning_rate": 5e-05, "loss": 0.9485, "mean_token_accuracy": 0.8732418417930603, "num_tokens": 858201634.0, "step": 22600 }, { "epoch": 2.875079506424119, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.641698360443115, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8719877004623413, "num_tokens": 858241618.0, "step": 22601 }, { "epoch": 2.8752067167027096, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.056137084960938, "learning_rate": 5e-05, "loss": 0.9406, "mean_token_accuracy": 0.8776230812072754, "num_tokens": 858275393.0, "step": 22602 }, { "epoch": 2.8753339269813, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.169247627258301, "learning_rate": 5e-05, "loss": 0.834, "mean_token_accuracy": 0.8758262395858765, "num_tokens": 858316470.0, "step": 22603 }, { "epoch": 2.8754611372598906, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.450456619262695, "learning_rate": 5e-05, "loss": 0.8667, "mean_token_accuracy": 0.8857122659683228, "num_tokens": 858350567.0, "step": 22604 }, { "epoch": 2.875588347538481, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 7.796924114227295, "learning_rate": 5e-05, "loss": 0.8947, "mean_token_accuracy": 0.8780678510665894, "num_tokens": 858393998.0, "step": 22605 }, { "epoch": 2.8757155578170717, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.55247688293457, "learning_rate": 5e-05, "loss": 0.7705, "mean_token_accuracy": 0.8895101547241211, "num_tokens": 858430987.0, "step": 22606 }, { "epoch": 2.875842768095662, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.25777530670166, "learning_rate": 5e-05, "loss": 0.8769, "mean_token_accuracy": 0.8740203976631165, "num_tokens": 858464230.0, "step": 22607 }, { "epoch": 2.8759699783742527, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.019002914428711, "learning_rate": 5e-05, "loss": 0.7581, "mean_token_accuracy": 0.8743835091590881, "num_tokens": 858505965.0, "step": 22608 }, { "epoch": 2.8760971886528433, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.440442085266113, "learning_rate": 5e-05, "loss": 0.8454, "mean_token_accuracy": 0.8824747800827026, "num_tokens": 858542629.0, "step": 22609 }, { "epoch": 2.876224398931434, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.892952919006348, "learning_rate": 5e-05, "loss": 0.6866, "mean_token_accuracy": 0.8910597562789917, "num_tokens": 858583377.0, "step": 22610 }, { "epoch": 2.8763516092100243, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.490915298461914, "learning_rate": 5e-05, "loss": 0.8262, "mean_token_accuracy": 0.8915769457817078, "num_tokens": 858620546.0, "step": 22611 }, { "epoch": 2.876478819488615, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.921449184417725, "learning_rate": 5e-05, "loss": 0.7548, "mean_token_accuracy": 0.8728825449943542, "num_tokens": 858659469.0, "step": 22612 }, { "epoch": 2.8766060297672054, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.473026275634766, "learning_rate": 5e-05, "loss": 0.8585, "mean_token_accuracy": 0.8773161172866821, "num_tokens": 858694519.0, "step": 22613 }, { "epoch": 2.8767332400457954, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.940852642059326, "learning_rate": 5e-05, "loss": 0.714, "mean_token_accuracy": 0.8887679576873779, "num_tokens": 858728371.0, "step": 22614 }, { "epoch": 2.8768604503243864, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.464600563049316, "learning_rate": 5e-05, "loss": 0.8443, "mean_token_accuracy": 0.8868417739868164, "num_tokens": 858765195.0, "step": 22615 }, { "epoch": 2.8769876606029765, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.965384006500244, "learning_rate": 5e-05, "loss": 0.7175, "mean_token_accuracy": 0.8883473873138428, "num_tokens": 858804129.0, "step": 22616 }, { "epoch": 2.8771148708815675, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.470625877380371, "learning_rate": 5e-05, "loss": 0.8451, "mean_token_accuracy": 0.8880085945129395, "num_tokens": 858841267.0, "step": 22617 }, { "epoch": 2.8772420811601576, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.94140625, "learning_rate": 5e-05, "loss": 0.7397, "mean_token_accuracy": 0.8797789812088013, "num_tokens": 858882405.0, "step": 22618 }, { "epoch": 2.877369291438748, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.52859115600586, "learning_rate": 5e-05, "loss": 0.8715, "mean_token_accuracy": 0.8818161487579346, "num_tokens": 858920176.0, "step": 22619 }, { "epoch": 2.8774965017173386, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.919981479644775, "learning_rate": 5e-05, "loss": 0.7627, "mean_token_accuracy": 0.8743276000022888, "num_tokens": 858959065.0, "step": 22620 }, { "epoch": 2.877623711995929, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.572402000427246, "learning_rate": 5e-05, "loss": 0.8737, "mean_token_accuracy": 0.8879755139350891, "num_tokens": 858997497.0, "step": 22621 }, { "epoch": 2.8777509222745197, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.9772419929504395, "learning_rate": 5e-05, "loss": 0.7449, "mean_token_accuracy": 0.8861243724822998, "num_tokens": 859036425.0, "step": 22622 }, { "epoch": 2.87787813255311, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.523015975952148, "learning_rate": 5e-05, "loss": 0.8963, "mean_token_accuracy": 0.8777036666870117, "num_tokens": 859073602.0, "step": 22623 }, { "epoch": 2.8780053428317007, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.185258388519287, "learning_rate": 5e-05, "loss": 0.8085, "mean_token_accuracy": 0.8747062683105469, "num_tokens": 859107576.0, "step": 22624 }, { "epoch": 2.8781325531102913, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.216817855834961, "learning_rate": 5e-05, "loss": 0.8838, "mean_token_accuracy": 0.872788667678833, "num_tokens": 859141866.0, "step": 22625 }, { "epoch": 2.878259763388882, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.451467990875244, "learning_rate": 5e-05, "loss": 0.788, "mean_token_accuracy": 0.8873441219329834, "num_tokens": 859180592.0, "step": 22626 }, { "epoch": 2.8783869736674723, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.803934574127197, "learning_rate": 5e-05, "loss": 0.8424, "mean_token_accuracy": 0.8754790425300598, "num_tokens": 859216365.0, "step": 22627 }, { "epoch": 2.878514183946063, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.653285026550293, "learning_rate": 5e-05, "loss": 0.8089, "mean_token_accuracy": 0.8769856691360474, "num_tokens": 859254167.0, "step": 22628 }, { "epoch": 2.8786413942246534, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.553832530975342, "learning_rate": 5e-05, "loss": 0.7493, "mean_token_accuracy": 0.8859177827835083, "num_tokens": 859288980.0, "step": 22629 }, { "epoch": 2.878768604503244, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.6579060554504395, "learning_rate": 5e-05, "loss": 0.7417, "mean_token_accuracy": 0.894798755645752, "num_tokens": 859326741.0, "step": 22630 }, { "epoch": 2.8788958147818344, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.533879280090332, "learning_rate": 5e-05, "loss": 0.7724, "mean_token_accuracy": 0.8746943473815918, "num_tokens": 859369530.0, "step": 22631 }, { "epoch": 2.879023025060425, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.52816104888916, "learning_rate": 5e-05, "loss": 0.8039, "mean_token_accuracy": 0.8629348278045654, "num_tokens": 859409870.0, "step": 22632 }, { "epoch": 2.8791502353390155, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.666679382324219, "learning_rate": 5e-05, "loss": 0.799, "mean_token_accuracy": 0.8644353747367859, "num_tokens": 859443419.0, "step": 22633 }, { "epoch": 2.879277445617606, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.281338214874268, "learning_rate": 5e-05, "loss": 0.7091, "mean_token_accuracy": 0.8851935863494873, "num_tokens": 859474832.0, "step": 22634 }, { "epoch": 2.8794046558961965, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.876809597015381, "learning_rate": 5e-05, "loss": 0.7939, "mean_token_accuracy": 0.8766730427742004, "num_tokens": 859517668.0, "step": 22635 }, { "epoch": 2.879531866174787, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 7.065697193145752, "learning_rate": 5e-05, "loss": 0.7032, "mean_token_accuracy": 0.8848347663879395, "num_tokens": 859555954.0, "step": 22636 }, { "epoch": 2.8796590764533776, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 8.094660758972168, "learning_rate": 5e-05, "loss": 0.8115, "mean_token_accuracy": 0.873630166053772, "num_tokens": 859591221.0, "step": 22637 }, { "epoch": 2.879786286731968, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.836366653442383, "learning_rate": 5e-05, "loss": 0.6576, "mean_token_accuracy": 0.8933670520782471, "num_tokens": 859631879.0, "step": 22638 }, { "epoch": 2.879913497010558, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.380303382873535, "learning_rate": 5e-05, "loss": 0.8444, "mean_token_accuracy": 0.8740501999855042, "num_tokens": 859669259.0, "step": 22639 }, { "epoch": 2.880040707289149, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.601351261138916, "learning_rate": 5e-05, "loss": 0.6715, "mean_token_accuracy": 0.8869384527206421, "num_tokens": 859706168.0, "step": 22640 }, { "epoch": 2.8801679175677393, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.716897010803223, "learning_rate": 5e-05, "loss": 0.8412, "mean_token_accuracy": 0.8828040957450867, "num_tokens": 859743182.0, "step": 22641 }, { "epoch": 2.8802951278463302, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.453622817993164, "learning_rate": 5e-05, "loss": 0.6901, "mean_token_accuracy": 0.8832257390022278, "num_tokens": 859786120.0, "step": 22642 }, { "epoch": 2.8804223381249203, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.93868350982666, "learning_rate": 5e-05, "loss": 0.8844, "mean_token_accuracy": 0.884222686290741, "num_tokens": 859824097.0, "step": 22643 }, { "epoch": 2.880549548403511, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.654546737670898, "learning_rate": 5e-05, "loss": 0.7574, "mean_token_accuracy": 0.8686259388923645, "num_tokens": 859860081.0, "step": 22644 }, { "epoch": 2.8806767586821014, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.661249160766602, "learning_rate": 5e-05, "loss": 0.9102, "mean_token_accuracy": 0.8734216690063477, "num_tokens": 859898731.0, "step": 22645 }, { "epoch": 2.880803968960692, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.304846286773682, "learning_rate": 5e-05, "loss": 0.7991, "mean_token_accuracy": 0.8796980977058411, "num_tokens": 859938406.0, "step": 22646 }, { "epoch": 2.8809311792392824, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.733608245849609, "learning_rate": 5e-05, "loss": 0.7555, "mean_token_accuracy": 0.8913638591766357, "num_tokens": 859974270.0, "step": 22647 }, { "epoch": 2.881058389517873, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.189995765686035, "learning_rate": 5e-05, "loss": 0.8431, "mean_token_accuracy": 0.8827788829803467, "num_tokens": 860007261.0, "step": 22648 }, { "epoch": 2.8811855997964635, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.616770267486572, "learning_rate": 5e-05, "loss": 0.7025, "mean_token_accuracy": 0.8794039487838745, "num_tokens": 860051170.0, "step": 22649 }, { "epoch": 2.881312810075054, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.193284034729004, "learning_rate": 5e-05, "loss": 0.9171, "mean_token_accuracy": 0.8840137124061584, "num_tokens": 860082771.0, "step": 22650 }, { "epoch": 2.8814400203536445, "ewc_loss": 0.33203125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002536773681640625, "grad_norm": 5.896712303161621, "learning_rate": 5e-05, "loss": 0.7151, "mean_token_accuracy": 0.8685057163238525, "num_tokens": 860124184.0, "step": 22651 }, { "epoch": 2.881567230632235, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 10.225309371948242, "learning_rate": 5e-05, "loss": 1.0555, "mean_token_accuracy": 0.8680688142776489, "num_tokens": 860158453.0, "step": 22652 }, { "epoch": 2.8816944409108256, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 6.395772457122803, "learning_rate": 5e-05, "loss": 0.771, "mean_token_accuracy": 0.8853468894958496, "num_tokens": 860195403.0, "step": 22653 }, { "epoch": 2.881821651189416, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.474675178527832, "learning_rate": 5e-05, "loss": 1.1415, "mean_token_accuracy": 0.8697941303253174, "num_tokens": 860232813.0, "step": 22654 }, { "epoch": 2.8819488614680067, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 8.264227867126465, "learning_rate": 5e-05, "loss": 1.0639, "mean_token_accuracy": 0.8756086230278015, "num_tokens": 860269542.0, "step": 22655 }, { "epoch": 2.882076071746597, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 8.953262329101562, "learning_rate": 5e-05, "loss": 0.9754, "mean_token_accuracy": 0.8735979199409485, "num_tokens": 860305386.0, "step": 22656 }, { "epoch": 2.8822032820251877, "ewc_loss": 0.75, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 9.254678726196289, "learning_rate": 5e-05, "loss": 1.0961, "mean_token_accuracy": 0.8793810606002808, "num_tokens": 860343137.0, "step": 22657 }, { "epoch": 2.8823304923037782, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.191932201385498, "learning_rate": 5e-05, "loss": 0.8286, "mean_token_accuracy": 0.8725065588951111, "num_tokens": 860382968.0, "step": 22658 }, { "epoch": 2.8824577025823688, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.27586841583252, "learning_rate": 5e-05, "loss": 1.0993, "mean_token_accuracy": 0.8746121525764465, "num_tokens": 860426985.0, "step": 22659 }, { "epoch": 2.8825849128609593, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 7.3223395347595215, "learning_rate": 5e-05, "loss": 0.9434, "mean_token_accuracy": 0.8694862127304077, "num_tokens": 860460950.0, "step": 22660 }, { "epoch": 2.88271212313955, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.235934257507324, "learning_rate": 5e-05, "loss": 1.0302, "mean_token_accuracy": 0.88526850938797, "num_tokens": 860499328.0, "step": 22661 }, { "epoch": 2.88283933341814, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 8.306441307067871, "learning_rate": 5e-05, "loss": 1.0423, "mean_token_accuracy": 0.8798829317092896, "num_tokens": 860532411.0, "step": 22662 }, { "epoch": 2.882966543696731, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.14855670928955, "learning_rate": 5e-05, "loss": 0.9287, "mean_token_accuracy": 0.8899539113044739, "num_tokens": 860567565.0, "step": 22663 }, { "epoch": 2.883093753975321, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.483885765075684, "learning_rate": 5e-05, "loss": 1.017, "mean_token_accuracy": 0.8729978799819946, "num_tokens": 860611687.0, "step": 22664 }, { "epoch": 2.883220964253912, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.293418884277344, "learning_rate": 5e-05, "loss": 0.9408, "mean_token_accuracy": 0.8655515909194946, "num_tokens": 860652220.0, "step": 22665 }, { "epoch": 2.883348174532502, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.242172241210938, "learning_rate": 5e-05, "loss": 0.8954, "mean_token_accuracy": 0.8877778053283691, "num_tokens": 860689641.0, "step": 22666 }, { "epoch": 2.883475384811093, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.031241416931152, "learning_rate": 5e-05, "loss": 0.8775, "mean_token_accuracy": 0.8821501731872559, "num_tokens": 860727490.0, "step": 22667 }, { "epoch": 2.883602595089683, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.745236396789551, "learning_rate": 5e-05, "loss": 0.8257, "mean_token_accuracy": 0.8744820356369019, "num_tokens": 860768625.0, "step": 22668 }, { "epoch": 2.8837298053682736, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.133536338806152, "learning_rate": 5e-05, "loss": 0.8446, "mean_token_accuracy": 0.8867712616920471, "num_tokens": 860806294.0, "step": 22669 }, { "epoch": 2.883857015646864, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.29028844833374, "learning_rate": 5e-05, "loss": 0.8121, "mean_token_accuracy": 0.8629447817802429, "num_tokens": 860842672.0, "step": 22670 }, { "epoch": 2.8839842259254547, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.35439395904541, "learning_rate": 5e-05, "loss": 0.8803, "mean_token_accuracy": 0.8707561492919922, "num_tokens": 860879882.0, "step": 22671 }, { "epoch": 2.884111436204045, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.963879108428955, "learning_rate": 5e-05, "loss": 0.7234, "mean_token_accuracy": 0.8786939382553101, "num_tokens": 860916099.0, "step": 22672 }, { "epoch": 2.8842386464826357, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.6114501953125, "learning_rate": 5e-05, "loss": 0.8702, "mean_token_accuracy": 0.8798401951789856, "num_tokens": 860957982.0, "step": 22673 }, { "epoch": 2.8843658567612263, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.7765374183654785, "learning_rate": 5e-05, "loss": 0.6997, "mean_token_accuracy": 0.8904701471328735, "num_tokens": 860995528.0, "step": 22674 }, { "epoch": 2.884493067039817, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.822720527648926, "learning_rate": 5e-05, "loss": 0.9042, "mean_token_accuracy": 0.8766754865646362, "num_tokens": 861034676.0, "step": 22675 }, { "epoch": 2.8846202773184073, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 6.771726131439209, "learning_rate": 5e-05, "loss": 0.7844, "mean_token_accuracy": 0.8707505464553833, "num_tokens": 861069767.0, "step": 22676 }, { "epoch": 2.884747487596998, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.834075927734375, "learning_rate": 5e-05, "loss": 0.893, "mean_token_accuracy": 0.8814761638641357, "num_tokens": 861109787.0, "step": 22677 }, { "epoch": 2.8848746978755884, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 6.965193748474121, "learning_rate": 5e-05, "loss": 0.7707, "mean_token_accuracy": 0.8825358748435974, "num_tokens": 861144603.0, "step": 22678 }, { "epoch": 2.885001908154179, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.566278457641602, "learning_rate": 5e-05, "loss": 0.8586, "mean_token_accuracy": 0.887779176235199, "num_tokens": 861184750.0, "step": 22679 }, { "epoch": 2.8851291184327694, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.263057708740234, "learning_rate": 5e-05, "loss": 0.818, "mean_token_accuracy": 0.8778409361839294, "num_tokens": 861223591.0, "step": 22680 }, { "epoch": 2.88525632871136, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.173492431640625, "learning_rate": 5e-05, "loss": 0.8573, "mean_token_accuracy": 0.8745501041412354, "num_tokens": 861260456.0, "step": 22681 }, { "epoch": 2.8853835389899505, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.498899936676025, "learning_rate": 5e-05, "loss": 0.8246, "mean_token_accuracy": 0.8772737979888916, "num_tokens": 861298266.0, "step": 22682 }, { "epoch": 2.885510749268541, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.77700662612915, "learning_rate": 5e-05, "loss": 0.7832, "mean_token_accuracy": 0.8857225179672241, "num_tokens": 861334646.0, "step": 22683 }, { "epoch": 2.8856379595471315, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.682281494140625, "learning_rate": 5e-05, "loss": 0.8203, "mean_token_accuracy": 0.8728427886962891, "num_tokens": 861370734.0, "step": 22684 }, { "epoch": 2.885765169825722, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.479793071746826, "learning_rate": 5e-05, "loss": 0.7132, "mean_token_accuracy": 0.8950141668319702, "num_tokens": 861403111.0, "step": 22685 }, { "epoch": 2.8858923801043126, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.782440185546875, "learning_rate": 5e-05, "loss": 0.7727, "mean_token_accuracy": 0.8829896450042725, "num_tokens": 861440668.0, "step": 22686 }, { "epoch": 2.8860195903829027, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.355180740356445, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8701928853988647, "num_tokens": 861478586.0, "step": 22687 }, { "epoch": 2.8861468006614936, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.807079792022705, "learning_rate": 5e-05, "loss": 0.7653, "mean_token_accuracy": 0.8792264461517334, "num_tokens": 861511637.0, "step": 22688 }, { "epoch": 2.8862740109400837, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.298250675201416, "learning_rate": 5e-05, "loss": 0.7315, "mean_token_accuracy": 0.8782710433006287, "num_tokens": 861545421.0, "step": 22689 }, { "epoch": 2.8864012212186747, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.819432735443115, "learning_rate": 5e-05, "loss": 0.7857, "mean_token_accuracy": 0.8758476972579956, "num_tokens": 861586230.0, "step": 22690 }, { "epoch": 2.886528431497265, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.205004692077637, "learning_rate": 5e-05, "loss": 0.7041, "mean_token_accuracy": 0.8876665830612183, "num_tokens": 861620556.0, "step": 22691 }, { "epoch": 2.8866556417758558, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.918547630310059, "learning_rate": 5e-05, "loss": 0.7645, "mean_token_accuracy": 0.8781581521034241, "num_tokens": 861655826.0, "step": 22692 }, { "epoch": 2.886782852054446, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.061460018157959, "learning_rate": 5e-05, "loss": 0.6966, "mean_token_accuracy": 0.8826231956481934, "num_tokens": 861699160.0, "step": 22693 }, { "epoch": 2.8869100623330364, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 8.108564376831055, "learning_rate": 5e-05, "loss": 0.8191, "mean_token_accuracy": 0.8717730045318604, "num_tokens": 861736784.0, "step": 22694 }, { "epoch": 2.887037272611627, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.805631637573242, "learning_rate": 5e-05, "loss": 0.6655, "mean_token_accuracy": 0.891207754611969, "num_tokens": 861775359.0, "step": 22695 }, { "epoch": 2.8871644828902174, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.486842155456543, "learning_rate": 5e-05, "loss": 0.8587, "mean_token_accuracy": 0.8701591491699219, "num_tokens": 861813852.0, "step": 22696 }, { "epoch": 2.887291693168808, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.40499210357666, "learning_rate": 5e-05, "loss": 0.6893, "mean_token_accuracy": 0.8772276639938354, "num_tokens": 861853730.0, "step": 22697 }, { "epoch": 2.8874189034473985, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 9.107097625732422, "learning_rate": 5e-05, "loss": 0.8804, "mean_token_accuracy": 0.8773969411849976, "num_tokens": 861891405.0, "step": 22698 }, { "epoch": 2.887546113725989, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.068452835083008, "learning_rate": 5e-05, "loss": 0.6546, "mean_token_accuracy": 0.887948751449585, "num_tokens": 861929595.0, "step": 22699 }, { "epoch": 2.8876733240045795, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.784029960632324, "learning_rate": 5e-05, "loss": 1.005, "mean_token_accuracy": 0.8690621852874756, "num_tokens": 861963263.0, "step": 22700 }, { "epoch": 2.88780053428317, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.407399654388428, "learning_rate": 5e-05, "loss": 0.7828, "mean_token_accuracy": 0.8726327419281006, "num_tokens": 862002259.0, "step": 22701 }, { "epoch": 2.8879277445617606, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 9.809632301330566, "learning_rate": 5e-05, "loss": 1.0795, "mean_token_accuracy": 0.8611401319503784, "num_tokens": 862039869.0, "step": 22702 }, { "epoch": 2.888054954840351, "ewc_loss": 0.625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 7.753593921661377, "learning_rate": 5e-05, "loss": 0.9663, "mean_token_accuracy": 0.8785749673843384, "num_tokens": 862068979.0, "step": 22703 }, { "epoch": 2.8881821651189417, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.369168281555176, "learning_rate": 5e-05, "loss": 0.9379, "mean_token_accuracy": 0.8664796352386475, "num_tokens": 862117384.0, "step": 22704 }, { "epoch": 2.888309375397532, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 8.98970890045166, "learning_rate": 5e-05, "loss": 1.0317, "mean_token_accuracy": 0.8700665235519409, "num_tokens": 862157961.0, "step": 22705 }, { "epoch": 2.8884365856761227, "ewc_loss": 0.375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.497895240783691, "learning_rate": 5e-05, "loss": 0.7165, "mean_token_accuracy": 0.8822882175445557, "num_tokens": 862199023.0, "step": 22706 }, { "epoch": 2.8885637959547132, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 10.649944305419922, "learning_rate": 5e-05, "loss": 1.1049, "mean_token_accuracy": 0.8786869645118713, "num_tokens": 862237839.0, "step": 22707 }, { "epoch": 2.8886910062333038, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 6.672662734985352, "learning_rate": 5e-05, "loss": 0.8346, "mean_token_accuracy": 0.8753138780593872, "num_tokens": 862273060.0, "step": 22708 }, { "epoch": 2.8888182165118943, "ewc_loss": 0.92578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00084686279296875, "grad_norm": 11.678975105285645, "learning_rate": 5e-05, "loss": 1.2909, "mean_token_accuracy": 0.8705706596374512, "num_tokens": 862316469.0, "step": 22709 }, { "epoch": 2.888945426790485, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 9.198264122009277, "learning_rate": 5e-05, "loss": 1.2615, "mean_token_accuracy": 0.8742752075195312, "num_tokens": 862359367.0, "step": 22710 }, { "epoch": 2.8890726370690754, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 10.421031951904297, "learning_rate": 5e-05, "loss": 1.0113, "mean_token_accuracy": 0.8853548169136047, "num_tokens": 862397113.0, "step": 22711 }, { "epoch": 2.8891998473476654, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 9.310561180114746, "learning_rate": 5e-05, "loss": 1.1826, "mean_token_accuracy": 0.8721946477890015, "num_tokens": 862435627.0, "step": 22712 }, { "epoch": 2.8893270576262564, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.447330474853516, "learning_rate": 5e-05, "loss": 1.016, "mean_token_accuracy": 0.878145694732666, "num_tokens": 862476948.0, "step": 22713 }, { "epoch": 2.8894542679048465, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.729656219482422, "learning_rate": 5e-05, "loss": 0.999, "mean_token_accuracy": 0.8761228322982788, "num_tokens": 862513429.0, "step": 22714 }, { "epoch": 2.8895814781834375, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.961360931396484, "learning_rate": 5e-05, "loss": 1.0196, "mean_token_accuracy": 0.8678176403045654, "num_tokens": 862548570.0, "step": 22715 }, { "epoch": 2.8897086884620276, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.877297401428223, "learning_rate": 5e-05, "loss": 0.8652, "mean_token_accuracy": 0.8788659572601318, "num_tokens": 862586265.0, "step": 22716 }, { "epoch": 2.889835898740618, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.23202896118164, "learning_rate": 5e-05, "loss": 1.0322, "mean_token_accuracy": 0.8645389080047607, "num_tokens": 862620552.0, "step": 22717 }, { "epoch": 2.8899631090192086, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.2320098876953125, "learning_rate": 5e-05, "loss": 0.7638, "mean_token_accuracy": 0.8895552158355713, "num_tokens": 862656449.0, "step": 22718 }, { "epoch": 2.890090319297799, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.848541259765625, "learning_rate": 5e-05, "loss": 1.0483, "mean_token_accuracy": 0.8722217679023743, "num_tokens": 862697612.0, "step": 22719 }, { "epoch": 2.8902175295763897, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 7.420831203460693, "learning_rate": 5e-05, "loss": 0.8989, "mean_token_accuracy": 0.8684774041175842, "num_tokens": 862742304.0, "step": 22720 }, { "epoch": 2.89034473985498, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.973348617553711, "learning_rate": 5e-05, "loss": 1.0373, "mean_token_accuracy": 0.8805292844772339, "num_tokens": 862781738.0, "step": 22721 }, { "epoch": 2.8904719501335707, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.080547332763672, "learning_rate": 5e-05, "loss": 1.0244, "mean_token_accuracy": 0.854750394821167, "num_tokens": 862823124.0, "step": 22722 }, { "epoch": 2.8905991604121613, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.444154739379883, "learning_rate": 5e-05, "loss": 1.0299, "mean_token_accuracy": 0.8694298267364502, "num_tokens": 862857706.0, "step": 22723 }, { "epoch": 2.890726370690752, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.315218925476074, "learning_rate": 5e-05, "loss": 1.015, "mean_token_accuracy": 0.86036217212677, "num_tokens": 862899966.0, "step": 22724 }, { "epoch": 2.8908535809693423, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.722232818603516, "learning_rate": 5e-05, "loss": 0.9114, "mean_token_accuracy": 0.8805651664733887, "num_tokens": 862931090.0, "step": 22725 }, { "epoch": 2.890980791247933, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.171979904174805, "learning_rate": 5e-05, "loss": 0.8585, "mean_token_accuracy": 0.8865703344345093, "num_tokens": 862962833.0, "step": 22726 }, { "epoch": 2.8911080015265234, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.298956871032715, "learning_rate": 5e-05, "loss": 0.91, "mean_token_accuracy": 0.8752287030220032, "num_tokens": 863003769.0, "step": 22727 }, { "epoch": 2.891235211805114, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.70841121673584, "learning_rate": 5e-05, "loss": 0.8093, "mean_token_accuracy": 0.8788261413574219, "num_tokens": 863038130.0, "step": 22728 }, { "epoch": 2.8913624220837044, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.531322479248047, "learning_rate": 5e-05, "loss": 0.8669, "mean_token_accuracy": 0.891895592212677, "num_tokens": 863077943.0, "step": 22729 }, { "epoch": 2.891489632362295, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.132559299468994, "learning_rate": 5e-05, "loss": 0.7656, "mean_token_accuracy": 0.8761223554611206, "num_tokens": 863113853.0, "step": 22730 }, { "epoch": 2.8916168426408855, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.116244316101074, "learning_rate": 5e-05, "loss": 0.9391, "mean_token_accuracy": 0.8842493891716003, "num_tokens": 863149407.0, "step": 22731 }, { "epoch": 2.891744052919476, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.9970245361328125, "learning_rate": 5e-05, "loss": 0.7894, "mean_token_accuracy": 0.8724209070205688, "num_tokens": 863183324.0, "step": 22732 }, { "epoch": 2.8918712631980665, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.417936325073242, "learning_rate": 5e-05, "loss": 0.9494, "mean_token_accuracy": 0.8933128714561462, "num_tokens": 863220400.0, "step": 22733 }, { "epoch": 2.891998473476657, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 7.540046691894531, "learning_rate": 5e-05, "loss": 0.8317, "mean_token_accuracy": 0.8848997354507446, "num_tokens": 863259565.0, "step": 22734 }, { "epoch": 2.8921256837552476, "ewc_loss": 0.625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.97762680053711, "learning_rate": 5e-05, "loss": 0.9863, "mean_token_accuracy": 0.8735493421554565, "num_tokens": 863293480.0, "step": 22735 }, { "epoch": 2.892252894033838, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.160521507263184, "learning_rate": 5e-05, "loss": 0.9198, "mean_token_accuracy": 0.878404974937439, "num_tokens": 863331702.0, "step": 22736 }, { "epoch": 2.892380104312428, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.905543804168701, "learning_rate": 5e-05, "loss": 0.8826, "mean_token_accuracy": 0.8733360767364502, "num_tokens": 863366195.0, "step": 22737 }, { "epoch": 2.892507314591019, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.442340850830078, "learning_rate": 5e-05, "loss": 0.908, "mean_token_accuracy": 0.8723260164260864, "num_tokens": 863406674.0, "step": 22738 }, { "epoch": 2.8926345248696093, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.170499324798584, "learning_rate": 5e-05, "loss": 0.7751, "mean_token_accuracy": 0.8772485256195068, "num_tokens": 863441837.0, "step": 22739 }, { "epoch": 2.8927617351482002, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 8.480722427368164, "learning_rate": 5e-05, "loss": 0.829, "mean_token_accuracy": 0.8864268064498901, "num_tokens": 863476461.0, "step": 22740 }, { "epoch": 2.8928889454267903, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.1259236335754395, "learning_rate": 5e-05, "loss": 0.7682, "mean_token_accuracy": 0.8722200393676758, "num_tokens": 863513715.0, "step": 22741 }, { "epoch": 2.893016155705381, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.199207305908203, "learning_rate": 5e-05, "loss": 0.8034, "mean_token_accuracy": 0.8849139213562012, "num_tokens": 863553172.0, "step": 22742 }, { "epoch": 2.8931433659839714, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.355477333068848, "learning_rate": 5e-05, "loss": 0.7788, "mean_token_accuracy": 0.8697493076324463, "num_tokens": 863585353.0, "step": 22743 }, { "epoch": 2.893270576262562, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.8921918869018555, "learning_rate": 5e-05, "loss": 0.8117, "mean_token_accuracy": 0.8725870847702026, "num_tokens": 863627510.0, "step": 22744 }, { "epoch": 2.8933977865411524, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.4779052734375, "learning_rate": 5e-05, "loss": 0.7805, "mean_token_accuracy": 0.8719359636306763, "num_tokens": 863670683.0, "step": 22745 }, { "epoch": 2.893524996819743, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.717730522155762, "learning_rate": 5e-05, "loss": 0.7784, "mean_token_accuracy": 0.873753547668457, "num_tokens": 863710793.0, "step": 22746 }, { "epoch": 2.8936522070983335, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.554431438446045, "learning_rate": 5e-05, "loss": 0.7169, "mean_token_accuracy": 0.8861525058746338, "num_tokens": 863745963.0, "step": 22747 }, { "epoch": 2.893779417376924, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.583873271942139, "learning_rate": 5e-05, "loss": 0.7138, "mean_token_accuracy": 0.8892478346824646, "num_tokens": 863783729.0, "step": 22748 }, { "epoch": 2.8939066276555145, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.577844619750977, "learning_rate": 5e-05, "loss": 0.7307, "mean_token_accuracy": 0.8821637034416199, "num_tokens": 863820070.0, "step": 22749 }, { "epoch": 2.894033837934105, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.594751834869385, "learning_rate": 5e-05, "loss": 0.722, "mean_token_accuracy": 0.8818831443786621, "num_tokens": 863853438.0, "step": 22750 }, { "epoch": 2.8941610482126956, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.4310808181762695, "learning_rate": 5e-05, "loss": 0.7252, "mean_token_accuracy": 0.8821377754211426, "num_tokens": 863894151.0, "step": 22751 }, { "epoch": 2.894288258491286, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.7941412925720215, "learning_rate": 5e-05, "loss": 0.7276, "mean_token_accuracy": 0.8871235251426697, "num_tokens": 863929712.0, "step": 22752 }, { "epoch": 2.8944154687698767, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.133576393127441, "learning_rate": 5e-05, "loss": 0.757, "mean_token_accuracy": 0.8629786372184753, "num_tokens": 863970236.0, "step": 22753 }, { "epoch": 2.894542679048467, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.107588768005371, "learning_rate": 5e-05, "loss": 0.8031, "mean_token_accuracy": 0.8696010112762451, "num_tokens": 864009164.0, "step": 22754 }, { "epoch": 2.8946698893270577, "ewc_loss": 0.345703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.801184177398682, "learning_rate": 5e-05, "loss": 0.6982, "mean_token_accuracy": 0.87507164478302, "num_tokens": 864048941.0, "step": 22755 }, { "epoch": 2.8947970996056482, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.583915710449219, "learning_rate": 5e-05, "loss": 0.8367, "mean_token_accuracy": 0.8739786148071289, "num_tokens": 864084041.0, "step": 22756 }, { "epoch": 2.8949243098842388, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002422332763671875, "grad_norm": 6.3135199546813965, "learning_rate": 5e-05, "loss": 0.6469, "mean_token_accuracy": 0.8821154236793518, "num_tokens": 864118370.0, "step": 22757 }, { "epoch": 2.8950515201628293, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 9.318524360656738, "learning_rate": 5e-05, "loss": 0.8557, "mean_token_accuracy": 0.8919105529785156, "num_tokens": 864148888.0, "step": 22758 }, { "epoch": 2.89517873044142, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 5.926105499267578, "learning_rate": 5e-05, "loss": 0.656, "mean_token_accuracy": 0.8809214234352112, "num_tokens": 864185850.0, "step": 22759 }, { "epoch": 2.89530594072001, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 10.116700172424316, "learning_rate": 5e-05, "loss": 1.0546, "mean_token_accuracy": 0.8644525408744812, "num_tokens": 864224914.0, "step": 22760 }, { "epoch": 2.895433150998601, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 6.543407917022705, "learning_rate": 5e-05, "loss": 0.7626, "mean_token_accuracy": 0.8751238584518433, "num_tokens": 864266430.0, "step": 22761 }, { "epoch": 2.895560361277191, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.040061950683594, "learning_rate": 5e-05, "loss": 1.0433, "mean_token_accuracy": 0.8741371035575867, "num_tokens": 864299769.0, "step": 22762 }, { "epoch": 2.895687571555782, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.122454643249512, "learning_rate": 5e-05, "loss": 1.0012, "mean_token_accuracy": 0.8741220235824585, "num_tokens": 864340713.0, "step": 22763 }, { "epoch": 2.895814781834372, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.413736343383789, "learning_rate": 5e-05, "loss": 0.9207, "mean_token_accuracy": 0.8704814314842224, "num_tokens": 864382930.0, "step": 22764 }, { "epoch": 2.895941992112963, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.235894203186035, "learning_rate": 5e-05, "loss": 0.9857, "mean_token_accuracy": 0.8879938721656799, "num_tokens": 864424632.0, "step": 22765 }, { "epoch": 2.896069202391553, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.731713771820068, "learning_rate": 5e-05, "loss": 0.7104, "mean_token_accuracy": 0.8896875977516174, "num_tokens": 864463444.0, "step": 22766 }, { "epoch": 2.8961964126701436, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.457201957702637, "learning_rate": 5e-05, "loss": 1.1362, "mean_token_accuracy": 0.8597289323806763, "num_tokens": 864506651.0, "step": 22767 }, { "epoch": 2.896323622948734, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.070558547973633, "learning_rate": 5e-05, "loss": 0.8206, "mean_token_accuracy": 0.8852506279945374, "num_tokens": 864544075.0, "step": 22768 }, { "epoch": 2.8964508332273247, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.716033935546875, "learning_rate": 5e-05, "loss": 1.0499, "mean_token_accuracy": 0.8949765563011169, "num_tokens": 864579814.0, "step": 22769 }, { "epoch": 2.896578043505915, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 8.612356185913086, "learning_rate": 5e-05, "loss": 1.1255, "mean_token_accuracy": 0.860430121421814, "num_tokens": 864616690.0, "step": 22770 }, { "epoch": 2.8967052537845057, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.505071640014648, "learning_rate": 5e-05, "loss": 0.9506, "mean_token_accuracy": 0.8837077617645264, "num_tokens": 864662035.0, "step": 22771 }, { "epoch": 2.8968324640630962, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 8.869623184204102, "learning_rate": 5e-05, "loss": 1.0579, "mean_token_accuracy": 0.8689239025115967, "num_tokens": 864695219.0, "step": 22772 }, { "epoch": 2.8969596743416868, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.5079984664917, "learning_rate": 5e-05, "loss": 0.8889, "mean_token_accuracy": 0.8789522647857666, "num_tokens": 864731990.0, "step": 22773 }, { "epoch": 2.8970868846202773, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 8.499039649963379, "learning_rate": 5e-05, "loss": 0.9181, "mean_token_accuracy": 0.8824928402900696, "num_tokens": 864769642.0, "step": 22774 }, { "epoch": 2.897214094898868, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.295738220214844, "learning_rate": 5e-05, "loss": 0.9056, "mean_token_accuracy": 0.8743704557418823, "num_tokens": 864805329.0, "step": 22775 }, { "epoch": 2.8973413051774584, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.740912437438965, "learning_rate": 5e-05, "loss": 0.7894, "mean_token_accuracy": 0.8880892992019653, "num_tokens": 864847748.0, "step": 22776 }, { "epoch": 2.897468515456049, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.554065704345703, "learning_rate": 5e-05, "loss": 0.9266, "mean_token_accuracy": 0.8654381036758423, "num_tokens": 864884621.0, "step": 22777 }, { "epoch": 2.8975957257346394, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.08046293258667, "learning_rate": 5e-05, "loss": 0.7256, "mean_token_accuracy": 0.884192943572998, "num_tokens": 864927398.0, "step": 22778 }, { "epoch": 2.89772293601323, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.002805709838867, "learning_rate": 5e-05, "loss": 0.8925, "mean_token_accuracy": 0.8882814645767212, "num_tokens": 864961038.0, "step": 22779 }, { "epoch": 2.8978501462918205, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.809296607971191, "learning_rate": 5e-05, "loss": 0.7513, "mean_token_accuracy": 0.8737454414367676, "num_tokens": 865000307.0, "step": 22780 }, { "epoch": 2.897977356570411, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.366724014282227, "learning_rate": 5e-05, "loss": 0.9178, "mean_token_accuracy": 0.8886129260063171, "num_tokens": 865035898.0, "step": 22781 }, { "epoch": 2.8981045668490015, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 6.975991249084473, "learning_rate": 5e-05, "loss": 0.7295, "mean_token_accuracy": 0.9005489349365234, "num_tokens": 865076325.0, "step": 22782 }, { "epoch": 2.898231777127592, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.414105415344238, "learning_rate": 5e-05, "loss": 0.9783, "mean_token_accuracy": 0.8749130964279175, "num_tokens": 865117290.0, "step": 22783 }, { "epoch": 2.8983589874061826, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.429370880126953, "learning_rate": 5e-05, "loss": 0.8375, "mean_token_accuracy": 0.8846274614334106, "num_tokens": 865153003.0, "step": 22784 }, { "epoch": 2.8984861976847727, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.010794639587402, "learning_rate": 5e-05, "loss": 0.9584, "mean_token_accuracy": 0.8748147487640381, "num_tokens": 865195113.0, "step": 22785 }, { "epoch": 2.8986134079633636, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.770321846008301, "learning_rate": 5e-05, "loss": 0.8613, "mean_token_accuracy": 0.8856258392333984, "num_tokens": 865231319.0, "step": 22786 }, { "epoch": 2.8987406182419537, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 8.398597717285156, "learning_rate": 5e-05, "loss": 0.868, "mean_token_accuracy": 0.8807408213615417, "num_tokens": 865265862.0, "step": 22787 }, { "epoch": 2.8988678285205447, "ewc_loss": 0.5, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.809357643127441, "learning_rate": 5e-05, "loss": 0.8608, "mean_token_accuracy": 0.8738845586776733, "num_tokens": 865303873.0, "step": 22788 }, { "epoch": 2.898995038799135, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.098689079284668, "learning_rate": 5e-05, "loss": 0.8232, "mean_token_accuracy": 0.8872677087783813, "num_tokens": 865335163.0, "step": 22789 }, { "epoch": 2.8991222490777258, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.578554153442383, "learning_rate": 5e-05, "loss": 0.8022, "mean_token_accuracy": 0.8771246075630188, "num_tokens": 865373498.0, "step": 22790 }, { "epoch": 2.899249459356316, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.110498428344727, "learning_rate": 5e-05, "loss": 0.8331, "mean_token_accuracy": 0.879352867603302, "num_tokens": 865411042.0, "step": 22791 }, { "epoch": 2.8993766696349064, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.3244452476501465, "learning_rate": 5e-05, "loss": 0.7512, "mean_token_accuracy": 0.8811224699020386, "num_tokens": 865447676.0, "step": 22792 }, { "epoch": 2.899503879913497, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.25973129272461, "learning_rate": 5e-05, "loss": 0.8601, "mean_token_accuracy": 0.8749989867210388, "num_tokens": 865490905.0, "step": 22793 }, { "epoch": 2.8996310901920874, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.102716445922852, "learning_rate": 5e-05, "loss": 0.7394, "mean_token_accuracy": 0.8819754719734192, "num_tokens": 865530572.0, "step": 22794 }, { "epoch": 2.899758300470678, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.485003471374512, "learning_rate": 5e-05, "loss": 0.8471, "mean_token_accuracy": 0.8825034499168396, "num_tokens": 865565999.0, "step": 22795 }, { "epoch": 2.8998855107492685, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.93404483795166, "learning_rate": 5e-05, "loss": 0.737, "mean_token_accuracy": 0.8741252422332764, "num_tokens": 865602455.0, "step": 22796 }, { "epoch": 2.900012721027859, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.681779861450195, "learning_rate": 5e-05, "loss": 0.8315, "mean_token_accuracy": 0.9000891447067261, "num_tokens": 865637723.0, "step": 22797 }, { "epoch": 2.9001399313064495, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 6.976365566253662, "learning_rate": 5e-05, "loss": 0.7548, "mean_token_accuracy": 0.8709701895713806, "num_tokens": 865669668.0, "step": 22798 }, { "epoch": 2.90026714158504, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000476837158203125, "grad_norm": 8.614818572998047, "learning_rate": 5e-05, "loss": 0.8967, "mean_token_accuracy": 0.8778961896896362, "num_tokens": 865708255.0, "step": 22799 }, { "epoch": 2.9003943518636306, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.25804328918457, "learning_rate": 5e-05, "loss": 0.81, "mean_token_accuracy": 0.8678170442581177, "num_tokens": 865749607.0, "step": 22800 }, { "epoch": 2.900521562142221, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.226040840148926, "learning_rate": 5e-05, "loss": 0.8958, "mean_token_accuracy": 0.8718407154083252, "num_tokens": 865786372.0, "step": 22801 }, { "epoch": 2.9006487724208116, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.594568729400635, "learning_rate": 5e-05, "loss": 0.7896, "mean_token_accuracy": 0.881536602973938, "num_tokens": 865822939.0, "step": 22802 }, { "epoch": 2.900775982699402, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.723842620849609, "learning_rate": 5e-05, "loss": 0.7567, "mean_token_accuracy": 0.8895559906959534, "num_tokens": 865859012.0, "step": 22803 }, { "epoch": 2.9009031929779927, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.8626179695129395, "learning_rate": 5e-05, "loss": 0.7817, "mean_token_accuracy": 0.8859450817108154, "num_tokens": 865897641.0, "step": 22804 }, { "epoch": 2.9010304032565832, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.331324577331543, "learning_rate": 5e-05, "loss": 0.7562, "mean_token_accuracy": 0.8780190944671631, "num_tokens": 865938270.0, "step": 22805 }, { "epoch": 2.9011576135351738, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.062435150146484, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8835921883583069, "num_tokens": 865971066.0, "step": 22806 }, { "epoch": 2.9012848238137643, "ewc_loss": 0.375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.085545063018799, "learning_rate": 5e-05, "loss": 0.6907, "mean_token_accuracy": 0.8887271881103516, "num_tokens": 866008595.0, "step": 22807 }, { "epoch": 2.901412034092355, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.186184883117676, "learning_rate": 5e-05, "loss": 0.7996, "mean_token_accuracy": 0.8830951452255249, "num_tokens": 866046650.0, "step": 22808 }, { "epoch": 2.9015392443709453, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.974637031555176, "learning_rate": 5e-05, "loss": 0.735, "mean_token_accuracy": 0.8709339499473572, "num_tokens": 866086305.0, "step": 22809 }, { "epoch": 2.9016664546495354, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.27797794342041, "learning_rate": 5e-05, "loss": 0.8092, "mean_token_accuracy": 0.8772777318954468, "num_tokens": 866125770.0, "step": 22810 }, { "epoch": 2.9017936649281264, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.914699554443359, "learning_rate": 5e-05, "loss": 0.679, "mean_token_accuracy": 0.8879287838935852, "num_tokens": 866160777.0, "step": 22811 }, { "epoch": 2.9019208752067165, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.338153839111328, "learning_rate": 5e-05, "loss": 0.7811, "mean_token_accuracy": 0.8875972032546997, "num_tokens": 866195212.0, "step": 22812 }, { "epoch": 2.9020480854853075, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.866551876068115, "learning_rate": 5e-05, "loss": 0.6942, "mean_token_accuracy": 0.884308934211731, "num_tokens": 866236252.0, "step": 22813 }, { "epoch": 2.9021752957638975, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.417837142944336, "learning_rate": 5e-05, "loss": 0.7913, "mean_token_accuracy": 0.8909876346588135, "num_tokens": 866275705.0, "step": 22814 }, { "epoch": 2.902302506042488, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.872811317443848, "learning_rate": 5e-05, "loss": 0.7266, "mean_token_accuracy": 0.8773112297058105, "num_tokens": 866316458.0, "step": 22815 }, { "epoch": 2.9024297163210786, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 8.424603462219238, "learning_rate": 5e-05, "loss": 0.8507, "mean_token_accuracy": 0.873361349105835, "num_tokens": 866351107.0, "step": 22816 }, { "epoch": 2.902556926599669, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.946402549743652, "learning_rate": 5e-05, "loss": 0.6756, "mean_token_accuracy": 0.8964808583259583, "num_tokens": 866385381.0, "step": 22817 }, { "epoch": 2.9026841368782597, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.345593452453613, "learning_rate": 5e-05, "loss": 0.8129, "mean_token_accuracy": 0.8854284286499023, "num_tokens": 866427018.0, "step": 22818 }, { "epoch": 2.90281134715685, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.0825958251953125, "learning_rate": 5e-05, "loss": 0.6849, "mean_token_accuracy": 0.8936204314231873, "num_tokens": 866463973.0, "step": 22819 }, { "epoch": 2.9029385574354407, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.17151927947998, "learning_rate": 5e-05, "loss": 0.8371, "mean_token_accuracy": 0.8755525350570679, "num_tokens": 866500900.0, "step": 22820 }, { "epoch": 2.9030657677140312, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.236477851867676, "learning_rate": 5e-05, "loss": 0.6999, "mean_token_accuracy": 0.8953620791435242, "num_tokens": 866539757.0, "step": 22821 }, { "epoch": 2.9031929779926218, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.972895622253418, "learning_rate": 5e-05, "loss": 0.798, "mean_token_accuracy": 0.879487156867981, "num_tokens": 866577454.0, "step": 22822 }, { "epoch": 2.9033201882712123, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.336867332458496, "learning_rate": 5e-05, "loss": 0.711, "mean_token_accuracy": 0.8908089399337769, "num_tokens": 866615032.0, "step": 22823 }, { "epoch": 2.903447398549803, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.882758140563965, "learning_rate": 5e-05, "loss": 0.7954, "mean_token_accuracy": 0.8737456202507019, "num_tokens": 866649700.0, "step": 22824 }, { "epoch": 2.9035746088283934, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.309250831604004, "learning_rate": 5e-05, "loss": 0.7439, "mean_token_accuracy": 0.8757436871528625, "num_tokens": 866684720.0, "step": 22825 }, { "epoch": 2.903701819106984, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.920872688293457, "learning_rate": 5e-05, "loss": 0.7901, "mean_token_accuracy": 0.8800430297851562, "num_tokens": 866722127.0, "step": 22826 }, { "epoch": 2.9038290293855744, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.1891584396362305, "learning_rate": 5e-05, "loss": 0.7176, "mean_token_accuracy": 0.882482647895813, "num_tokens": 866757093.0, "step": 22827 }, { "epoch": 2.903956239664165, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.062174797058105, "learning_rate": 5e-05, "loss": 0.8186, "mean_token_accuracy": 0.8721270561218262, "num_tokens": 866796229.0, "step": 22828 }, { "epoch": 2.9040834499427555, "ewc_loss": 0.375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.009437084197998, "learning_rate": 5e-05, "loss": 0.7062, "mean_token_accuracy": 0.883013129234314, "num_tokens": 866831867.0, "step": 22829 }, { "epoch": 2.904210660221346, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.260163307189941, "learning_rate": 5e-05, "loss": 0.8244, "mean_token_accuracy": 0.8700479865074158, "num_tokens": 866876246.0, "step": 22830 }, { "epoch": 2.9043378704999365, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.854769229888916, "learning_rate": 5e-05, "loss": 0.7122, "mean_token_accuracy": 0.8798580169677734, "num_tokens": 866909555.0, "step": 22831 }, { "epoch": 2.904465080778527, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.440882682800293, "learning_rate": 5e-05, "loss": 0.8264, "mean_token_accuracy": 0.8777097463607788, "num_tokens": 866945208.0, "step": 22832 }, { "epoch": 2.9045922910571176, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.7940168380737305, "learning_rate": 5e-05, "loss": 0.6973, "mean_token_accuracy": 0.8831720948219299, "num_tokens": 866976832.0, "step": 22833 }, { "epoch": 2.904719501335708, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.448403358459473, "learning_rate": 5e-05, "loss": 0.8445, "mean_token_accuracy": 0.87601238489151, "num_tokens": 867013964.0, "step": 22834 }, { "epoch": 2.904846711614298, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.989724159240723, "learning_rate": 5e-05, "loss": 0.7333, "mean_token_accuracy": 0.8780912756919861, "num_tokens": 867054361.0, "step": 22835 }, { "epoch": 2.904973921892889, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.139608383178711, "learning_rate": 5e-05, "loss": 0.7608, "mean_token_accuracy": 0.8969942331314087, "num_tokens": 867098732.0, "step": 22836 }, { "epoch": 2.9051011321714793, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.413179874420166, "learning_rate": 5e-05, "loss": 0.8034, "mean_token_accuracy": 0.8637651801109314, "num_tokens": 867133839.0, "step": 22837 }, { "epoch": 2.9052283424500702, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.5148468017578125, "learning_rate": 5e-05, "loss": 0.8095, "mean_token_accuracy": 0.8595743179321289, "num_tokens": 867175419.0, "step": 22838 }, { "epoch": 2.9053555527286603, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.009413719177246, "learning_rate": 5e-05, "loss": 0.8163, "mean_token_accuracy": 0.8730440139770508, "num_tokens": 867216585.0, "step": 22839 }, { "epoch": 2.905482763007251, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.811922073364258, "learning_rate": 5e-05, "loss": 0.6857, "mean_token_accuracy": 0.8840954899787903, "num_tokens": 867257228.0, "step": 22840 }, { "epoch": 2.9056099732858414, "ewc_loss": 0.5, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.714027404785156, "learning_rate": 5e-05, "loss": 0.8552, "mean_token_accuracy": 0.8784937262535095, "num_tokens": 867297504.0, "step": 22841 }, { "epoch": 2.905737183564432, "ewc_loss": 0.318359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000240325927734375, "grad_norm": 6.149180889129639, "learning_rate": 5e-05, "loss": 0.6611, "mean_token_accuracy": 0.8805152177810669, "num_tokens": 867336799.0, "step": 22842 }, { "epoch": 2.9058643938430224, "ewc_loss": 0.58984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00051116943359375, "grad_norm": 9.605191230773926, "learning_rate": 5e-05, "loss": 0.9348, "mean_token_accuracy": 0.8787990808486938, "num_tokens": 867371337.0, "step": 22843 }, { "epoch": 2.905991604121613, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 5.767795562744141, "learning_rate": 5e-05, "loss": 0.6666, "mean_token_accuracy": 0.8831730484962463, "num_tokens": 867411870.0, "step": 22844 }, { "epoch": 2.9061188144002035, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.45413589477539, "learning_rate": 5e-05, "loss": 1.0844, "mean_token_accuracy": 0.8684297800064087, "num_tokens": 867450278.0, "step": 22845 }, { "epoch": 2.906246024678794, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 6.814285755157471, "learning_rate": 5e-05, "loss": 0.8804, "mean_token_accuracy": 0.8722343444824219, "num_tokens": 867490316.0, "step": 22846 }, { "epoch": 2.9063732349573845, "ewc_loss": 0.71875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000640869140625, "grad_norm": 10.089407920837402, "learning_rate": 5e-05, "loss": 1.0378, "mean_token_accuracy": 0.885350227355957, "num_tokens": 867526243.0, "step": 22847 }, { "epoch": 2.906500445235975, "ewc_loss": 0.76171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000682830810546875, "grad_norm": 8.624463081359863, "learning_rate": 5e-05, "loss": 1.1209, "mean_token_accuracy": 0.8754045367240906, "num_tokens": 867568757.0, "step": 22848 }, { "epoch": 2.9066276555145656, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.969577789306641, "learning_rate": 5e-05, "loss": 0.8278, "mean_token_accuracy": 0.8851749300956726, "num_tokens": 867602164.0, "step": 22849 }, { "epoch": 2.906754865793156, "ewc_loss": 0.78125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 10.061100006103516, "learning_rate": 5e-05, "loss": 1.1391, "mean_token_accuracy": 0.8732084631919861, "num_tokens": 867640722.0, "step": 22850 }, { "epoch": 2.9068820760717466, "ewc_loss": 0.435546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003566741943359375, "grad_norm": 6.584923267364502, "learning_rate": 5e-05, "loss": 0.7716, "mean_token_accuracy": 0.884196937084198, "num_tokens": 867679568.0, "step": 22851 }, { "epoch": 2.907009286350337, "ewc_loss": 0.953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000873565673828125, "grad_norm": 11.86508846282959, "learning_rate": 5e-05, "loss": 1.3101, "mean_token_accuracy": 0.8728393316268921, "num_tokens": 867714991.0, "step": 22852 }, { "epoch": 2.9071364966289277, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 8.696592330932617, "learning_rate": 5e-05, "loss": 1.1573, "mean_token_accuracy": 0.8730040788650513, "num_tokens": 867753180.0, "step": 22853 }, { "epoch": 2.9072637069075182, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000732421875, "grad_norm": 11.476874351501465, "learning_rate": 5e-05, "loss": 1.1457, "mean_token_accuracy": 0.8806617259979248, "num_tokens": 867789535.0, "step": 22854 }, { "epoch": 2.9073909171861088, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 9.399457931518555, "learning_rate": 5e-05, "loss": 1.2236, "mean_token_accuracy": 0.8814781904220581, "num_tokens": 867821830.0, "step": 22855 }, { "epoch": 2.9075181274646993, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 10.881110191345215, "learning_rate": 5e-05, "loss": 1.0857, "mean_token_accuracy": 0.8822200298309326, "num_tokens": 867859878.0, "step": 22856 }, { "epoch": 2.90764533774329, "ewc_loss": 0.6796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00060272216796875, "grad_norm": 8.674054145812988, "learning_rate": 5e-05, "loss": 1.0275, "mean_token_accuracy": 0.8756994605064392, "num_tokens": 867897212.0, "step": 22857 }, { "epoch": 2.90777254802188, "ewc_loss": 0.796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0007171630859375, "grad_norm": 10.690853118896484, "learning_rate": 5e-05, "loss": 1.1734, "mean_token_accuracy": 0.8665193915367126, "num_tokens": 867929899.0, "step": 22858 }, { "epoch": 2.907899758300471, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.235331535339355, "learning_rate": 5e-05, "loss": 0.9243, "mean_token_accuracy": 0.8846673965454102, "num_tokens": 867962313.0, "step": 22859 }, { "epoch": 2.908026968579061, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 10.977734565734863, "learning_rate": 5e-05, "loss": 1.2548, "mean_token_accuracy": 0.856008768081665, "num_tokens": 867998886.0, "step": 22860 }, { "epoch": 2.908154178857652, "ewc_loss": 0.6640625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00058746337890625, "grad_norm": 8.649563789367676, "learning_rate": 5e-05, "loss": 1.0074, "mean_token_accuracy": 0.877178430557251, "num_tokens": 868039151.0, "step": 22861 }, { "epoch": 2.908281389136242, "ewc_loss": 0.8125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000732421875, "grad_norm": 10.920119285583496, "learning_rate": 5e-05, "loss": 1.145, "mean_token_accuracy": 0.8831372261047363, "num_tokens": 868071021.0, "step": 22862 }, { "epoch": 2.908408599414833, "ewc_loss": 0.72265625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000644683837890625, "grad_norm": 9.167140007019043, "learning_rate": 5e-05, "loss": 1.0573, "mean_token_accuracy": 0.8822870254516602, "num_tokens": 868109872.0, "step": 22863 }, { "epoch": 2.908535809693423, "ewc_loss": 0.734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0006561279296875, "grad_norm": 10.264276504516602, "learning_rate": 5e-05, "loss": 1.0781, "mean_token_accuracy": 0.8801392316818237, "num_tokens": 868150258.0, "step": 22864 }, { "epoch": 2.9086630199720136, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.019371032714844, "learning_rate": 5e-05, "loss": 1.0304, "mean_token_accuracy": 0.8678745031356812, "num_tokens": 868195202.0, "step": 22865 }, { "epoch": 2.908790230250604, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.687163352966309, "learning_rate": 5e-05, "loss": 1.0027, "mean_token_accuracy": 0.888803243637085, "num_tokens": 868226458.0, "step": 22866 }, { "epoch": 2.9089174405291947, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.57250690460205, "learning_rate": 5e-05, "loss": 0.9361, "mean_token_accuracy": 0.8778246641159058, "num_tokens": 868262908.0, "step": 22867 }, { "epoch": 2.909044650807785, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.431303977966309, "learning_rate": 5e-05, "loss": 1.0092, "mean_token_accuracy": 0.8692107200622559, "num_tokens": 868300013.0, "step": 22868 }, { "epoch": 2.9091718610863757, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.0805025100708, "learning_rate": 5e-05, "loss": 0.8642, "mean_token_accuracy": 0.8796088695526123, "num_tokens": 868342794.0, "step": 22869 }, { "epoch": 2.9092990713649662, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 9.543853759765625, "learning_rate": 5e-05, "loss": 1.0102, "mean_token_accuracy": 0.875937819480896, "num_tokens": 868382009.0, "step": 22870 }, { "epoch": 2.9094262816435568, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.823561191558838, "learning_rate": 5e-05, "loss": 0.8925, "mean_token_accuracy": 0.8675528764724731, "num_tokens": 868417770.0, "step": 22871 }, { "epoch": 2.9095534919221473, "ewc_loss": 0.6875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0006103515625, "grad_norm": 9.81786060333252, "learning_rate": 5e-05, "loss": 1.0115, "mean_token_accuracy": 0.8841217160224915, "num_tokens": 868455447.0, "step": 22872 }, { "epoch": 2.909680702200738, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.993839263916016, "learning_rate": 5e-05, "loss": 0.8972, "mean_token_accuracy": 0.8749616146087646, "num_tokens": 868500388.0, "step": 22873 }, { "epoch": 2.9098079124793284, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.816180229187012, "learning_rate": 5e-05, "loss": 1.0479, "mean_token_accuracy": 0.878985583782196, "num_tokens": 868535707.0, "step": 22874 }, { "epoch": 2.909935122757919, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.424527168273926, "learning_rate": 5e-05, "loss": 0.9822, "mean_token_accuracy": 0.863410472869873, "num_tokens": 868576634.0, "step": 22875 }, { "epoch": 2.9100623330365094, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.22059440612793, "learning_rate": 5e-05, "loss": 0.9572, "mean_token_accuracy": 0.8849606513977051, "num_tokens": 868619914.0, "step": 22876 }, { "epoch": 2.9101895433151, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.437674522399902, "learning_rate": 5e-05, "loss": 0.9084, "mean_token_accuracy": 0.882660984992981, "num_tokens": 868652279.0, "step": 22877 }, { "epoch": 2.9103167535936905, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.644389152526855, "learning_rate": 5e-05, "loss": 0.9419, "mean_token_accuracy": 0.8721655011177063, "num_tokens": 868689418.0, "step": 22878 }, { "epoch": 2.910443963872281, "ewc_loss": 0.5, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.050190925598145, "learning_rate": 5e-05, "loss": 0.7925, "mean_token_accuracy": 0.8946403861045837, "num_tokens": 868726695.0, "step": 22879 }, { "epoch": 2.9105711741508715, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.51359748840332, "learning_rate": 5e-05, "loss": 0.9097, "mean_token_accuracy": 0.8749950528144836, "num_tokens": 868766120.0, "step": 22880 }, { "epoch": 2.910698384429462, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.575287342071533, "learning_rate": 5e-05, "loss": 0.7516, "mean_token_accuracy": 0.8906595706939697, "num_tokens": 868806041.0, "step": 22881 }, { "epoch": 2.9108255947080526, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.727160453796387, "learning_rate": 5e-05, "loss": 0.9152, "mean_token_accuracy": 0.8782575726509094, "num_tokens": 868852980.0, "step": 22882 }, { "epoch": 2.9109528049866427, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.290878772735596, "learning_rate": 5e-05, "loss": 0.7916, "mean_token_accuracy": 0.8711122274398804, "num_tokens": 868890951.0, "step": 22883 }, { "epoch": 2.9110800152652336, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.983524322509766, "learning_rate": 5e-05, "loss": 0.9667, "mean_token_accuracy": 0.8712236285209656, "num_tokens": 868926412.0, "step": 22884 }, { "epoch": 2.9112072255438237, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.333337783813477, "learning_rate": 5e-05, "loss": 0.803, "mean_token_accuracy": 0.8711899518966675, "num_tokens": 868965181.0, "step": 22885 }, { "epoch": 2.9113344358224147, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.979748725891113, "learning_rate": 5e-05, "loss": 0.929, "mean_token_accuracy": 0.8840569853782654, "num_tokens": 869004635.0, "step": 22886 }, { "epoch": 2.9114616461010048, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.627589702606201, "learning_rate": 5e-05, "loss": 0.8078, "mean_token_accuracy": 0.8827159404754639, "num_tokens": 869043833.0, "step": 22887 }, { "epoch": 2.9115888563795957, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.606654167175293, "learning_rate": 5e-05, "loss": 0.9325, "mean_token_accuracy": 0.8726253509521484, "num_tokens": 869078868.0, "step": 22888 }, { "epoch": 2.911716066658186, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.882902145385742, "learning_rate": 5e-05, "loss": 0.822, "mean_token_accuracy": 0.8839271068572998, "num_tokens": 869113444.0, "step": 22889 }, { "epoch": 2.9118432769367764, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.025528907775879, "learning_rate": 5e-05, "loss": 0.8263, "mean_token_accuracy": 0.8855244517326355, "num_tokens": 869147741.0, "step": 22890 }, { "epoch": 2.911970487215367, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.93977689743042, "learning_rate": 5e-05, "loss": 0.8066, "mean_token_accuracy": 0.8768981695175171, "num_tokens": 869180247.0, "step": 22891 }, { "epoch": 2.9120976974939574, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.6776862144470215, "learning_rate": 5e-05, "loss": 0.7453, "mean_token_accuracy": 0.8945789337158203, "num_tokens": 869221602.0, "step": 22892 }, { "epoch": 2.912224907772548, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.8407721519470215, "learning_rate": 5e-05, "loss": 0.7723, "mean_token_accuracy": 0.8794301152229309, "num_tokens": 869259601.0, "step": 22893 }, { "epoch": 2.9123521180511385, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.629279613494873, "learning_rate": 5e-05, "loss": 0.7259, "mean_token_accuracy": 0.8918601870536804, "num_tokens": 869303366.0, "step": 22894 }, { "epoch": 2.912479328329729, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.709063529968262, "learning_rate": 5e-05, "loss": 0.7989, "mean_token_accuracy": 0.8636921644210815, "num_tokens": 869338579.0, "step": 22895 }, { "epoch": 2.9126065386083195, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.625406265258789, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.8665742874145508, "num_tokens": 869384129.0, "step": 22896 }, { "epoch": 2.91273374888691, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.682078838348389, "learning_rate": 5e-05, "loss": 0.7391, "mean_token_accuracy": 0.8799185752868652, "num_tokens": 869421752.0, "step": 22897 }, { "epoch": 2.9128609591655006, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.545597553253174, "learning_rate": 5e-05, "loss": 0.7452, "mean_token_accuracy": 0.8749123215675354, "num_tokens": 869459309.0, "step": 22898 }, { "epoch": 2.912988169444091, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.68088436126709, "learning_rate": 5e-05, "loss": 0.7118, "mean_token_accuracy": 0.8897980451583862, "num_tokens": 869497718.0, "step": 22899 }, { "epoch": 2.9131153797226816, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.532606601715088, "learning_rate": 5e-05, "loss": 0.7509, "mean_token_accuracy": 0.8724676966667175, "num_tokens": 869540522.0, "step": 22900 }, { "epoch": 2.913242590001272, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.587093830108643, "learning_rate": 5e-05, "loss": 0.7565, "mean_token_accuracy": 0.8724716901779175, "num_tokens": 869579610.0, "step": 22901 }, { "epoch": 2.9133698002798627, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.629202365875244, "learning_rate": 5e-05, "loss": 0.7074, "mean_token_accuracy": 0.8880943655967712, "num_tokens": 869619186.0, "step": 22902 }, { "epoch": 2.9134970105584532, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.463624000549316, "learning_rate": 5e-05, "loss": 0.7363, "mean_token_accuracy": 0.8718101978302002, "num_tokens": 869657415.0, "step": 22903 }, { "epoch": 2.9136242208370438, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.719282150268555, "learning_rate": 5e-05, "loss": 0.7488, "mean_token_accuracy": 0.8750288486480713, "num_tokens": 869698319.0, "step": 22904 }, { "epoch": 2.9137514311156343, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.356667518615723, "learning_rate": 5e-05, "loss": 0.6893, "mean_token_accuracy": 0.8875914812088013, "num_tokens": 869738630.0, "step": 22905 }, { "epoch": 2.913878641394225, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.8426690101623535, "learning_rate": 5e-05, "loss": 0.7722, "mean_token_accuracy": 0.867129921913147, "num_tokens": 869780713.0, "step": 22906 }, { "epoch": 2.9140058516728153, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 7.155704975128174, "learning_rate": 5e-05, "loss": 0.6677, "mean_token_accuracy": 0.8914675116539001, "num_tokens": 869817046.0, "step": 22907 }, { "epoch": 2.9141330619514054, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 8.10954761505127, "learning_rate": 5e-05, "loss": 0.7437, "mean_token_accuracy": 0.8869115710258484, "num_tokens": 869855099.0, "step": 22908 }, { "epoch": 2.9142602722299964, "ewc_loss": 0.333984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.853541374206543, "learning_rate": 5e-05, "loss": 0.6899, "mean_token_accuracy": 0.8796316385269165, "num_tokens": 869893207.0, "step": 22909 }, { "epoch": 2.9143874825085865, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.501184463500977, "learning_rate": 5e-05, "loss": 0.7942, "mean_token_accuracy": 0.8799570798873901, "num_tokens": 869929925.0, "step": 22910 }, { "epoch": 2.9145146927871775, "ewc_loss": 0.3203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00024127960205078125, "grad_norm": 6.464171886444092, "learning_rate": 5e-05, "loss": 0.6562, "mean_token_accuracy": 0.8821543455123901, "num_tokens": 869974607.0, "step": 22911 }, { "epoch": 2.9146419030657675, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 9.10712718963623, "learning_rate": 5e-05, "loss": 0.8537, "mean_token_accuracy": 0.8800669312477112, "num_tokens": 870013056.0, "step": 22912 }, { "epoch": 2.914769113344358, "ewc_loss": 0.31640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002384185791015625, "grad_norm": 6.080286026000977, "learning_rate": 5e-05, "loss": 0.6824, "mean_token_accuracy": 0.8707380890846252, "num_tokens": 870053735.0, "step": 22913 }, { "epoch": 2.9148963236229486, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.77009391784668, "learning_rate": 5e-05, "loss": 0.9461, "mean_token_accuracy": 0.8859618902206421, "num_tokens": 870093607.0, "step": 22914 }, { "epoch": 2.915023533901539, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.346794605255127, "learning_rate": 5e-05, "loss": 0.7025, "mean_token_accuracy": 0.8788358569145203, "num_tokens": 870133118.0, "step": 22915 }, { "epoch": 2.9151507441801296, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.703554153442383, "learning_rate": 5e-05, "loss": 0.9852, "mean_token_accuracy": 0.8800774812698364, "num_tokens": 870169782.0, "step": 22916 }, { "epoch": 2.91527795445872, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 7.680637359619141, "learning_rate": 5e-05, "loss": 0.8707, "mean_token_accuracy": 0.879719614982605, "num_tokens": 870206891.0, "step": 22917 }, { "epoch": 2.9154051647373107, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.13789176940918, "learning_rate": 5e-05, "loss": 0.8825, "mean_token_accuracy": 0.8710815906524658, "num_tokens": 870253035.0, "step": 22918 }, { "epoch": 2.9155323750159012, "ewc_loss": 0.625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.092328071594238, "learning_rate": 5e-05, "loss": 0.9719, "mean_token_accuracy": 0.878333330154419, "num_tokens": 870292286.0, "step": 22919 }, { "epoch": 2.9156595852944918, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.283527851104736, "learning_rate": 5e-05, "loss": 0.7283, "mean_token_accuracy": 0.868172824382782, "num_tokens": 870335594.0, "step": 22920 }, { "epoch": 2.9157867955730823, "ewc_loss": 0.7109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00063323974609375, "grad_norm": 10.735550880432129, "learning_rate": 5e-05, "loss": 1.0351, "mean_token_accuracy": 0.8821814656257629, "num_tokens": 870379465.0, "step": 22921 }, { "epoch": 2.915914005851673, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 6.601953029632568, "learning_rate": 5e-05, "loss": 0.7517, "mean_token_accuracy": 0.8892765641212463, "num_tokens": 870418060.0, "step": 22922 }, { "epoch": 2.9160412161302633, "ewc_loss": 0.84765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00077056884765625, "grad_norm": 11.475594520568848, "learning_rate": 5e-05, "loss": 1.158, "mean_token_accuracy": 0.8863401412963867, "num_tokens": 870455565.0, "step": 22923 }, { "epoch": 2.916168426408854, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 9.096881866455078, "learning_rate": 5e-05, "loss": 1.1368, "mean_token_accuracy": 0.8770595192909241, "num_tokens": 870492609.0, "step": 22924 }, { "epoch": 2.9162956366874444, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.924833297729492, "learning_rate": 5e-05, "loss": 0.9839, "mean_token_accuracy": 0.8764965534210205, "num_tokens": 870528150.0, "step": 22925 }, { "epoch": 2.916422846966035, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 9.447510719299316, "learning_rate": 5e-05, "loss": 1.0668, "mean_token_accuracy": 0.8832350969314575, "num_tokens": 870568188.0, "step": 22926 }, { "epoch": 2.9165500572446255, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 8.610694885253906, "learning_rate": 5e-05, "loss": 0.8959, "mean_token_accuracy": 0.8831588625907898, "num_tokens": 870601459.0, "step": 22927 }, { "epoch": 2.916677267523216, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.342233657836914, "learning_rate": 5e-05, "loss": 0.9855, "mean_token_accuracy": 0.8818591833114624, "num_tokens": 870638976.0, "step": 22928 }, { "epoch": 2.9168044778018065, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.27836799621582, "learning_rate": 5e-05, "loss": 0.9438, "mean_token_accuracy": 0.8657675981521606, "num_tokens": 870671163.0, "step": 22929 }, { "epoch": 2.916931688080397, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.489773750305176, "learning_rate": 5e-05, "loss": 0.8801, "mean_token_accuracy": 0.8791305422782898, "num_tokens": 870705806.0, "step": 22930 }, { "epoch": 2.917058898358987, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.559349060058594, "learning_rate": 5e-05, "loss": 0.9193, "mean_token_accuracy": 0.8743295073509216, "num_tokens": 870743045.0, "step": 22931 }, { "epoch": 2.917186108637578, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.471473693847656, "learning_rate": 5e-05, "loss": 0.7432, "mean_token_accuracy": 0.8871780633926392, "num_tokens": 870780965.0, "step": 22932 }, { "epoch": 2.917313318916168, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 9.091567993164062, "learning_rate": 5e-05, "loss": 0.8846, "mean_token_accuracy": 0.8935738801956177, "num_tokens": 870814694.0, "step": 22933 }, { "epoch": 2.917440529194759, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.161884307861328, "learning_rate": 5e-05, "loss": 0.7737, "mean_token_accuracy": 0.8731851577758789, "num_tokens": 870845654.0, "step": 22934 }, { "epoch": 2.9175677394733492, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.104740142822266, "learning_rate": 5e-05, "loss": 0.9242, "mean_token_accuracy": 0.8792001008987427, "num_tokens": 870881720.0, "step": 22935 }, { "epoch": 2.91769494975194, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.356939792633057, "learning_rate": 5e-05, "loss": 0.7726, "mean_token_accuracy": 0.8844171166419983, "num_tokens": 870927480.0, "step": 22936 }, { "epoch": 2.9178221600305303, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004634857177734375, "grad_norm": 8.699331283569336, "learning_rate": 5e-05, "loss": 0.8815, "mean_token_accuracy": 0.8845075368881226, "num_tokens": 870962407.0, "step": 22937 }, { "epoch": 2.917949370309121, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.618751049041748, "learning_rate": 5e-05, "loss": 0.7911, "mean_token_accuracy": 0.8886121511459351, "num_tokens": 870994638.0, "step": 22938 }, { "epoch": 2.9180765805877114, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.106522560119629, "learning_rate": 5e-05, "loss": 0.8089, "mean_token_accuracy": 0.8841427564620972, "num_tokens": 871032638.0, "step": 22939 }, { "epoch": 2.918203790866302, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.848018646240234, "learning_rate": 5e-05, "loss": 0.8095, "mean_token_accuracy": 0.8791694045066833, "num_tokens": 871070955.0, "step": 22940 }, { "epoch": 2.9183310011448924, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.672299861907959, "learning_rate": 5e-05, "loss": 0.7908, "mean_token_accuracy": 0.8774536848068237, "num_tokens": 871109282.0, "step": 22941 }, { "epoch": 2.918458211423483, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.878462314605713, "learning_rate": 5e-05, "loss": 0.797, "mean_token_accuracy": 0.8714839816093445, "num_tokens": 871151214.0, "step": 22942 }, { "epoch": 2.9185854217020735, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.671943187713623, "learning_rate": 5e-05, "loss": 0.8033, "mean_token_accuracy": 0.8703147172927856, "num_tokens": 871192909.0, "step": 22943 }, { "epoch": 2.918712631980664, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.549201011657715, "learning_rate": 5e-05, "loss": 0.7622, "mean_token_accuracy": 0.8759509325027466, "num_tokens": 871234952.0, "step": 22944 }, { "epoch": 2.9188398422592545, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.049689292907715, "learning_rate": 5e-05, "loss": 0.7881, "mean_token_accuracy": 0.8796893954277039, "num_tokens": 871270789.0, "step": 22945 }, { "epoch": 2.918967052537845, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.020847797393799, "learning_rate": 5e-05, "loss": 0.7344, "mean_token_accuracy": 0.8719208836555481, "num_tokens": 871312425.0, "step": 22946 }, { "epoch": 2.9190942628164356, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.617996215820312, "learning_rate": 5e-05, "loss": 0.7912, "mean_token_accuracy": 0.8923754692077637, "num_tokens": 871347685.0, "step": 22947 }, { "epoch": 2.919221473095026, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.501131534576416, "learning_rate": 5e-05, "loss": 0.6864, "mean_token_accuracy": 0.8799024820327759, "num_tokens": 871381789.0, "step": 22948 }, { "epoch": 2.9193486833736166, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 9.351818084716797, "learning_rate": 5e-05, "loss": 0.8704, "mean_token_accuracy": 0.891667366027832, "num_tokens": 871417578.0, "step": 22949 }, { "epoch": 2.919475893652207, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.160930156707764, "learning_rate": 5e-05, "loss": 0.693, "mean_token_accuracy": 0.8762021660804749, "num_tokens": 871454816.0, "step": 22950 }, { "epoch": 2.9196031039307977, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 10.062368392944336, "learning_rate": 5e-05, "loss": 0.9846, "mean_token_accuracy": 0.8794921040534973, "num_tokens": 871488898.0, "step": 22951 }, { "epoch": 2.9197303142093882, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 6.6833109855651855, "learning_rate": 5e-05, "loss": 0.7572, "mean_token_accuracy": 0.8861587047576904, "num_tokens": 871531895.0, "step": 22952 }, { "epoch": 2.9198575244879788, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 10.074458122253418, "learning_rate": 5e-05, "loss": 1.0178, "mean_token_accuracy": 0.8859759569168091, "num_tokens": 871566151.0, "step": 22953 }, { "epoch": 2.9199847347665693, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 7.949624061584473, "learning_rate": 5e-05, "loss": 0.9583, "mean_token_accuracy": 0.876395583152771, "num_tokens": 871603777.0, "step": 22954 }, { "epoch": 2.92011194504516, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.96330451965332, "learning_rate": 5e-05, "loss": 0.9231, "mean_token_accuracy": 0.8843889236450195, "num_tokens": 871636118.0, "step": 22955 }, { "epoch": 2.92023915532375, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 8.62520694732666, "learning_rate": 5e-05, "loss": 0.9816, "mean_token_accuracy": 0.8783885836601257, "num_tokens": 871670948.0, "step": 22956 }, { "epoch": 2.920366365602341, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 7.718159198760986, "learning_rate": 5e-05, "loss": 0.8432, "mean_token_accuracy": 0.8716274499893188, "num_tokens": 871707987.0, "step": 22957 }, { "epoch": 2.920493575880931, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.816208839416504, "learning_rate": 5e-05, "loss": 0.9257, "mean_token_accuracy": 0.8858063817024231, "num_tokens": 871753861.0, "step": 22958 }, { "epoch": 2.920620786159522, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.379614353179932, "learning_rate": 5e-05, "loss": 0.8078, "mean_token_accuracy": 0.8754265308380127, "num_tokens": 871792667.0, "step": 22959 }, { "epoch": 2.920747996438112, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.482254981994629, "learning_rate": 5e-05, "loss": 0.9374, "mean_token_accuracy": 0.8639694452285767, "num_tokens": 871830994.0, "step": 22960 }, { "epoch": 2.920875206716703, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 7.545778751373291, "learning_rate": 5e-05, "loss": 0.8376, "mean_token_accuracy": 0.8659491539001465, "num_tokens": 871869984.0, "step": 22961 }, { "epoch": 2.921002416995293, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.066903114318848, "learning_rate": 5e-05, "loss": 0.8633, "mean_token_accuracy": 0.8684760928153992, "num_tokens": 871906757.0, "step": 22962 }, { "epoch": 2.9211296272738836, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.60565710067749, "learning_rate": 5e-05, "loss": 0.776, "mean_token_accuracy": 0.8854607343673706, "num_tokens": 871945387.0, "step": 22963 }, { "epoch": 2.921256837552474, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.953099250793457, "learning_rate": 5e-05, "loss": 0.7995, "mean_token_accuracy": 0.8800926208496094, "num_tokens": 871981456.0, "step": 22964 }, { "epoch": 2.9213840478310646, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.4311113357543945, "learning_rate": 5e-05, "loss": 0.7307, "mean_token_accuracy": 0.8879345059394836, "num_tokens": 872022953.0, "step": 22965 }, { "epoch": 2.921511258109655, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.068442344665527, "learning_rate": 5e-05, "loss": 0.7888, "mean_token_accuracy": 0.8796043395996094, "num_tokens": 872059103.0, "step": 22966 }, { "epoch": 2.9216384683882457, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.2308349609375, "learning_rate": 5e-05, "loss": 0.7203, "mean_token_accuracy": 0.881213903427124, "num_tokens": 872091910.0, "step": 22967 }, { "epoch": 2.9217656786668362, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 8.241316795349121, "learning_rate": 5e-05, "loss": 0.7976, "mean_token_accuracy": 0.8795658349990845, "num_tokens": 872130863.0, "step": 22968 }, { "epoch": 2.9218928889454268, "ewc_loss": 0.375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.020082950592041, "learning_rate": 5e-05, "loss": 0.7504, "mean_token_accuracy": 0.8698482513427734, "num_tokens": 872168104.0, "step": 22969 }, { "epoch": 2.9220200992240173, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.44899845123291, "learning_rate": 5e-05, "loss": 0.8377, "mean_token_accuracy": 0.8733524084091187, "num_tokens": 872197554.0, "step": 22970 }, { "epoch": 2.922147309502608, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.835445880889893, "learning_rate": 5e-05, "loss": 0.7098, "mean_token_accuracy": 0.8791013956069946, "num_tokens": 872232283.0, "step": 22971 }, { "epoch": 2.9222745197811983, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.648567199707031, "learning_rate": 5e-05, "loss": 0.8429, "mean_token_accuracy": 0.8806767463684082, "num_tokens": 872263459.0, "step": 22972 }, { "epoch": 2.922401730059789, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.790892124176025, "learning_rate": 5e-05, "loss": 0.6789, "mean_token_accuracy": 0.8913394212722778, "num_tokens": 872303056.0, "step": 22973 }, { "epoch": 2.9225289403383794, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.701444625854492, "learning_rate": 5e-05, "loss": 0.8826, "mean_token_accuracy": 0.8749666810035706, "num_tokens": 872341030.0, "step": 22974 }, { "epoch": 2.92265615061697, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.922879695892334, "learning_rate": 5e-05, "loss": 0.7527, "mean_token_accuracy": 0.8757204413414001, "num_tokens": 872384006.0, "step": 22975 }, { "epoch": 2.9227833608955605, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.526161193847656, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8695428371429443, "num_tokens": 872425575.0, "step": 22976 }, { "epoch": 2.922910571174151, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.194586277008057, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.8835408687591553, "num_tokens": 872463483.0, "step": 22977 }, { "epoch": 2.9230377814527415, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.175093650817871, "learning_rate": 5e-05, "loss": 0.8324, "mean_token_accuracy": 0.8775526881217957, "num_tokens": 872496066.0, "step": 22978 }, { "epoch": 2.923164991731332, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.521940231323242, "learning_rate": 5e-05, "loss": 0.7758, "mean_token_accuracy": 0.8819980621337891, "num_tokens": 872529330.0, "step": 22979 }, { "epoch": 2.9232922020099226, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.695732593536377, "learning_rate": 5e-05, "loss": 0.7594, "mean_token_accuracy": 0.8839999437332153, "num_tokens": 872563989.0, "step": 22980 }, { "epoch": 2.9234194122885127, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.866713523864746, "learning_rate": 5e-05, "loss": 0.773, "mean_token_accuracy": 0.8862165212631226, "num_tokens": 872601506.0, "step": 22981 }, { "epoch": 2.9235466225671036, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.289506435394287, "learning_rate": 5e-05, "loss": 0.7156, "mean_token_accuracy": 0.8829771280288696, "num_tokens": 872633258.0, "step": 22982 }, { "epoch": 2.9236738328456937, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.155479431152344, "learning_rate": 5e-05, "loss": 0.8287, "mean_token_accuracy": 0.8673334121704102, "num_tokens": 872665803.0, "step": 22983 }, { "epoch": 2.9238010431242847, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.981637954711914, "learning_rate": 5e-05, "loss": 0.7118, "mean_token_accuracy": 0.8773530721664429, "num_tokens": 872703607.0, "step": 22984 }, { "epoch": 2.9239282534028748, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.445645332336426, "learning_rate": 5e-05, "loss": 0.8307, "mean_token_accuracy": 0.8767749071121216, "num_tokens": 872741966.0, "step": 22985 }, { "epoch": 2.9240554636814657, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.724740505218506, "learning_rate": 5e-05, "loss": 0.6923, "mean_token_accuracy": 0.8808634281158447, "num_tokens": 872773052.0, "step": 22986 }, { "epoch": 2.924182673960056, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.770223617553711, "learning_rate": 5e-05, "loss": 0.8826, "mean_token_accuracy": 0.8688310384750366, "num_tokens": 872813667.0, "step": 22987 }, { "epoch": 2.9243098842386464, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.560011386871338, "learning_rate": 5e-05, "loss": 0.7472, "mean_token_accuracy": 0.8685373067855835, "num_tokens": 872854320.0, "step": 22988 }, { "epoch": 2.924437094517237, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 9.00551986694336, "learning_rate": 5e-05, "loss": 0.8906, "mean_token_accuracy": 0.8794234395027161, "num_tokens": 872895190.0, "step": 22989 }, { "epoch": 2.9245643047958274, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.663095951080322, "learning_rate": 5e-05, "loss": 0.7224, "mean_token_accuracy": 0.8809393048286438, "num_tokens": 872932288.0, "step": 22990 }, { "epoch": 2.924691515074418, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.921585083007812, "learning_rate": 5e-05, "loss": 0.9069, "mean_token_accuracy": 0.883247435092926, "num_tokens": 872963617.0, "step": 22991 }, { "epoch": 2.9248187253530085, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.139483451843262, "learning_rate": 5e-05, "loss": 0.8166, "mean_token_accuracy": 0.8721276521682739, "num_tokens": 872997635.0, "step": 22992 }, { "epoch": 2.924945935631599, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.33261775970459, "learning_rate": 5e-05, "loss": 0.92, "mean_token_accuracy": 0.8628209233283997, "num_tokens": 873037275.0, "step": 22993 }, { "epoch": 2.9250731459101895, "ewc_loss": 0.5, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 7.7330217361450195, "learning_rate": 5e-05, "loss": 0.8618, "mean_token_accuracy": 0.8726766705513, "num_tokens": 873074791.0, "step": 22994 }, { "epoch": 2.92520035618878, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 7.561344623565674, "learning_rate": 5e-05, "loss": 0.7647, "mean_token_accuracy": 0.8865348696708679, "num_tokens": 873111521.0, "step": 22995 }, { "epoch": 2.9253275664673706, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.145004272460938, "learning_rate": 5e-05, "loss": 0.8916, "mean_token_accuracy": 0.8601747751235962, "num_tokens": 873150007.0, "step": 22996 }, { "epoch": 2.925454776745961, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.213287353515625, "learning_rate": 5e-05, "loss": 0.7435, "mean_token_accuracy": 0.8793562054634094, "num_tokens": 873188360.0, "step": 22997 }, { "epoch": 2.9255819870245516, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.108304023742676, "learning_rate": 5e-05, "loss": 0.7771, "mean_token_accuracy": 0.8867120742797852, "num_tokens": 873222962.0, "step": 22998 }, { "epoch": 2.925709197303142, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.367124080657959, "learning_rate": 5e-05, "loss": 0.7168, "mean_token_accuracy": 0.8853340744972229, "num_tokens": 873260579.0, "step": 22999 }, { "epoch": 2.9258364075817327, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.690057754516602, "learning_rate": 5e-05, "loss": 0.7243, "mean_token_accuracy": 0.8893342018127441, "num_tokens": 873292862.0, "step": 23000 }, { "epoch": 2.925963617860323, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.789827823638916, "learning_rate": 5e-05, "loss": 0.8411, "mean_token_accuracy": 0.8536442518234253, "num_tokens": 873331824.0, "step": 23001 }, { "epoch": 2.9260908281389137, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.194881916046143, "learning_rate": 5e-05, "loss": 0.7222, "mean_token_accuracy": 0.8782435059547424, "num_tokens": 873364902.0, "step": 23002 }, { "epoch": 2.9262180384175043, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.228740692138672, "learning_rate": 5e-05, "loss": 0.7503, "mean_token_accuracy": 0.8928906917572021, "num_tokens": 873398523.0, "step": 23003 }, { "epoch": 2.926345248696095, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.776176452636719, "learning_rate": 5e-05, "loss": 0.7015, "mean_token_accuracy": 0.8800022006034851, "num_tokens": 873435779.0, "step": 23004 }, { "epoch": 2.9264724589746853, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.700739860534668, "learning_rate": 5e-05, "loss": 0.8109, "mean_token_accuracy": 0.889918327331543, "num_tokens": 873476706.0, "step": 23005 }, { "epoch": 2.9265996692532754, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.3881402015686035, "learning_rate": 5e-05, "loss": 0.6828, "mean_token_accuracy": 0.8810104727745056, "num_tokens": 873507762.0, "step": 23006 }, { "epoch": 2.9267268795318664, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00048828125, "grad_norm": 9.256634712219238, "learning_rate": 5e-05, "loss": 0.8765, "mean_token_accuracy": 0.888216495513916, "num_tokens": 873545692.0, "step": 23007 }, { "epoch": 2.9268540898104565, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.206395626068115, "learning_rate": 5e-05, "loss": 0.6905, "mean_token_accuracy": 0.8812860250473022, "num_tokens": 873589131.0, "step": 23008 }, { "epoch": 2.9269813000890474, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.67429256439209, "learning_rate": 5e-05, "loss": 1.0237, "mean_token_accuracy": 0.8636090755462646, "num_tokens": 873628593.0, "step": 23009 }, { "epoch": 2.9271085103676375, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 6.722855091094971, "learning_rate": 5e-05, "loss": 0.8079, "mean_token_accuracy": 0.8739480972290039, "num_tokens": 873671525.0, "step": 23010 }, { "epoch": 2.927235720646228, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 9.318705558776855, "learning_rate": 5e-05, "loss": 0.9948, "mean_token_accuracy": 0.8794078230857849, "num_tokens": 873710618.0, "step": 23011 }, { "epoch": 2.9273629309248186, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 7.795350074768066, "learning_rate": 5e-05, "loss": 0.9785, "mean_token_accuracy": 0.8642101883888245, "num_tokens": 873751731.0, "step": 23012 }, { "epoch": 2.927490141203409, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.003423690795898, "learning_rate": 5e-05, "loss": 0.865, "mean_token_accuracy": 0.8850976228713989, "num_tokens": 873791948.0, "step": 23013 }, { "epoch": 2.9276173514819996, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.662708282470703, "learning_rate": 5e-05, "loss": 0.9688, "mean_token_accuracy": 0.8725820779800415, "num_tokens": 873834193.0, "step": 23014 }, { "epoch": 2.92774456176059, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.898464202880859, "learning_rate": 5e-05, "loss": 0.7694, "mean_token_accuracy": 0.87331223487854, "num_tokens": 873870035.0, "step": 23015 }, { "epoch": 2.9278717720391807, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.094785690307617, "learning_rate": 5e-05, "loss": 0.9371, "mean_token_accuracy": 0.8858482241630554, "num_tokens": 873902007.0, "step": 23016 }, { "epoch": 2.9279989823177712, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 6.831061363220215, "learning_rate": 5e-05, "loss": 0.8276, "mean_token_accuracy": 0.8570041656494141, "num_tokens": 873947063.0, "step": 23017 }, { "epoch": 2.9281261925963618, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.800589561462402, "learning_rate": 5e-05, "loss": 0.9651, "mean_token_accuracy": 0.868310272693634, "num_tokens": 873987542.0, "step": 23018 }, { "epoch": 2.9282534028749523, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.374662399291992, "learning_rate": 5e-05, "loss": 0.8359, "mean_token_accuracy": 0.879870593547821, "num_tokens": 874020492.0, "step": 23019 }, { "epoch": 2.928380613153543, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.086426734924316, "learning_rate": 5e-05, "loss": 0.8197, "mean_token_accuracy": 0.8883266448974609, "num_tokens": 874061902.0, "step": 23020 }, { "epoch": 2.9285078234321333, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.774868488311768, "learning_rate": 5e-05, "loss": 0.8036, "mean_token_accuracy": 0.8932545185089111, "num_tokens": 874093277.0, "step": 23021 }, { "epoch": 2.928635033710724, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.671234607696533, "learning_rate": 5e-05, "loss": 0.7709, "mean_token_accuracy": 0.8860592842102051, "num_tokens": 874136543.0, "step": 23022 }, { "epoch": 2.9287622439893144, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.786798477172852, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8894079923629761, "num_tokens": 874180933.0, "step": 23023 }, { "epoch": 2.928889454267905, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.671065807342529, "learning_rate": 5e-05, "loss": 0.7672, "mean_token_accuracy": 0.8828223943710327, "num_tokens": 874221980.0, "step": 23024 }, { "epoch": 2.9290166645464955, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.5424017906188965, "learning_rate": 5e-05, "loss": 0.758, "mean_token_accuracy": 0.8823563456535339, "num_tokens": 874255814.0, "step": 23025 }, { "epoch": 2.929143874825086, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.8455023765563965, "learning_rate": 5e-05, "loss": 0.8168, "mean_token_accuracy": 0.8705057501792908, "num_tokens": 874296230.0, "step": 23026 }, { "epoch": 2.9292710851036765, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.2726054191589355, "learning_rate": 5e-05, "loss": 0.7692, "mean_token_accuracy": 0.8665145635604858, "num_tokens": 874331854.0, "step": 23027 }, { "epoch": 2.929398295382267, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.074543952941895, "learning_rate": 5e-05, "loss": 0.782, "mean_token_accuracy": 0.8833647966384888, "num_tokens": 874373348.0, "step": 23028 }, { "epoch": 2.929525505660857, "ewc_loss": 0.375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.022992134094238, "learning_rate": 5e-05, "loss": 0.7009, "mean_token_accuracy": 0.8807870745658875, "num_tokens": 874415226.0, "step": 23029 }, { "epoch": 2.929652715939448, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.316109657287598, "learning_rate": 5e-05, "loss": 0.8751, "mean_token_accuracy": 0.8609775304794312, "num_tokens": 874451894.0, "step": 23030 }, { "epoch": 2.929779926218038, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.771671295166016, "learning_rate": 5e-05, "loss": 0.6944, "mean_token_accuracy": 0.8830823302268982, "num_tokens": 874490248.0, "step": 23031 }, { "epoch": 2.929907136496629, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.651494979858398, "learning_rate": 5e-05, "loss": 0.8402, "mean_token_accuracy": 0.8837311267852783, "num_tokens": 874527421.0, "step": 23032 }, { "epoch": 2.9300343467752192, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 6.518154621124268, "learning_rate": 5e-05, "loss": 0.7202, "mean_token_accuracy": 0.8725169897079468, "num_tokens": 874570834.0, "step": 23033 }, { "epoch": 2.93016155705381, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 9.064658164978027, "learning_rate": 5e-05, "loss": 0.9259, "mean_token_accuracy": 0.8702801465988159, "num_tokens": 874607956.0, "step": 23034 }, { "epoch": 2.9302887673324003, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.378448963165283, "learning_rate": 5e-05, "loss": 0.7241, "mean_token_accuracy": 0.8721863031387329, "num_tokens": 874644599.0, "step": 23035 }, { "epoch": 2.930415977610991, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.355785369873047, "learning_rate": 5e-05, "loss": 0.9805, "mean_token_accuracy": 0.8734065294265747, "num_tokens": 874682214.0, "step": 23036 }, { "epoch": 2.9305431878895813, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.714663505554199, "learning_rate": 5e-05, "loss": 0.808, "mean_token_accuracy": 0.8647747039794922, "num_tokens": 874719298.0, "step": 23037 }, { "epoch": 2.930670398168172, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.085098266601562, "learning_rate": 5e-05, "loss": 0.9418, "mean_token_accuracy": 0.8893648982048035, "num_tokens": 874757391.0, "step": 23038 }, { "epoch": 2.9307976084467624, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.525603294372559, "learning_rate": 5e-05, "loss": 0.8627, "mean_token_accuracy": 0.884038507938385, "num_tokens": 874788307.0, "step": 23039 }, { "epoch": 2.930924818725353, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.086138725280762, "learning_rate": 5e-05, "loss": 0.8671, "mean_token_accuracy": 0.877974808216095, "num_tokens": 874825648.0, "step": 23040 }, { "epoch": 2.9310520290039435, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.261557579040527, "learning_rate": 5e-05, "loss": 0.9185, "mean_token_accuracy": 0.8749128580093384, "num_tokens": 874862275.0, "step": 23041 }, { "epoch": 2.931179239282534, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.100192546844482, "learning_rate": 5e-05, "loss": 0.775, "mean_token_accuracy": 0.8735089898109436, "num_tokens": 874898411.0, "step": 23042 }, { "epoch": 2.9313064495611245, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.732622146606445, "learning_rate": 5e-05, "loss": 0.8936, "mean_token_accuracy": 0.8799872994422913, "num_tokens": 874935599.0, "step": 23043 }, { "epoch": 2.931433659839715, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 6.765331745147705, "learning_rate": 5e-05, "loss": 0.739, "mean_token_accuracy": 0.877263069152832, "num_tokens": 874966302.0, "step": 23044 }, { "epoch": 2.9315608701183056, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.834617614746094, "learning_rate": 5e-05, "loss": 0.8748, "mean_token_accuracy": 0.892380952835083, "num_tokens": 875003158.0, "step": 23045 }, { "epoch": 2.931688080396896, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.9441819190979, "learning_rate": 5e-05, "loss": 0.7664, "mean_token_accuracy": 0.8812689781188965, "num_tokens": 875042163.0, "step": 23046 }, { "epoch": 2.9318152906754866, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.591221809387207, "learning_rate": 5e-05, "loss": 0.9089, "mean_token_accuracy": 0.8768405318260193, "num_tokens": 875081957.0, "step": 23047 }, { "epoch": 2.931942500954077, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.240489482879639, "learning_rate": 5e-05, "loss": 0.8, "mean_token_accuracy": 0.8796302080154419, "num_tokens": 875122286.0, "step": 23048 }, { "epoch": 2.9320697112326677, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.300657272338867, "learning_rate": 5e-05, "loss": 0.9225, "mean_token_accuracy": 0.8639477491378784, "num_tokens": 875163320.0, "step": 23049 }, { "epoch": 2.932196921511258, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.4126715660095215, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.8866868019104004, "num_tokens": 875203819.0, "step": 23050 }, { "epoch": 2.9323241317898487, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.095315933227539, "learning_rate": 5e-05, "loss": 0.8274, "mean_token_accuracy": 0.8806586861610413, "num_tokens": 875239028.0, "step": 23051 }, { "epoch": 2.9324513420684393, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.401819229125977, "learning_rate": 5e-05, "loss": 0.8264, "mean_token_accuracy": 0.8702840209007263, "num_tokens": 875276324.0, "step": 23052 }, { "epoch": 2.93257855234703, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.083678245544434, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8857232332229614, "num_tokens": 875316918.0, "step": 23053 }, { "epoch": 2.93270576262562, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.243077278137207, "learning_rate": 5e-05, "loss": 0.7728, "mean_token_accuracy": 0.8731029033660889, "num_tokens": 875351009.0, "step": 23054 }, { "epoch": 2.932832972904211, "ewc_loss": 0.482421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 8.228081703186035, "learning_rate": 5e-05, "loss": 0.794, "mean_token_accuracy": 0.8873329162597656, "num_tokens": 875386988.0, "step": 23055 }, { "epoch": 2.932960183182801, "ewc_loss": 0.390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 7.01239538192749, "learning_rate": 5e-05, "loss": 0.7215, "mean_token_accuracy": 0.8824577927589417, "num_tokens": 875429607.0, "step": 23056 }, { "epoch": 2.933087393461392, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.462911605834961, "learning_rate": 5e-05, "loss": 0.8353, "mean_token_accuracy": 0.8871567845344543, "num_tokens": 875461071.0, "step": 23057 }, { "epoch": 2.933214603739982, "ewc_loss": 0.375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 6.814817428588867, "learning_rate": 5e-05, "loss": 0.707, "mean_token_accuracy": 0.8833896517753601, "num_tokens": 875498146.0, "step": 23058 }, { "epoch": 2.933341814018573, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.667546272277832, "learning_rate": 5e-05, "loss": 0.8733, "mean_token_accuracy": 0.8797868490219116, "num_tokens": 875536155.0, "step": 23059 }, { "epoch": 2.933469024297163, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 6.801955223083496, "learning_rate": 5e-05, "loss": 0.7378, "mean_token_accuracy": 0.8749626278877258, "num_tokens": 875569868.0, "step": 23060 }, { "epoch": 2.9335962345757536, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.67963981628418, "learning_rate": 5e-05, "loss": 0.8951, "mean_token_accuracy": 0.8781026601791382, "num_tokens": 875609298.0, "step": 23061 }, { "epoch": 2.933723444854344, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.031606674194336, "learning_rate": 5e-05, "loss": 0.7256, "mean_token_accuracy": 0.8901350498199463, "num_tokens": 875642837.0, "step": 23062 }, { "epoch": 2.9338506551329346, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.387621879577637, "learning_rate": 5e-05, "loss": 0.8926, "mean_token_accuracy": 0.8746623396873474, "num_tokens": 875683273.0, "step": 23063 }, { "epoch": 2.933977865411525, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.427211761474609, "learning_rate": 5e-05, "loss": 0.8476, "mean_token_accuracy": 0.864677369594574, "num_tokens": 875717461.0, "step": 23064 }, { "epoch": 2.9341050756901157, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 7.828866958618164, "learning_rate": 5e-05, "loss": 0.8332, "mean_token_accuracy": 0.8704873323440552, "num_tokens": 875756975.0, "step": 23065 }, { "epoch": 2.934232285968706, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 7.815395355224609, "learning_rate": 5e-05, "loss": 0.7995, "mean_token_accuracy": 0.8809466361999512, "num_tokens": 875793414.0, "step": 23066 }, { "epoch": 2.9343594962472968, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.3321990966796875, "learning_rate": 5e-05, "loss": 0.797, "mean_token_accuracy": 0.8604452610015869, "num_tokens": 875822477.0, "step": 23067 }, { "epoch": 2.9344867065258873, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.084671974182129, "learning_rate": 5e-05, "loss": 0.8062, "mean_token_accuracy": 0.8768250942230225, "num_tokens": 875856220.0, "step": 23068 }, { "epoch": 2.934613916804478, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.05902624130249, "learning_rate": 5e-05, "loss": 0.7201, "mean_token_accuracy": 0.8797404170036316, "num_tokens": 875890572.0, "step": 23069 }, { "epoch": 2.9347411270830683, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.228689193725586, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.8867236971855164, "num_tokens": 875928231.0, "step": 23070 }, { "epoch": 2.934868337361659, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.9509148597717285, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.8713563680648804, "num_tokens": 875966211.0, "step": 23071 }, { "epoch": 2.9349955476402494, "ewc_loss": 0.470703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000392913818359375, "grad_norm": 8.29545783996582, "learning_rate": 5e-05, "loss": 0.8061, "mean_token_accuracy": 0.8809118270874023, "num_tokens": 876004857.0, "step": 23072 }, { "epoch": 2.93512275791884, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.910791873931885, "learning_rate": 5e-05, "loss": 0.7547, "mean_token_accuracy": 0.8644206523895264, "num_tokens": 876041966.0, "step": 23073 }, { "epoch": 2.9352499681974304, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.366632461547852, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.8836368322372437, "num_tokens": 876075732.0, "step": 23074 }, { "epoch": 2.935377178476021, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.879708766937256, "learning_rate": 5e-05, "loss": 0.7344, "mean_token_accuracy": 0.87242192029953, "num_tokens": 876116582.0, "step": 23075 }, { "epoch": 2.9355043887546115, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.415735244750977, "learning_rate": 5e-05, "loss": 0.7867, "mean_token_accuracy": 0.8931161165237427, "num_tokens": 876150122.0, "step": 23076 }, { "epoch": 2.935631599033202, "ewc_loss": 0.375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.9060468673706055, "learning_rate": 5e-05, "loss": 0.6946, "mean_token_accuracy": 0.8881834745407104, "num_tokens": 876187902.0, "step": 23077 }, { "epoch": 2.9357588093117926, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.369892120361328, "learning_rate": 5e-05, "loss": 0.85, "mean_token_accuracy": 0.877133846282959, "num_tokens": 876224304.0, "step": 23078 }, { "epoch": 2.9358860195903826, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.0477070808410645, "learning_rate": 5e-05, "loss": 0.7842, "mean_token_accuracy": 0.8623754978179932, "num_tokens": 876261436.0, "step": 23079 }, { "epoch": 2.9360132298689736, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.168176651000977, "learning_rate": 5e-05, "loss": 0.8411, "mean_token_accuracy": 0.8712273836135864, "num_tokens": 876299996.0, "step": 23080 }, { "epoch": 2.9361404401475637, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.259323596954346, "learning_rate": 5e-05, "loss": 0.7746, "mean_token_accuracy": 0.8714610934257507, "num_tokens": 876337793.0, "step": 23081 }, { "epoch": 2.9362676504261547, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.903273105621338, "learning_rate": 5e-05, "loss": 0.7981, "mean_token_accuracy": 0.8805886507034302, "num_tokens": 876371837.0, "step": 23082 }, { "epoch": 2.9363948607047448, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.422615051269531, "learning_rate": 5e-05, "loss": 0.7421, "mean_token_accuracy": 0.8877221345901489, "num_tokens": 876414402.0, "step": 23083 }, { "epoch": 2.9365220709833357, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.7283830642700195, "learning_rate": 5e-05, "loss": 0.7276, "mean_token_accuracy": 0.8931950926780701, "num_tokens": 876451790.0, "step": 23084 }, { "epoch": 2.936649281261926, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.46368932723999, "learning_rate": 5e-05, "loss": 0.7868, "mean_token_accuracy": 0.8634511828422546, "num_tokens": 876486891.0, "step": 23085 }, { "epoch": 2.9367764915405163, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.708863258361816, "learning_rate": 5e-05, "loss": 0.7308, "mean_token_accuracy": 0.8877816200256348, "num_tokens": 876525560.0, "step": 23086 }, { "epoch": 2.936903701819107, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.384503364562988, "learning_rate": 5e-05, "loss": 0.7507, "mean_token_accuracy": 0.8728873133659363, "num_tokens": 876569037.0, "step": 23087 }, { "epoch": 2.9370309120976974, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.804275035858154, "learning_rate": 5e-05, "loss": 0.769, "mean_token_accuracy": 0.8759576678276062, "num_tokens": 876605354.0, "step": 23088 }, { "epoch": 2.937158122376288, "ewc_loss": 0.375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.179597854614258, "learning_rate": 5e-05, "loss": 0.6931, "mean_token_accuracy": 0.8860595226287842, "num_tokens": 876644251.0, "step": 23089 }, { "epoch": 2.9372853326548785, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 8.054502487182617, "learning_rate": 5e-05, "loss": 0.7562, "mean_token_accuracy": 0.8878365755081177, "num_tokens": 876686558.0, "step": 23090 }, { "epoch": 2.937412542933469, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.875586032867432, "learning_rate": 5e-05, "loss": 0.6959, "mean_token_accuracy": 0.8796757459640503, "num_tokens": 876718629.0, "step": 23091 }, { "epoch": 2.9375397532120595, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.425716400146484, "learning_rate": 5e-05, "loss": 0.8349, "mean_token_accuracy": 0.8664068579673767, "num_tokens": 876758037.0, "step": 23092 }, { "epoch": 2.93766696349065, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.51569128036499, "learning_rate": 5e-05, "loss": 0.6851, "mean_token_accuracy": 0.8807957172393799, "num_tokens": 876795201.0, "step": 23093 }, { "epoch": 2.9377941737692406, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.95267391204834, "learning_rate": 5e-05, "loss": 0.83, "mean_token_accuracy": 0.8913509845733643, "num_tokens": 876833465.0, "step": 23094 }, { "epoch": 2.937921384047831, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 6.2096686363220215, "learning_rate": 5e-05, "loss": 0.6924, "mean_token_accuracy": 0.877313494682312, "num_tokens": 876869915.0, "step": 23095 }, { "epoch": 2.9380485943264216, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.471616744995117, "learning_rate": 5e-05, "loss": 0.9331, "mean_token_accuracy": 0.8856600522994995, "num_tokens": 876906811.0, "step": 23096 }, { "epoch": 2.938175804605012, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 6.434879302978516, "learning_rate": 5e-05, "loss": 0.7602, "mean_token_accuracy": 0.8693975210189819, "num_tokens": 876944247.0, "step": 23097 }, { "epoch": 2.9383030148836027, "ewc_loss": 0.625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.32773208618164, "learning_rate": 5e-05, "loss": 0.9649, "mean_token_accuracy": 0.878430187702179, "num_tokens": 876978557.0, "step": 23098 }, { "epoch": 2.938430225162193, "ewc_loss": 0.546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 7.515811443328857, "learning_rate": 5e-05, "loss": 0.9328, "mean_token_accuracy": 0.8634545803070068, "num_tokens": 877021241.0, "step": 23099 }, { "epoch": 2.9385574354407837, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 7.863751411437988, "learning_rate": 5e-05, "loss": 0.8762, "mean_token_accuracy": 0.8701263666152954, "num_tokens": 877060777.0, "step": 23100 }, { "epoch": 2.9386846457193743, "ewc_loss": 0.625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.933295249938965, "learning_rate": 5e-05, "loss": 1.001, "mean_token_accuracy": 0.8700751066207886, "num_tokens": 877096955.0, "step": 23101 }, { "epoch": 2.938811855997965, "ewc_loss": 0.33984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002613067626953125, "grad_norm": 6.056490421295166, "learning_rate": 5e-05, "loss": 0.7049, "mean_token_accuracy": 0.8763786554336548, "num_tokens": 877130202.0, "step": 23102 }, { "epoch": 2.9389390662765553, "ewc_loss": 0.75, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 10.875619888305664, "learning_rate": 5e-05, "loss": 1.1094, "mean_token_accuracy": 0.872100830078125, "num_tokens": 877168777.0, "step": 23103 }, { "epoch": 2.9390662765551454, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 6.117952346801758, "learning_rate": 5e-05, "loss": 0.778, "mean_token_accuracy": 0.8760089874267578, "num_tokens": 877209737.0, "step": 23104 }, { "epoch": 2.9391934868337364, "ewc_loss": 1.0703125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0009918212890625, "grad_norm": 12.569733619689941, "learning_rate": 5e-05, "loss": 1.4087, "mean_token_accuracy": 0.8808280229568481, "num_tokens": 877245826.0, "step": 23105 }, { "epoch": 2.9393206971123265, "ewc_loss": 1.0234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000942230224609375, "grad_norm": 9.720815658569336, "learning_rate": 5e-05, "loss": 1.3679, "mean_token_accuracy": 0.8831534385681152, "num_tokens": 877285249.0, "step": 23106 }, { "epoch": 2.9394479073909174, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 11.153578758239746, "learning_rate": 5e-05, "loss": 1.0338, "mean_token_accuracy": 0.8947888612747192, "num_tokens": 877321971.0, "step": 23107 }, { "epoch": 2.9395751176695075, "ewc_loss": 0.94921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0008697509765625, "grad_norm": 9.755542755126953, "learning_rate": 5e-05, "loss": 1.3333, "mean_token_accuracy": 0.873685359954834, "num_tokens": 877361479.0, "step": 23108 }, { "epoch": 2.939702327948098, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.29647159576416, "learning_rate": 5e-05, "loss": 1.0843, "mean_token_accuracy": 0.8771184682846069, "num_tokens": 877400166.0, "step": 23109 }, { "epoch": 2.9398295382266886, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 9.095945358276367, "learning_rate": 5e-05, "loss": 1.0421, "mean_token_accuracy": 0.8804722428321838, "num_tokens": 877437755.0, "step": 23110 }, { "epoch": 2.939956748505279, "ewc_loss": 0.671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.277669906616211, "learning_rate": 5e-05, "loss": 1.063, "mean_token_accuracy": 0.8639159202575684, "num_tokens": 877471496.0, "step": 23111 }, { "epoch": 2.9400839587838696, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.734024047851562, "learning_rate": 5e-05, "loss": 0.9413, "mean_token_accuracy": 0.8923792839050293, "num_tokens": 877504576.0, "step": 23112 }, { "epoch": 2.94021116906246, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.578310012817383, "learning_rate": 5e-05, "loss": 0.958, "mean_token_accuracy": 0.8730775713920593, "num_tokens": 877542795.0, "step": 23113 }, { "epoch": 2.9403383793410507, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.359051704406738, "learning_rate": 5e-05, "loss": 0.9316, "mean_token_accuracy": 0.8744763135910034, "num_tokens": 877581731.0, "step": 23114 }, { "epoch": 2.940465589619641, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.147022247314453, "learning_rate": 5e-05, "loss": 0.8909, "mean_token_accuracy": 0.8729110956192017, "num_tokens": 877618405.0, "step": 23115 }, { "epoch": 2.9405927998982317, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.12990665435791, "learning_rate": 5e-05, "loss": 0.8845, "mean_token_accuracy": 0.8811341524124146, "num_tokens": 877657792.0, "step": 23116 }, { "epoch": 2.9407200101768223, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.749330043792725, "learning_rate": 5e-05, "loss": 0.7722, "mean_token_accuracy": 0.8910127878189087, "num_tokens": 877688965.0, "step": 23117 }, { "epoch": 2.940847220455413, "ewc_loss": 0.5, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.132774353027344, "learning_rate": 5e-05, "loss": 0.8142, "mean_token_accuracy": 0.8890507817268372, "num_tokens": 877726595.0, "step": 23118 }, { "epoch": 2.9409744307340033, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.366861343383789, "learning_rate": 5e-05, "loss": 0.7652, "mean_token_accuracy": 0.8781770467758179, "num_tokens": 877771713.0, "step": 23119 }, { "epoch": 2.941101641012594, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.343644142150879, "learning_rate": 5e-05, "loss": 0.822, "mean_token_accuracy": 0.8836731910705566, "num_tokens": 877805048.0, "step": 23120 }, { "epoch": 2.9412288512911844, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.973321437835693, "learning_rate": 5e-05, "loss": 0.7029, "mean_token_accuracy": 0.8830933570861816, "num_tokens": 877840469.0, "step": 23121 }, { "epoch": 2.941356061569775, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.802403450012207, "learning_rate": 5e-05, "loss": 0.8986, "mean_token_accuracy": 0.8689408302307129, "num_tokens": 877876747.0, "step": 23122 }, { "epoch": 2.9414832718483654, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.492166519165039, "learning_rate": 5e-05, "loss": 0.6913, "mean_token_accuracy": 0.8835150003433228, "num_tokens": 877913871.0, "step": 23123 }, { "epoch": 2.941610482126956, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.658942222595215, "learning_rate": 5e-05, "loss": 1.0575, "mean_token_accuracy": 0.8644590377807617, "num_tokens": 877954195.0, "step": 23124 }, { "epoch": 2.9417376924055465, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.360812187194824, "learning_rate": 5e-05, "loss": 0.7602, "mean_token_accuracy": 0.8656324744224548, "num_tokens": 877989124.0, "step": 23125 }, { "epoch": 2.941864902684137, "ewc_loss": 0.74609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000667572021484375, "grad_norm": 10.554133415222168, "learning_rate": 5e-05, "loss": 1.0958, "mean_token_accuracy": 0.8783101439476013, "num_tokens": 878030188.0, "step": 23126 }, { "epoch": 2.941992112962727, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.553964614868164, "learning_rate": 5e-05, "loss": 0.8897, "mean_token_accuracy": 0.896425187587738, "num_tokens": 878073839.0, "step": 23127 }, { "epoch": 2.942119323241318, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.291446685791016, "learning_rate": 5e-05, "loss": 1.0741, "mean_token_accuracy": 0.8800028562545776, "num_tokens": 878111446.0, "step": 23128 }, { "epoch": 2.942246533519908, "ewc_loss": 0.75, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00067138671875, "grad_norm": 8.879183769226074, "learning_rate": 5e-05, "loss": 1.1456, "mean_token_accuracy": 0.8603653311729431, "num_tokens": 878154543.0, "step": 23129 }, { "epoch": 2.942373743798499, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 8.680752754211426, "learning_rate": 5e-05, "loss": 0.9072, "mean_token_accuracy": 0.8850078582763672, "num_tokens": 878194265.0, "step": 23130 }, { "epoch": 2.9425009540770892, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 9.369794845581055, "learning_rate": 5e-05, "loss": 1.1018, "mean_token_accuracy": 0.8634809851646423, "num_tokens": 878231475.0, "step": 23131 }, { "epoch": 2.94262816435568, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.313708305358887, "learning_rate": 5e-05, "loss": 0.8161, "mean_token_accuracy": 0.8804126381874084, "num_tokens": 878269492.0, "step": 23132 }, { "epoch": 2.9427553746342703, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.988840103149414, "learning_rate": 5e-05, "loss": 1.0752, "mean_token_accuracy": 0.8671206831932068, "num_tokens": 878307047.0, "step": 23133 }, { "epoch": 2.942882584912861, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 7.289880752563477, "learning_rate": 5e-05, "loss": 0.8713, "mean_token_accuracy": 0.8747021555900574, "num_tokens": 878343971.0, "step": 23134 }, { "epoch": 2.9430097951914513, "ewc_loss": 0.70703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000629425048828125, "grad_norm": 10.080608367919922, "learning_rate": 5e-05, "loss": 1.0358, "mean_token_accuracy": 0.8842679262161255, "num_tokens": 878380858.0, "step": 23135 }, { "epoch": 2.943137005470042, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.099076271057129, "learning_rate": 5e-05, "loss": 0.9815, "mean_token_accuracy": 0.8849945068359375, "num_tokens": 878420854.0, "step": 23136 }, { "epoch": 2.9432642157486324, "ewc_loss": 0.625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.282215118408203, "learning_rate": 5e-05, "loss": 0.9997, "mean_token_accuracy": 0.8693119287490845, "num_tokens": 878455465.0, "step": 23137 }, { "epoch": 2.943391426027223, "ewc_loss": 0.65234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000576019287109375, "grad_norm": 8.439927101135254, "learning_rate": 5e-05, "loss": 0.9652, "mean_token_accuracy": 0.8901865482330322, "num_tokens": 878487122.0, "step": 23138 }, { "epoch": 2.9435186363058135, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004425048828125, "grad_norm": 8.343542098999023, "learning_rate": 5e-05, "loss": 0.9112, "mean_token_accuracy": 0.8653873801231384, "num_tokens": 878523881.0, "step": 23139 }, { "epoch": 2.943645846584404, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.41185188293457, "learning_rate": 5e-05, "loss": 0.9725, "mean_token_accuracy": 0.8690273761749268, "num_tokens": 878564351.0, "step": 23140 }, { "epoch": 2.9437730568629945, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.787023544311523, "learning_rate": 5e-05, "loss": 0.7874, "mean_token_accuracy": 0.8891422152519226, "num_tokens": 878605759.0, "step": 23141 }, { "epoch": 2.943900267141585, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.3494291305542, "learning_rate": 5e-05, "loss": 0.8713, "mean_token_accuracy": 0.8883590698242188, "num_tokens": 878644975.0, "step": 23142 }, { "epoch": 2.9440274774201756, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.469882488250732, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8850225806236267, "num_tokens": 878679882.0, "step": 23143 }, { "epoch": 2.944154687698766, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.41014575958252, "learning_rate": 5e-05, "loss": 0.8282, "mean_token_accuracy": 0.892287015914917, "num_tokens": 878723539.0, "step": 23144 }, { "epoch": 2.9442818979773566, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.196028232574463, "learning_rate": 5e-05, "loss": 0.7379, "mean_token_accuracy": 0.8882604837417603, "num_tokens": 878761993.0, "step": 23145 }, { "epoch": 2.944409108255947, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.641267776489258, "learning_rate": 5e-05, "loss": 0.8713, "mean_token_accuracy": 0.8787368535995483, "num_tokens": 878794355.0, "step": 23146 }, { "epoch": 2.9445363185345377, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 6.990684509277344, "learning_rate": 5e-05, "loss": 0.8064, "mean_token_accuracy": 0.8613473176956177, "num_tokens": 878828477.0, "step": 23147 }, { "epoch": 2.944663528813128, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.823246955871582, "learning_rate": 5e-05, "loss": 0.9141, "mean_token_accuracy": 0.8704747557640076, "num_tokens": 878868134.0, "step": 23148 }, { "epoch": 2.9447907390917187, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 6.986894607543945, "learning_rate": 5e-05, "loss": 0.7636, "mean_token_accuracy": 0.8841623067855835, "num_tokens": 878908912.0, "step": 23149 }, { "epoch": 2.9449179493703093, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004749298095703125, "grad_norm": 8.813536643981934, "learning_rate": 5e-05, "loss": 0.899, "mean_token_accuracy": 0.8772283792495728, "num_tokens": 878947354.0, "step": 23150 }, { "epoch": 2.9450451596489, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.166329383850098, "learning_rate": 5e-05, "loss": 0.7914, "mean_token_accuracy": 0.8837333917617798, "num_tokens": 878985538.0, "step": 23151 }, { "epoch": 2.94517236992749, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 8.602714538574219, "learning_rate": 5e-05, "loss": 0.9041, "mean_token_accuracy": 0.8718157410621643, "num_tokens": 879028884.0, "step": 23152 }, { "epoch": 2.945299580206081, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.393764495849609, "learning_rate": 5e-05, "loss": 0.8167, "mean_token_accuracy": 0.8811427354812622, "num_tokens": 879064388.0, "step": 23153 }, { "epoch": 2.945426790484671, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.20491886138916, "learning_rate": 5e-05, "loss": 0.8074, "mean_token_accuracy": 0.8838230967521667, "num_tokens": 879100353.0, "step": 23154 }, { "epoch": 2.945554000763262, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 7.613016605377197, "learning_rate": 5e-05, "loss": 0.8135, "mean_token_accuracy": 0.885722279548645, "num_tokens": 879138447.0, "step": 23155 }, { "epoch": 2.945681211041852, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.820608139038086, "learning_rate": 5e-05, "loss": 0.8275, "mean_token_accuracy": 0.8665779232978821, "num_tokens": 879172825.0, "step": 23156 }, { "epoch": 2.945808421320443, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.758906364440918, "learning_rate": 5e-05, "loss": 0.8237, "mean_token_accuracy": 0.8726040124893188, "num_tokens": 879212086.0, "step": 23157 }, { "epoch": 2.945935631599033, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.58253288269043, "learning_rate": 5e-05, "loss": 0.7455, "mean_token_accuracy": 0.8824408054351807, "num_tokens": 879247689.0, "step": 23158 }, { "epoch": 2.9460628418776236, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.8135986328125, "learning_rate": 5e-05, "loss": 0.7664, "mean_token_accuracy": 0.8871662020683289, "num_tokens": 879283510.0, "step": 23159 }, { "epoch": 2.946190052156214, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.436820983886719, "learning_rate": 5e-05, "loss": 0.7427, "mean_token_accuracy": 0.8796817660331726, "num_tokens": 879322037.0, "step": 23160 }, { "epoch": 2.9463172624348046, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.893983840942383, "learning_rate": 5e-05, "loss": 0.7999, "mean_token_accuracy": 0.8707137107849121, "num_tokens": 879366400.0, "step": 23161 }, { "epoch": 2.946444472713395, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.282769203186035, "learning_rate": 5e-05, "loss": 0.7848, "mean_token_accuracy": 0.8607885241508484, "num_tokens": 879397768.0, "step": 23162 }, { "epoch": 2.9465716829919857, "ewc_loss": 0.439453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003604888916015625, "grad_norm": 8.023534774780273, "learning_rate": 5e-05, "loss": 0.7824, "mean_token_accuracy": 0.878829836845398, "num_tokens": 879436751.0, "step": 23163 }, { "epoch": 2.946698893270576, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.132171630859375, "learning_rate": 5e-05, "loss": 0.7697, "mean_token_accuracy": 0.8620513677597046, "num_tokens": 879477750.0, "step": 23164 }, { "epoch": 2.9468261035491667, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.181443214416504, "learning_rate": 5e-05, "loss": 0.791, "mean_token_accuracy": 0.8756229281425476, "num_tokens": 879516049.0, "step": 23165 }, { "epoch": 2.9469533138277573, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.975562572479248, "learning_rate": 5e-05, "loss": 0.7457, "mean_token_accuracy": 0.8655718564987183, "num_tokens": 879553176.0, "step": 23166 }, { "epoch": 2.947080524106348, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 8.358251571655273, "learning_rate": 5e-05, "loss": 0.7668, "mean_token_accuracy": 0.8905258774757385, "num_tokens": 879588435.0, "step": 23167 }, { "epoch": 2.9472077343849383, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.820576190948486, "learning_rate": 5e-05, "loss": 0.6697, "mean_token_accuracy": 0.890810489654541, "num_tokens": 879623232.0, "step": 23168 }, { "epoch": 2.947334944663529, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.566784858703613, "learning_rate": 5e-05, "loss": 0.861, "mean_token_accuracy": 0.8698193430900574, "num_tokens": 879669470.0, "step": 23169 }, { "epoch": 2.9474621549421194, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.722569942474365, "learning_rate": 5e-05, "loss": 0.6828, "mean_token_accuracy": 0.8905824422836304, "num_tokens": 879708173.0, "step": 23170 }, { "epoch": 2.94758936522071, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.695717811584473, "learning_rate": 5e-05, "loss": 0.8837, "mean_token_accuracy": 0.872970700263977, "num_tokens": 879749086.0, "step": 23171 }, { "epoch": 2.9477165754993004, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 6.776974678039551, "learning_rate": 5e-05, "loss": 0.7383, "mean_token_accuracy": 0.8773499131202698, "num_tokens": 879786521.0, "step": 23172 }, { "epoch": 2.947843785777891, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 8.638882637023926, "learning_rate": 5e-05, "loss": 0.8413, "mean_token_accuracy": 0.8853231072425842, "num_tokens": 879820332.0, "step": 23173 }, { "epoch": 2.9479709960564815, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.040221214294434, "learning_rate": 5e-05, "loss": 0.7466, "mean_token_accuracy": 0.8767699003219604, "num_tokens": 879859224.0, "step": 23174 }, { "epoch": 2.948098206335072, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.27318286895752, "learning_rate": 5e-05, "loss": 0.8073, "mean_token_accuracy": 0.8840367197990417, "num_tokens": 879892177.0, "step": 23175 }, { "epoch": 2.9482254166136626, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.447174549102783, "learning_rate": 5e-05, "loss": 0.7264, "mean_token_accuracy": 0.8982646465301514, "num_tokens": 879924195.0, "step": 23176 }, { "epoch": 2.9483526268922526, "ewc_loss": 0.427734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003490447998046875, "grad_norm": 7.704777240753174, "learning_rate": 5e-05, "loss": 0.7846, "mean_token_accuracy": 0.8736116886138916, "num_tokens": 879965679.0, "step": 23177 }, { "epoch": 2.9484798371708436, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.888322830200195, "learning_rate": 5e-05, "loss": 0.7721, "mean_token_accuracy": 0.885416567325592, "num_tokens": 880006441.0, "step": 23178 }, { "epoch": 2.9486070474494337, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.20694637298584, "learning_rate": 5e-05, "loss": 0.7599, "mean_token_accuracy": 0.8668532371520996, "num_tokens": 880042823.0, "step": 23179 }, { "epoch": 2.9487342577280247, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.201128005981445, "learning_rate": 5e-05, "loss": 0.7979, "mean_token_accuracy": 0.8796507716178894, "num_tokens": 880073949.0, "step": 23180 }, { "epoch": 2.9488614680066147, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.9050703048706055, "learning_rate": 5e-05, "loss": 0.6557, "mean_token_accuracy": 0.8964698314666748, "num_tokens": 880112921.0, "step": 23181 }, { "epoch": 2.9489886782852053, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.443864822387695, "learning_rate": 5e-05, "loss": 0.8505, "mean_token_accuracy": 0.8713424205780029, "num_tokens": 880152384.0, "step": 23182 }, { "epoch": 2.949115888563796, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.780611515045166, "learning_rate": 5e-05, "loss": 0.7228, "mean_token_accuracy": 0.8771189451217651, "num_tokens": 880187107.0, "step": 23183 }, { "epoch": 2.9492430988423863, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.506418228149414, "learning_rate": 5e-05, "loss": 0.8227, "mean_token_accuracy": 0.8821017742156982, "num_tokens": 880226072.0, "step": 23184 }, { "epoch": 2.949370309120977, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.844027042388916, "learning_rate": 5e-05, "loss": 0.7089, "mean_token_accuracy": 0.8794305920600891, "num_tokens": 880259607.0, "step": 23185 }, { "epoch": 2.9494975193995674, "ewc_loss": 0.486328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000408172607421875, "grad_norm": 8.421520233154297, "learning_rate": 5e-05, "loss": 0.8527, "mean_token_accuracy": 0.8726222515106201, "num_tokens": 880301238.0, "step": 23186 }, { "epoch": 2.949624729678158, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.024497985839844, "learning_rate": 5e-05, "loss": 0.7428, "mean_token_accuracy": 0.8711438775062561, "num_tokens": 880336449.0, "step": 23187 }, { "epoch": 2.9497519399567484, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.190189361572266, "learning_rate": 5e-05, "loss": 0.8315, "mean_token_accuracy": 0.8745124340057373, "num_tokens": 880373527.0, "step": 23188 }, { "epoch": 2.949879150235339, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.256659030914307, "learning_rate": 5e-05, "loss": 0.7402, "mean_token_accuracy": 0.8810387253761292, "num_tokens": 880412676.0, "step": 23189 }, { "epoch": 2.9500063605139295, "ewc_loss": 0.447265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000370025634765625, "grad_norm": 7.907388210296631, "learning_rate": 5e-05, "loss": 0.7802, "mean_token_accuracy": 0.8819519281387329, "num_tokens": 880448171.0, "step": 23190 }, { "epoch": 2.95013357079252, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.47130012512207, "learning_rate": 5e-05, "loss": 0.7574, "mean_token_accuracy": 0.8758726716041565, "num_tokens": 880487056.0, "step": 23191 }, { "epoch": 2.9502607810711106, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.668613433837891, "learning_rate": 5e-05, "loss": 0.7036, "mean_token_accuracy": 0.896927535533905, "num_tokens": 880520990.0, "step": 23192 }, { "epoch": 2.950387991349701, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.56088399887085, "learning_rate": 5e-05, "loss": 0.7802, "mean_token_accuracy": 0.8669219613075256, "num_tokens": 880557919.0, "step": 23193 }, { "epoch": 2.9505152016282916, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 7.581706523895264, "learning_rate": 5e-05, "loss": 0.7061, "mean_token_accuracy": 0.8910348415374756, "num_tokens": 880590578.0, "step": 23194 }, { "epoch": 2.950642411906882, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.509716510772705, "learning_rate": 5e-05, "loss": 0.7253, "mean_token_accuracy": 0.8814575672149658, "num_tokens": 880631607.0, "step": 23195 }, { "epoch": 2.9507696221854727, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.685650825500488, "learning_rate": 5e-05, "loss": 0.7332, "mean_token_accuracy": 0.8820013999938965, "num_tokens": 880674301.0, "step": 23196 }, { "epoch": 2.950896832464063, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 7.3013458251953125, "learning_rate": 5e-05, "loss": 0.7256, "mean_token_accuracy": 0.8762500286102295, "num_tokens": 880714254.0, "step": 23197 }, { "epoch": 2.9510240427426537, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 7.94015645980835, "learning_rate": 5e-05, "loss": 0.6975, "mean_token_accuracy": 0.8974936008453369, "num_tokens": 880751562.0, "step": 23198 }, { "epoch": 2.9511512530212443, "ewc_loss": 0.353515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000274658203125, "grad_norm": 6.9556074142456055, "learning_rate": 5e-05, "loss": 0.7289, "mean_token_accuracy": 0.8676367402076721, "num_tokens": 880785062.0, "step": 23199 }, { "epoch": 2.951278463299835, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 8.380545616149902, "learning_rate": 5e-05, "loss": 0.8053, "mean_token_accuracy": 0.8754214644432068, "num_tokens": 880820673.0, "step": 23200 }, { "epoch": 2.9514056735784253, "ewc_loss": 0.328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002498626708984375, "grad_norm": 6.472466468811035, "learning_rate": 5e-05, "loss": 0.6818, "mean_token_accuracy": 0.8756335973739624, "num_tokens": 880860707.0, "step": 23201 }, { "epoch": 2.9515328838570154, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 9.082789421081543, "learning_rate": 5e-05, "loss": 0.8511, "mean_token_accuracy": 0.8833068609237671, "num_tokens": 880903504.0, "step": 23202 }, { "epoch": 2.9516600941356064, "ewc_loss": 0.310546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000232696533203125, "grad_norm": 5.936174392700195, "learning_rate": 5e-05, "loss": 0.6197, "mean_token_accuracy": 0.8898657560348511, "num_tokens": 880943414.0, "step": 23203 }, { "epoch": 2.9517873044141965, "ewc_loss": 0.640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000560760498046875, "grad_norm": 10.015069961547852, "learning_rate": 5e-05, "loss": 0.971, "mean_token_accuracy": 0.8848764896392822, "num_tokens": 880977736.0, "step": 23204 }, { "epoch": 2.9519145146927874, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.053396701812744, "learning_rate": 5e-05, "loss": 0.6647, "mean_token_accuracy": 0.8939691781997681, "num_tokens": 881019793.0, "step": 23205 }, { "epoch": 2.9520417249713775, "ewc_loss": 0.73046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000652313232421875, "grad_norm": 10.451610565185547, "learning_rate": 5e-05, "loss": 1.0442, "mean_token_accuracy": 0.8904139399528503, "num_tokens": 881057703.0, "step": 23206 }, { "epoch": 2.952168935249968, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 7.742908000946045, "learning_rate": 5e-05, "loss": 0.9551, "mean_token_accuracy": 0.8850473165512085, "num_tokens": 881093649.0, "step": 23207 }, { "epoch": 2.9522961455285586, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 9.062775611877441, "learning_rate": 5e-05, "loss": 1.0009, "mean_token_accuracy": 0.8611361980438232, "num_tokens": 881125620.0, "step": 23208 }, { "epoch": 2.952423355807149, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 9.2611083984375, "learning_rate": 5e-05, "loss": 1.0465, "mean_token_accuracy": 0.8906662464141846, "num_tokens": 881165313.0, "step": 23209 }, { "epoch": 2.9525505660857396, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.703085899353027, "learning_rate": 5e-05, "loss": 0.7214, "mean_token_accuracy": 0.8803746700286865, "num_tokens": 881195390.0, "step": 23210 }, { "epoch": 2.95267777636433, "ewc_loss": 0.8984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000820159912109375, "grad_norm": 11.612558364868164, "learning_rate": 5e-05, "loss": 1.2653, "mean_token_accuracy": 0.8723852634429932, "num_tokens": 881242005.0, "step": 23211 }, { "epoch": 2.9528049866429207, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 7.501461029052734, "learning_rate": 5e-05, "loss": 0.9162, "mean_token_accuracy": 0.8816138505935669, "num_tokens": 881284587.0, "step": 23212 }, { "epoch": 2.952932196921511, "ewc_loss": 1.0, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000926971435546875, "grad_norm": 12.936081886291504, "learning_rate": 5e-05, "loss": 1.3313, "mean_token_accuracy": 0.8834381103515625, "num_tokens": 881324174.0, "step": 23213 }, { "epoch": 2.9530594072001017, "ewc_loss": 1.0859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.001007080078125, "grad_norm": 10.552628517150879, "learning_rate": 5e-05, "loss": 1.4801, "mean_token_accuracy": 0.8605104684829712, "num_tokens": 881360727.0, "step": 23214 }, { "epoch": 2.9531866174786923, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 11.737727165222168, "learning_rate": 5e-05, "loss": 1.1429, "mean_token_accuracy": 0.878456711769104, "num_tokens": 881400482.0, "step": 23215 }, { "epoch": 2.953313827757283, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 9.570664405822754, "learning_rate": 5e-05, "loss": 1.1301, "mean_token_accuracy": 0.8799546360969543, "num_tokens": 881440157.0, "step": 23216 }, { "epoch": 2.9534410380358733, "ewc_loss": 0.90625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000827789306640625, "grad_norm": 11.971098899841309, "learning_rate": 5e-05, "loss": 1.229, "mean_token_accuracy": 0.8876379728317261, "num_tokens": 881486814.0, "step": 23217 }, { "epoch": 2.953568248314464, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.050628662109375, "learning_rate": 5e-05, "loss": 1.0238, "mean_token_accuracy": 0.8739086389541626, "num_tokens": 881526997.0, "step": 23218 }, { "epoch": 2.9536954585930544, "ewc_loss": 0.9453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000865936279296875, "grad_norm": 12.185142517089844, "learning_rate": 5e-05, "loss": 1.2857, "mean_token_accuracy": 0.8798988461494446, "num_tokens": 881569103.0, "step": 23219 }, { "epoch": 2.953822668871645, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 10.05778980255127, "learning_rate": 5e-05, "loss": 1.1251, "mean_token_accuracy": 0.8827624320983887, "num_tokens": 881604734.0, "step": 23220 }, { "epoch": 2.9539498791502354, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 11.422262191772461, "learning_rate": 5e-05, "loss": 1.114, "mean_token_accuracy": 0.8815207481384277, "num_tokens": 881645788.0, "step": 23221 }, { "epoch": 2.954077089428826, "ewc_loss": 0.77734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000698089599609375, "grad_norm": 10.038141250610352, "learning_rate": 5e-05, "loss": 1.0781, "mean_token_accuracy": 0.8910698890686035, "num_tokens": 881674738.0, "step": 23222 }, { "epoch": 2.9542042997074165, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.805707931518555, "learning_rate": 5e-05, "loss": 1.0695, "mean_token_accuracy": 0.8785611391067505, "num_tokens": 881711685.0, "step": 23223 }, { "epoch": 2.954331509986007, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.357115745544434, "learning_rate": 5e-05, "loss": 0.9614, "mean_token_accuracy": 0.8923105001449585, "num_tokens": 881748165.0, "step": 23224 }, { "epoch": 2.954458720264597, "ewc_loss": 0.69921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000621795654296875, "grad_norm": 10.573076248168945, "learning_rate": 5e-05, "loss": 1.0245, "mean_token_accuracy": 0.8860572576522827, "num_tokens": 881781332.0, "step": 23225 }, { "epoch": 2.954585930543188, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.828177452087402, "learning_rate": 5e-05, "loss": 0.9236, "mean_token_accuracy": 0.881840169429779, "num_tokens": 881814526.0, "step": 23226 }, { "epoch": 2.954713140821778, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 10.641483306884766, "learning_rate": 5e-05, "loss": 1.098, "mean_token_accuracy": 0.8679102063179016, "num_tokens": 881856625.0, "step": 23227 }, { "epoch": 2.954840351100369, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.64814281463623, "learning_rate": 5e-05, "loss": 0.9111, "mean_token_accuracy": 0.8850551843643188, "num_tokens": 881895092.0, "step": 23228 }, { "epoch": 2.954967561378959, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 10.873616218566895, "learning_rate": 5e-05, "loss": 1.0749, "mean_token_accuracy": 0.8830698728561401, "num_tokens": 881936789.0, "step": 23229 }, { "epoch": 2.95509477165755, "ewc_loss": 0.609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.915329933166504, "learning_rate": 5e-05, "loss": 0.928, "mean_token_accuracy": 0.8901773691177368, "num_tokens": 881973471.0, "step": 23230 }, { "epoch": 2.9552219819361403, "ewc_loss": 0.7421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00066375732421875, "grad_norm": 10.808492660522461, "learning_rate": 5e-05, "loss": 1.0961, "mean_token_accuracy": 0.8752103447914124, "num_tokens": 882011039.0, "step": 23231 }, { "epoch": 2.955349192214731, "ewc_loss": 0.62890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00054931640625, "grad_norm": 9.228832244873047, "learning_rate": 5e-05, "loss": 0.9657, "mean_token_accuracy": 0.8805543184280396, "num_tokens": 882049221.0, "step": 23232 }, { "epoch": 2.9554764024933213, "ewc_loss": 0.703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0006256103515625, "grad_norm": 10.313409805297852, "learning_rate": 5e-05, "loss": 1.0288, "mean_token_accuracy": 0.8849521279335022, "num_tokens": 882086527.0, "step": 23233 }, { "epoch": 2.955603612771912, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.138503074645996, "learning_rate": 5e-05, "loss": 0.8879, "mean_token_accuracy": 0.9008831977844238, "num_tokens": 882118375.0, "step": 23234 }, { "epoch": 2.9557308230505024, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.752304077148438, "learning_rate": 5e-05, "loss": 0.9913, "mean_token_accuracy": 0.8781484961509705, "num_tokens": 882150604.0, "step": 23235 }, { "epoch": 2.955858033329093, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.714345932006836, "learning_rate": 5e-05, "loss": 0.8913, "mean_token_accuracy": 0.8791971206665039, "num_tokens": 882185822.0, "step": 23236 }, { "epoch": 2.9559852436076834, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.482897758483887, "learning_rate": 5e-05, "loss": 0.9543, "mean_token_accuracy": 0.882786214351654, "num_tokens": 882224829.0, "step": 23237 }, { "epoch": 2.956112453886274, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.187247276306152, "learning_rate": 5e-05, "loss": 0.8642, "mean_token_accuracy": 0.8641767501831055, "num_tokens": 882263107.0, "step": 23238 }, { "epoch": 2.9562396641648645, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.648727416992188, "learning_rate": 5e-05, "loss": 0.9739, "mean_token_accuracy": 0.8788572549819946, "num_tokens": 882294302.0, "step": 23239 }, { "epoch": 2.956366874443455, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.940412521362305, "learning_rate": 5e-05, "loss": 0.789, "mean_token_accuracy": 0.8844116926193237, "num_tokens": 882327081.0, "step": 23240 }, { "epoch": 2.9564940847220456, "ewc_loss": 0.66796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000591278076171875, "grad_norm": 9.949782371520996, "learning_rate": 5e-05, "loss": 0.9972, "mean_token_accuracy": 0.8821904063224792, "num_tokens": 882361793.0, "step": 23241 }, { "epoch": 2.956621295000636, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.223576545715332, "learning_rate": 5e-05, "loss": 0.7818, "mean_token_accuracy": 0.8971513509750366, "num_tokens": 882395255.0, "step": 23242 }, { "epoch": 2.9567485052792266, "ewc_loss": 0.66015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000583648681640625, "grad_norm": 9.802106857299805, "learning_rate": 5e-05, "loss": 1.0005, "mean_token_accuracy": 0.8824210166931152, "num_tokens": 882431915.0, "step": 23243 }, { "epoch": 2.956875715557817, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.609436988830566, "learning_rate": 5e-05, "loss": 0.8864, "mean_token_accuracy": 0.876127302646637, "num_tokens": 882472688.0, "step": 23244 }, { "epoch": 2.9570029258364077, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 9.092456817626953, "learning_rate": 5e-05, "loss": 0.9166, "mean_token_accuracy": 0.8818645477294922, "num_tokens": 882506851.0, "step": 23245 }, { "epoch": 2.957130136114998, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.473860740661621, "learning_rate": 5e-05, "loss": 0.8867, "mean_token_accuracy": 0.8719019889831543, "num_tokens": 882541935.0, "step": 23246 }, { "epoch": 2.9572573463935887, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.5213623046875, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.8857236504554749, "num_tokens": 882577860.0, "step": 23247 }, { "epoch": 2.9573845566721793, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.030162811279297, "learning_rate": 5e-05, "loss": 0.8468, "mean_token_accuracy": 0.865197479724884, "num_tokens": 882616182.0, "step": 23248 }, { "epoch": 2.95751176695077, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.400664329528809, "learning_rate": 5e-05, "loss": 0.9004, "mean_token_accuracy": 0.8620284199714661, "num_tokens": 882658847.0, "step": 23249 }, { "epoch": 2.95763897722936, "ewc_loss": 0.416015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003376007080078125, "grad_norm": 7.5744309425354, "learning_rate": 5e-05, "loss": 0.7606, "mean_token_accuracy": 0.8765158653259277, "num_tokens": 882699842.0, "step": 23250 }, { "epoch": 2.957766187507951, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.50377368927002, "learning_rate": 5e-05, "loss": 0.8213, "mean_token_accuracy": 0.88680100440979, "num_tokens": 882733143.0, "step": 23251 }, { "epoch": 2.957893397786541, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.351648330688477, "learning_rate": 5e-05, "loss": 0.735, "mean_token_accuracy": 0.8745671510696411, "num_tokens": 882763233.0, "step": 23252 }, { "epoch": 2.958020608065132, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.584534645080566, "learning_rate": 5e-05, "loss": 0.8456, "mean_token_accuracy": 0.8815299868583679, "num_tokens": 882807612.0, "step": 23253 }, { "epoch": 2.958147818343722, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.2388715744018555, "learning_rate": 5e-05, "loss": 0.6565, "mean_token_accuracy": 0.901708722114563, "num_tokens": 882842135.0, "step": 23254 }, { "epoch": 2.958275028622313, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.687474250793457, "learning_rate": 5e-05, "loss": 0.8433, "mean_token_accuracy": 0.8845683336257935, "num_tokens": 882876687.0, "step": 23255 }, { "epoch": 2.958402238900903, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.260536193847656, "learning_rate": 5e-05, "loss": 0.7338, "mean_token_accuracy": 0.877595067024231, "num_tokens": 882919142.0, "step": 23256 }, { "epoch": 2.9585294491794936, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004444122314453125, "grad_norm": 8.673675537109375, "learning_rate": 5e-05, "loss": 0.8445, "mean_token_accuracy": 0.8890026211738586, "num_tokens": 882955301.0, "step": 23257 }, { "epoch": 2.958656659458084, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.469593048095703, "learning_rate": 5e-05, "loss": 0.7015, "mean_token_accuracy": 0.8945457935333252, "num_tokens": 882985436.0, "step": 23258 }, { "epoch": 2.9587838697366746, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 8.2902250289917, "learning_rate": 5e-05, "loss": 0.8383, "mean_token_accuracy": 0.8769248723983765, "num_tokens": 883028082.0, "step": 23259 }, { "epoch": 2.958911080015265, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.757309913635254, "learning_rate": 5e-05, "loss": 0.7665, "mean_token_accuracy": 0.877853274345398, "num_tokens": 883063771.0, "step": 23260 }, { "epoch": 2.9590382902938557, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.737543106079102, "learning_rate": 5e-05, "loss": 0.7425, "mean_token_accuracy": 0.8831223249435425, "num_tokens": 883096742.0, "step": 23261 }, { "epoch": 2.959165500572446, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.975731372833252, "learning_rate": 5e-05, "loss": 0.7801, "mean_token_accuracy": 0.8754756450653076, "num_tokens": 883137916.0, "step": 23262 }, { "epoch": 2.9592927108510367, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.34002685546875, "learning_rate": 5e-05, "loss": 0.7349, "mean_token_accuracy": 0.8757413029670715, "num_tokens": 883172730.0, "step": 23263 }, { "epoch": 2.9594199211296273, "ewc_loss": 0.431640625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 8.2048921585083, "learning_rate": 5e-05, "loss": 0.7767, "mean_token_accuracy": 0.8785227537155151, "num_tokens": 883212139.0, "step": 23264 }, { "epoch": 2.959547131408218, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.998770236968994, "learning_rate": 5e-05, "loss": 0.677, "mean_token_accuracy": 0.8881816267967224, "num_tokens": 883244866.0, "step": 23265 }, { "epoch": 2.9596743416868083, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.556329727172852, "learning_rate": 5e-05, "loss": 0.8048, "mean_token_accuracy": 0.8839669227600098, "num_tokens": 883277595.0, "step": 23266 }, { "epoch": 2.959801551965399, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002689361572265625, "grad_norm": 6.757327079772949, "learning_rate": 5e-05, "loss": 0.6942, "mean_token_accuracy": 0.8774663209915161, "num_tokens": 883311246.0, "step": 23267 }, { "epoch": 2.9599287622439894, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.771127700805664, "learning_rate": 5e-05, "loss": 0.852, "mean_token_accuracy": 0.8793827295303345, "num_tokens": 883353351.0, "step": 23268 }, { "epoch": 2.96005597252258, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.898470878601074, "learning_rate": 5e-05, "loss": 0.6873, "mean_token_accuracy": 0.8852144479751587, "num_tokens": 883392245.0, "step": 23269 }, { "epoch": 2.9601831828011704, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 8.52591609954834, "learning_rate": 5e-05, "loss": 0.8071, "mean_token_accuracy": 0.8860714435577393, "num_tokens": 883430404.0, "step": 23270 }, { "epoch": 2.960310393079761, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.397242069244385, "learning_rate": 5e-05, "loss": 0.7966, "mean_token_accuracy": 0.8595373034477234, "num_tokens": 883477894.0, "step": 23271 }, { "epoch": 2.9604376033583515, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.862646579742432, "learning_rate": 5e-05, "loss": 0.812, "mean_token_accuracy": 0.8632586002349854, "num_tokens": 883520396.0, "step": 23272 }, { "epoch": 2.960564813636942, "ewc_loss": 0.423828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.916004180908203, "learning_rate": 5e-05, "loss": 0.7817, "mean_token_accuracy": 0.8756043910980225, "num_tokens": 883559120.0, "step": 23273 }, { "epoch": 2.9606920239155325, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 7.25041389465332, "learning_rate": 5e-05, "loss": 0.7315, "mean_token_accuracy": 0.87474524974823, "num_tokens": 883603061.0, "step": 23274 }, { "epoch": 2.9608192341941226, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 8.317500114440918, "learning_rate": 5e-05, "loss": 0.788, "mean_token_accuracy": 0.87593013048172, "num_tokens": 883637674.0, "step": 23275 }, { "epoch": 2.9609464444727136, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.915615558624268, "learning_rate": 5e-05, "loss": 0.6855, "mean_token_accuracy": 0.883948564529419, "num_tokens": 883674631.0, "step": 23276 }, { "epoch": 2.9610736547513037, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.491482734680176, "learning_rate": 5e-05, "loss": 0.8045, "mean_token_accuracy": 0.8785613179206848, "num_tokens": 883716823.0, "step": 23277 }, { "epoch": 2.9612008650298947, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.949000358581543, "learning_rate": 5e-05, "loss": 0.7269, "mean_token_accuracy": 0.8757843971252441, "num_tokens": 883751849.0, "step": 23278 }, { "epoch": 2.9613280753084847, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 8.299129486083984, "learning_rate": 5e-05, "loss": 0.7668, "mean_token_accuracy": 0.8882585167884827, "num_tokens": 883790698.0, "step": 23279 }, { "epoch": 2.9614552855870753, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.28243350982666, "learning_rate": 5e-05, "loss": 0.6809, "mean_token_accuracy": 0.8937718868255615, "num_tokens": 883823238.0, "step": 23280 }, { "epoch": 2.961582495865666, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.853570938110352, "learning_rate": 5e-05, "loss": 0.756, "mean_token_accuracy": 0.879694938659668, "num_tokens": 883863036.0, "step": 23281 }, { "epoch": 2.9617097061442563, "ewc_loss": 0.396484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003185272216796875, "grad_norm": 7.680145740509033, "learning_rate": 5e-05, "loss": 0.7589, "mean_token_accuracy": 0.8684923648834229, "num_tokens": 883902556.0, "step": 23282 }, { "epoch": 2.961836916422847, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.434739589691162, "learning_rate": 5e-05, "loss": 0.761, "mean_token_accuracy": 0.8643353581428528, "num_tokens": 883937039.0, "step": 23283 }, { "epoch": 2.9619641267014374, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.948893070220947, "learning_rate": 5e-05, "loss": 0.7443, "mean_token_accuracy": 0.8792085647583008, "num_tokens": 883969047.0, "step": 23284 }, { "epoch": 2.962091336980028, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.247586250305176, "learning_rate": 5e-05, "loss": 0.7253, "mean_token_accuracy": 0.874975323677063, "num_tokens": 884013431.0, "step": 23285 }, { "epoch": 2.9622185472586184, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 8.001006126403809, "learning_rate": 5e-05, "loss": 0.7582, "mean_token_accuracy": 0.8768033385276794, "num_tokens": 884053714.0, "step": 23286 }, { "epoch": 2.962345757537209, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.255996227264404, "learning_rate": 5e-05, "loss": 0.7286, "mean_token_accuracy": 0.8720669150352478, "num_tokens": 884094624.0, "step": 23287 }, { "epoch": 2.9624729678157995, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 7.874061584472656, "learning_rate": 5e-05, "loss": 0.7468, "mean_token_accuracy": 0.8772299289703369, "num_tokens": 884131759.0, "step": 23288 }, { "epoch": 2.96260017809439, "ewc_loss": 0.373046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.408934593200684, "learning_rate": 5e-05, "loss": 0.7004, "mean_token_accuracy": 0.8843910694122314, "num_tokens": 884175333.0, "step": 23289 }, { "epoch": 2.9627273883729806, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.672119617462158, "learning_rate": 5e-05, "loss": 0.6813, "mean_token_accuracy": 0.8922241926193237, "num_tokens": 884209778.0, "step": 23290 }, { "epoch": 2.962854598651571, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.588283061981201, "learning_rate": 5e-05, "loss": 0.7145, "mean_token_accuracy": 0.8803107738494873, "num_tokens": 884245581.0, "step": 23291 }, { "epoch": 2.9629818089301616, "ewc_loss": 0.375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00029754638671875, "grad_norm": 7.484357833862305, "learning_rate": 5e-05, "loss": 0.6879, "mean_token_accuracy": 0.8872331380844116, "num_tokens": 884284925.0, "step": 23292 }, { "epoch": 2.963109019208752, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.719250679016113, "learning_rate": 5e-05, "loss": 0.7034, "mean_token_accuracy": 0.8857530355453491, "num_tokens": 884326056.0, "step": 23293 }, { "epoch": 2.9632362294873427, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 7.374722003936768, "learning_rate": 5e-05, "loss": 0.7008, "mean_token_accuracy": 0.8790105581283569, "num_tokens": 884368169.0, "step": 23294 }, { "epoch": 2.963363439765933, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.77329158782959, "learning_rate": 5e-05, "loss": 0.7352, "mean_token_accuracy": 0.8795827627182007, "num_tokens": 884402785.0, "step": 23295 }, { "epoch": 2.9634906500445237, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.3408660888671875, "learning_rate": 5e-05, "loss": 0.6839, "mean_token_accuracy": 0.8811976909637451, "num_tokens": 884444755.0, "step": 23296 }, { "epoch": 2.9636178603231143, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.789351940155029, "learning_rate": 5e-05, "loss": 0.716, "mean_token_accuracy": 0.8843055963516235, "num_tokens": 884486299.0, "step": 23297 }, { "epoch": 2.963745070601705, "ewc_loss": 0.35546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002765655517578125, "grad_norm": 7.301141262054443, "learning_rate": 5e-05, "loss": 0.7546, "mean_token_accuracy": 0.8594155311584473, "num_tokens": 884532076.0, "step": 23298 }, { "epoch": 2.9638722808802953, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.883450031280518, "learning_rate": 5e-05, "loss": 0.7366, "mean_token_accuracy": 0.8800899982452393, "num_tokens": 884569518.0, "step": 23299 }, { "epoch": 2.9639994911588854, "ewc_loss": 0.341796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000263214111328125, "grad_norm": 7.1359357833862305, "learning_rate": 5e-05, "loss": 0.6751, "mean_token_accuracy": 0.879981517791748, "num_tokens": 884601173.0, "step": 23300 }, { "epoch": 2.9641267014374764, "ewc_loss": 0.412109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003337860107421875, "grad_norm": 8.10610294342041, "learning_rate": 5e-05, "loss": 0.7505, "mean_token_accuracy": 0.8810634016990662, "num_tokens": 884637558.0, "step": 23301 }, { "epoch": 2.9642539117160664, "ewc_loss": 0.32421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002460479736328125, "grad_norm": 6.806163787841797, "learning_rate": 5e-05, "loss": 0.6615, "mean_token_accuracy": 0.8805311918258667, "num_tokens": 884673377.0, "step": 23302 }, { "epoch": 2.9643811219946574, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.627464294433594, "learning_rate": 5e-05, "loss": 0.7577, "mean_token_accuracy": 0.894575834274292, "num_tokens": 884710334.0, "step": 23303 }, { "epoch": 2.9645083322732475, "ewc_loss": 0.296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002193450927734375, "grad_norm": 6.159586429595947, "learning_rate": 5e-05, "loss": 0.6059, "mean_token_accuracy": 0.8907694220542908, "num_tokens": 884741515.0, "step": 23304 }, { "epoch": 2.964635542551838, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 9.709796905517578, "learning_rate": 5e-05, "loss": 0.9217, "mean_token_accuracy": 0.8706006407737732, "num_tokens": 884782204.0, "step": 23305 }, { "epoch": 2.9647627528304286, "ewc_loss": 0.271484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00019359588623046875, "grad_norm": 5.29046106338501, "learning_rate": 5e-05, "loss": 0.5872, "mean_token_accuracy": 0.888335108757019, "num_tokens": 884819613.0, "step": 23306 }, { "epoch": 2.964889963109019, "ewc_loss": 0.82421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 11.604963302612305, "learning_rate": 5e-05, "loss": 1.1763, "mean_token_accuracy": 0.8757334351539612, "num_tokens": 884857381.0, "step": 23307 }, { "epoch": 2.9650171733876096, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 6.350548267364502, "learning_rate": 5e-05, "loss": 0.8711, "mean_token_accuracy": 0.8579325675964355, "num_tokens": 884893562.0, "step": 23308 }, { "epoch": 2.9651443836662, "ewc_loss": 1.140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00106048583984375, "grad_norm": 13.268850326538086, "learning_rate": 5e-05, "loss": 1.4874, "mean_token_accuracy": 0.8794665336608887, "num_tokens": 884934359.0, "step": 23309 }, { "epoch": 2.9652715939447907, "ewc_loss": 1.1328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0010528564453125, "grad_norm": 10.415224075317383, "learning_rate": 5e-05, "loss": 1.4457, "mean_token_accuracy": 0.8908425569534302, "num_tokens": 884974446.0, "step": 23310 }, { "epoch": 2.965398804223381, "ewc_loss": 0.76953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000690460205078125, "grad_norm": 11.700569152832031, "learning_rate": 5e-05, "loss": 1.1368, "mean_token_accuracy": 0.8726804852485657, "num_tokens": 885016447.0, "step": 23311 }, { "epoch": 2.9655260145019717, "ewc_loss": 0.984375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00090789794921875, "grad_norm": 10.070592880249023, "learning_rate": 5e-05, "loss": 1.3425, "mean_token_accuracy": 0.8724956512451172, "num_tokens": 885055064.0, "step": 23312 }, { "epoch": 2.9656532247805623, "ewc_loss": 0.78125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000701904296875, "grad_norm": 11.046222686767578, "learning_rate": 5e-05, "loss": 1.1213, "mean_token_accuracy": 0.8832035064697266, "num_tokens": 885092380.0, "step": 23313 }, { "epoch": 2.965780435059153, "ewc_loss": 0.71484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000637054443359375, "grad_norm": 9.107451438903809, "learning_rate": 5e-05, "loss": 1.0342, "mean_token_accuracy": 0.8867996335029602, "num_tokens": 885130717.0, "step": 23314 }, { "epoch": 2.9659076453377433, "ewc_loss": 0.69140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000614166259765625, "grad_norm": 9.987009048461914, "learning_rate": 5e-05, "loss": 1.0329, "mean_token_accuracy": 0.8805159330368042, "num_tokens": 885168291.0, "step": 23315 }, { "epoch": 2.966034855616334, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 8.721835136413574, "learning_rate": 5e-05, "loss": 1.0334, "mean_token_accuracy": 0.8625086545944214, "num_tokens": 885203839.0, "step": 23316 }, { "epoch": 2.9661620658949244, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.167305946350098, "learning_rate": 5e-05, "loss": 0.9849, "mean_token_accuracy": 0.8701471090316772, "num_tokens": 885239336.0, "step": 23317 }, { "epoch": 2.966289276173515, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.389948844909668, "learning_rate": 5e-05, "loss": 0.9686, "mean_token_accuracy": 0.8635066151618958, "num_tokens": 885275964.0, "step": 23318 }, { "epoch": 2.9664164864521054, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.564538955688477, "learning_rate": 5e-05, "loss": 0.8865, "mean_token_accuracy": 0.8770252466201782, "num_tokens": 885307444.0, "step": 23319 }, { "epoch": 2.966543696730696, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000453948974609375, "grad_norm": 8.16611099243164, "learning_rate": 5e-05, "loss": 0.8462, "mean_token_accuracy": 0.889999508857727, "num_tokens": 885345379.0, "step": 23320 }, { "epoch": 2.9666709070092865, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.190025329589844, "learning_rate": 5e-05, "loss": 0.8225, "mean_token_accuracy": 0.8848457932472229, "num_tokens": 885384105.0, "step": 23321 }, { "epoch": 2.966798117287877, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 7.904576778411865, "learning_rate": 5e-05, "loss": 0.816, "mean_token_accuracy": 0.8790420889854431, "num_tokens": 885420374.0, "step": 23322 }, { "epoch": 2.966925327566467, "ewc_loss": 0.47265625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003948211669921875, "grad_norm": 8.07883358001709, "learning_rate": 5e-05, "loss": 0.8128, "mean_token_accuracy": 0.8823679685592651, "num_tokens": 885453450.0, "step": 23323 }, { "epoch": 2.967052537845058, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.575376033782959, "learning_rate": 5e-05, "loss": 0.7756, "mean_token_accuracy": 0.883266031742096, "num_tokens": 885486331.0, "step": 23324 }, { "epoch": 2.967179748123648, "ewc_loss": 0.474609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000396728515625, "grad_norm": 8.197259902954102, "learning_rate": 5e-05, "loss": 0.7886, "mean_token_accuracy": 0.8914586305618286, "num_tokens": 885520474.0, "step": 23325 }, { "epoch": 2.967306958402239, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.203172206878662, "learning_rate": 5e-05, "loss": 0.7514, "mean_token_accuracy": 0.8765398263931274, "num_tokens": 885563128.0, "step": 23326 }, { "epoch": 2.967434168680829, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.545472145080566, "learning_rate": 5e-05, "loss": 0.7871, "mean_token_accuracy": 0.8990726470947266, "num_tokens": 885604219.0, "step": 23327 }, { "epoch": 2.96756137895942, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002841949462890625, "grad_norm": 6.7295241355896, "learning_rate": 5e-05, "loss": 0.6961, "mean_token_accuracy": 0.8850671052932739, "num_tokens": 885639932.0, "step": 23328 }, { "epoch": 2.9676885892380103, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.246744155883789, "learning_rate": 5e-05, "loss": 0.9518, "mean_token_accuracy": 0.8662947416305542, "num_tokens": 885675914.0, "step": 23329 }, { "epoch": 2.967815799516601, "ewc_loss": 0.357421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.247999668121338, "learning_rate": 5e-05, "loss": 0.6794, "mean_token_accuracy": 0.8864762783050537, "num_tokens": 885711910.0, "step": 23330 }, { "epoch": 2.9679430097951913, "ewc_loss": 0.6953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00061798095703125, "grad_norm": 10.326279640197754, "learning_rate": 5e-05, "loss": 1.0073, "mean_token_accuracy": 0.8882027864456177, "num_tokens": 885751615.0, "step": 23331 }, { "epoch": 2.968070220073782, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 6.752987861633301, "learning_rate": 5e-05, "loss": 0.7981, "mean_token_accuracy": 0.883715033531189, "num_tokens": 885788394.0, "step": 23332 }, { "epoch": 2.9681974303523724, "ewc_loss": 0.82421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0007476806640625, "grad_norm": 10.945558547973633, "learning_rate": 5e-05, "loss": 1.1455, "mean_token_accuracy": 0.8891522884368896, "num_tokens": 885820462.0, "step": 23333 }, { "epoch": 2.968324640630963, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 8.745352745056152, "learning_rate": 5e-05, "loss": 1.0783, "mean_token_accuracy": 0.889633059501648, "num_tokens": 885853714.0, "step": 23334 }, { "epoch": 2.9684518509095534, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.49577522277832, "learning_rate": 5e-05, "loss": 0.9868, "mean_token_accuracy": 0.8773424625396729, "num_tokens": 885887508.0, "step": 23335 }, { "epoch": 2.968579061188144, "ewc_loss": 0.79296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000713348388671875, "grad_norm": 9.523908615112305, "learning_rate": 5e-05, "loss": 1.1434, "mean_token_accuracy": 0.8747297525405884, "num_tokens": 885923314.0, "step": 23336 }, { "epoch": 2.9687062714667345, "ewc_loss": 0.466796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00038909912109375, "grad_norm": 7.612825870513916, "learning_rate": 5e-05, "loss": 0.7776, "mean_token_accuracy": 0.8892319202423096, "num_tokens": 885964466.0, "step": 23337 }, { "epoch": 2.968833481745325, "ewc_loss": 0.7890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00070953369140625, "grad_norm": 10.534083366394043, "learning_rate": 5e-05, "loss": 1.1327, "mean_token_accuracy": 0.8779363632202148, "num_tokens": 886001483.0, "step": 23338 }, { "epoch": 2.9689606920239155, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 7.439188480377197, "learning_rate": 5e-05, "loss": 0.8716, "mean_token_accuracy": 0.8817201852798462, "num_tokens": 886040130.0, "step": 23339 }, { "epoch": 2.969087902302506, "ewc_loss": 0.80859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000728607177734375, "grad_norm": 11.120511054992676, "learning_rate": 5e-05, "loss": 1.1439, "mean_token_accuracy": 0.8823412656784058, "num_tokens": 886076291.0, "step": 23340 }, { "epoch": 2.9692151125810966, "ewc_loss": 0.75390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000675201416015625, "grad_norm": 8.892183303833008, "learning_rate": 5e-05, "loss": 1.1437, "mean_token_accuracy": 0.8663491010665894, "num_tokens": 886109430.0, "step": 23341 }, { "epoch": 2.969342322859687, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.13289737701416, "learning_rate": 5e-05, "loss": 1.0068, "mean_token_accuracy": 0.8845595717430115, "num_tokens": 886147158.0, "step": 23342 }, { "epoch": 2.9694695331382777, "ewc_loss": 0.73828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000659942626953125, "grad_norm": 9.050545692443848, "learning_rate": 5e-05, "loss": 1.0833, "mean_token_accuracy": 0.8806689977645874, "num_tokens": 886184810.0, "step": 23343 }, { "epoch": 2.969596743416868, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.148268699645996, "learning_rate": 5e-05, "loss": 0.9266, "mean_token_accuracy": 0.8835700154304504, "num_tokens": 886225511.0, "step": 23344 }, { "epoch": 2.9697239536954587, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 8.699347496032715, "learning_rate": 5e-05, "loss": 0.969, "mean_token_accuracy": 0.8801814317703247, "num_tokens": 886265185.0, "step": 23345 }, { "epoch": 2.9698511639740492, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004711151123046875, "grad_norm": 8.502346992492676, "learning_rate": 5e-05, "loss": 0.8961, "mean_token_accuracy": 0.8781798481941223, "num_tokens": 886304897.0, "step": 23346 }, { "epoch": 2.9699783742526398, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000469207763671875, "grad_norm": 8.286730766296387, "learning_rate": 5e-05, "loss": 0.9128, "mean_token_accuracy": 0.8715410828590393, "num_tokens": 886343097.0, "step": 23347 }, { "epoch": 2.97010558453123, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.203022003173828, "learning_rate": 5e-05, "loss": 0.8633, "mean_token_accuracy": 0.8787537813186646, "num_tokens": 886377428.0, "step": 23348 }, { "epoch": 2.970232794809821, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 7.8878326416015625, "learning_rate": 5e-05, "loss": 0.8276, "mean_token_accuracy": 0.8741863369941711, "num_tokens": 886418344.0, "step": 23349 }, { "epoch": 2.970360005088411, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 8.107016563415527, "learning_rate": 5e-05, "loss": 0.909, "mean_token_accuracy": 0.8613665103912354, "num_tokens": 886458983.0, "step": 23350 }, { "epoch": 2.970487215367002, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.598570823669434, "learning_rate": 5e-05, "loss": 0.7293, "mean_token_accuracy": 0.8927065134048462, "num_tokens": 886492174.0, "step": 23351 }, { "epoch": 2.970614425645592, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 8.111286163330078, "learning_rate": 5e-05, "loss": 0.8399, "mean_token_accuracy": 0.8693907856941223, "num_tokens": 886537145.0, "step": 23352 }, { "epoch": 2.970741635924183, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.366319179534912, "learning_rate": 5e-05, "loss": 0.7897, "mean_token_accuracy": 0.8705981373786926, "num_tokens": 886573420.0, "step": 23353 }, { "epoch": 2.970868846202773, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 8.242439270019531, "learning_rate": 5e-05, "loss": 0.7907, "mean_token_accuracy": 0.8822219371795654, "num_tokens": 886610489.0, "step": 23354 }, { "epoch": 2.9709960564813636, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 7.114744663238525, "learning_rate": 5e-05, "loss": 0.6859, "mean_token_accuracy": 0.8880332708358765, "num_tokens": 886646954.0, "step": 23355 }, { "epoch": 2.971123266759954, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 8.491037368774414, "learning_rate": 5e-05, "loss": 0.7925, "mean_token_accuracy": 0.8894929885864258, "num_tokens": 886686979.0, "step": 23356 }, { "epoch": 2.9712504770385446, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.8701324462890625, "learning_rate": 5e-05, "loss": 0.8187, "mean_token_accuracy": 0.8537890315055847, "num_tokens": 886719734.0, "step": 23357 }, { "epoch": 2.971377687317135, "ewc_loss": 0.515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.756495475769043, "learning_rate": 5e-05, "loss": 0.8373, "mean_token_accuracy": 0.8843926191329956, "num_tokens": 886754762.0, "step": 23358 }, { "epoch": 2.9715048975957257, "ewc_loss": 0.369140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002918243408203125, "grad_norm": 6.724228382110596, "learning_rate": 5e-05, "loss": 0.6823, "mean_token_accuracy": 0.8853257894515991, "num_tokens": 886794830.0, "step": 23359 }, { "epoch": 2.971632107874316, "ewc_loss": 0.5625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004825592041015625, "grad_norm": 9.036285400390625, "learning_rate": 5e-05, "loss": 0.9164, "mean_token_accuracy": 0.8746035695075989, "num_tokens": 886834952.0, "step": 23360 }, { "epoch": 2.9717593181529067, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 6.720240592956543, "learning_rate": 5e-05, "loss": 0.7625, "mean_token_accuracy": 0.8674719333648682, "num_tokens": 886874236.0, "step": 23361 }, { "epoch": 2.9718865284314973, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.189229965209961, "learning_rate": 5e-05, "loss": 0.9668, "mean_token_accuracy": 0.8706638813018799, "num_tokens": 886907723.0, "step": 23362 }, { "epoch": 2.972013738710088, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.050573825836182, "learning_rate": 5e-05, "loss": 0.809, "mean_token_accuracy": 0.8764092922210693, "num_tokens": 886939914.0, "step": 23363 }, { "epoch": 2.9721409489886783, "ewc_loss": 0.5859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000507354736328125, "grad_norm": 8.844507217407227, "learning_rate": 5e-05, "loss": 0.953, "mean_token_accuracy": 0.8714689016342163, "num_tokens": 886970533.0, "step": 23364 }, { "epoch": 2.972268159267269, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 7.652542591094971, "learning_rate": 5e-05, "loss": 0.9052, "mean_token_accuracy": 0.86524498462677, "num_tokens": 887012103.0, "step": 23365 }, { "epoch": 2.9723953695458594, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004138946533203125, "grad_norm": 7.98276424407959, "learning_rate": 5e-05, "loss": 0.8677, "mean_token_accuracy": 0.8684405088424683, "num_tokens": 887055789.0, "step": 23366 }, { "epoch": 2.97252257982445, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.190784454345703, "learning_rate": 5e-05, "loss": 0.8391, "mean_token_accuracy": 0.8879978060722351, "num_tokens": 887094505.0, "step": 23367 }, { "epoch": 2.9726497901030404, "ewc_loss": 0.404296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003261566162109375, "grad_norm": 7.236802577972412, "learning_rate": 5e-05, "loss": 0.718, "mean_token_accuracy": 0.8897953033447266, "num_tokens": 887131009.0, "step": 23368 }, { "epoch": 2.972777000381631, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00045013427734375, "grad_norm": 8.562536239624023, "learning_rate": 5e-05, "loss": 0.8212, "mean_token_accuracy": 0.8924465179443359, "num_tokens": 887170969.0, "step": 23369 }, { "epoch": 2.9729042106602215, "ewc_loss": 0.376953125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002994537353515625, "grad_norm": 6.880308151245117, "learning_rate": 5e-05, "loss": 0.6885, "mean_token_accuracy": 0.8869971632957458, "num_tokens": 887205123.0, "step": 23370 }, { "epoch": 2.973031420938812, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.80587100982666, "learning_rate": 5e-05, "loss": 0.8492, "mean_token_accuracy": 0.890005350112915, "num_tokens": 887235908.0, "step": 23371 }, { "epoch": 2.9731586312174025, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 6.837405681610107, "learning_rate": 5e-05, "loss": 0.6957, "mean_token_accuracy": 0.8877014517784119, "num_tokens": 887273929.0, "step": 23372 }, { "epoch": 2.9732858414959926, "ewc_loss": 0.5546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.833929061889648, "learning_rate": 5e-05, "loss": 0.8788, "mean_token_accuracy": 0.8842348456382751, "num_tokens": 887309379.0, "step": 23373 }, { "epoch": 2.9734130517745836, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.031975269317627, "learning_rate": 5e-05, "loss": 0.807, "mean_token_accuracy": 0.8709042072296143, "num_tokens": 887346108.0, "step": 23374 }, { "epoch": 2.9735402620531737, "ewc_loss": 0.53515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000457763671875, "grad_norm": 8.64840030670166, "learning_rate": 5e-05, "loss": 0.847, "mean_token_accuracy": 0.8925039768218994, "num_tokens": 887380066.0, "step": 23375 }, { "epoch": 2.9736674723317646, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.309900283813477, "learning_rate": 5e-05, "loss": 0.8098, "mean_token_accuracy": 0.8774820566177368, "num_tokens": 887414043.0, "step": 23376 }, { "epoch": 2.9737946826103547, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.331570625305176, "learning_rate": 5e-05, "loss": 0.8611, "mean_token_accuracy": 0.8755817413330078, "num_tokens": 887451921.0, "step": 23377 }, { "epoch": 2.9739218928889453, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 7.476503372192383, "learning_rate": 5e-05, "loss": 0.8212, "mean_token_accuracy": 0.8792646527290344, "num_tokens": 887483977.0, "step": 23378 }, { "epoch": 2.974049103167536, "ewc_loss": 0.4765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003986358642578125, "grad_norm": 8.121826171875, "learning_rate": 5e-05, "loss": 0.8058, "mean_token_accuracy": 0.8803879022598267, "num_tokens": 887516795.0, "step": 23379 }, { "epoch": 2.9741763134461263, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.448781490325928, "learning_rate": 5e-05, "loss": 0.7925, "mean_token_accuracy": 0.8752378225326538, "num_tokens": 887554791.0, "step": 23380 }, { "epoch": 2.974303523724717, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.143159866333008, "learning_rate": 5e-05, "loss": 0.8029, "mean_token_accuracy": 0.8846926689147949, "num_tokens": 887598451.0, "step": 23381 }, { "epoch": 2.9744307340033074, "ewc_loss": 0.40625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00032806396484375, "grad_norm": 7.2481770515441895, "learning_rate": 5e-05, "loss": 0.706, "mean_token_accuracy": 0.8945280909538269, "num_tokens": 887633242.0, "step": 23382 }, { "epoch": 2.974557944281898, "ewc_loss": 0.490234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.34695816040039, "learning_rate": 5e-05, "loss": 0.8048, "mean_token_accuracy": 0.8870866298675537, "num_tokens": 887665084.0, "step": 23383 }, { "epoch": 2.9746851545604884, "ewc_loss": 0.384765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003070831298828125, "grad_norm": 7.028543949127197, "learning_rate": 5e-05, "loss": 0.8133, "mean_token_accuracy": 0.8532195091247559, "num_tokens": 887706086.0, "step": 23384 }, { "epoch": 2.974812364839079, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 8.526681900024414, "learning_rate": 5e-05, "loss": 0.8729, "mean_token_accuracy": 0.8717421293258667, "num_tokens": 887747395.0, "step": 23385 }, { "epoch": 2.9749395751176695, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.949546813964844, "learning_rate": 5e-05, "loss": 0.7534, "mean_token_accuracy": 0.8709957599639893, "num_tokens": 887789834.0, "step": 23386 }, { "epoch": 2.97506678539626, "ewc_loss": 0.51171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004329681396484375, "grad_norm": 8.583968162536621, "learning_rate": 5e-05, "loss": 0.9027, "mean_token_accuracy": 0.8624923229217529, "num_tokens": 887831416.0, "step": 23387 }, { "epoch": 2.9751939956748505, "ewc_loss": 0.392578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003147125244140625, "grad_norm": 7.050775527954102, "learning_rate": 5e-05, "loss": 0.7462, "mean_token_accuracy": 0.8704005479812622, "num_tokens": 887869781.0, "step": 23388 }, { "epoch": 2.975321205953441, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.408151626586914, "learning_rate": 5e-05, "loss": 0.8496, "mean_token_accuracy": 0.8822544813156128, "num_tokens": 887906844.0, "step": 23389 }, { "epoch": 2.9754484162320316, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.3152289390563965, "learning_rate": 5e-05, "loss": 0.7898, "mean_token_accuracy": 0.8714061975479126, "num_tokens": 887939514.0, "step": 23390 }, { "epoch": 2.975575626510622, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.018486022949219, "learning_rate": 5e-05, "loss": 0.8542, "mean_token_accuracy": 0.8649504780769348, "num_tokens": 887976041.0, "step": 23391 }, { "epoch": 2.9757028367892127, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.652435302734375, "learning_rate": 5e-05, "loss": 0.8283, "mean_token_accuracy": 0.8680174946784973, "num_tokens": 888010007.0, "step": 23392 }, { "epoch": 2.975830047067803, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.586692810058594, "learning_rate": 5e-05, "loss": 0.7433, "mean_token_accuracy": 0.8827407360076904, "num_tokens": 888045654.0, "step": 23393 }, { "epoch": 2.9759572573463937, "ewc_loss": 0.443359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003643035888671875, "grad_norm": 7.955686569213867, "learning_rate": 5e-05, "loss": 0.7714, "mean_token_accuracy": 0.8847342133522034, "num_tokens": 888084075.0, "step": 23394 }, { "epoch": 2.9760844676249842, "ewc_loss": 0.380859375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 7.206481456756592, "learning_rate": 5e-05, "loss": 0.7339, "mean_token_accuracy": 0.8761435151100159, "num_tokens": 888124891.0, "step": 23395 }, { "epoch": 2.9762116779035748, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 8.20844841003418, "learning_rate": 5e-05, "loss": 0.7814, "mean_token_accuracy": 0.8879111409187317, "num_tokens": 888161078.0, "step": 23396 }, { "epoch": 2.9763388881821653, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 6.973983287811279, "learning_rate": 5e-05, "loss": 0.69, "mean_token_accuracy": 0.8828202486038208, "num_tokens": 888197013.0, "step": 23397 }, { "epoch": 2.9764660984607554, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.408258438110352, "learning_rate": 5e-05, "loss": 0.8208, "mean_token_accuracy": 0.8722165822982788, "num_tokens": 888231676.0, "step": 23398 }, { "epoch": 2.9765933087393464, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.812407970428467, "learning_rate": 5e-05, "loss": 0.686, "mean_token_accuracy": 0.8844326138496399, "num_tokens": 888272676.0, "step": 23399 }, { "epoch": 2.9767205190179364, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.578744888305664, "learning_rate": 5e-05, "loss": 0.817, "mean_token_accuracy": 0.8831920027732849, "num_tokens": 888310064.0, "step": 23400 }, { "epoch": 2.9768477292965274, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000286102294921875, "grad_norm": 6.754588603973389, "learning_rate": 5e-05, "loss": 0.7023, "mean_token_accuracy": 0.881446361541748, "num_tokens": 888344103.0, "step": 23401 }, { "epoch": 2.9769749395751175, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.646174430847168, "learning_rate": 5e-05, "loss": 0.8746, "mean_token_accuracy": 0.8714130520820618, "num_tokens": 888379284.0, "step": 23402 }, { "epoch": 2.977102149853708, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.84138298034668, "learning_rate": 5e-05, "loss": 0.6904, "mean_token_accuracy": 0.889321506023407, "num_tokens": 888414058.0, "step": 23403 }, { "epoch": 2.9772293601322986, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000431060791015625, "grad_norm": 8.538212776184082, "learning_rate": 5e-05, "loss": 0.8728, "mean_token_accuracy": 0.870098888874054, "num_tokens": 888454756.0, "step": 23404 }, { "epoch": 2.977356570410889, "ewc_loss": 0.39453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000316619873046875, "grad_norm": 7.110833644866943, "learning_rate": 5e-05, "loss": 0.7357, "mean_token_accuracy": 0.8799706697463989, "num_tokens": 888491995.0, "step": 23405 }, { "epoch": 2.9774837806894796, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 8.191649436950684, "learning_rate": 5e-05, "loss": 0.8851, "mean_token_accuracy": 0.8618296384811401, "num_tokens": 888533073.0, "step": 23406 }, { "epoch": 2.97761099096807, "ewc_loss": 0.421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.420068740844727, "learning_rate": 5e-05, "loss": 0.7018, "mean_token_accuracy": 0.8992495536804199, "num_tokens": 888570392.0, "step": 23407 }, { "epoch": 2.9777382012466607, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.841681480407715, "learning_rate": 5e-05, "loss": 0.7876, "mean_token_accuracy": 0.8804808855056763, "num_tokens": 888612078.0, "step": 23408 }, { "epoch": 2.977865411525251, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.62058687210083, "learning_rate": 5e-05, "loss": 0.7469, "mean_token_accuracy": 0.8865015506744385, "num_tokens": 888653603.0, "step": 23409 }, { "epoch": 2.9779926218038417, "ewc_loss": 0.419921875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.6495256423950195, "learning_rate": 5e-05, "loss": 0.79, "mean_token_accuracy": 0.872347354888916, "num_tokens": 888687089.0, "step": 23410 }, { "epoch": 2.9781198320824323, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.6262946128845215, "learning_rate": 5e-05, "loss": 0.7792, "mean_token_accuracy": 0.8650020360946655, "num_tokens": 888732221.0, "step": 23411 }, { "epoch": 2.978247042361023, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 7.677873134613037, "learning_rate": 5e-05, "loss": 0.7079, "mean_token_accuracy": 0.8929375410079956, "num_tokens": 888768992.0, "step": 23412 }, { "epoch": 2.9783742526396133, "ewc_loss": 0.38671875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000308990478515625, "grad_norm": 7.430577278137207, "learning_rate": 5e-05, "loss": 0.6987, "mean_token_accuracy": 0.8896167278289795, "num_tokens": 888809834.0, "step": 23413 }, { "epoch": 2.978501462918204, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 7.9162373542785645, "learning_rate": 5e-05, "loss": 0.7529, "mean_token_accuracy": 0.8836501836776733, "num_tokens": 888851943.0, "step": 23414 }, { "epoch": 2.9786286731967944, "ewc_loss": 0.365234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002880096435546875, "grad_norm": 7.105094909667969, "learning_rate": 5e-05, "loss": 0.6934, "mean_token_accuracy": 0.8811956644058228, "num_tokens": 888890465.0, "step": 23415 }, { "epoch": 2.978755883475385, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 8.290173530578613, "learning_rate": 5e-05, "loss": 0.8019, "mean_token_accuracy": 0.8794489502906799, "num_tokens": 888932954.0, "step": 23416 }, { "epoch": 2.9788830937539754, "ewc_loss": 0.34375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002651214599609375, "grad_norm": 6.71226692199707, "learning_rate": 5e-05, "loss": 0.6848, "mean_token_accuracy": 0.8795851469039917, "num_tokens": 888973916.0, "step": 23417 }, { "epoch": 2.979010304032566, "ewc_loss": 0.498046875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00041961669921875, "grad_norm": 8.788537979125977, "learning_rate": 5e-05, "loss": 0.8992, "mean_token_accuracy": 0.8592877388000488, "num_tokens": 889017202.0, "step": 23418 }, { "epoch": 2.9791375143111565, "ewc_loss": 0.337890625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002593994140625, "grad_norm": 6.32162618637085, "learning_rate": 5e-05, "loss": 0.7266, "mean_token_accuracy": 0.8660366535186768, "num_tokens": 889053604.0, "step": 23419 }, { "epoch": 2.979264724589747, "ewc_loss": 0.57421875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00049591064453125, "grad_norm": 9.37670612335205, "learning_rate": 5e-05, "loss": 0.9176, "mean_token_accuracy": 0.8782588243484497, "num_tokens": 889087328.0, "step": 23420 }, { "epoch": 2.979391934868337, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.232951641082764, "learning_rate": 5e-05, "loss": 0.7217, "mean_token_accuracy": 0.8757665157318115, "num_tokens": 889125209.0, "step": 23421 }, { "epoch": 2.979519145146928, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.622862815856934, "learning_rate": 5e-05, "loss": 0.9948, "mean_token_accuracy": 0.8730894327163696, "num_tokens": 889160968.0, "step": 23422 }, { "epoch": 2.979646355425518, "ewc_loss": 0.478515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000400543212890625, "grad_norm": 7.025525093078613, "learning_rate": 5e-05, "loss": 0.8517, "mean_token_accuracy": 0.8742408752441406, "num_tokens": 889201710.0, "step": 23423 }, { "epoch": 2.979773565704109, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.767617225646973, "learning_rate": 5e-05, "loss": 1.0036, "mean_token_accuracy": 0.862107515335083, "num_tokens": 889237003.0, "step": 23424 }, { "epoch": 2.979900775982699, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.284979820251465, "learning_rate": 5e-05, "loss": 0.9806, "mean_token_accuracy": 0.8696575164794922, "num_tokens": 889278775.0, "step": 23425 }, { "epoch": 2.98002798626129, "ewc_loss": 0.400390625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003223419189453125, "grad_norm": 6.955870628356934, "learning_rate": 5e-05, "loss": 0.7741, "mean_token_accuracy": 0.8725718259811401, "num_tokens": 889320888.0, "step": 23426 }, { "epoch": 2.9801551965398803, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.742324829101562, "learning_rate": 5e-05, "loss": 0.9897, "mean_token_accuracy": 0.8809263706207275, "num_tokens": 889356746.0, "step": 23427 }, { "epoch": 2.980282406818471, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 5.8963236808776855, "learning_rate": 5e-05, "loss": 0.6739, "mean_token_accuracy": 0.8873953223228455, "num_tokens": 889400771.0, "step": 23428 }, { "epoch": 2.9804096170970613, "ewc_loss": 0.828125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000751495361328125, "grad_norm": 11.230793952941895, "learning_rate": 5e-05, "loss": 1.1929, "mean_token_accuracy": 0.8720203638076782, "num_tokens": 889440459.0, "step": 23429 }, { "epoch": 2.980536827375652, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 7.410015106201172, "learning_rate": 5e-05, "loss": 0.9458, "mean_token_accuracy": 0.8847562670707703, "num_tokens": 889479782.0, "step": 23430 }, { "epoch": 2.9806640376542424, "ewc_loss": 0.8359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000759124755859375, "grad_norm": 11.246098518371582, "learning_rate": 5e-05, "loss": 1.1248, "mean_token_accuracy": 0.8951744437217712, "num_tokens": 889520167.0, "step": 23431 }, { "epoch": 2.980791247932833, "ewc_loss": 0.93359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0008544921875, "grad_norm": 9.438718795776367, "learning_rate": 5e-05, "loss": 1.2771, "mean_token_accuracy": 0.8815251588821411, "num_tokens": 889567589.0, "step": 23432 }, { "epoch": 2.9809184582114234, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.522431373596191, "learning_rate": 5e-05, "loss": 0.9919, "mean_token_accuracy": 0.8766495585441589, "num_tokens": 889607703.0, "step": 23433 }, { "epoch": 2.981045668490014, "ewc_loss": 0.80078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000720977783203125, "grad_norm": 9.66380786895752, "learning_rate": 5e-05, "loss": 1.1539, "mean_token_accuracy": 0.8747698664665222, "num_tokens": 889645735.0, "step": 23434 }, { "epoch": 2.9811728787686045, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.227880477905273, "learning_rate": 5e-05, "loss": 0.918, "mean_token_accuracy": 0.8834255337715149, "num_tokens": 889680035.0, "step": 23435 }, { "epoch": 2.981300089047195, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.789756774902344, "learning_rate": 5e-05, "loss": 1.029, "mean_token_accuracy": 0.87751305103302, "num_tokens": 889720234.0, "step": 23436 }, { "epoch": 2.9814272993257855, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.011194229125977, "learning_rate": 5e-05, "loss": 0.9177, "mean_token_accuracy": 0.8887463808059692, "num_tokens": 889756070.0, "step": 23437 }, { "epoch": 2.981554509604376, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 9.165351867675781, "learning_rate": 5e-05, "loss": 0.9627, "mean_token_accuracy": 0.8804904222488403, "num_tokens": 889796499.0, "step": 23438 }, { "epoch": 2.9816817198829666, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 8.217369079589844, "learning_rate": 5e-05, "loss": 1.0191, "mean_token_accuracy": 0.8598676919937134, "num_tokens": 889830804.0, "step": 23439 }, { "epoch": 2.981808930161557, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.216294288635254, "learning_rate": 5e-05, "loss": 0.8961, "mean_token_accuracy": 0.8674019575119019, "num_tokens": 889865228.0, "step": 23440 }, { "epoch": 2.9819361404401477, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 8.384950637817383, "learning_rate": 5e-05, "loss": 0.9158, "mean_token_accuracy": 0.8763176202774048, "num_tokens": 889900780.0, "step": 23441 }, { "epoch": 2.982063350718738, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003833770751953125, "grad_norm": 7.600707530975342, "learning_rate": 5e-05, "loss": 0.8107, "mean_token_accuracy": 0.8792848587036133, "num_tokens": 889944366.0, "step": 23442 }, { "epoch": 2.9821905609973287, "ewc_loss": 0.53125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004558563232421875, "grad_norm": 8.518842697143555, "learning_rate": 5e-05, "loss": 0.819, "mean_token_accuracy": 0.896192729473114, "num_tokens": 889979739.0, "step": 23443 }, { "epoch": 2.9823177712759192, "ewc_loss": 0.4375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00035858154296875, "grad_norm": 7.2871012687683105, "learning_rate": 5e-05, "loss": 0.7684, "mean_token_accuracy": 0.8802435994148254, "num_tokens": 890017288.0, "step": 23444 }, { "epoch": 2.9824449815545098, "ewc_loss": 0.5234375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 8.553698539733887, "learning_rate": 5e-05, "loss": 0.8384, "mean_token_accuracy": 0.8849847316741943, "num_tokens": 890050848.0, "step": 23445 }, { "epoch": 2.9825721918331, "ewc_loss": 0.44140625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000362396240234375, "grad_norm": 7.232172966003418, "learning_rate": 5e-05, "loss": 0.775, "mean_token_accuracy": 0.8802401423454285, "num_tokens": 890087985.0, "step": 23446 }, { "epoch": 2.982699402111691, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00042724609375, "grad_norm": 8.406606674194336, "learning_rate": 5e-05, "loss": 0.8785, "mean_token_accuracy": 0.8694363832473755, "num_tokens": 890128292.0, "step": 23447 }, { "epoch": 2.982826612390281, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.338511943817139, "learning_rate": 5e-05, "loss": 0.7885, "mean_token_accuracy": 0.8805165886878967, "num_tokens": 890176130.0, "step": 23448 }, { "epoch": 2.982953822668872, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.071762084960938, "learning_rate": 5e-05, "loss": 0.8085, "mean_token_accuracy": 0.8787621855735779, "num_tokens": 890212547.0, "step": 23449 }, { "epoch": 2.983081032947462, "ewc_loss": 0.455078125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.5839314460754395, "learning_rate": 5e-05, "loss": 0.8205, "mean_token_accuracy": 0.8744124174118042, "num_tokens": 890257423.0, "step": 23450 }, { "epoch": 2.983208243226053, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.658170223236084, "learning_rate": 5e-05, "loss": 0.7705, "mean_token_accuracy": 0.8774027824401855, "num_tokens": 890299465.0, "step": 23451 }, { "epoch": 2.983335453504643, "ewc_loss": 0.451171875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.883389949798584, "learning_rate": 5e-05, "loss": 0.7719, "mean_token_accuracy": 0.8854877948760986, "num_tokens": 890340010.0, "step": 23452 }, { "epoch": 2.9834626637832335, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.292327404022217, "learning_rate": 5e-05, "loss": 0.7354, "mean_token_accuracy": 0.8764466047286987, "num_tokens": 890380865.0, "step": 23453 }, { "epoch": 2.983589874061824, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 8.194391250610352, "learning_rate": 5e-05, "loss": 0.7676, "mean_token_accuracy": 0.8915812969207764, "num_tokens": 890416710.0, "step": 23454 }, { "epoch": 2.9837170843404146, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.946805000305176, "learning_rate": 5e-05, "loss": 0.6816, "mean_token_accuracy": 0.8902563452720642, "num_tokens": 890452963.0, "step": 23455 }, { "epoch": 2.983844294619005, "ewc_loss": 0.494140625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.582724571228027, "learning_rate": 5e-05, "loss": 0.8241, "mean_token_accuracy": 0.8831923007965088, "num_tokens": 890493308.0, "step": 23456 }, { "epoch": 2.9839715048975957, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.650995254516602, "learning_rate": 5e-05, "loss": 0.7242, "mean_token_accuracy": 0.8737291097640991, "num_tokens": 890529279.0, "step": 23457 }, { "epoch": 2.984098715176186, "ewc_loss": 0.53125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004520416259765625, "grad_norm": 8.983728408813477, "learning_rate": 5e-05, "loss": 0.9279, "mean_token_accuracy": 0.861041784286499, "num_tokens": 890568866.0, "step": 23458 }, { "epoch": 2.9842259254547767, "ewc_loss": 0.3671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0002899169921875, "grad_norm": 6.493415355682373, "learning_rate": 5e-05, "loss": 0.6997, "mean_token_accuracy": 0.8831920623779297, "num_tokens": 890613077.0, "step": 23459 }, { "epoch": 2.9843531357333672, "ewc_loss": 0.59375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 9.258413314819336, "learning_rate": 5e-05, "loss": 0.935, "mean_token_accuracy": 0.880460798740387, "num_tokens": 890653599.0, "step": 23460 }, { "epoch": 2.9844803460119578, "ewc_loss": 0.408203125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003299713134765625, "grad_norm": 6.685294151306152, "learning_rate": 5e-05, "loss": 0.7485, "mean_token_accuracy": 0.8796931505203247, "num_tokens": 890694581.0, "step": 23461 }, { "epoch": 2.9846075562905483, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 9.208558082580566, "learning_rate": 5e-05, "loss": 0.9215, "mean_token_accuracy": 0.8861777782440186, "num_tokens": 890725912.0, "step": 23462 }, { "epoch": 2.984734766569139, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 7.237473487854004, "learning_rate": 5e-05, "loss": 0.8571, "mean_token_accuracy": 0.8775641918182373, "num_tokens": 890763028.0, "step": 23463 }, { "epoch": 2.9848619768477294, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.00048828125, "grad_norm": 8.617783546447754, "learning_rate": 5e-05, "loss": 0.9128, "mean_token_accuracy": 0.8770102262496948, "num_tokens": 890803318.0, "step": 23464 }, { "epoch": 2.98498918712632, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000461578369140625, "grad_norm": 7.785027027130127, "learning_rate": 5e-05, "loss": 0.8979, "mean_token_accuracy": 0.8743934631347656, "num_tokens": 890841405.0, "step": 23465 }, { "epoch": 2.9851163974049104, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004177093505859375, "grad_norm": 7.859002590179443, "learning_rate": 5e-05, "loss": 0.8205, "mean_token_accuracy": 0.8846448659896851, "num_tokens": 890884271.0, "step": 23466 }, { "epoch": 2.985243607683501, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.025009155273438, "learning_rate": 5e-05, "loss": 0.8322, "mean_token_accuracy": 0.8870681524276733, "num_tokens": 890917747.0, "step": 23467 }, { "epoch": 2.9853708179620915, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.576743125915527, "learning_rate": 5e-05, "loss": 0.7803, "mean_token_accuracy": 0.8831292390823364, "num_tokens": 890958659.0, "step": 23468 }, { "epoch": 2.985498028240682, "ewc_loss": 0.458984375, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.771529674530029, "learning_rate": 5e-05, "loss": 0.7626, "mean_token_accuracy": 0.893474817276001, "num_tokens": 890993792.0, "step": 23469 }, { "epoch": 2.9856252385192725, "ewc_loss": 0.462890625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.000385284423828125, "grad_norm": 7.884792327880859, "learning_rate": 5e-05, "loss": 0.7895, "mean_token_accuracy": 0.8829014301300049, "num_tokens": 891030873.0, "step": 23470 }, { "epoch": 2.9857524487978626, "ewc_loss": 0.388671875, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.183990478515625, "learning_rate": 5e-05, "loss": 0.6916, "mean_token_accuracy": 0.8939362168312073, "num_tokens": 891068062.0, "step": 23471 }, { "epoch": 2.9858796590764536, "ewc_loss": 0.5, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004215240478515625, "grad_norm": 8.481620788574219, "learning_rate": 5e-05, "loss": 0.8741, "mean_token_accuracy": 0.8687630891799927, "num_tokens": 891108749.0, "step": 23472 }, { "epoch": 2.9860068693550437, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.613985538482666, "learning_rate": 5e-05, "loss": 0.7089, "mean_token_accuracy": 0.8771162033081055, "num_tokens": 891144749.0, "step": 23473 }, { "epoch": 2.9861340796336346, "ewc_loss": 0.5703125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000492095947265625, "grad_norm": 9.145000457763672, "learning_rate": 5e-05, "loss": 0.9571, "mean_token_accuracy": 0.8680320978164673, "num_tokens": 891180795.0, "step": 23474 }, { "epoch": 2.9862612899122247, "ewc_loss": 0.361328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.3441643714904785, "learning_rate": 5e-05, "loss": 0.718, "mean_token_accuracy": 0.8742227554321289, "num_tokens": 891221333.0, "step": 23475 }, { "epoch": 2.9863885001908153, "ewc_loss": 0.6328125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000553131103515625, "grad_norm": 9.6069974899292, "learning_rate": 5e-05, "loss": 0.9776, "mean_token_accuracy": 0.8798646926879883, "num_tokens": 891258414.0, "step": 23476 }, { "epoch": 2.986515710469406, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.0003757476806640625, "grad_norm": 6.888568878173828, "learning_rate": 5e-05, "loss": 0.8227, "mean_token_accuracy": 0.8704222440719604, "num_tokens": 891296864.0, "step": 23477 }, { "epoch": 2.9866429207479963, "ewc_loss": 0.625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.378029823303223, "learning_rate": 5e-05, "loss": 0.922, "mean_token_accuracy": 0.8938419818878174, "num_tokens": 891338191.0, "step": 23478 }, { "epoch": 2.986770131026587, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.000782012939453125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 7.925469875335693, "learning_rate": 5e-05, "loss": 0.9523, "mean_token_accuracy": 0.87525475025177, "num_tokens": 891377354.0, "step": 23479 }, { "epoch": 2.9868973413051774, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.00078582763671875, "ewc_loss_parallel": 0.0004253387451171875, "grad_norm": 7.919284343719482, "learning_rate": 5e-05, "loss": 0.8355, "mean_token_accuracy": 0.8829627633094788, "num_tokens": 891415868.0, "step": 23480 }, { "epoch": 2.987024551583768, "ewc_loss": 0.62109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00054168701171875, "grad_norm": 8.906713485717773, "learning_rate": 5e-05, "loss": 0.9446, "mean_token_accuracy": 0.883598268032074, "num_tokens": 891449548.0, "step": 23481 }, { "epoch": 2.9871517618623584, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003032684326171875, "grad_norm": 6.626028537750244, "learning_rate": 5e-05, "loss": 0.7802, "mean_token_accuracy": 0.860571026802063, "num_tokens": 891485937.0, "step": 23482 }, { "epoch": 2.987278972140949, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 9.767268180847168, "learning_rate": 5e-05, "loss": 0.9737, "mean_token_accuracy": 0.884581446647644, "num_tokens": 891522458.0, "step": 23483 }, { "epoch": 2.9874061824195395, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 6.654810428619385, "learning_rate": 5e-05, "loss": 0.8017, "mean_token_accuracy": 0.8695755004882812, "num_tokens": 891560804.0, "step": 23484 }, { "epoch": 2.98753339269813, "ewc_loss": 0.68359375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000606536865234375, "grad_norm": 9.88347053527832, "learning_rate": 5e-05, "loss": 0.9976, "mean_token_accuracy": 0.889968991279602, "num_tokens": 891595692.0, "step": 23485 }, { "epoch": 2.9876606029767205, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.755268096923828, "learning_rate": 5e-05, "loss": 0.923, "mean_token_accuracy": 0.8870577812194824, "num_tokens": 891632007.0, "step": 23486 }, { "epoch": 2.987787813255311, "ewc_loss": 0.609375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000530242919921875, "grad_norm": 8.90485954284668, "learning_rate": 5e-05, "loss": 0.9367, "mean_token_accuracy": 0.8840532898902893, "num_tokens": 891671605.0, "step": 23487 }, { "epoch": 2.9879150235339016, "ewc_loss": 0.625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 8.404380798339844, "learning_rate": 5e-05, "loss": 1.0032, "mean_token_accuracy": 0.8673702478408813, "num_tokens": 891708369.0, "step": 23488 }, { "epoch": 2.988042233812492, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.963067531585693, "learning_rate": 5e-05, "loss": 0.8302, "mean_token_accuracy": 0.8875900506973267, "num_tokens": 891744673.0, "step": 23489 }, { "epoch": 2.9881694440910826, "ewc_loss": 0.56640625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004863739013671875, "grad_norm": 8.248543739318848, "learning_rate": 5e-05, "loss": 0.9187, "mean_token_accuracy": 0.875815212726593, "num_tokens": 891781971.0, "step": 23490 }, { "epoch": 2.988296654369673, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00043487548828125, "grad_norm": 7.986410140991211, "learning_rate": 5e-05, "loss": 0.8376, "mean_token_accuracy": 0.8873801231384277, "num_tokens": 891814079.0, "step": 23491 }, { "epoch": 2.9884238646482637, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003662109375, "grad_norm": 7.445576190948486, "learning_rate": 5e-05, "loss": 0.7778, "mean_token_accuracy": 0.8850470781326294, "num_tokens": 891852188.0, "step": 23492 }, { "epoch": 2.9885510749268542, "ewc_loss": 0.54296875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00046539306640625, "grad_norm": 8.587667465209961, "learning_rate": 5e-05, "loss": 0.895, "mean_token_accuracy": 0.8750776052474976, "num_tokens": 891894048.0, "step": 23493 }, { "epoch": 2.9886782852054448, "ewc_loss": 0.37890625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000301361083984375, "grad_norm": 6.797905921936035, "learning_rate": 5e-05, "loss": 0.728, "mean_token_accuracy": 0.8777299523353577, "num_tokens": 891935199.0, "step": 23494 }, { "epoch": 2.9888054954840353, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 9.087270736694336, "learning_rate": 5e-05, "loss": 0.9376, "mean_token_accuracy": 0.8811021447181702, "num_tokens": 891977686.0, "step": 23495 }, { "epoch": 2.9889327057626254, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00031280517578125, "grad_norm": 6.704050064086914, "learning_rate": 5e-05, "loss": 0.7645, "mean_token_accuracy": 0.8692249059677124, "num_tokens": 892017501.0, "step": 23496 }, { "epoch": 2.9890599160412163, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 9.21376895904541, "learning_rate": 5e-05, "loss": 0.9891, "mean_token_accuracy": 0.8674883842468262, "num_tokens": 892057369.0, "step": 23497 }, { "epoch": 2.9891871263198064, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003795623779296875, "grad_norm": 7.017506122589111, "learning_rate": 5e-05, "loss": 0.8041, "mean_token_accuracy": 0.8762438893318176, "num_tokens": 892093998.0, "step": 23498 }, { "epoch": 2.9893143365983974, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.083503723144531, "learning_rate": 5e-05, "loss": 0.9259, "mean_token_accuracy": 0.8908273577690125, "num_tokens": 892125835.0, "step": 23499 }, { "epoch": 2.9894415468769875, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 7.375931262969971, "learning_rate": 5e-05, "loss": 0.865, "mean_token_accuracy": 0.8810688257217407, "num_tokens": 892160340.0, "step": 23500 }, { "epoch": 2.989568757155578, "ewc_loss": 0.59765625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000518798828125, "grad_norm": 8.75621509552002, "learning_rate": 5e-05, "loss": 0.945, "mean_token_accuracy": 0.8767423629760742, "num_tokens": 892202096.0, "step": 23501 }, { "epoch": 2.9896959674341685, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004482269287109375, "grad_norm": 7.617026329040527, "learning_rate": 5e-05, "loss": 0.8495, "mean_token_accuracy": 0.888629138469696, "num_tokens": 892241942.0, "step": 23502 }, { "epoch": 2.989823177712759, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004787445068359375, "grad_norm": 8.343671798706055, "learning_rate": 5e-05, "loss": 0.9321, "mean_token_accuracy": 0.8682489395141602, "num_tokens": 892284724.0, "step": 23503 }, { "epoch": 2.9899503879913496, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 7.69804048538208, "learning_rate": 5e-05, "loss": 0.8318, "mean_token_accuracy": 0.8902603983879089, "num_tokens": 892324761.0, "step": 23504 }, { "epoch": 2.99007759826994, "ewc_loss": 0.515625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004367828369140625, "grad_norm": 8.0474853515625, "learning_rate": 5e-05, "loss": 0.8213, "mean_token_accuracy": 0.8884584903717041, "num_tokens": 892358128.0, "step": 23505 }, { "epoch": 2.9902048085485307, "ewc_loss": 0.484375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.648504734039307, "learning_rate": 5e-05, "loss": 0.8475, "mean_token_accuracy": 0.8748815059661865, "num_tokens": 892400380.0, "step": 23506 }, { "epoch": 2.990332018827121, "ewc_loss": 0.48046875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004024505615234375, "grad_norm": 7.91636323928833, "learning_rate": 5e-05, "loss": 0.8529, "mean_token_accuracy": 0.8702411651611328, "num_tokens": 892439141.0, "step": 23507 }, { "epoch": 2.9904592291057117, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.6121368408203125, "learning_rate": 5e-05, "loss": 0.7943, "mean_token_accuracy": 0.8794040679931641, "num_tokens": 892480983.0, "step": 23508 }, { "epoch": 2.9905864393843022, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.775703430175781, "learning_rate": 5e-05, "loss": 0.8035, "mean_token_accuracy": 0.8774107098579407, "num_tokens": 892518220.0, "step": 23509 }, { "epoch": 2.9907136496628928, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000354766845703125, "grad_norm": 7.615924835205078, "learning_rate": 5e-05, "loss": 0.7705, "mean_token_accuracy": 0.8786866068840027, "num_tokens": 892549308.0, "step": 23510 }, { "epoch": 2.9908408599414833, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.687417030334473, "learning_rate": 5e-05, "loss": 0.7804, "mean_token_accuracy": 0.879848062992096, "num_tokens": 892590154.0, "step": 23511 }, { "epoch": 2.990968070220074, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003414154052734375, "grad_norm": 7.5969462394714355, "learning_rate": 5e-05, "loss": 0.7125, "mean_token_accuracy": 0.896073579788208, "num_tokens": 892630951.0, "step": 23512 }, { "epoch": 2.9910952804986644, "ewc_loss": 0.421875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00034332275390625, "grad_norm": 7.688731670379639, "learning_rate": 5e-05, "loss": 0.7793, "mean_token_accuracy": 0.8754434585571289, "num_tokens": 892675345.0, "step": 23513 }, { "epoch": 2.991222490777255, "ewc_loss": 0.40234375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000324249267578125, "grad_norm": 7.532208442687988, "learning_rate": 5e-05, "loss": 0.7679, "mean_token_accuracy": 0.8720540404319763, "num_tokens": 892718129.0, "step": 23514 }, { "epoch": 2.9913497010558454, "ewc_loss": 0.4140625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000335693359375, "grad_norm": 7.716014862060547, "learning_rate": 5e-05, "loss": 0.7608, "mean_token_accuracy": 0.8802204132080078, "num_tokens": 892757159.0, "step": 23515 }, { "epoch": 2.991476911334436, "ewc_loss": 0.390625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0003108978271484375, "grad_norm": 7.432734489440918, "learning_rate": 5e-05, "loss": 0.7313, "mean_token_accuracy": 0.8797131180763245, "num_tokens": 892797872.0, "step": 23516 }, { "epoch": 2.9916041216130265, "ewc_loss": 0.41015625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000331878662109375, "grad_norm": 7.822413444519043, "learning_rate": 5e-05, "loss": 0.7784, "mean_token_accuracy": 0.8703173398971558, "num_tokens": 892837213.0, "step": 23517 }, { "epoch": 2.991731331891617, "ewc_loss": 0.375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002956390380859375, "grad_norm": 7.258562088012695, "learning_rate": 5e-05, "loss": 0.7281, "mean_token_accuracy": 0.8750523924827576, "num_tokens": 892878651.0, "step": 23518 }, { "epoch": 2.991858542170207, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000347137451171875, "grad_norm": 8.021007537841797, "learning_rate": 5e-05, "loss": 0.7292, "mean_token_accuracy": 0.8875733613967896, "num_tokens": 892913739.0, "step": 23519 }, { "epoch": 2.991985752448798, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 7.024574279785156, "learning_rate": 5e-05, "loss": 0.7075, "mean_token_accuracy": 0.8772169351577759, "num_tokens": 892949649.0, "step": 23520 }, { "epoch": 2.992112962727388, "ewc_loss": 0.453125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 8.307878494262695, "learning_rate": 5e-05, "loss": 0.789, "mean_token_accuracy": 0.8832322955131531, "num_tokens": 892988831.0, "step": 23521 }, { "epoch": 2.992240173005979, "ewc_loss": 0.34765625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00026702880859375, "grad_norm": 6.718026161193848, "learning_rate": 5e-05, "loss": 0.6879, "mean_token_accuracy": 0.882554829120636, "num_tokens": 893025311.0, "step": 23522 }, { "epoch": 2.992367383284569, "ewc_loss": 0.49609375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000415802001953125, "grad_norm": 8.765790939331055, "learning_rate": 5e-05, "loss": 0.8616, "mean_token_accuracy": 0.8710527420043945, "num_tokens": 893072461.0, "step": 23523 }, { "epoch": 2.99249459356316, "ewc_loss": 0.3359375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000255584716796875, "grad_norm": 6.314703941345215, "learning_rate": 5e-05, "loss": 0.7017, "mean_token_accuracy": 0.8729803562164307, "num_tokens": 893109268.0, "step": 23524 }, { "epoch": 2.9926218038417502, "ewc_loss": 0.578125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 9.408170700073242, "learning_rate": 5e-05, "loss": 0.9497, "mean_token_accuracy": 0.8751774430274963, "num_tokens": 893139033.0, "step": 23525 }, { "epoch": 2.992749014120341, "ewc_loss": 0.359375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000278472900390625, "grad_norm": 6.157711505889893, "learning_rate": 5e-05, "loss": 0.6796, "mean_token_accuracy": 0.8858509063720703, "num_tokens": 893180056.0, "step": 23526 }, { "epoch": 2.9928762243989313, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.838396072387695, "learning_rate": 5e-05, "loss": 1.0075, "mean_token_accuracy": 0.8767277002334595, "num_tokens": 893214164.0, "step": 23527 }, { "epoch": 2.993003434677522, "ewc_loss": 0.48828125, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0004100799560546875, "grad_norm": 6.976179122924805, "learning_rate": 5e-05, "loss": 0.827, "mean_token_accuracy": 0.8781634569168091, "num_tokens": 893255030.0, "step": 23528 }, { "epoch": 2.9931306449561124, "ewc_loss": 0.625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000545501708984375, "grad_norm": 9.137386322021484, "learning_rate": 5e-05, "loss": 0.9829, "mean_token_accuracy": 0.8733112215995789, "num_tokens": 893287661.0, "step": 23529 }, { "epoch": 2.993257855234703, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 8.320414543151855, "learning_rate": 5e-05, "loss": 1.0487, "mean_token_accuracy": 0.8567928075790405, "num_tokens": 893328461.0, "step": 23530 }, { "epoch": 2.9933850655132934, "ewc_loss": 0.41796875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.000339508056640625, "grad_norm": 7.230828762054443, "learning_rate": 5e-05, "loss": 0.7487, "mean_token_accuracy": 0.8813562393188477, "num_tokens": 893367543.0, "step": 23531 }, { "epoch": 2.993512275791884, "ewc_loss": 0.671875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.0005950927734375, "grad_norm": 9.778380393981934, "learning_rate": 5e-05, "loss": 0.9886, "mean_token_accuracy": 0.8884996175765991, "num_tokens": 893401180.0, "step": 23532 }, { "epoch": 2.9936394860704745, "ewc_loss": 0.359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002803802490234375, "grad_norm": 6.0239996910095215, "learning_rate": 5e-05, "loss": 0.6479, "mean_token_accuracy": 0.893202543258667, "num_tokens": 893429823.0, "step": 23533 }, { "epoch": 2.993766696349065, "ewc_loss": 0.85546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0007781982421875, "grad_norm": 11.365729331970215, "learning_rate": 5e-05, "loss": 1.1915, "mean_token_accuracy": 0.8773252964019775, "num_tokens": 893472730.0, "step": 23534 }, { "epoch": 2.9938939066276555, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 7.714331150054932, "learning_rate": 5e-05, "loss": 0.969, "mean_token_accuracy": 0.881513774394989, "num_tokens": 893516506.0, "step": 23535 }, { "epoch": 2.994021116906246, "ewc_loss": 0.81640625, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00074005126953125, "grad_norm": 11.261792182922363, "learning_rate": 5e-05, "loss": 1.1667, "mean_token_accuracy": 0.876855731010437, "num_tokens": 893560539.0, "step": 23536 }, { "epoch": 2.9941483271848366, "ewc_loss": 0.92578125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00084686279296875, "grad_norm": 9.602216720581055, "learning_rate": 5e-05, "loss": 1.2675, "mean_token_accuracy": 0.8778722882270813, "num_tokens": 893595986.0, "step": 23537 }, { "epoch": 2.994275537463427, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.5770845413208, "learning_rate": 5e-05, "loss": 0.956, "mean_token_accuracy": 0.8860834836959839, "num_tokens": 893632997.0, "step": 23538 }, { "epoch": 2.9944027477420176, "ewc_loss": 0.7578125, "ewc_loss_diag": 0.000797271728515625, "ewc_loss_parallel": 0.00067901611328125, "grad_norm": 9.617534637451172, "learning_rate": 5e-05, "loss": 1.1245, "mean_token_accuracy": 0.8746212124824524, "num_tokens": 893671889.0, "step": 23539 }, { "epoch": 2.994529958020608, "ewc_loss": 0.60546875, "ewc_loss_diag": 0.00079345703125, "ewc_loss_parallel": 0.00052642822265625, "grad_norm": 8.458512306213379, "learning_rate": 5e-05, "loss": 0.9224, "mean_token_accuracy": 0.8883516788482666, "num_tokens": 893708781.0, "step": 23540 }, { "epoch": 2.9946571682991987, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00057220458984375, "grad_norm": 9.381224632263184, "learning_rate": 5e-05, "loss": 0.9954, "mean_token_accuracy": 0.8753452301025391, "num_tokens": 893745590.0, "step": 23541 }, { "epoch": 2.9947843785777892, "ewc_loss": 0.6171875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000537872314453125, "grad_norm": 8.259851455688477, "learning_rate": 5e-05, "loss": 0.9481, "mean_token_accuracy": 0.8838883638381958, "num_tokens": 893783599.0, "step": 23542 }, { "epoch": 2.9949115888563798, "ewc_loss": 0.546875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004673004150390625, "grad_norm": 8.503868103027344, "learning_rate": 5e-05, "loss": 0.836, "mean_token_accuracy": 0.8946928977966309, "num_tokens": 893822603.0, "step": 23543 }, { "epoch": 2.99503879913497, "ewc_loss": 0.59375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000514984130859375, "grad_norm": 8.488106727600098, "learning_rate": 5e-05, "loss": 0.9293, "mean_token_accuracy": 0.8823452591896057, "num_tokens": 893857318.0, "step": 23544 }, { "epoch": 2.995166009413561, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.602810382843018, "learning_rate": 5e-05, "loss": 0.7859, "mean_token_accuracy": 0.8763670325279236, "num_tokens": 893896269.0, "step": 23545 }, { "epoch": 2.995293219692151, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.822141647338867, "learning_rate": 5e-05, "loss": 0.957, "mean_token_accuracy": 0.8668549060821533, "num_tokens": 893936877.0, "step": 23546 }, { "epoch": 2.995420429970742, "ewc_loss": 0.4296875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003509521484375, "grad_norm": 7.1969218254089355, "learning_rate": 5e-05, "loss": 0.8021, "mean_token_accuracy": 0.8764868378639221, "num_tokens": 893970014.0, "step": 23547 }, { "epoch": 2.995547640249332, "ewc_loss": 0.578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000499725341796875, "grad_norm": 8.986323356628418, "learning_rate": 5e-05, "loss": 0.9195, "mean_token_accuracy": 0.8816743493080139, "num_tokens": 894005568.0, "step": 23548 }, { "epoch": 2.995674850527923, "ewc_loss": 0.453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00037384033203125, "grad_norm": 7.164227485656738, "learning_rate": 5e-05, "loss": 0.8005, "mean_token_accuracy": 0.8777745366096497, "num_tokens": 894053880.0, "step": 23549 }, { "epoch": 2.995802060806513, "ewc_loss": 0.55859375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00048065185546875, "grad_norm": 8.877556800842285, "learning_rate": 5e-05, "loss": 0.8606, "mean_token_accuracy": 0.8914937376976013, "num_tokens": 894083981.0, "step": 23550 }, { "epoch": 2.9959292710851035, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00040435791015625, "grad_norm": 7.347987174987793, "learning_rate": 5e-05, "loss": 0.8203, "mean_token_accuracy": 0.882621705532074, "num_tokens": 894125315.0, "step": 23551 }, { "epoch": 2.996056481363694, "ewc_loss": 0.52734375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000446319580078125, "grad_norm": 8.489795684814453, "learning_rate": 5e-05, "loss": 0.889, "mean_token_accuracy": 0.8759807348251343, "num_tokens": 894165195.0, "step": 23552 }, { "epoch": 2.9961836916422846, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 7.569127559661865, "learning_rate": 5e-05, "loss": 0.8554, "mean_token_accuracy": 0.8717275857925415, "num_tokens": 894206443.0, "step": 23553 }, { "epoch": 2.996310901920875, "ewc_loss": 0.46875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003910064697265625, "grad_norm": 8.048853874206543, "learning_rate": 5e-05, "loss": 0.801, "mean_token_accuracy": 0.8828232288360596, "num_tokens": 894240897.0, "step": 23554 }, { "epoch": 2.9964381121994657, "ewc_loss": 0.45703125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000377655029296875, "grad_norm": 7.672604560852051, "learning_rate": 5e-05, "loss": 0.7695, "mean_token_accuracy": 0.8884280920028687, "num_tokens": 894277475.0, "step": 23555 }, { "epoch": 2.996565322478056, "ewc_loss": 0.4453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003681182861328125, "grad_norm": 7.853948593139648, "learning_rate": 5e-05, "loss": 0.817, "mean_token_accuracy": 0.8704690337181091, "num_tokens": 894320231.0, "step": 23556 }, { "epoch": 2.9966925327566467, "ewc_loss": 0.42578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003452301025390625, "grad_norm": 7.548218250274658, "learning_rate": 5e-05, "loss": 0.7671, "mean_token_accuracy": 0.8793023228645325, "num_tokens": 894359196.0, "step": 23557 }, { "epoch": 2.9968197430352372, "ewc_loss": 0.44921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003719329833984375, "grad_norm": 7.979328632354736, "learning_rate": 5e-05, "loss": 0.8096, "mean_token_accuracy": 0.8734267354011536, "num_tokens": 894401362.0, "step": 23558 }, { "epoch": 2.9969469533138278, "ewc_loss": 0.3828125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00030517578125, "grad_norm": 7.25070333480835, "learning_rate": 5e-05, "loss": 0.6976, "mean_token_accuracy": 0.8865298628807068, "num_tokens": 894435445.0, "step": 23559 }, { "epoch": 2.9970741635924183, "ewc_loss": 0.46484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003871917724609375, "grad_norm": 8.292983055114746, "learning_rate": 5e-05, "loss": 0.8175, "mean_token_accuracy": 0.876007080078125, "num_tokens": 894475968.0, "step": 23560 }, { "epoch": 2.997201373871009, "ewc_loss": 0.36328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00028228759765625, "grad_norm": 6.824448585510254, "learning_rate": 5e-05, "loss": 0.6883, "mean_token_accuracy": 0.8834019303321838, "num_tokens": 894511692.0, "step": 23561 }, { "epoch": 2.9973285841495994, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004405975341796875, "grad_norm": 8.863992691040039, "learning_rate": 5e-05, "loss": 0.8874, "mean_token_accuracy": 0.875656247138977, "num_tokens": 894550886.0, "step": 23562 }, { "epoch": 2.99745579442819, "ewc_loss": 0.3515625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0002727508544921875, "grad_norm": 6.345479965209961, "learning_rate": 5e-05, "loss": 0.6602, "mean_token_accuracy": 0.891409158706665, "num_tokens": 894584811.0, "step": 23563 }, { "epoch": 2.9975830047067804, "ewc_loss": 0.61328125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005340576171875, "grad_norm": 9.633790016174316, "learning_rate": 5e-05, "loss": 0.9668, "mean_token_accuracy": 0.8768467307090759, "num_tokens": 894622866.0, "step": 23564 }, { "epoch": 2.997710214985371, "ewc_loss": 0.37109375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000293731689453125, "grad_norm": 6.316112041473389, "learning_rate": 5e-05, "loss": 0.681, "mean_token_accuracy": 0.8930283784866333, "num_tokens": 894663513.0, "step": 23565 }, { "epoch": 2.9978374252639615, "ewc_loss": 0.67578125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000598907470703125, "grad_norm": 10.003772735595703, "learning_rate": 5e-05, "loss": 1.0333, "mean_token_accuracy": 0.8752345442771912, "num_tokens": 894702873.0, "step": 23566 }, { "epoch": 2.997964635542552, "ewc_loss": 0.5390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004596710205078125, "grad_norm": 7.353824615478516, "learning_rate": 5e-05, "loss": 0.8706, "mean_token_accuracy": 0.8820174932479858, "num_tokens": 894737990.0, "step": 23567 }, { "epoch": 2.9980918458211425, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.25438404083252, "learning_rate": 5e-05, "loss": 0.9718, "mean_token_accuracy": 0.8836385607719421, "num_tokens": 894775408.0, "step": 23568 }, { "epoch": 2.9982190560997326, "ewc_loss": 0.6484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000568389892578125, "grad_norm": 8.45877456665039, "learning_rate": 5e-05, "loss": 0.9969, "mean_token_accuracy": 0.8755175471305847, "num_tokens": 894815702.0, "step": 23569 }, { "epoch": 2.9983462663783236, "ewc_loss": 0.4609375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003814697265625, "grad_norm": 7.67432165145874, "learning_rate": 5e-05, "loss": 0.8031, "mean_token_accuracy": 0.880597710609436, "num_tokens": 894849116.0, "step": 23570 }, { "epoch": 2.9984734766569137, "ewc_loss": 0.63671875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.00055694580078125, "grad_norm": 9.215208053588867, "learning_rate": 5e-05, "loss": 0.9456, "mean_token_accuracy": 0.8892025947570801, "num_tokens": 894882020.0, "step": 23571 }, { "epoch": 2.9986006869355046, "ewc_loss": 0.3984375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003204345703125, "grad_norm": 6.799327850341797, "learning_rate": 5e-05, "loss": 0.7368, "mean_token_accuracy": 0.8820600509643555, "num_tokens": 894916994.0, "step": 23572 }, { "epoch": 2.9987278972140947, "ewc_loss": 0.65625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000579833984375, "grad_norm": 9.728473663330078, "learning_rate": 5e-05, "loss": 1.0239, "mean_token_accuracy": 0.8727539777755737, "num_tokens": 894952763.0, "step": 23573 }, { "epoch": 2.9988551074926852, "ewc_loss": 0.484375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004062652587890625, "grad_norm": 7.1527838706970215, "learning_rate": 5e-05, "loss": 0.819, "mean_token_accuracy": 0.8817635178565979, "num_tokens": 894983427.0, "step": 23574 }, { "epoch": 2.9989823177712758, "ewc_loss": 0.64453125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005645751953125, "grad_norm": 9.433609962463379, "learning_rate": 5e-05, "loss": 1.0376, "mean_token_accuracy": 0.8668633103370667, "num_tokens": 895019082.0, "step": 23575 }, { "epoch": 2.9991095280498663, "ewc_loss": 0.6015625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000522613525390625, "grad_norm": 7.912653923034668, "learning_rate": 5e-05, "loss": 0.9106, "mean_token_accuracy": 0.8902484178543091, "num_tokens": 895054051.0, "step": 23576 }, { "epoch": 2.999236738328457, "ewc_loss": 0.55078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004730224609375, "grad_norm": 8.593212127685547, "learning_rate": 5e-05, "loss": 0.8882, "mean_token_accuracy": 0.8801284432411194, "num_tokens": 895090185.0, "step": 23577 }, { "epoch": 2.9993639486070474, "ewc_loss": 0.58203125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0005035400390625, "grad_norm": 8.110030174255371, "learning_rate": 5e-05, "loss": 0.9334, "mean_token_accuracy": 0.8781440258026123, "num_tokens": 895130159.0, "step": 23578 }, { "epoch": 2.999491158885638, "ewc_loss": 0.5078125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004291534423828125, "grad_norm": 8.087014198303223, "learning_rate": 5e-05, "loss": 0.8524, "mean_token_accuracy": 0.8791050910949707, "num_tokens": 895169639.0, "step": 23579 }, { "epoch": 2.9996183691642284, "ewc_loss": 0.51953125, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000438690185546875, "grad_norm": 7.849671363830566, "learning_rate": 5e-05, "loss": 0.8167, "mean_token_accuracy": 0.8941831588745117, "num_tokens": 895209939.0, "step": 23580 }, { "epoch": 2.999745579442819, "ewc_loss": 0.4921875, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0004119873046875, "grad_norm": 8.103150367736816, "learning_rate": 5e-05, "loss": 0.854, "mean_token_accuracy": 0.870932400226593, "num_tokens": 895253073.0, "step": 23581 }, { "epoch": 2.9998727897214095, "ewc_loss": 0.43359375, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.0003528594970703125, "grad_norm": 7.409012317657471, "learning_rate": 5e-05, "loss": 0.7895, "mean_token_accuracy": 0.8772554397583008, "num_tokens": 895295058.0, "step": 23582 }, { "epoch": 3.0, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000423431396484375, "grad_norm": 8.309475898742676, "learning_rate": 5e-05, "loss": 0.8289, "mean_token_accuracy": 0.8851231336593628, "num_tokens": 895335272.0, "step": 23583 }, { "epoch": 3.0, "ewc_loss": 0.50390625, "ewc_loss_diag": 0.000789642333984375, "ewc_loss_parallel": 0.000423431396484375, "step": 23583, "total_flos": 5.156325133756924e+19, "train_loss": 0.8269113934204859, "train_runtime": 50783.0573, "train_samples_per_second": 7.43, "train_steps_per_second": 0.464 } ], "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 }